Webinars

Use the buttons to expand or collapse the section.

• Socio-economic factors associated with the type of treatment received by men with prostate cancer

  31 August, Dr Mei Ling Yap

  Play video

  Read transcript

  Transcript

  yeah yeah thanks nikita uh emailing yap is a clinical researcher whose focus is on

  achieving equity in cancer control locally and globally she's a senior staff specialist a radiation oncologist

  based in southwestern sydney local health districts as well as a senior research fellow at the georgia institute

  for global health she's a conjoined senior lecturer at unsw and western sydney university and

  an adult senior lecturer at sydney university she is

  she's working to support cancer services uh capacity building in low and middle income countries

  um in asia pacific as co-chair of the asia pacific special interest group of

  the royal australian and new zealand college of radiologists so please give a warm welcome to mainly and i'll hand

  the screen over to you thanks mainly thanks serena for the introduction

  um let me just share my screen

  okay what's that is that good we can say it thank you great

  okay yeah so thank you to richard and serena for inviting me to speak today i've recently presented at a prostate

  cancer meeting which richard attended and he reached out and said that there were potentially some synergies with

  some of the work that your team is doing so i'm just going to present some work

  which was done as part of my phd thesis and it was done together with colleagues

  from the daffodil center and i can see that david goldsberry is on the call as well and he was central

  to to this work so prostate cancer is the most common

  cancer diagnosed in men in australia you can see from this graph that is far and

  away the most common cancer diagnosed when we look at deaths um there were 3

  500 in in estimated for this year as compared to almost 25 000 diagnoses

  so the takeaway point is that most men will survive they will be cured of their cancer

  and there are if many thousands of men walking around the community who have either been cured or still

  living with prostate cancer meaning that the treatments that these men choose

  have potential longer term implications for the rest of their life

  with regards to side effects so the choice of treatment that these men have is very important

  in terms of the main treatments for curative intent that is men who have

  prostate cancer that hasn't spread you know further from the prostate area to other parts of the body

  the main treatments include firstly surgery and the surgery that's done is a radical

  prostatectomy where the entire prostate is removed and in australia this is mostly done

  using a robotic technique which is shown in this picture here where the surgeon uses this robotic equipment to do the

  procedure most of the surgery done for prostate cancer is done

  in the private sector and men typically have out-of-pocket costs associated with this even if they use

  their private health insurance and the main side effects from surgery

  include urinary incontinence so they can have leakage and also impotence of the inability to

  gain erections and with surgery of course it requires a general anesthetic and patients might

  stay in hospital for around a week or so and so the more elderly patients this can also

  be a bit of a challenge and and maybe a reason to to look at other treatments

  radiation therapy is another potentially curative treatment for prostate cancer

  and radiation therapy or radiotherapy it says the same thing it uses ionizing radiation to damage

  cancer cell dna and the radiation is targeted to the prostate

  so mostly radiation therapy is delivered as external beam radiotherapy so delivered using a machine similar to one

  in this picture a linear accelerator where a patient will lie on the what we call the radiation couch although it

  doesn't really look that uncomfortable so they're lying on this couch and then the radiation machine will move around

  them and a visible x-ray will go out to to target the prostate cancer

  so that's external beam radiotherapy is delivered as an outpatient so people come to and

  from home usually for treatment in terms of the side effects tends to have lower urinary and impotent

  side effects compared to surgery although a higher rate of bowel toxicity

  and also because it's done as an outpatient people coming to and from home they

  don't need a general anesthetic to have it done traditionally radiation treatment for

  prostate cancer was given over many weeks up to about nine weeks of treatment monday to friday with the

  weekends off in the last few years there's emerged a newer treatment

  technique called stereotactic radiation therapy which is a more focused type of

  treatment it's still delivered through the same machine the linear accelerator but it has meant that treatment

  schedules can be reduced to over one to two weeks um so that has

  you know been been a great development for our patients particularly those who are working where previously it'd been

  quite an inconvenience to have to come for treatment over such a long period of time

  in new south wales most radiation therapy is delivered in the public sector

  there is another type of radiation treatment which we're not going to focus on in this talk but just to let you know

  brachytherapy is radiation therapy delivered internally into the prostate through

  radiation sources and when brachytherapy is given sometimes a

  general anaesthetic will be needed but generally this treatment now is less common in australia um

  it's i i guess with the the developments that have been made with the external beam treatment uh we see brachytherapy

  use less and is typically only used in a few specialist centers around australia

  and then there is a third management option for curative intent treatment particularly for what we call low-risk

  prostate cancer where it's it's diagnosed at an early stage and there's this criteria in which

  street criteria in which patients can qualify to have what we call active surveillance and that's essentially

  monitoring the prostate cancer very closely doing a number of different tests

  including potentially repeat biopsies and only if the prostate cancer then

  changes into a more aggressive looking cancer from these tests will the patient then go and have a

  discussion about the treatment options and active surveillance um

  again i think it's really been over the last decade or so that that's becoming more popular and it certainly is a a

  good management option for some patients who qualify

  so is one treatment better than the other um i think for

  in certainly in the setting of localized prostate cancer the evidence is very clear that survival

  is the same really if you undertake surgery radiation treatment or active

  surveillance as seen by this paper from the new england journal that's now six years old there really isn't any

  difference when you look at these three treatments with regards to survival

  however when we look at the patterns of care of prostate cancer in australia this paper

  looked at victoria and south australia and what we can see here is that surgery

  was by far and away the most common treatment that was given and with radiation treatment being

  really about half the number of men receiving radiation treatment and over time those numbers are dropping

  so it doesn't reflect really the evidence that we see in terms of the survival being equivalent

  a few years ago this issue about you know the type of treatment for men with prostate cancer and the fact that

  surgery is being given more commonly led to a number of

  articles within the mainstream press including a quite

  lively debate on late line between a radiation oncologist and a urologist

  with some really i guess heated discussions about

  could there be potentially other reasons why men with prostate cancer in australia

  are receiving surgery more often than radiation therapy and there was a lot of

  dialogue around the potential out-of-pocket costs that patients were receiving particularly in the context of

  having surgery so in light of this we decided to undertake this study which

  was to firstly describe the patterns of care for prostate cancer patients in the 45 and up study in new south wales

  and secondly to ascertain which factors and these include socioeconomic factors that we'll discuss are associated with

  treatment for prostate cancer so

  we used data from the 45 and up study so this is a large scale cohort study that

  was performed in new south wales and is still on still ongoing um which recruited

  people who were well people aged 45 years and over from the medicare database

  and so over 250 000 participants were recruited in new south wales age 45 years and over

  and importantly these participants not only filled out a very extensive survey on

  their health their lifestyle their social demographics but they also then gave permission for

  their data to be linked to routinely collected data sets which many of you will be familiar with

  including the medicare database um pbs

  inpatient admissions through apdc and cancer registry

  and also birth deaths and marriages so this gave us the opportunity to be

  able to study for the first time really on a granular level

  people's access to radiation therapy and what social demographic factors may be associated with this

  in terms of the linked databases which were used there they're shown here so we

  included any man who had developed prostate cancer after being recruited on study

  and in order to identify those who'd received radiation therapy we used both

  mbs and apdc records similarly for surgery we use those data

  sets for chemotherapy we use those two and in addition the pbs data

  and we also look to antigen deprivation therapy which we use the pbs data for

  in terms of the social demographic and socio-economic factors that we looked at

  firstly we looked at measures of social isolation their place of residence the distance from the nearest radiotherapy

  centre and their ethnicity we also looked at health related factors

  including comorbidities and their physical functioning and in terms of measures of socio-economic status we looked at the

  annual household income area level socio-economic status the individuals education level their health

  insurance status and whether or not they worked

  and we use a competing risk regression to compare the receipt of radical prostatectomy versus no prostatectomy

  across the categories of social demographic factors we adjusted this for age and stage

  and similarly we used computing risk regression to compare the receipt of external beam radiotherapy versus no

  external beam radiotherapy and we did multi-variable models with purposeful selection of covariates

  so in terms of the linkage um you can see that there was there were millions of records that were

  linked we excluded men who had a department of veteran affairs card

  because there was some concern that their records may be incomplete certainly for pbs and possibly for mbs

  we also excluded those who had been diagnosed with multiple cancers after being recruited on study

  and we also excluded those who'd had any previous radiation therapy

  and this we ended up with just over 4 000 men who'd been diagnosed with prostate

  cancer after recruitment on study with a median follow-up of five years

  so when we looked at the treatment received by the men in our cohort

  similar to the victorian and south australian data you can see that surgery was the most common treatment

  and really about half the men who'd received surgery half the number

  then we had half the amount of men who received radiation therapy

  one in ten received just hormone treatment alone and we had one in five who didn't have

  an active treatment such as surgery or radiation therapy although we weren't able to ascertain if they were actually

  on that active surveillance regimen as opposed to having no treatment

  when we look at localized prostate cancer and localized prostate cancer low risk cancer these are the patients where

  there's clear evidence that surgery and radiation therapy have equivalent outcomes you can see that

  it's actually even more marked with almost half the men receiving surgery

  and only one in five receiving radiation therapy

  and again we have smaller numbers receiving just hormone therapy and one in five with no active

  treatment interestingly our data

  is very different to what we see in the us and in in the us the surgery and

  radiation therapy rates are actually very similar in the us they have very different

  models of reimbursement and and actually there are that the typical setup for a center

  which is delivering radiation therapy is that there'll be urologists who own

  the center and that are referring within their center to radiation therapists

  thereby really um i guess the reimbursements are different in the sense that the urologist

  will also have potentially benefits in um referring to the radiation oncologist

  within their center because often they will own that that treatment center

  which is obviously very different to what the model is within australia and

  similarly in the uk we see a different picture whereby radiation therapy uptake is much higher

  and in the uk of course things are quite different with majority of treatment being through nhs

  publicly provided treatment both for radiation therapy and surgery so the

  patterns of care we see in australia certainly contrast very differently

  significantly to the us and the uk

  we also were able to look at the number of patients who had radiation on calls

  oncology consultations prior to surgery because we had the mbs data

  so we looked at you know patients who went ahead with surgery for their for the potential cure of their treatment

  whether or not they actually saw a radiation oncologist in consultation prior to that to

  potentially discuss another option and you can see that those rates were very low where

  13 of men who had surgery had a formal consultation with the radiation oncologist prior to going ahead with

  their operation then we looked at the factors related to prostatectomy receipt

  so you can see that age certainly a factor where elderly patients are less likely to have surgery

  um also we found that those patients who have more localized or regional disease

  are less likely to are more likely to receive an operation patients who had a higher body mass

  index so who were obese or less likely to have an operation and those patients who had poorer

  physical functioning also less likely to have an operation interestingly we found that patients who

  were single as opposed to partnered or married were less likely to have surgery

  patients who lived more than 100 kilometers from the nearest radiotherapy center um were more likely to have

  surgery patients who were in more disadvantaged socioeconomic status

  quintiles are less likely to receive surgery and interestingly the patients who had

  no private health insurance or less likely to receive an operation as well

  when we move on to radiation therapy again we found that age was predictive with

  patients who were in the 70 to 79 age group more likely to receive radiation

  therapy patients who had regional or distant disease are more likely to receive radiation therapy

  patients who had more comorbidities were less likely to receive it patients who lived further from a

  nearest radiotherapy center again less likely to receive radiation therapy and i guess as a complement to what we

  saw with surgery we found that patients who did not have private health insurance were more likely to receive

  radiation therapy we also found that patients who were working were less likely to

  receive radiation therapy so just to summarize in terms of the

  factors predicting for the type of treatments i think it was that the health factors that were predictive are

  quite i think self-explanatory and makes sense those who have more committed more co-morbidities high bmi

  etc are less likely to receive surgery and radiation therapy the socio-demographic factors some of

  them i guess were less intuitive to us partnered patients we found were more

  likely to receive surgery which i guess questions the role of the

  partner in decision making in that those who were partnered were less likely to

  go ahead with it and patients living more than a hundred kilometers from the nearest radiation

  center were less likely to receive radiotherapy and conversely were more likely to receive surgery which highlights

  the poor access to radiotherapy for some patients and patients who work full time are less

  likely to access radiation therapy as mentioned the cohort from this study

  would have had treatment at the time when most treatment was given over seven to nine

  weeks however more recently there are newer shorter treatment schedules

  which hopefully will mean that patients who work will be less put out by potentially accessing

  radiation therapy we also found that surgery was more common for patients who could afford it

  so men living in disadvantaged areas were less likely to have surgery men who had private health insurance

  more likely to have surgery and less likely to have radiation therapy

  importantly radiation therapy is not covered by private health insurance patients who access private

  uh who access radiation therapy in the private sector um so

  you know go to a private center and have their treatment they i think i've heard that your group

  is doing some research in this area but typically medicare will get billed but then

  there'll be additional out-of-pocket costs for that patient but they're not able to actually use their private health insurance for the radiation

  treatment why is the choice of treatment important if you know the survival is the same

  whether it be surgery or radiation therapy so it's important because we know that

  their treatment decisions for prostate cancer will lead to potentially

  long-term lifelong toxicities as mentioned for surgery these could be

  urinary and sexual side effects for radiation therapy these could be bowel side effects

  and there has been a work that has shown that men with prostate cancer often experience

  decision regret despite this we found that patients were twice as likely to receive surgery compared to

  radiation therapy and only one in eight men saw a radiation oncologist prior to

  undergoing their surgery for prostate cancer the choice of treatment is also

  important because there can be significant financial toxicity from prostate cancer treatment within

  australia with data showing that out-of-pocket costs can be

  for men with 75 spending up to 17 000

  for their prostate cancer treatment and that the out-of-pocket costs for men with prostate cancer were found to be

  higher for those with private health insurance in our study within our cohort

  three-quarters of men who had surgery received it within a private hospital however in for radiation therapy in new

  south wales most is delivered in public hospitals and there's no additional out-of-pocket costs given to the

  treatment centre so this is an interesting qualitative

  study which was done um which which interviewed they interviewed

  men with prostate cancer who were actually able to see both a surgeon and a radiation oncologist

  within a combined euro-oncology clinic and

  the study was interesting because it suggested that treatment choices were predominantly influenced by the

  clinician and most commonly that's the urologist's recommendations because all these men who went to this combine

  clinic had already been seen by a urologist prior to this for the diagnosis

  patients mistakenly believe that surgery provides a more definitive cure than radiotherapy

  patients concluded that they valued seeing both a urologist and radiation oncologist and that patients

  need information about all treatments earlier in the treatment decision making because there were some men who went to

  this combined euro oncology clinic who had already decided they were going to undertake surgery based on

  their initial consultations with the urologist who is the person who makes the diagnosis of prostate cancer

  the limitations of our study were that the sample population is the 45 and up study cohort which is not the whole of

  new south wales and it meant that there were future fewer culturally linguistically diverse patients

  and likely more health literate patients because they had to fill out that um quite extensive survey

  and this study also purposely oversampled elderly and rural patients

  in terms of the clinical data we had access to there were high proportion of patients with unknown stage which we see

  in the new south wales cancer registry due to the way that they actually stage patients

  and we did not have what we call tnm categories which again is a more granular way of

  staging patients and we didn't have gleason score which is important in risk stratification for prostate cancer

  we were also unable to differentiate whether patients who were having no active treatment were actually on active

  surveillance and of course there was a time lag in the availability of data um based on the

  the cancer registry the the how recent their data is

  so to conclude despite the evidence patients with prostate cancer had twice the utilization of prostatectomy than

  radiotherapy with a low percentage of surgical patients having a radiation oncology concept prior

  and that raises the question of whether patients have full understanding of the options we found that patients who have

  partnered were more likely to receive surgery and that distance from the radiation therapy center was predicted of

  treatment and lastly that the ability to pay appears to be influencing the type of

  treatment received in terms of future solutions i think we

  need a better understanding of the patient experience and there's certainly a lot of work in this area

  patients in rural and remote regions they need further initiatives to help

  their ability to access both surgery and radiation therapy

  euro neuro-oncology specialist clinics uh ideal where a patient could see a urologist and a radiation oncologist at

  the same time um however there are very few of these in australia whereas they're very common

  in the united kingdom we need more education of patients partners

  caregivers and doctors on the options for prostate cancer and decision aids could be helpful

  and lastly we we need to review the policies that not only affect access to care but that also affect the

  reimbursement schedules for the different treatment options for men with prostate cancer in australia

  um a positive thing that has come out in the last few years is that there's been a change i mean

  working around because of this issue that we talk about in terms of referrals

  to radiation oncologists before surgery there's actually been a change to the mbs

  wording for the codes for prostatectomy that say that there should be multi-disciplinary management including

  team review documented in writing and provided to the patient and that men should be encouraged to see

  both a urologist and a radiation oncologist to discuss their options

  if you want to have a look at this work in more detail this was published in the mja last year and with us with a

  editorial that was associated with it with both a surgeon and a radiation oncologist giving their opinion

  so i'd just like to acknowledge all the collaborators on this work from the daffodil center as well as from secor

  ingham institute and of course all the participants in the 45 and up study and it's just open to any questions

  thank you

  thank you mailing for that really fascinating presentation did anyone have any

  questions

  i'll change your channel i'm jumping in um

  i wondered if you could look at the active surveillance by looking

  in the um mbs data for uh repeated psa tests or or the um

  or indeed a biopsy whether that would give you a better indication

  so that that's the first question because that follow-up is very interesting the other

  thing we might discuss though is is sort of radiotherapy um and how you ensure equitable access

  given we're a big country with with a lot of remote

  and difficult rural areas when you've got something that is so capital intensive and i wonder if you'd like to

  expand on that yep so the first question yes that's a

  good point that i think in terms of active surveillance you know there are ways to potentially capture that using

  mbs it was outside the scope of our study um but yes i think biopsies and psas would

  certainly be a way to potentially capture that for future work with regards to access to radiation

  therapy in regional and remote areas it certainly is an ongoing issue it has

  been in there have been some improvements um in the last decade so um it's new centers that have popped

  up in the north coast lismore port macquarie coffs harbour um in the south coast in nara

  um so you know there has been some improvements but of course there will be patients where it's far too difficult

  for them to travel um and you know it wasn't until i think

  15 or 20 years ago that darwin had a radiation center built and then

  before then patients within the northern territory all had to fly to adelaide to have treatment

  so we're still not you know able to have each patient within 100 kilometers

  so again you know having things like tele telehealth for their initial appointments and accommodation and

  transport as well as you know ways to cover their financial costs is important for patients

  if they choose to have radiation treatment i think for those living further away

  when there's an alternative option of surgery they may often choose to

  undertake surgery for other cancer types whereby you know we don't

  have the same evidence for surgery and radiation having equitable out outcomes then it'll be more challenging

  for patients where they actually do need radiation therapy as their only treatment option

  then we then we run into those issues about accommodation transport

  you know how to make sure that they still have enough family support social supports

  so all of that is very important to ensure that our rural and regional patients are looked after

  yeah i i had a follow-up question uh and this is going to uh

  demonstrate my uh ignorance but uh this this um

  division between the public and the private sector here seem very stark and so

  from from jane's comment i'm guessing that the problem is just the setup costs

  associated with uh um so i'm wondering why there isn't a bigger

  this is not not not a good thing to say when we're worried about equity but why isn't there a bigger private sector for radiotherapy

  right so because it then becomes a you know one treatments associated with the private sector the other treatments

  associated with the public sector and so there's you know there's there's really no

  yeah that's the choice right so yeah it's confounding the the problems that

  you've got that one treatment uh that they're the treatments themselves you know

  are equally as good it's a good question so

  um when i first started training which was quite quite some time ago there there

  really weren't many private radiation centers available um but over the last decade there really

  has been a huge increase in the number of private radiation centers

  i guess for these private centers to have you know a model which is going to be financially beneficial they need to

  have a certain number of patients who are going to come through each year

  so you need you know you need to have have an area where firstly there

  there are enough diagnoses of cancer and secondly there's not necessarily a good

  public option nearby um so there have been some areas where

  private machines have gone in recently in new south wales fruits be for example is one um

  and but i think we're probably going to see that increasing more and more and there have

  been um there's at least one case i know of of a public center in australia that was

  taken over by the private sector whereby i mean i can only imagine that the private companies will approach the

  public and say that we can do everything that you're doing for much cheaper so i think we're going to see the lands

  get changing more but i guess the the the differences remain still at the

  moment within in terms of um private health insurance that those

  patients who go ahead with radiation therapy in the private they can't use their private health insurance

  they will end up have medicare will be will be billed and then they'll pay the out-of-pocket costs which is different

  for patients who receive chemotherapy they can use their private health insurance to

  to undertake chemotherapy so it's it's it's odd it's the way that it's set up but that's the

  way it is at the moment i think part of it has to do with my

  understanding is that because radiation therapy is an outpatient treatment that's why it's it it's it's medicare

  rather than private health insurance whereas even if the chemotherapy the people receiving chemotherapy only

  are sitting in there for a you know a few hours i think they're getting billed as potentially

  similar to an inpatient so then they can use private health insurance is my understanding

  anyone else i might jump in with one i think you touched on this in one of the papers you referred to but i was

  wondering if you could reflect a bit more on uh to what extent patients ar

Use the buttons to expand or collapse the section.

• Using stated preferences methods to explore broader concepts of value in cancer care

  10 August, Associate Professor Richard De Abreu Lourenço

  Play video

  Read transcript

  Transcript

  
  Kees van Gool  00:00
  So, let me now introduce the speaker. Richard is Associate Professor with the central health economics research and evaluation at UTSA. And he's one of the CRE's chief investigators. He is the senior evaluator of the feedback team and coordinator of the Cancer Research Economic Support Team. CREST are supported by Cancer Australia, which had commenced his career as an economist with the Reserve Bank of Australia and made the jump through the brilliant but short-lived New South Wales Health Economics training program where in fact, that's where I met Richard for the first time well over two decades ago. He holds a Master's of economics honours from Sydney University and a Bachelor of economics honours from Murdoch University, and a PhD from UTS. Richard is a highly experienced and highly regarded health economist, having worked in the public and private sectors as long as I can date. He has a keen interest in applied economic evaluations, the economics of specialty health areas and patient preferences and quality of life priority setting. Richard's primary interests and focus in specialty health has been the fields of oncology and hematology across a broad range of indications and clinical settings. Today, he's going to bring a few of these primary interests together in his presentations, entitled, using state of preferences methods to explore broader concepts of value for cancer care, please take it away, Richard,

  Richard De Abreu Lourenco  01:18
  Thank you, Kees and welcome, everybody to this webinar. It's a real pleasure for me to be able to speak to you all today, here virtually. So I'm just making sure that that's working. As Kees said, I'm more than happy for this to be a bit of an interactive session; it would be just as boring for you as it is for me to listen to me speak for 30 to 45 minutes without some interaction. So please stop me ask questions along the way. As Kees said, the ideas I'm going to speak to about today do bring together some of the research I've been looking at for a number of years, and and an interest that I have in how we make decisions, patient preference, cancer. And importantly, these, these very critical concepts of value. Here are my general disclaimers; the most important one is the one at the bottom, which I will try to limit for the purposes of this afternoon. So I thought I'd start the day with highlighting an article that many of you may know, as it helps to ground much of the discussions that we are having now about value-based care and value in cancer care. And for me, one of the takeaways that that that come from this article is that when we're looking at this idea of value and value-based care, we should be thinking about it as being grounded and always being defined around the patient. What is important to the patient in in accessing care and in what they're taking away from that care. And it isn't just about what is important. It's not only about what goes into the back care or the costs of producing care, but it's how those costs are achieved, relative to the outcomes. So how do we balance outcomes relative to costs, which is a pretty fundamental concept to people like myself in case and many others on this call, who are always thinking about that value proposition and how we weigh up, what goes into and comes out of the provision of care. And we're also recognising that outcomes. And what we're producing is multidimensional, there are many aspects have many facets to those outcomes of care, and that they vary across care. They pretend there isn't a one size fits all approach in what we're thinking about as outcomes, and in what we're valuing. So how might we as a population come at this notion of, of value in determining value? Well, as many of us, many of us know, then the idea of value is often reflected by what we're prepared to pay. And in a public system like Australia, that comes about in terms of what we're prepared to pay it as reflected in something like an incremental cost-effectiveness ratio, we look at how much we're prepared to pay relative to, to its outputs. Compared to what we're doing at the moment, and committees like the PBASand MSAC do that all the time. That's not a notion that is only restricted to Australia. But it's something that's happening more and more around the world. Obviously, the UK has been doing it for a long time. New Zealand does it Canada does it. It happens in Thailand and even Japan is now doing it to a more or less extent. Even in the United States where we don't have public reimbursement per se. We have we have ISO which has adopted its own value assessment framework, where they are looking at these ideas of how we can incorporate value to look at what we're paying for care and ensuring that it represents some notion of value. Interestingly, for eisah, they're not just looking at that assessment and based on the comparison of costs and outcomes, but they are establishing some notion of a willingness to pay, and looking at prices that are required to achieve that actual willingness to pay threshold as one of their metrics. What's common to many of these approaches is that the outcome that they're interested in is the quality. So that combination of how long people are living and their quality of life, the quality adjusted life year. But that's brought some in the industry, to thinking about well, is the quality broad enough to measure that multi dimensionality that we that I mentioned earlier, when we're thinking about value based care, is that nexus between the length of life and quality of life enough when we're thinking about value propositions, and indeed the that the group that the article I just showed you here, led by Lu garrison from esport has produced a taxonomy of potential additions, we might like to add to our conceptualization of outcomes, when looking at this value proposition. and expand on the quality. And think about how we can include other outcomes beyond quality of life and potentially length of life, in economic evaluations. So some of the things that we're thinking about values around motion information, for example, so the value of knowing so for meaning diagnostic tests, there is value in understanding what our risks of particular conditions are, indeed, in the face of a pandemic, the value of knowing what our what our risk of contagion might be, whether we are going to face increased, increased risk of becoming ill, there's peace of mind from certain treatments and the value of insurance from one treatment versus another, we might want to include disease severity, when looking at the value of one of a new treatment, how that might influence the the idea of what we're willing to willing to pay more and more in cancer particularly for in advanced conditions, there is the notion of can we pay for the value of hope, particularly with with with some conditions like pancreatic cancer or brain tumors, there is an increasing call to factoring whether or not treatments are adding value in terms of adding hope to conditions where there might not otherwise have been treatments, or treatments available. We can think about potential value is also arising from this notion of an option value. That is by by extending life by making people live longer, we are giving them the potential to benefit from future treatments, is there a value attached to that? There might also be a sense in which we want to expand outcomes measurement to allow for better equity of access. And there is a there is a wide movement at the moment, particularly in the health economics, space about how we can adapt economic evaluations to take that into account. Finally, there are some more process related measures around outcomes and how we can change those to take into account the effects of innovations in moving from one treatment to another. And how advances in one area can impact on another area. And indeed, more directly how the patient's experience of care might lead to value in and of itself. are patients finding care more or less cumbersome? Are they benefiting from more convenient care? Are they engaging in more in more shared decision making, and experiencing greater autonomy? And are those aspects or part of the value proposition? So we were interested in this question of whether or not those sorts of broader value propositions were actually being utilized by by researchers and individuals, when assessing value and we looked at the literature and found that within publications of economic evaluations that people were actually using some of these concepts to try and get out this notion of value, either through a willingness to pay metric or through a quality of life metric and that is by trying to put quality of life weights on some of these concepts like information, convenience or process of receiving care like shared decision making. And of the papers that we reviewed, what we found was that quite a large proportion, around 28 of those papers used a technique called stated preference or discrete choice experiments as the means of attaching value to the outcomes that they were investigating. So that we have this broader notion of outcomes. And they were trying to value those broader notions using these discrete choice experiments. And in the next few slides, what I'd like to do is give you a bit of a flavor for what a DC actually is. And then I'll take you through a couple of examples. So many of you on this, I can't actually see you at the moment because you know, I can only see keys. But for many of you on this webinar, you will actually have a much more deeper technical understanding of what a DCE that is the novel ever. But this is a survey technique that we use, where we can actually observe in a market how choices are being made. Ordinarily, we can observe in markets how people are making choices, but discrete choice experiments allow us to construct surveys to observe how people might behave in conditions where we can't observe a certain situation. They are therefore based on what we call a hypothetical hypothetical, but realistic scenario that describes a product or service in terms of its underlying characteristics. And I'll show you what I mean in a moment. And we vary those underlying characteristics or attributes between plausible ranges and ask people to make choices about what they would do in varying scenarios. And we use those choices to understand what it is how, how people are trading off, and essentially, what it is that people are valuing in those different contexts and in different circumstances. The key assumptions underpin underpinning this type of survey based assessment of how people make choices and in understanding preferences and values is that we can understand choices by comparing the underlying attributes of a product or service that's on offer. This relies on the premise that any good or service can be broken down into those attributes. So for example, we can think about a car as being broken down by whether it's four wheel drive or two wheel drive, its color, how fast it goes, its cost, those sorts of things. They're all attributes. And an individual will always choose the combination of attributes on offer that they expect to maximize their well being, which which combination of these attributes that they're being offered, would make someone feel most well off. So here's a bit of an example a bit topical from today. So imagine that we're offering someone asking someone whether or not they would participate in a vaccination program. The characteristics or attributes that we might use to explain that vaccination program are things like whether or not there are side effects, who administers that vaccine, that vaccine and where, who should be vaccinated within the community, the proportion of the population that's likely to be vaccinated once it's rolled out whether or not the vaccine is efficacious in avoiding infection, how much it might cost the individual. And at the moment, whether or not it's going to impact on lock downs and allow everybody to go out and have some fun. Okay, so these are the sorts of ways that we might conceptualize a program. And hopefully, you can start to see that we could think about most programs or products or services in this type of way and breaking it down into its attributes or components. Setting up a DCE and thinking about the problem that we're going to answer is, is quite critical. And it really does frame the results that you will be able to get out of your out of your research. So if trying to use it to value a particular outcome of interest, then the way that the question is framed is really, really critical. One of the outcomes I alluded to earlier in that broader framework from garrison was something around the value of hope. And if you could, if you could think about that for a moment and think about well, could you describe a or could you construct a discrete choice experiment that was looking at hope as something that we might try and value? and important thing to consider in thinking about that sort of question is whether or not hope is extrinsic or intrinsic. And what I mean by that is, can you in the context of an experiment and And a survey to ask someone to imagine that they are feeling more or less hopeful, or that by using a particular product, they will feel more or less hopeful. And I'll let you think about that. And the answer might be no, in most cases, that there are some characteristics or attributes that are very difficult to ask people to imagine. And so we have to think about a different way of potentially getting at those outcomes. And I'll show you an example in a moment. So setting up the problem is very critical. And you can see here that the the steps we'd go through, I'm not intending to give you a lecture on how to construct the DCE. But these are the steps. The key takeaway from this slide, is that how we quantify the question is critical. How we describe the attributes and levels, it's critical, and important input input at this point, is to actually have some robust qualitative research, usually with the people who are going to be affected by the program. So if it's about how patients are interacting with the healthcare system, then I would greatly suggest that having patients involved in this qualitative research is critical. And having the patient voice embedded in understanding what the outcome is, and, and in understanding how the outcome is described, would really help set this phase of this research up. really well. The much of the much of the driving force and the engine behind the discrete choice experiment comes with its mathematical design. And it comes in ensuring that what as you can imagine, is a very complex combination of, of attributes and their levels, is presented in a way that can be understood by respondents, but importantly, allows the the analytical framework to investigate the relationships that we're interested in. And to potentially produce or to investigate whether the value propositions that we're interested in can be can be backed out. And so getting this design right, with the help of people who have strong mathematical backgrounds, and have experience in experimental design is a really important element of framing of developing a discrete choice experiment. It helps to set up how the experiment will be rolled out or is critical in in setting up how the experiment will be rolled out, but also in how the analysis will also will be conducted. actually implementing a DCE, who you're going to implement it with will determine the types of information that will be produced. A critical question here in assessing value and how that value is going to be used is who you're going to ask. So if we're interested in coming at values for cancer care, while patients will help us to understand what is critical in terms of the description of the outcomes that we're interested in, in the descriptions of the service, if those if the values that come out of this sort of exercise are going to be used in public decision making, then perhaps it's this it's societal preferences, society's views on these discrete choice questions that are going to be more important, and we could have a discussion in the debate about that, at the end. Getting the surveys completed. With enough individuals and having the right number of completions is also critical. That's something that can be worked out during the DCE design phase. Happily, these days, we can do these surveys online so that large surveys can be completed relatively quickly. Of course, if we are doing with doing surveys with patients, it can take much longer to recruit the patient samples required. Nonetheless, they are it is generally much faster than doing surveys by paper analysis of all these data for number crunchers and geeks like me can be a lot of fun, and allows us to interrogate not only what is driving choice within a particular choice, context or service in and of itself. But if it's been designed appropriately, it also allows us to look at whether or not value propositions and what is influencing value differs between between different groups. And we can potentially assess how something like the value of information or the value of convenience differs between different patient characteristics are different socio different socio demographic goods. As I said that really depends on making sure that we've designed it correctly to allow us to do that, but it is possible Now, thanks to developments by people like Nick bands back, we can also not only look at value in this in the context of dollar values, but we can also start to look at value here in terms of quality of life weights, which starts to bring us back to that the notion that the quality can be expanded to bring in some of these broader concepts of what represents value. How we apply this, all of this information that comes out of these sorts of experiments, I've talked about willingness to pay, I've talked about quality of life weights, one of the things I've not talked about is actually looking at uptake. And that's really not something I was going to talk about a lot today. But these experiments do tell us about the probability of use, because they're telling us about when people might choose one program one service over another. And if you're designing a healthcare program, understanding those probabilities of use can be very powerful in helping to set up programs in a way that will be potentially used to the most in a maximum way. Sorry, that's my home phone, very old school. So let me take you through one brief case study. How embarrassing. Here's a, a, an example that we looked at a couple of years ago, where we were looking at the situation of women who were managing breast cancer risk following an initial diagnosis of breast cancer. And we want to understand what though, in that particular situation, what the value was associated with reducing the fear of recurrence in those women, once they if they were undergoing a contralateral prophylactic mastectomy. So much of this in the same way as it's difficult to imagine telling someone that they are feeling more or less hopeful. It's also very difficult to ask someone to imagine that they're feeling more or less fearful. The experience of fear is something that's very intrinsic to an individual. So in conceptualizing the data, the discrete choice experiment here, which was looking at the choice between undergoing a contralateral prophylactic mastectomy, or, and or continuing to undergo usual monitoring following a breast cancer diagnosis, we couldn't really conceptualize of an attribute that was about how fearful someone felt. But we could have a question after the discrete choice experiment that assessed how the level of fear or cancer concern that people were feeling while they were completing the choice question is, and that's what we did, because we wanted to utilize that that rating about cancer concern to then do some subgroup analysis. So you can see here the schemer of the overall project, it was grounded in doing qualitative work with women who had undergone contralateral prophylactic mastectomy or not piloted and then implemented in in, in looking at whether or not sorry, looking at the choice between CPM or contralateral prophylactic mastectomy, or regular management, regular monitoring for breast cancer risk. What we saw, and what we saw is that when we looked at the willingness, marginal willingness to pay for our attributes, and we had attributes around the impact on breast sensitivity, the frequency of undergoing scanning the types of scanning that women underwent, and the impact on breast cancer risk overall, what we saw was that in women who expressed some degree of concern about the cancer coming back, they placed a higher value in avoiding cancer risk, and wanted more frequent checkups and wanted to be involved in their decision making compared to women who didn't express a cancer concern. So when we use that idea of fear of cancer recurrence or cancer concern, to try and stratify, or to look at how value differed between these groups, we could see some notion of difference. Now the cane beans among you will say I look but you know, many of your values there are different, that's due to our sample size, but we at least starting to see some signals here. That driven by that are according to difference in these in these cans of concern levels. And so we can start to see some differences there from from that source. Another case study I wanted to run was a recently completed and recently published study, looking at genomic approaches to Children's Cancer Care. And here we want, we were looking at the value associated with choosing to participate in genomic medicine in children with cancer. And in particular, we were interested in looking at how the factors that influence that participation differed between the parents of children who had an advanced pediatric cancer, healthcare practitioners, and the general community health care protect health care practitioners in this context was a broad, inclusive statement. So it wasn't it wasn't only medical practitioners, it included supportive nursing and allied health. Again, this work was informed by extensive pre qualitative work with parents and healthcare practitioners. And they had identified that hope was something that they worked, they were all interested in, you know, they wanted to do anything that they could for their children, and a desire to do what it takes, often motivated actions. The attributes that were included in this DCE were things around the survival prognosis, so how sick was the child before they underwent the test, the benefit that they might, they might experience from having a treatment based on a test whether or not a target could be located as the source of of treatment. So could we could a mutation be found that could be treated, the impact on quality of life of being treated, whether there was a recommendation from a clinician to treat or whether the family supported the treatment. Now, that was a reflexive attribute. So clinicians only ever saw whether or not the family supported the treatment and families saw whether or not clinicians were recommending treatment. We then had a couple of attributes around cost, which looked diff, which looked separately at who was paying and how much was being paid. And finally, an attribute around the need for a biopsy. What we've found across those three groups was that increasing the chance of people participating in, in genomic medicine increased when we had a higher probability of baseline survival, a higher probability of children living longer if they did find a target, improved quality of life. So improving quality of life following treatment meant people were more likely upon a recommendation to participate in in targeted, a targeted approach and finding a target. But that was only important for medical practitioners. What we did find, though, was that the importance of those things differ across the groups. So for families and and the general population, we found that the quality of life was the most important attribute. So they were really influenced by what was happening around kids, kids potential quality of life, whereas for healthcare providers, they were all around, let's let's make sure that these children are going to live longer. And so we see a slight difference between the two groups in that, now health care providers were, were also influenced by quality of life. And you can see that there. But we don't see the same influence of survival benefits for parents and the general community. They were more influenced by quality of life, as I said, and then by things around recommendation, and who pays. What we took away from this is that we start to see a little bit of a hierarchy of, of outcomes in, you know, in terms of what people were willing to pay. And these these are based on the general community sample, where we have quality of life being something that is very highly valued. shared decision making, is also quite highly valued. But we found this interesting idea that there was some level of baseline survival at which people actually were not willing to undergo or recommend that their children be treated, where it was almost the inverse of hope. But it was potentially reflecting some notion of utility, that where children were where the prognosis for a child was so poor, and could potentially couldn't be improved through quality of life. They didn't want to participate. Pain in treatment, reflecting a potential notion of why would we put kids through this if the game is so small, which we thought was a very interesting finding in this context. So, from those two case studies, I wanted to link this back a little bit to some of the discussions that Porter raised raised through his work. And that is that we can use these sorts of sorts of studies to look at how people are making decisions and making choices, but also to start to think about how outcomes might be valued through some sort of hierarchy. And I raised this because through these case studies, and through other theses that I've seen either been involved with or I've seen, you start to see that there are outcomes, which we might call high level outcomes. You know, things like efficacy, risk reductions, and quality of life. These are things that are potential drivers, or what I've called an influencer, in today's parlance of value, they are things that we can observe quite readily, and have a large impact on people's behavior, and a large contributors to value. But there are other attributes such as shared decision making, and who pays for care, that also influenced choice and value. And that perhaps we shouldn't be ignoring, when we are thinking about how we pay for care, and how and when we're thinking about what constitutes the value that that we attach to care when we're when we're putting forward our value propositions for payments. The other thing is that we can differentiate between groups in terms of how much people are in terms of what people value and how they might value things differently. And that's, again, another another important motion to keep in mind, when we are looking at what what we're paying for care. And thinking about the payment systems, we might put in place. An interesting proposition, and this is hopefully to foment some some discussion is whether or not we can use these types of studies to investigate how we might pay for care. And indeed, there have been some studies of some VCs looking at payment systems. I've come across some, or there have been some here locally, but most of the ones I've seen, we're looking at payment systems in in LCMS, looking at how we might structure payment systems for primary care, but could we conceive of constructing a discrete choice experiment around a value based healthcare system? And if we could, what might that look like? Who would be involved? And you know, how would we go? Whose preferences would we go? Go about eliciting? Julie, I'm nearly finished, and then we'll come to your question. That's all right? So my my musings because this is my last slide, Julia promise, is that the and I hope you've you've taken this away is that this notion of value is broader than then what we often measure as a final outcome in clinical studies, and clearly goes beyond what we put into care in terms of costs. These fees are a useful tool in helping us to get at some of those different notions of value, they can be very, they can be a very powerful tool, we do need to be careful in how they are applied, and utilized the way that they designed it is very, very critical. I think I think there is a lot of potential to utilize them in understanding what matters in the way we assess value and potentially, in helping us to put together programs that better reflect what's value and potentially to design those payment systems. But there's a lot to knock out in how we might do that. And as my little icon men there says thank you, everybody. And now we have a couple of questions case you want me to go straight to those?

  Kees van Gool  34:30
  Yes. Would you like me to read them? Yes, and if you like, or you might be worthwhile reading enough for everybody else who can't see the other chat rooms. So Julie asks, How much variation is there across different countries or cultures?

  Richard De Abreu Lourenco  34:48
  So Julie, do you mean within a discrete choice experiment, or do you mean in the practice of applying?

  34:56
  Hi, Richard, great to see you. Thank you. I mean, Is my video anybody know now? I meant to replicate the same research study in a different country? Yeah, do it in terms of, you know, the high level things that matter and the other things that matter left or the individual results really anything that you can say at all about variation across different countries and whether or not we can apply results from one country to or one one group of participants to a different group.

  Richard De Abreu Lourenco  35:32
  Great question. I think we always we, we should always be a little bit cautious about directly applying results from one country to another, particularly where we're dealing with some concepts that aren't where we're dealing with concepts that might not be concrete. So if we're dealing with something like survival, for example, that might be transferable. But when we're starting to deal with things that are less tangible, and might be more subject as to cultural context, then we need to be very careful, we need to ensure that the the ideas and concepts are going to be going to translate. So I think local, local adaptation of a study is always a good idea. I run it in the local context, if you can, if you can't do that, then think about what how applicable other methods that have been used there to o

• Provider behaviour and payment reform: Evidence from the withdrawal of public

  11 June, Dr. Olukorede Abiona

  Play video

  Read transcript

  Transcript

  Jane Hall  00:00
  Well, in case there's any body on who doesn't know me, my name is Jane Hall. And I'm one of the professors of health economics in CHERE at UTS and part of the CRE, I want to formally start the meeting by recognizing and acknowledging the traditional owners of the lands on which we're all seated today many different lands as we're in very different parts of Australia. Let me just start with the usual housekeeping about these sorts of zoom meetings. If you would, please mute your microphone that will help with any background noise or other interventions. If you wish to ask a question, please do put something in the indicate in the chat function or by raised hand. And there will be the opportunity for questions at the end of Olu's presentation. Before I formally introduce Olu , first of all, a word from our sponsors, this series of webinars is very much about the work we're doing under our NH and MRC Center for Research Excellence funded program, which is about paying for value and improving payment methods, particularly around cancer. So you can hear more about that work either by emailing us and the email address is on that slide there by following us on Twitter. And, or getting in contact with any of the individuals from from the CHERE team at any stage. The webinar is being recorded. And it will be available on our website, as indeed our our previous seminars webinars are. So thank you for all of that. And it's my pleasure to introduce Olukorede Abiona. Olu has a PhD in economics in Applied Economics means lots of quantitative stuff. So all it comes to us with those skills from the University of Leicester. He joined share in 2017. He's been involved in a range of projects. But today, he's going to talk to us you can go to the next slide Olu, he's going to talk to us about provider behavior and payment reform, a very important topic because so often we assume that by withdrawing a subsidy for something, the thing will go away. And this is a very important topic to explore in terms of understanding how providers respond to payment reform. So thanks, Olu and over to you.

  Olu Abiona  03:06
  Thanks a lot, Jane. I would like to start by saying this paper is co authored with five other beauteous members, including ourselves, Dan Lui , Maryam, Phil, and Kees and we're delighted to have our next panel member in the name of Sarah Neville, as well. So in this paper in this project that we have now that we have been doing, Robin, Robin investigating for about three years now, we're looking at the unintended impacts of subsidy removal in Australia. And by way of introduction, Medicare provides approximately 5700 eltz services that are listed on MBS, which means that there is a fixed term ribbit. Our service so far are for kids. So out of hospital services, patients need to pay out of pocket, which is the gap between the the provider fee that is unregulated, and the repeat that is paid by the government. Also, for the illustrative services. The Medicare rebate is supplemented by private health insurance as well as out of out of pocket costs. installing it, there are a few regulatory policies and that's included as well. Removal of MBS item from this scheme. And in this paper in this project we're focusing on the removal of joint injection which is item 50124. We wanted to investigate the potential on intended on on expected behavioral changes that this remove all bring brought forth and this particularly relates to income compacity apparatuses as we know, that when physicians experience loss in income, they definitely will not fold their arms, they will statistically make some changes. Okay? The joint injection services before starting on 10 are charged in combination with attendance items by the physicians that are providing this service. And the MPs fee for this item is actually $23.25 per session in the year 2009. from Southern I think the joint injection service is not obvious, no more available for rebates, but physicians are willing to provide the service are allowed to continue to do so in the new environments. And this simply means that this decisions will lose out on the extra $23.25 cent that joint injection service fetches them. And this cost, this definitely, maybe may lead to pushing up costs to the patient, or to the government in terms of rebates, in terms of claims or claims from other items, will capture this set of physicians by the continued prescription of steroids for steel of zero to a patient that are affected by direct injection afterwards. And I will shed more light on that next few slides. So this policy constitutes only a very small fraction of the of the MBS that are provided by the affected physicians. For example, from the data that we're using, we can see that the composition is just about 2% of all the MBS items that are charged. And in comparison to the to the palm tree and consultation items, you can see that the joint injection itself is very, very, very, very low in terms of dollar value. So level B items in the 1009, costs $34.30 level see items cost $65.20 and level, the items cost $95.95. So we can see that relative to one consultation items, or attendance items is quite is quite low. So this costs a very small shock. But the question remains, will this small shot lead to any bigger changes? I want to quantify that change? That is the reason. That's the main reason for this research. The overarching objective of this policy is that the government is trying to make some savings from perceived low value k. And our research and question is to quantify the potential unintended consequences of this removal. There is actually literally no paper that has looked at complete removal of subsidy. In the next slide, I will show you a few papers from an abundance of literature that I've looked at different policy Medicare policies across the world, especially in the US and Australia, but nothing about subsidy removal. So basically, we're looking at unique contributions and this contribution is threefold. The first is to extend knowledge frontiers on physician physicians responses or behavioral responses in economics. The second is to use different components of our initiative, MBS data set for deeper insights. And the third would be to provide some evidence that could guide proactive policy designs relating to prospective return to prove perspective of impact shock on priority This is in Australia. Okay, so, a few policy tools that are being used across the world includes cost sharing, benefit cap restriction, which we investigated I ourselves at UTS benefit eligibility revolution, price transparency, and treasured for our benefits. Amongst others, well look at cup restrictions in one of our previous papers, and we have looked at treasured for benefits. This purpose actually looked at use reports econometric analysis across studies, and they're able to interpret the isolate as causal and causal impacts. Some studies also looked at supply side impact of private health insurance experiments. And all this sits within the DVR as a theory in economics. So mostly due to a symmetric information that is that is skewed towards providers or physicians that are not necessarily available to a patient or consumers. Most of the evidence emerged from the US and I'll just stick on just so I'm in the jeep Naxalite ch paper. The basis for this bigger impact is actually demand investment. professes because we know that physicians are able to induce demand. And also in the destruction on fire paper, there could be of coding strategy. And of coding simply means that physicians are able to charge for the highest level item when they're losing income from from assets are frozen aspects of the of the of the services we took apart from China. And we have our very own people from one from Australia as well, which looked at the impact of physician behaviour, your responses from threshold for eligibility for 45 age. By way of introduction, of course, and also on the policy, I'll just show you what happened to steroid prescription steroid I used, which during injection during transcription services. So we noticed that before the policy is was fairly stable, but afterwards, we find that there is a decline in the number of thyroid that has been prescribed and is the National representative datasets. Of course, this has not completely gone away. And that as an implication for searches simply means that there are some people that are still providing the joint injection, even after the policy took place in 2010. And that's why this policy for this project is valid. And complements what I just showed in the previous slide, we can see as well that attracts attract, this continues to be treated. So you can see the pattern for knee replacements and admissions in just two to four also national everything that was considered they still continue to go up without any decline despite the policy taken off in 2010. And simply simply means that they are still provide us for this service despite losing that aspect of income. So we explore the natural experiment that this policy provides to us. And the questions that we're asking is what is the impact of the removal of this subsidy on the bill in practice? For gap consultation items, we focus on three, and the three categories, which are the level be the short consultation, medium consultation, which is level C and long consultation, which is level D. And also we go to the other side, which is the payments button. All those this policy, subsidy removal, influence or impact fees that are charged by the physicians, the out of pocket costs paid by the IOM by the patient, and even the benefits claim patterns, but that the physicians put out to the government? And the supplementary question would be, what is the distribution of the policy impact according to the exposure level of GPs to the income shock. So we know that the contract is going to be very small in general, but it will be bigger for some GPs compared to other GPs in terms of the number of direct injection services that are provided before the policy took place in 2010. political context, so one MBS item affected by the subsidy removal in 2010, providing a unitary religious setting at that time, because up until that time, there were no policies that are directly affecting GP MBs services. Now, we'll know that joint injection services are provided by general practitioners, GPS and specialists. In this clip, I will focus just on the GPS to preserve homogeneous nature of positions in this category. And that's also absolutely trackability of our analysis. Australia GPS have extensive scope to change in terms of response to software reform of this kind due to fee for service payment system that will practice it. Can this change lead to billing practice changing as well, especially in relation to consultation atom, which is jointly charged with with the joint injection before we move on? Secondly, can this, can they respond to respond by adjusting the payment structure in terms of the fee that they charge on which component of this fee charge changes? Is it the out of pocket costs or the benefit payments? So how did you define the treatment and control groups? So we focus on joint injection for the as the policy item, and we use this to capture the treatment GPS, so we restrict our datasets to GPs  that are prescribing steroid drugs pre and post policy and the reason is because we can only track those who prescribe steroids after the policy took place in 2010. We're plugged in for joint injection as I said earlier, on Because this is a joint good, which, which was trilogy, so we're able to capture, those will still continue to provide a joint injection afterwards for our control groups will link our control group to three MDS items that are comparable to direct injection and that's diagnostic biopsy, I will use throw the items as for robustness of our results for the focus the main result come from DB diagnostic biopsy, due to the non mutually exclusive nature of the policy, we have different treatment GPS, because some GPS that provide joint injection also provide diagnostic biopsy. So we have in some cases exclusive treatment group where we only exclusively focus on joint injection and not diagnostic biopsy. Of course, we have the alternative one which would be the interactive treatment group where the Japanese are allowed to charge for joint injection on diagnostic basit and the third one is a combination of of the first two. So we look at these three then treatment categories separately show that our results are not necessarily sensitive to the kind of treatment group that we use. Not that control groups are completely isolated from only just GP services so did not interact with the with the joint injection, they do not interact with authentication, they only focus on diagnostic biopsy. Results remain the same, roughly roughly the same across the alternative comparison groups and treatment called control classifications. For this paper, in the analysis, we use for the five an O which is which constitute our 260 7000 community based sample of people live in in New South Wales. MDS data sets of focus is on 2008 2011. It includes our provider unique identifier, our specialty group, fee, charge out of pocket cost and benefit payment for each MDS claims. And we also needed the ABS data sets for the same for searight. In this case, because we will make reference to the date of thyroid prescription in our analysis. And so in that regard, and what we do is that we have we have to make reference to this because we use this to capture the time of timeline for loss in income that is experienced by by the by the physicians. So we use a very rigid time frame which is same day to the time of sera prescription. And we use very flexible one, which is three day to the type of steroid prescription. We used in this paper a difference in difference method to estimate causal impacts. And we also included provider fixed effects, so we can't program so you can show that it's within provide an impact that we're actually estimating, not between providers. So before I go to the to the results, we have a very small conceptual framework that were designed for this for this paper. And the summary of this framework is that we're investigating two conflicting side motives of of each physician on and we're covered up on the average. And for the so this is a sign utility function. And we have the the selfish parameter, which is one minus alpha and the altruistic parameter. So for the selfish parameter, we simply look at the price of the of the service provided to the to the patient minus the cost, which of course signals the profit across all the patients. And when we solve this together, we kill it by the selfish parameter, which is one minus alpha, I'll call it profit motive as well. So for the altruistic convicts, which is quite interesting, we know that some providers or some physicians as well, they're kind to the patient in terms of being very careful about the LT utility that is derived. And so the measure that alongside the price that each of the patients have to pay. And so when we remove the price the patient has to pay from the utility from the LT utility of each patient was killed us by the altruistic motive. And we saw this over all the patients that are attended to by the physician as killed by the altruistic. And so we combine these two together to show that we can indeed have more altruistic or less altruistic positions and that would definitely play a role in this In the parameter in the in the magnitude of the parameters, which should scale which is which is sum of one in this analysis, there are requirements for the two components of the of the provider utility is the fees charged by the, by the physician or by the provider, and the government benefits, which is representing the selfish responsive hypothesis. Whereas when we look at the out of pocket cost that is paid by each patient, of course, that should signal some average of altruistic response responses behavioral or physiological. And going forward from that we add to the size and some predictions. And we believe that to satisfy the income capacity hypothesis, physicians may want to change their billing from short to long term consultation items. Secondly, demand to show an increase in out of pocket costs and for a non autistic Jeepers. And indeed, heterogeneous. physicians will definitely heterogeneity official will definitely play a role in the out of pocket costs or even the fees that are charged or service across across board. I interference design can be showcased in this table, we use policy period, two years, eight or nine and the post policy period seven and 10 and 11. So, our treatment GPS, as I described earlier on, does refer to the joint injection with steroid prescription. And we have three categories for those and the control GPS for the diagnostic biopsy. So we have MBs claims pre policy. This includes the joint injection force policy excludes the joint injection for the treatment group, and we have the treatment differential for the control group, we have the MBs Cliff pre and post policy and we have the control differential. Of course, this excludes in both cases, the joint injection however treatment effects of different things. So our outcome variable various across our billion practice and payment patterns. Our treatment effects is the five different different this our regression specification will include patient and also level characteristics that we have from our 45 data sets. And also include on physician Level Indicators as well, just to note that our arrow downs are clustered at the postcode level to capture spatial correlation of Shan charging videos. And we have our first preliminary results on a one. So, on the left plate, we have our billing practices, and on the right layer, we have our payments from the physicians. So, we also have two panels for results the panel a captures the 3d MDS claims from the date of steroid prescription, while panel B is same day like I said the same day is the rigid time frame, which which is looking at responsibility on the same day that the last time took place. Whereas for the panel, it's a generous one which is looking at three days from that. So, focusing on panel a, we we do not serve for short consultation item, which is level big item, it simply means that we do not find any impacts of the policy on short consultation items or invariably level B charging for level B, when we move to the middle medium consultation item, we can say that our results is significant at 5%. And it simply shows that there is 1.14 percentage points increase in the likelihood to charge for the level c items after the policy took place for the treatment GPS compared to the control depicts therefore the long consultation our result is much smaller compared to the medium coalition which is cost level they are usually very small in terms of composition to level See, but we estimate a statistically significant impacts and that is just is 0.42 percentage points increase in likely to charge for level D items amongst all ordered items for the treatment cheapest compared to the control GPs. Okay, so when we move to the payments section, we also did not estimate any impact for our out of pocket costs variable However, for the fee charged by the providers, we can See that there is indeed a statistically significant impact, which is around 3.6% increase in the face charged for the treatment GPs after the policy compared to the policy proposal period, compared to the control group, and the same very similar results when we look at the benefits page, also statistically significant. Of course, when we compare these two together, they're much more bigger than the out of pocket costs. For the same day claims, of course, we have reduced numbers by n MDS claims in terms of observations, we have now around 88,000 observations compared to after five in the three days, but we can still see very similar pattern. So for example, level, the items, of course now increases by 0.8 percentage points compared to the proposal period, and compared to the control group, and she just got out 5%. And we can see that our provider fee has increased by 7%, as well, and which is very similar to the magnitude of impact estimated for the benefit page. So in very simplistic terms, what this result means is that, on the average, one additional level, the item for the treatment chip is relative to control group for 100 MBs items provided after the policy. And under the billing practice result I will have is that on the average, we have four additional level See, from 1000 MBs item provided after the policy, were asked for the payment estimates, it simply means that on the average, we have 4% increase in fees and benefits for treatment GPS relative to the control GPS after the policy. This results, like I said earlier on within provide a fixed estimator would provide our fixed effects, signaling signaling that we have within providers. So it seems it's also of the tone of the term within provider. So it's relative to what the providers were doing. Prior prior to the policy and not between providers. To estimate I heterogeneous results were compared for different quarters of the treatment and control groups. And what we did is that we looked at GPs that are more exposed in terms of composition of their services, in terms of the fraction that of services that they provided before the fraction of injection services provided before the policy took place in 2010. To focus on just selling online in this case. And we distributed this into four different quarters, which falcata for being the is composition wire cutter one is the least composition. We did the same thing for the control group is where we're focusing on Joe diagnostic biopsy for the distribution of cutouts. And in this case, we compare Kotel GPS, and GPS to cuddled up for the treatment under control groups. And we can see so we're focusing here just on level C level, the item in the billing practice results, where we do not find any statistically significant results. But we can see that RSI estimates increases with quarters as well, which seems . which simply means that this has increased the quota for those who are more exposed, have more responsive impacts in terms of changes to their billing patterns, but no statistical precision. Whereas when we look at provider fe results, we can see statistical precision for the quarter four segments, which is actually much bigger than the baseline results that we reported earlier. So our baseline results is just around 7%. Whereas our results in this case, is actually 16% twice more than twice the estimated impacts for the baseline results. And when we go to the benefit pay, we seem to find similar pattern, which we'll find in provider phase for whatever the result was a little bit not as big as what we have done, but is still on us almost twice the baseline result that we that we estimated and this simply suggests that the result actually driven by those who are exposed more to the policy shock compared to those who are exposed on less. So just to check that nothing was going on Prior to the introduction of the policy. We try the placebo I'm test and we focus on stolen Five, six, and seven, eight us placebo per policy and post policy. And in this case, we will also use the three day category and, and the one day. So we use the three day category, in both cases for bundling of our results and use of MBs items. And we can see that our estimates do not provide any statistically significant results across all all the estimates out present in this case, I can see that as they are actually very smaller in terms of the magnitude as well compared to the estimated results in the baseline. So the summary of our findings is that we can see that there are direct price mechanisms. And that's why estimates from the increase in average fee and benefits for the remainder items. So there is a 4% increase as a result of the policy. There are also greater impacts of GPUs that are relatively more exposed to the policy. When we look at the provider fee and benefits claims. We do not however, receive maximum impact for the out of pocket costs. And the indirect mechanism is of course from the billion practices. And that should that case, we show that we are 1.4 and 0.4 percentage points increasingly likely to build for level C, and level the gap attendances roughly the same results using alternative control groups that I that I highlighted earlier on, on different comparison items. So one implication of our results from the direct price mechanism is that we may say that it seems that the physicians are actually more altruistic in terms of the out of pocket costs, but they're also competing, selfish motives. So they're just the result of shows both cases that there can be both altruistic and selfish of GP responses in in, in response to response to remove out of subsidy. And what are the policy implication of this, of this our results, we show that providers do capitalize on reports of regulatory framework. We attribute that to the market power. They enable physicians to exert price or fee control. Our findings have broader implications on current or previous policy policy initiatives, which includes MPs phrase or Medicare review price transparency. In essence, there is greater need for policy design that takes into consideration potential provider responses. Further, we believe that fee for service payment structure is critically important barriers to achieve efficient healthcare delivery in Australia, and that is also causing unintended unintended impacts of that is true unintended impacts of payment reforms, which may affect also patient I am worth outcomes. So in general, our car projects were trying to investigate the effectiveness of value based systems, such as bundling of the payment structure, and the result that would provide may provide some useful insights in sought to tackle this problem altogether. We have some further research items that we have, that we intend to intend to accomplish in the next few weeks or months. The project aims to explore alternative empirical models, which should strengthen the policy interpretation of our results. And one of this is that we want to bundle the services for each provider from one day or three day timelines. And what we aim to do is that we hope that we'll be able to ask directly the research question that relates to not apparatuses been changing average fee being equal to a reduction of $3.25, which is the price per session of the of the poly of the subsidy before it was removed, and the alternative hypothesis of the average fee being greater than this reduction. And the new research design will support evidence on structural responsive behavioral responses. We also hope that at some point we would have access to even national representative data sets that is beyond New South Wales. As we know that 45 and up as that limitation, and provide representative volume of services, for research. So we have knowledge on funding from care nhmrc, CRE grant number 1171749. And we also acknowledge data custodians for this research, which is sucks institutes, and also acknowledge services Australia for providing the MPs and the PBS data sets. Thank you very much for joining us this afternoon. And I look forward to receiving feedback and comments on this paper. Thank you.

  Jane Hall  35:37
  Thank you. Thank you Olu. And I think for me, there are two things that come out of this presentation. And one is how, how challenging it is to get those robust estimates that really, we need quite complex research designs to be to have confidence in the sorts of results that you've been able to show us this afternoon. I think that was illustrated very well. The other I think, is that the health system is a very complex system of moving parts, and that if you push it in one part, something quite unintended might happen in another part. And I think this illustrates that very well. I'd like to open that for comments and questions. And again, perhaps if you stop screen sharing for a moment, I can see, I'm able to see a few more people. But if you could indicate in the chat or raise your hand, or by all means just just interrupt, and I'll do my best to moderate the discussion. But any comments or questions for all?

  Kees van Gool  36:54
  Perhaps we've got Bahare.

  Jane Hall  36:57
  Oh, thank you. Thank you. You may need to shout at me. Hello, please.

  Bahare  37:04
  Hello. Hi, Jane. And thanks. Thanks for giving the opportunity. Olu, that was a great presentation. Thank you, I really enjoyed your presentation. One comment or perhaps a question from, from my end, how to vote or confident we think that we can measure the patient's behavior in this study. And because I suspect part of this, it might come from a patient's demanding for these types of procedures. And I guess, part of that could be related to, for example, the level of the health frequency that are that are patient they have or potentially related to some of the factors that related to the suburbs, for example, like they access to other services like physiotherapy, or, like things like that, because we're basically looking at the behavior of the providers, but we're not looking at the behavior of the people that are demanding the service. And I think for me, it's always a challenge, because we might be at an out of pocket cost might be a way that we can measure it, because for example, if we say there is this much amount that patients they have to pay to get the service, and how much they might actually think twice when they go into the GPS, and they asking for such service.

  Unidentified speaker 38:29
  Thank you.

  Olu Abiona  38:30
  Thank you. Yes. So yes, we will go out and collect that jack would be an interaction between the provider behavior and patient behavior. And so one way like I said earlier on is that we could maybe sit down, try to walk across, but one thing we will also try to do now is to look at even the possibility for heterogeneous impact across the sort of patient level or socio economic group. So, this way, for example, if we distribute our results for low income households, and I income household, or rich and poor, also, we could, we will tease out the results more because, for example, if you find that the results are driven by just the rich segment of the patient is simply stem suggests that the physician seems to be only focusing on getting back the trying to compensate for the lost income from the rich also, and not from the poor. So we don't know. No apologies. I would definitely would take this further in what I believe my teammates were a

• Paying for health care: Problems and prospects

  13 May, Professor Jane Hall

  Play video

  Read transcript

  Transcript

  Kees van Gool 00:00
  I would also like to pay respect to the elders both past and present, acknowledge them as the traditional custodians of knowledge for this lapse. Just a few housekeeping notes. Before we get started. During the presentation, please put yourself on mute. We will have time for a q&a session straight after your presentation. But Jane is happy to take questions of clarification along the way. Please, if you have more substantive questions or comments, please leave before the q&a session. You can post questions or comments during the talk in the chat room. And we'll get to those in the q&a session. Or you can of course, just state your question during that at that time at the end. Please note that the session will be recorded and posted afterwards on our CRA website. So now I have the pleasure of introducing our speaker. Jane Hall is distinguished professor of health economics at BTS and 30 years ago, health economics research and evaluation. She was the centre's founding director and a role she carried out to 2012. Jane has made a lasting impact on the discipline of health economics with her work on patient preferences program and policy evaluation. She actively engages in a wide range of policy work including past and present contributions to M sack cancer, Australia, EPA, the Bureau of health information, local health networks, as well as numerous government inquiries and reviews into the funding and financing of health care. There are not many health economists in this country whose careers have not crossed paths with Jane and who have not benefited from the experience and dedication to building the research workforce capacity knows no bounds. And for that I will be eternally grateful. I'm delighted to hand over to Jane whose presentation on paying for health care. This presentation is titled paying for health care problems and prospects future.

  Jane Hall 01:52
  Thank you Kees for that rather overwhelming introduction. But thank you very much. I appreciate it. I too, would like to pay respects to the traditional owners of the lands on which we are all seated today. And remember them as the traditional custodians of knowledge. So what I want to do to pick up today, and this is very much a series because this will build on work that's been presented by my colleagues in the earlier part of the series is looking at problems and prospects around paying for healthcare. So next slide, please Nikita, you can see that this is an accommodation of my technological challenges. So what I thought I do in this in this talk is give you a short history of reform around payment, have a look at traditional methods of payment. Look at some of our recent experience with new ways of paying particularly with an emphasis on Value-Based Payment, and finish up with the challenges that we facing going forward. Next slide please. So this, you might recognise as a picture of the early settlement in Sydney Co. Now the first thing that happened when white settlers arrived in Sydney Cove was that they constructed a hospital out of out of tents. And it wasn't long before they were planning a more serious building. And that represents the the site of the first permanent hospital permanent if you can use that word hospital. So hospitals came first. And that stage, the doctors were all naval surgeons employed as salaried doctors. But it wasn't long before they'd got off those boats. And they started to petition the government for a change in the right to move to fee for service private practice. So I use this to say to you payment reform has always been with us. And we're unlikely to see an end to it. That for reasons that I'll explain, as I go on to the next slide, please. So if we think about reform milestones, I'm not going to go back before 1984 and the introduction of Medicare. But those early, early, at least in our lifetime, attempts at reform were very much around universal coverage, ensuring that everybody had some sort of financial coverage and had the right to healthcare. And the really big milestones in that, of course, were Medicare. And that has become such a valued part of the Australian social system. That it wasn't until 1996 with the private health insurance incentives that a Liberal government was returned. But those private health insurance incentives have stayed with us. And I do think it's rather than Interesting for those of you true will have seen this if you've trawl through the budget papers, that support for hospitals as you dig down through those papers, actually is an evaluation of the targeting of the private health insurance incentives. Next slide, please. So if that was sort of the reform efforts around universal coverage, we then moved on really to a much bigger emphasis on access to services. And we can see that there's been a lot of action there this century with a quite a large expansion of Medicare items in the early 2000s. Some of you may remember a strategy called a fairer Medicare, which was not necessarily a universally well-received. But out of that we got a whole lot of new medical schools to increase the number of doctors so that that would improve access. In 2007. Around the change of government, the policy focus moved very much to public hospitals. And Rod's mantra was to end the blame game between states and, and the Commonwealth. Of course, he set up the Health and Hospitals Reform Commission, which eventually led to the national health reform agreement with all states and territories, which was very much around access, but starting to be about efficiency. And of course, I can't resist noting that the 2014 budget, and did much of the nh era bile In fact, a lot of it's being put back. And again, I think one of the interesting things in this year's budget is that the government that did away with the national health prevention agency is now about to reinstate one. So from that was was really a focus about access to services, improving services. And we then move to the next slide and to talk about value. Now, if you go back to 1990, the national health strategy which was led by Brian how health minister in a Labour government, that was very much about getting better value for money, although it wasn't expressed as value-based purchasing or value-based care in those days. And it did come up with fairly strong recommendations around the purchase of provide a split, which never really went very far. In 1993, the Pharmaceutical Benefits Scheme, and I'm pleased to see Lloyd on this call, introduced the criterion of value for money. And that was shortly followed by the MBs. So we've been talking about value for money for a long time, it's not a new phenomenon. In 1993, ABF or casemix, funding was introduced, introduced in Victoria, worth noting that both that PBS value for money and casemix funding, we were the first country in the world to do that 2015, the NBS review around low-value care, again, very much about value for money. But by by 2020, the whole terminology of paying for value-based care has come to to the fore. And we see that enshrined in the current addendum to the national health reform agreement. Next slide. So there are various value-based approaches. The key to understanding value is this comparison of, of outcomes and cost. But if we go back in time, one of the most influential articles was Alan Enthoven in 1978. He didn't talk about value for money. But he did propose a consumer choice health plan, which was around competition to provide a more efficient healthcare market with the idea of what was valued, what would be at the the key of efficiency, being around consumer choice. And that was implemented in the Netherlands from 2006 on, so we have some experience with that. The next I think, seminal contribution to the area was Alan Williams article in 1985, where he looked at he was a UK health economist, looking at how should the National Health Service to allocate more funding, he took us His example, coronary artery bypass grafts. And he really laid the basis of the argument for the need for economic evaluation, to demonstrate value in where you put your resources, and the need for better measurement of outcomes. If you're going to make sense of that. So then another American, Michael Porter is very, whose book was published in 2006. very strongly now associated with the ideas of value based healthcare. And again, he's not that dissimilar to Enthoven, in terms of the idea that value-based payments should be driven by the market. And what we needed were more patient-centred institutions and measurement of outcomes. He was much more focused, though, on the providers of health care rather than the insurance arrangements around it. So if we look at that build-up of argument, what we can see is the notion of getting value for the money that we spend is not at all new, it has long historic forebears here. And that what we see really is that, if we're going to do that, we have to have a strong understanding of what is value and what we should pay for to get that value. But we also need that's interesting. We also need a strong emphasis on how to implement that. So we need knowledge and understanding. But we also need incentives. And those two pieces need to come together, if we're going to push value-based care. Now, Nikita, let's see if we can go forward rather than backward. Thank you. So thought, what I want to do now is have a look at our traditional payment problems, and the sort of incentives and problems that they throw up, and what experience we have of working with those, and how we've been ameliorating some of the problems. So if we start with global budgets at the institutional level, there is some fund holding and flexibility in how you use your funds, because you're sitting within a budget cap. The problem with that is that there's a strong incentive to limit activity to stay within budgets. And we've seen that happen with Australian public hospitals expenditures, was capped prospectively. And of course, the risk that that throws up is around access that because if you run out of funds, you you reduce your reduce your services. salaries are not unlike global budgets in that they don't really tell you how you should use your time but they're individual focused rather than institutional focused. So there isn't fund holding and flexibility. And again, the incentive is for the individual to limit effort rather than work longer or harder. If we look at fee for service, what we're paying for is services, obviously, but these are usually defined as as inputs to to the, to a treatment, often defined by schedule price. There isn't any fund holding there's very little flexibility at the incentive of courses to increase inputs. We see we have a lot of experience with fee for service in both in primary and specialist care in this country. I'm focusing on Australia, obviously, rather than looking at other countries. There is a financial risk that rests with the funder and the or the insurer because although they might define input prices, they're not capping the total budget. There can be a higher risk to patients because of the because of out of pocket payment costs if they're allowed, and they're not capped. And as we've seen in recent years, there actually is risk for providers. If there's a sudden and unexplained a ball. That's not the right word isn't if there's a sudden an unpredicted drop in demand. And I mean, we saw that with COVID and services ABF I've included in this table, because in my mind, it sits somewhere between our traditional payment methods. And the new ones that we're trying to move towards. It pays for a case treated the sum, fund holding some flexibility, there's a definite incentive to increase admissions, and to skimp on quality. And of course, we've seen that with Australian public hospitals over many years now. The volume risks still lies with the funder, but there is more financial risk for the provider, if they can't keep their average cost per case within the the the scheduled or determine price for those cases. So if we go to the next slide now, so given those given those problems and incentives, what can you do to ameliorate those? So for global budgets, what we saw was attempts to introduce shared shared savings, so that if people stayed below their global budget, they the institution was allowed to carry some of those savings forward and spend them in other ways. And we've seen a lot and still do see service targets and contracts around volume as a way of ameliorating the don't do too much incentives have a global budget. With fee for service. We've seen various strategies that have been add ons to fee for service payments, such as pay for performance, which was usually directed at quality, and more recently, limiting what in the input basket, which is the approach of the MBs review, looking at low-value care or ineffective care, with ABF. The way to manage some of the that has been cost-sharing between the Commonwealth and states. And I think, although there was a period of Commonwealth state reform, which was pushed by Why shouldn't the Commonwealth have everything all the states have everything? number of reasons why that wouldn't work. But one of them being if both the Commonwealth and the states have some financial exposure, it helps ameliorate that incentive to increase costs when they're shifted onto another party. We've seen the introduction of budget and volume caps in in more recent developments in the national health reform agreement. And of course, we're seeing a lot of work now. Moving towards rewarding quality, providing strong incentives for quality performance. So let's go to the next slide, NIKITA. So what do we know about pay for performance, we've got quite a lot of experience there in primary care. And I'm referring to a paper published in the BMJ A few years ago, where we looked at, we looked at income. So this was looking at primary care, yes, looking at a GPS income from sources other than fee for service. Around 5% of total GP income came from non-fee-for-service items. What we saw is that it was voluntary participation, it was very variable and very high turnover of practices, practices would come in, stay for a couple of years and then leave. Larger practices were more likely to participate. And our conclusion is that the more infrastructure you have, and the larger the practice, the more worthwhile it is to make the sorts of administrative and delivery changes that allow you to take advantage of those. So a very strong message about small payments don't don't result in large changes. Let's go on. If we look at activity-based funding, I think there are many lessons to be learned from the introduction of activity-based funding as a national approach to funding hospitals. The National Health Reform agreement was signed up by Commonwealth state and territory so all the jurisdictions came together with With support for activity-based funding, there had been experience in many jurisdictions with using casemix Arabia funding to a greater or lesser degree. But there was a lot of experience in working with the sorts of infrastructure that you need to make a system like that work. Part of that was very standardised cost collection. So there was a lot of a lot of data to work with, in terms of starting to, to manage what a nationally efficient price would be. The the price-setting authority was independent of all of those governments. And I should acknowledge that I have been a member of EPA since the beginning. And we worked on the basis that if they weren't all unhappy with us all of the time, we were probably doing a reasonable job. But a lot of reliance on evidence on the data. And as I say, we had the standardised costs and activity collections going back years, a lot of transparency about the methods by which costs are set, and very wide consultation around any changes that are made. What we saw were price increases, which had been quite high in those pre-2010 years stabilised, but volumes increased, as you would expect from the incentives in the system. We saw, we are seeing increasingly now new developments around quality and how to provide much stronger incentives for the improvement of quality. And of course, it has now embarked on exploring new funding models, although in the analysis we've done so far that will be for a minority group of patients ABF will still work for the vast majority of patients. Of course, all I haven't got on that slide is what was announced in the budget two nights ago, in that it is to be renamed and take on the funding of residential aged care as well. But that's a completely different topic. And I just take a deep breath every time I think about this. Let's go on the cater. Next slide, please. So why are we asking for health reform now. You've heard all this before better coordination outcomes that matter to patients diversity and innovation, stronger evidence, improve equity, deliver best practice care. And these are not new things, you go back to 1990 in the national health strategy that came that was also the impetus and the rationale for why we needed a national health strategy to look at all of the funding and organisational arrangements. So these themes are not new, they keep coming back to us. And they're back. They're back with us again. Next slide. So if we move on to Value-Based Payment. And what I want to do is take the approaches that are being promoted as better value-based payments, and look at them through that same lens that I've just been through with the more traditional methods of payment, look at what are they pay for? What are the incentives? What's the experience? What do we know about them? What we're looking for, of course, is an improved performance is driven by better outcomes. What I mean is, if you get better outcomes, you can see that as and and recognise that as improved performance. Care is coordinated around the person and across sectors, more flexible use of inputs, and to get the incentives right providers have to be to be rewarded for improved performance. Let's go on. So now you can see I've got ABF here again, and that's because to me, it sits somewhere between all of our traditional methods of funding and where we're moving to new methods of funding. So we've pretty much covered ABF cases treated we've looked at the incentives we've looked at the experience, except I've added there the risk of unintended consequences on hospital prevention. I don't know how big that risk is. But certainly many people criticise ABF for not being able to address the sorts of programs that would prevent hospital admissions. Now, that's only true to some extent, because actually ABF, as we implemented in this country at the moment, does allow for hospital prevention programs. So but however, the the three and again, they're not really new methods of payment, but the the three popular approaches to improving value-based care around episode-based funding for a capitation around chronic conditions and population capitation. So the episode of care really needs it, what you're paying for is an episode of treatment, it has to have a beginning and it has to have an end. It does, it does definitely allow fund holding. And it provides quite a deal of flexibility within that envelope. The problem is, again, that you would increase episodes and see skimping on quality. Now what we have is a reasonable amount, and certainly growing a body of evidence around internationally around episode-based funding. And this has been focused on orthopedic cases, because they, like a joint replacement has a clear beginning and an end. But also some cancer, but it tends to be cancer treatments, rather than the whole cancer treatment from diagnosis through to end, what we see with the evidence, there's certainly support for lower costs and better outcomes for patients. But it's not unequivocal. And there are a lot of shortcomings in the evaluations that have been done. If we look at capitation, for chronic conditions, what you're paying for is a person year, basically usually a person year for a chronic condition, but that might be monthly or less. Again, there's a fun holding aspect, there's flexibility in how that's approached, the problems are that there will cream skimming, but that providers will just choose the the cases that the easier to manage, and that there will be an increase in referrals. Because the chronic condition is an envelope, it doesn't cover all care for the person, it only covers care related to the chronic condition. And again, if things get too difficult, there's an incentive to refer on. Now we've seen that in Australian primary care, most recently with the health care homes model. And what I think it's fair to say from that, and others here may well have comments on this, I think it's fair to say it's it's quite a challenge to implement. I don't think there's a lot of resistance really from patients provided they have an ongoing primary care practitioner. But I think it's proved a lot more difficult for the practices that the need many of them had considered in the beginning. If we look at population capitation, the idea here is that there is a payment which will cover care, comprehensively all aspects of care. So very strong in terms of fund holding very strong in terms of flexibility. But of course, a a an incentive to limit activity, so that you stay within that that budget envelope or indeed are able to profit from that budget envelope. And again, there's actually a lot of international experience here. In the US, the Health Maintenance Organisations go back to, I think, to the 1960s and accountable care organisations more recently, but again, with this idea of appropriate population focus, a comprehensive budget. We've also seen GP fault fund holding particularly in the UK, which again, an attempt to have this population capitation. Now Health Maintenance Organisations I've always thought were rather interesting because they were seized on by a lot of enthusiasts who said well, if they're Health Maintenance Organisations, they must be doing a lot of prevention and That's not quite true, because unless the patient is trapped with the HMO for life, the incentive is to do short, certainly short term programs, but not to invest in those early early antecedents of disease, because your subscriber could well have moved on to a different insurer, by the time that the benefit comes around. Let's go on to the next slide. So, so we've got quite a lot of experience with all of these methods of payment, maybe not as much as we'd like. But we certainly have experienced, we're not just plunging into the dark here. So ABF have talked about that. We don't need to do that, again, with episode-based funding, the way that funders have introduced strategies to mitigate against the unintended consequences of those incentives is a very strong emphasis on quality monitoring, and a very strong reliance on the guideline, evidence-based guidelines for clinical care, to sit tariffs. And it's interesting in that you're you start off this journey by saying we need more flexibility and more innovation or how people are kept cared for. But then you end up with, but we've got to have guidelines, we've got to have some sort of standardised approach to care. So with chronic conditions, one of the ways to deal with the cream skimming, obviously, the way you deal with that is sophisticated risk adjustment models, which I apologise for not uploading in that table. But you also want to, if you want to provide a disincentive to increasing referrals, the budget needs to include referrals. But of course, once you do that, the the ability to control the consequential costs has moved away from the fund holder to the person to the provider who receives the referral. And again with population capitation. One of the big ways to deal with the the incentive to limit activity or to just pick selections of the population is to make it a geographically based rather than than voluntary. And maybe I won't go into the zombie of manage competition, but it may come up in the discussion and to set service standards. So again, this flexibility this, this aim for flexibility is being counterbalanced by a reliance on standardised approaches to ensure quality and act against skimping. Thanks, NIKITA, next slide. So let me say a little bit about Australian experience in trying to do some sort of fund holding. The coordinated care trials began in the mid-1990s. And it was notional funds, pooling, primary care, emphasis, lots of emphasis on coordination, it was going to it was something that we see in health care reform time and time again, something that seems like a good idea, but turns out to be really difficult to implement. And what we saw with the Coordinated Care trials are some promising results. A lot of unmet need a lot of expense that hadn't been predicted. And most of the funds holders, if they'd really been fat holding funds would have gone broke. Integrated Care has had a lot of emphasis on hospital avoidance and cost reduction. So we've seen lots of integrated care programs with the same sort of thing, same sorts of results about potentially promising outcomes, but not being able to achieve cost reduction. And of course, we're seeing it again now with with healthcare homes, the same sort of emphasis of some sort of bundle payment emphasis on flexibility team care. So there's a there's some really common themes running through all of these attempts to implement this. And what we see common problems around voluntary participation. So it waxes and wanes, people drop out. There's a need for complex administrative changes and a reluctance to to embark on those changes, particularly I think a lot of the Australian experience has been, it's tried for a year or two, and then it disappears and the funding for it disappears. So there has to be a lot of confidence in the sustained in sustaining the approach to make it worthwhile to make the changes. And of course, people need development and training and how to manage bundles rather than just having them arrive. Let's go to the next slide. So what do I think we need to do in terms of alternate models, EPA is, as I said, is working on future funding models. And there'll be a consultation document coming out in the next two to three months. From an analysis of hospital data, what we see is that ABF will cover most admissions, because they're pretty much one-offs. There are other the other approaches have some form of capitation, or some form of bundled payment will be suited to a small number of patients. But although it's a small number of patients, they're typically high-cost patients. So we're going to be looking at a funding landscape that uses different sorts of approaches for different problems. Now, the upper approach, of course, is very focused on hospitals, because that's our remit. And that's our data. So that's where we've been looking. And much care doesn't require hospitals. So I think there are far more challenges and thinking about primary care in thinking about mental health. And thinking, while I should say I'm thinking about a residential aged care, net, thanks, government, but certainly the interaction with aged care, the interaction between what is his health care and what is aged care. Indeed, that's not dissimilar to the interaction between health care and disability support. And let's go on, because I think I've just about finished, right. So the challenges as I see it to any further moves to Value-Based Payment, we, we need to have an agency that is a fund holder. And that's going to mean agencies that have expertise in coordination, that are able to develop and sustain relationships with other providers, because it's not going to be just one provider that that manages the full gamut of care. And they have to be able to sustain the sort of risk that they will expose to by moving to any sort of fund holding. If we're going to have outcomes that matter to people, we're going to be looking at more measurement, not less, we're going to need organisations that are centred around patients rather than at the moment our organisation's are very much determined by the funding flows rather than where the patient's go. And we're going to have to have strong purchasing and commissioning arrangements. Because we want to, although we want more flexibility, as I've said, over the last few minutes, although we want more flexibility, we're also seeking evidence-based care. So in my view, we're going to be looking at a lot more measurement rather than less measurement. And of course, we need these agencies to be able to manage risk, we're going to have to be able to set a fair price for the fund holder, we're going to need to adjust risk fairly. But also efficiently, excuse me. And we're going to have to have a lot of consideration about upside and downside risk. Who who's going to gain and who's going to lose another game to be limited to the gains and losses? Next slide. Thanks, NIKITA. So how do we change our funding structure? the sorts of groups we could work with as fund holders would be lhds and hospitals, but they're going to remain very hospital focused Primary Health networks but they've had a limited role in this so far. general practice, but many general practices are going to be very small and perhaps unable to sustain those risks, private private sector providers, potentially or private insurers, which of course is the basis of the managed competition proposal. But whatever we do, what we're going to be doing is redistributing around the health system risks, and also incomes. And as soon as you start to redistribute incomes, you set up a very strong incentive for inertia and a resistance to change. So let's go on because I think this is the last slide. So what are the prospects for payment reform from here? It's going to be challenging, we have to avoid having incentives that are too small compared to the change that's needed to successfully implement new schemes. I think if we look at where there's been successful implement implementation, and I would draw particularly on the Netherlands and I draw on our experience of ABA funding here. In both of those, there's been significant investment over time before you push the implementation onto the whole system. So enormous investment both in in data and and understanding, but that also develops human capital and capabilities to manage in a new system. I think that data collection requirements will increase skill and expertise requirements will increase, we're going to need a major change in funding structures, we can't work with the ones we have now. And I think something that I haven't mentioned so far, but it's a good point to finish on, is that we have a real challenge in the idea of universal health care about how we balance universality, and how with flexibility. So universality can has been so far in our system, interpreted as the same benefit for everyone. So the same NBS schedule item applies t

• Our experience in translating the international literature on bundled payments to Australia

  9 February, Dr. Philip Haywood

  Play video

  Read transcript

  Transcript

  KEES VAN GOOL 0:00

  That is leading the Centre for Research Excellence in value-based payments in cancer care. And as part of that, we hold a webinar series, which we try and do once a month we started at the end of last year. And we will proceed for 2021 and, and beyond. So thank you very much for joining us.

  And to start with, I'd like to acknowledge that I'm hosting this webinar from the lands of the Gadigal people of the year. I also acknowledge the traditional custodians of the various lands on which you will work today and the Aboriginal and Torres Strait Islander people participating in this webinar.

  Just a few housekeeping notes before we start, we'll have time for a Q&A session straight after the presentation. You can post questions or comments as they pop up in your mind during the during the talk in the chat room. But then off Also, you can obviously also just raise your hand or raise your voice in during the Q&A session. Please note that we are recording the session and we aim to post that shortly on our CRE website, which is coming very soon.

  Now, today's seminar will be hosted by Philip Haywood based on joint work with Lutfun Hossain, so it's my pleasure to introduce them both. Lutfun is a research fellow who joined CHERE in 2019 and she's been a key member of the of the of the projects working on the healthy homes evaluation for sustainability Victoria, the medical safety net project from Department of Health, and in her little spare time, the observatory work as part of the care insurance background is in pharmacy and previously worked as a clinical product specialist at NPS MedicineWise. 

  Phil Hayward is a senior lecturer at the Centre for Health Economics Research and Evaluation and is the research lead for the CRE on Value-Based Payments in Cancer Care. His unique background combines clinical economics, policy and academia with extensive cross-disciplinary experience in each of these fields. He's worked for over a decade in the emergency department at Newcastle's Calvary Mater Hospital and has also worked in New South Wales Health, the OECD and of course, with us at UTS. 

  Throughout much of his career, his focus has been on producing economic evidence to inform health policy. And today's seminar is no exception, as he delves into the ...oops..., as he delves into the literature on bundled payment performance oncology, and looks to draw key lessons for Australian Research. 

  DR PHILIP HAYWOOD 2:50

  Thank you, guys. 

  So I'll do the obligatory share screen and then check. It's working. So So first, have I shared the screen and is it in presentation mode?

  UNKNOWN SPEAKER 3:01

  Yes, good. Okay

  DR PHILIP HAYWOOD 3:03

  I'll advance the one slide and check that that's working. 

  Yep. And we on to the next slide. Yeah, yep. Perfect. Okay, so as Kees said, this seminar is based on work conducted by the CRE team, and lots of them myself will be here to hopefully answer the questions that you have. Um, this is this is an outline of our approach to researching the design and implementation of payment reform and oncology in Australia. 

  And the work that we're going to be presenting in this seminar is focused at the very earliest step, establishing the evidence and estimating the impact of changes in payment reform for oncology in Australia.

  There are four main components to this talk, a quick overview of bundled payments and outline of the oncology bundled payment schemes around the world. A synopsis of the evidence for impact of oncology bundled payment schemes, and then some discussion about what we think these results mean for Australia. 

  The reason why we're investigating the impact of different payment mechanisms is that we want to make it easier to deliver optimal high care, high-value care, payment methods, creating centres, and ideally, the incentives including the financial ones of the payer, be that the government or an insurer, the provider, whether it be a doctor, a pharmacist, a nurse or someone from allied health, and the patient are aligned, often they're not. And this may be an impediment to achieving optimal care. 

  There are lots of different payment mechanisms. And we've put a bit of a list together and they all come with their own strengths and weaknesses. One payment mechanism mechanism that we're particularly interested in for our work in oncology is bundled payments. 

  So what are bundled payments? I’m looking at this list, I know half of you know this, that that we might as well go through it just in case there's someone who doesn't know it. 

  Bundled payments of when one payment is made for a bundle of services for a specific episode of care. 

  And maybe one of the easiest ones to think about. And one of the most common bundled payment schemes is joint replacement, where instead of multiple practitioners being paid for each service, they deliver the operation, the anaesthetic, the post-op care, the pre-op care, the rehab, a single payment is made, that we still have to decide what is in the bundle. That is, is it just for the hospital stay is it for the pre-hospital stay as well, the post-hospital stay, and even in some cases and some Northern European countries, so multi-year warranties on the joint and what is outside the bundle and therefore paid for separately from the bundle. 

  One of the big impacts of moving to bundled payments from a system of fee for service where each service is paid for separately is that the provider is no longer financially rewarded for all the care that they deliver for simply doing things. 

  In a bundled payment, increasing the amount of ineffective or low-value care will reduce the financial return. They can also reap or financial reward if they're more efficient, do more with less or engage in preventable care, preventive care. And they also can be more flexible in how they go about things. They're not limited to the inputs being those things that are on an MBS schedule, they can choose to use different resources and combine them in different ways. But in allowing this flexibility, the provider is assuming a degree of financial risk. 

  If they can't manage to provide care within the size of the bundle, they will actually lose money to reward high-quality care and ensure that the incentive isn't to reduce care to a minimum, it's usually some measure of quality, and this might be tied to the payment. 

  So the variations of bundled payments already exist in Australia. And one of the easiest ways to see this as we're comparing public and private hospital patients. Public financing has the characteristics of bundled payments, where our payment was made for an entire episode of care, and private has characteristics more closely associated with fee for service where each each intervention can be itemised, and each provider paid separately. 

  So what do you have instant success for bundled payments. Rather conveniently in the last three years, there have been multiple reviews and evaluations of bundled payment schemes. And these in general have concluded that bundled payment schemes are promising without their benefit being conclusively proven. 

  And in general, there is more evidence suggesting that they may be effective and containing costs, then improving quality. 

  As most reviews acknowledge the strength of these conclusions as limited by the weak study design often being observational, and the lack of description of key features or some bundled payment schemes, which makes it difficult to be clear about what actually happened or occurred. 

  There are a few studies of bundled payments and oncology included in these reviews if there were three bundled payment schemes and oncology included within these reviews. 

  So we had some aims, we wanted to enumerate and describe the key design elements of bundled payment schemes that have been used in oncology. We were also interested in the evidence they have existed for bundled payment schemes and oncology's effectiveness. 

  And we were interested in whether the results of schemes could be transferred to Australia or translated to Australia. So we took a different approach for each of our aims. 

  To try and understand what underpayments schemes were out there and their design characteristics. We searched the peer-review literature, the grey, the grey literature, and eventually Google. Once we found a scheme, we tried to collect all the information that we could find out about it, including blogs, and newsletters, and interviews. 

  And then we've classified those schemes by the key design elements that we could identify. For the infinite estimation of impact. We restricted the studies to peer-reviewed studies that were either experimental or quasi-experimental design. So we used a much more restricted range of studies than most reviews for this step. We did this because for most things, but especially within colleges, things are changing all the time. This changes in technologies, this changes in costs, improving outcomes over time, and we thought that simple observational studies would be misleading. 

  So there had to be a comparator or an adjustment for these trends over time. 

  We did not expect to be able to conduct a meta-analysis because of the differences and measurements of the outcomes so we looked at a priority starting narrative synthesis. 

  We use the Cochrane effective practice and organisation of care tool to evaluate the quality and we use two factors which we're looking at myself and then the key to break the tie. 

  And, as opposed to a more traditional literature review. We've been repeating this process every six months. So we started a year ago did it six months ago, and then we've done it relatively recently again, to assist transferability we extracted information about the sittings and computer to Australia and made a judgment. 

  So this isn't a complete list of the design features we were interested in. But it's just to give a flavour of the potential choices that had to be made and designing a bundle scheme under payment scheme. And the information we attract attempted to extract, I'm just going to highlight a couple of them, because they they turn up later on is important. So the first is the issue of voluntary or not. 

  So is there a choice and entering into a bundled payment scheme? Or was it compulsory for everyone who was engaging in treatment and oncology for that particular bundle? And the second is how is the bundled payment scheme paid for? How is it actually implemented with regard to payment? 

  Is it a payment upfront, given to the to the bundle holder prospective payments, or, alternatively to payments continue as they were in a fee for service type arrangement, and there's a reconciliation with the provider gaining if the if the total payments for less than expected or having to make up a shortfall, and that's often called a retrospective payment. 

  So we found nine bundled payment schemes in oncology, eight of which were in America, and one of which was in Taiwan. five of these schemes had published evaluations for were in the US and one was in Taiwan. Four did not have peer-reviewed publications, and they were all set in the US. 

  Well go through the schemes, with publications and peer-reviewed literature associated with them first. 

  So the first is the ACM, the oncology care model, which was a five-year model launched in 2016. for Medicare and Medicaid patients, plus some private commercial insurance agencies and the aelius. participation was voluntary, and they were six months triggered by administration of chemotherapy. 

  If the total care during the episode was less than a historic benchmark amount, then the practice could qualifies for some performance-based payments. And the benchmark pipes were the benchmarks were extensively adjusted based on patient characteristics, cancer characteristics, and the area in previous trends of payments for the practice, the amount the practice could actually receive as part of these payments depended on what risk they were willing to adopt with if they were willing to adopt a greater level of financial risk, and they could receive potentially a larger payment. 

  And with the quality goals when quality outcomes include things like avoidance of ED, as well as guideline uses of antiemetics and, and other supportive care. 

  It was also required that there was a series of access provisions, they had to be 24 seven access to a clinician real-time access to practice records in a patient navigation service. 

  And additional payment was actually able to be billed for these services that are $160 a month. 

  So there are actually two financial interventions occurring at the same time, there was a change in the payment schedule, and there was a bundled payment. 

  This this bundle payment scheme was large, involving 10s of 1000s of patients a year and hundreds of practices. 

  The 21st century Humana radiation therapy model was at the other end of the scale, it's a one-year model for external beam radiation therapy only. 

  It was there were 13 different bundles designed for 13 different cancers, and cancer type a sorry cancer stages. 

  The model was the 90-day episodes, which started the first consultation, a single upfront payment was made, and each each bundle had a different payment, but there was no adjustment for quality and then other treatment expenses such as medications of arteries and other diagnostic imaging where were excluded. 

  The third model is the health key the United Healthcare model, which was an episode-based model between five large oncology medical groups and a single insurance payer. 

  In this model, a single episode-based payment was made to the oncologist on the initial visit, and drugs were paid using the average sales sales price in the US. 

  A standard payment was made for each of 19 different clinical presentations and breast colon and lung cancer. 

  The medical groups are free to choose the regimes the drug regimes and they were free to change them, but the episode payment for the oncologist would not change. All the other physician services were paid fee for service and payment started when the physician registered patient at the initial consult. 

  There were different lengths of time for different different bundles with edge event bundles been given for a defined period of time. And then they added 60 days to the scheduled end of the regime. 

  For patients with metastatic relapse cancer. They used an episode length of four months that could be rolled over again. There were 16 measures of quality in the model and physicians could increase their episode payments by increasing their improving their results. Either either increasing survival or decreasing the total cost of care. The fourth model was also a one-year model involving united healthcare. And that was the MD Anderson model, which was the head and neck cancer. And this this model includes a prospective payment at the start of the primary cancer treatment. 

  And it was to pay for services associated with the MD Anderson organisation. It was based on a retrospective financial analysis and a cost distribution of the patients with head and neck cancer, they produce for risk-adjusted bundles. 

  And these were on the basis of the treatment that was expected to be received surgery, chemo, radiation and also on the patient as to how many comorbidities they had as measured by the Charleston comorbidity index. 

  There were no quality metrics, and there were no changes for quality adjustment in the bundle. 

  And the non-American one was the pilot breast cancer multi-year scheme pay for performance scheme and Taiwan. With the national health insurance administration, reimburse hospitals have a sufficient size and quality with a bundled payment for a full cycle of care based for breast cancer. Rather than a fee for service arrangement which occurred beforehand. The full cycle of care varied, depending on the stage of the cancer between nine months and five years. 

  There were bonus payments to the institution at the end of each year, and partly on quality process measures and partly on survival outcomes and disease-free outcomes, which are standard set by the National Health Insurance Agency. So you can see what these five bundled payment models is considerable variation in the design of the model, their size, ranging from 10s of 1000s of patients a year in the ocm to less than 100 patients in the United Healthcare MD Anderson model. 

  There's differences in what services were included, how physicians were paid, how quality metrics were included, and whether the episode payments were adjusted on quality. And for some of them details of how the risk adjustment occurred and how the payments were altered over time occurred, they were of these five, only the ocm has comprehensive documentation. 

  The rest of them have more limited documentation, especially how the risk adjustment occurred. None of the schemes were compulsory, and the providers either had to make a decision to participate, or there was an end date after the a year and the potential to withdraw. 

  These are the small number of schemes that we identified with no supporting peer-reviewed literature. And not surprisingly, there's less information about each of these schemes. They they're all in America and they appear to be smaller. 

  So the Blue Cross of California was a bundled payment scheme based around breast cancer. 

  And it was about radiotherapy serve radiation therapy services provided to patients with early breast cancer. The Blue Cross and Blue Shield was a surgical bundle based around prostate cancer. 

  The Cancer Treatment Centers of America bundle appeared to be a bundle that was available to all all payers and is essentially a care planning bundle. So for prostate, breast, lung and colorectal cancers, and it's included seeing surgical medical radiation, oncologist, consultants, imaging and pathology services and consultation services. And at the end of the bundle, a a plan for care was produced that the patient could either stay with the Cancer Treatment Centers of America, or take that scale, take that plan somewhere else. 

  And then the horizon Blue Shield was another breast cancer payment scheme. It is unclear for any of the scans Were there any quality mystery matrix or risk adjustments in the bundled payment scheme. We just couldn't recover that information from the literature that we were that we were assessing. There were also two bundled payment programs which are in development and were proposed to be rolled out this year in 2021, but had been delayed until 2022, partly because of COVID. 

  So the first one the oncology kiya OCF is an extension to the ocm. But there's substantial alterations into into how the scheme is expected to run. And the first is rather than being a retrospective model like the ICM was changing it to a prospective model is also going to include a slightly larger range of services including pathology and imaging. And a number of changes have been made to cater for novel therapies and changes in technology. 

  The radiation oncology model, the proposed radiation oncology model is the only mandatory model we found it will be prospective And much like a number of the other radiation therapy funded payment models. It will be a 90-day episode for 16 cancer types. The as well as being compulsory, there's going to be a holdout of some of the money a small percentage of the bundle, the bundle that participating practices will be able to access this they make quality measures, especially around clinical data reporting and patient experience. So in total, we identified 1111 bundled payment schemes in oncology that range from quite small to enormous that range from just one modality of treatment. And one type until and to most services. 

  Now we're going to move to impact. And there's far less information about impact. Now, remember, we restricted what we considered a paper that dealt with the effectiveness of a bundled payment model to those which had a quasi-experimental or experimental structure. So we recovered five evaluations, and of those 453 recorded than last year. So there's an increasing amount of information there for were on the OSI model, and one was on the UnitedHealthcare model of the five for contained informations on costs and for contain information on some aspect of quality. Between the studies, there were different measurements of costs and quality. 

  So they can't be directly compared or summed up in a major analysis was what we expected. So no study showed a worsening of episode costs, there was either no change or an improvement, but there was an enormous variation in results. So the United Healthcare model reported a 34% reduction in payments associated with bundled payment, which which is very large, the ocm. The evaluations of ACM varied. So, in some of them, they did not measure the total costs, rather measured a subset of costs. So hospital costs and readmissions, and they in that one, they found substantial improvements that could be that could result from a reduction in an anticipated hospitalisations. 

  Another ocm evaluation compared to large practices, and they found overall that there was not statistically significant reduction in costs for all cancer types. But for some cancer types, there was in my specifically, there was a 30 30% reduction in the cost associated with prostate cancer, the evaluation of the entire ocm found and non-significant reduction in episode costs. So a small reduction in episode costs a couple of $100, which did not offset the participation payments for increased access so that $160 monthly payment outweighed the reduction that was not significant and costs associated with ACM and that evaluation. 

  So for those for that evaluation, the conclusion was adopting the ACM bundled payment cost money. 

  There was also significant heterogeneity, as we've discussed between the cancers and the types of treatment in general prostate cancer, or more intensive treatment was more likely to be associated with greater savings. In general, there was no improvement in quality, no message of improvement and quality over the four, the four evaluations that assess quality, but two of them did find a reduction in AD visits and unanticipated hospitalisations. 

  But that was that did not occur in all of them. Our conclusion is that there is not convincing evidence than bad or payments, save costs and improve quality and oncology. But this is because there's a limited availability of evidence at the moment. So as expected, we've got a small number of schemes evaluated in the literature. So only two of the nine schemes had an evaluation that was considered high quality enough to be included in this review. As we said, all the schemes are voluntary. So there's a substantial selection bias associated with this. 

  So that is, even though they tend to to adjust for it by comparing them to like practices, you still would be concerned that those practices that are participating in these in the bundled payment schemes feel that they would gain some benefit from them. 

  There was limited to no discussion about the potential of whether moving to bundled payments increased or decreased the amount of the amount of people who are actually receiving treatment under the different schemes. There was some else in the eye of the OSI model that suggested that the there didn't appear to be a change in the mix of patients who were attending for chemotherapy. 

  And that may give some reassurance that there wasn't any skimping of patients. All of them have been taken up the bundled payments intervention to be taken alongside other interventions either increased adherence to guidelines or payments for increasing Information Technology and care coordination. There's an enormous variation in the definitions of outcomes, different people are defining costs. And that there is also the majority of the cost outcomes are associated with the payments from from the from the payer, rather than the costs that may be experienced. 

  And as you would expect, because there's only a small number of schemes that can be included in that the impact of major design features cannot be estimated from these evaluations. 

  So what does that mean about translating the evidence of the impact to Australia, 

  in general, the results aren't directly translatable to Australia, the majority of bundled payments that were identified were in the United States sitting, most of them involving payers with comprehensive insurance. In contrast, in Australia, we have many different payers for different components here for any single patient. And so without the patient identifier, we may be limited in their ability to track patients through the system. Furthermore, physicians in the US have their income potentially directly linked to the supply of chemotherapy. 

  That is that part of the payment they receive as a percentage markup on chemotherapy. 

  And it's quite different in Australia, where most physician income is independent of any drug cost, which is the impact of the PBS. This is an important difference, because it means that the changes in costs and the changes and chemotherapy associated with the implementation of bundled payments in US may not have the same results in Australia. 

  There is also a lot of missing information about the bundled payment schemes, except for the ocm, which can make it difficult to ascertain which components of the model contribute to the success or the failure of schemes. However, I think there are a number of lessons that we can learn from, from the International literature for Australia. So the first is that there is definitely a demand for bandel payment models in oncology. There's an increasing amount of literature and is increasing rollout of large schemes, and high-quality evaluations of them. 

  I mean, to a certain extent, this is a derived demand, I don't think anyone's interested in bundled payments per se, interested in the changes that under payments produce. The other thing is that they are feasible and acceptable to develop and implement. And they actually can be relatively simple, occurring between a single-payer and a single provider group without complex risk mitigation strategies. 

  And that's important to know, because, as we've seen, those bundled payments, which tend to make those bundle payment schemes, which tend to make it into literature tend to be the larger ones with more complex risk mitigation strategies. And it's worthwhile knowing that it's acceptable to design something less complex. The other thing is that the international literature asks, allows us to ask some good questions about how we might implement bundled payments in Australia. So the first one is that it is what is the problem that the bundled payment is attempting to solve? 

  In America, a lot of it was related to the to the incentives associated with fee for service was thought to be driving low-value care and over-servicing in some situations. And there was also a concern that the that the financing of chemotherapy with a percentage markup going to the physician was driving the use of unnecessarily expensive chemotherapy agents. 

  In Australia, we might think that we're trying to solve a different set of problems, be they low-value care or lack of coordination or poor set of incentives. The other thing that I think we can get from the International literature as what are the possible key design elements, as you've seen, there's enormous variety in the potential bundled payment schemes about how these elements may be designed, how these elements might be implemented. 

  Now, we don't know which aspects of design are important in Australia, although we can take some guesses. And we will, we'll need to develop evidence to assist that. The other thing is that most bundled payment schemes are introduced alongside other initiatives in the guideline adherence, patient navigation, increased reporting requirements. 

  These can place an administrative burden on the physician or the practice or the hospital. 

  And as in the ocm. It may be that it's required that these be compensated for All the bonds of bundled payment programs we explored with voluntary participation and practices could withdraw. This means that the results could vary in Australia, if programs were made compulsory, increased risk mitigation was likely to be needed for compulsory schemes, as opposed to voluntary schemes, which has the risk mitigation strategy is stuffing the scheme? 

  I think one key issue that we have in Australia is what is the ultimate goal of introduction of bundled payment schemes? Is it to allow variable local uptake? Or is it expected to move to national implementation. And I think depending on which of those is likely to be the end goal would result in very different ways of approaching it. 

  Now, what can we do and by we, I mean, the CRM and a lot of the people sitting in this room. 

  So the first thing we we can do is we can we can achieve some degree of clarity about about the choices that have been made, and trying to potentially improve some of the issues to do with evaluation. As we've seen, designing a bundled payment scheme involves a lot of choices. 

  And it's worthwhile understanding those choices, understanding what is, what is potentially available, and why people have made the choice, the choices that they've done, is, as you've seen, there's a wide variety of quality metrics that could be included in a bundled payment scheme. And it might be useful that that was informed by an evidence base based both for its benefit for the bundled payment, and its alignment with people's value. The other thing is that evaluating bundled payments getting Jazzy, as you're saying can be difficult. 

  So it may be useful to develop some standardised methodology and outcomes cost and quality. One of its, as we suggested, one of the issues of bundled payment schemes in oncology is is the fact that everything's changing in oncology on a fairly rapid basis. 

  So we need to ensure comprehensive comparative information to allow some degree of confidence about what's actually happened with the bundled payment scheme. And I think in addition to the issues of assessing the impact on payments from a payer, it's also well worth thinking, what are the what are the changes that were caused by bundled payments on costs for institutions, and also in Australia, the impact

Use the buttons to expand or collapse the section.

• Improving the value of the Australian health care system

  10 December, Professor Jane Hall

  Play video

  Read transcript

  Transcript

  Kees van Gool:

  Center for Research Excellence in Value-Based Payments in Cancer Care Webinar Series. This is our second one. And so thank you for joining us. So I'd like to acknowledge the Gadigal people of the Eora nation upon which whose ancestral lands our City Campus now stands. I'd also like to pay respects to the elders both past and present, acknowledging them as traditional custodians of knowledge for this lands. Just a few housekeeping notes, before we get started. We'll have time for a Q and A session straight after the presentation. In fact, having had a sneak preview, I know that Jane has some questions for us to answer at the end of her presentation. That's it. You can post questions or comments during the talk in the chat room which should be available. Or of course, you can actually ask your question during the Q and A session at the end. Now, just to impose zoom etiquette, I think everybody has already done, so please put your cell phone on mute, before we get started.

  Kees van Gool:

  Please note we are recording this session and will be posted on the CHERE website. And from next year on the CRE website as well. So now I have the great pleasure of introducing our speaker. Jane Hall, who's Distinguished Professor of Health Economics at UTS. She was director of CHERE until 2012 a centre she established some 30 years ago. She is the current president of the Academy of Social Sciences in Australia and also a fellow of the Australian Academy of Health and Medical Sciences. And was a past president of the International Health Economics Association. Her contribution to health economics research and policy has been recognized for numerous awards, including the UTAS as vice Chancellor's Award for excellence in research leadership. The NHMRC Outstanding Contribution Award and the Inaugural Professional Award made by the Health Services Research Association of Australia and New Zealand. For her outstanding contributions to research, developing the fields and mentoring others. So without further ado, I'd like to hand over to Jane whose presentation asked the question is Australia's payment infrastructure capable of supporting Value-Based Payments? Over to you, Jane?

  Jane Hall:

  Thank you Casey. And this is a nerve wracking bit as I share screen and hope I can get the right things up. I'll just go to the slide presentation. Right. So yes, well, look thank you all for joining. I too, would like to pay my respects to the traditional owners of the many lands on which we sit as we meet. I think Casey is this slide in his first webinar. And it's really been a motif for the way we've proceeded with our work. Since the senior public servant said to us, "don't tell me what I've done wrong, because we're very good at evaluating things. Tell me how to get it right." And I think that's really a lot of the motivation for this CRE is not looking at what has been done, but looking at what should be done. The approach of the CRE, as we've explained before, is to establish evidence and estimate impact, understand current payment design limitations, analyze patterns of care and so on. And we're very much in the understanding of current payment design limitations at the moment.

  Jane Hall:

  Casey, in his seminar, set up the question; is Australia's Data Infrastructure capable of answering these research questions? And I think there was some interesting insights and a good answer to that. I want to take a slightly different but related tack, which is, is Australia's payment infrastructure, capable of supporting Value Based Payment? And particular to think about what sorts of organization models will be needed for a real move to Value Based Payment. But of course to do that, we have to understand what we mean by Value Based Payment. So that's the sort of territory I'm going to be covering in the next 30 to 40 minutes. So why do we need health reform? Well, there's an awful lot that's been written about that. But I think I've just gone back to the National Health Reform Agreement. The 2020 and addendum that was signed. So this is the sort of current view of all Australian Governments. We need better coordination across primary, secondary, disability and aged care. We need to improve health outcomes, we should be encouraging diversity and innovation. Whatever we do should be evidence based and they particularly emphasize clinical evidence and clinical outcomes.

  Jane Hall:

  We need to improve equity and we need to deliver best practice care. So that's really where the clinical indicators come in. And the strategies that they see for doing that and are committed to exploring, a paying for value and outcomes and joint planning and funding at the local level. But why the emphasis on payment? Well, the aim should be to align the financial incentives for the provider with the delivery of optimum care. So the incentives have a very powerful signal to move the provider to deliver the outcomes that we require. We can set up a model that shows providers having a mix of selfish and altruistic motivations and use that to explore provider behavior. But I think one of the starting points that we need to recognize that's not always recognized in policy development, is that not all providers are the same. There are many different providers in the healthcare system.

  Jane Hall:

  So what do we already know? We know that payment systems provide very strong incentives and they also redistribute risk. And I'm going to come back to that notion of risks several times in this talk. So the major payment methods, a fee for service, capitation and global budgets. Fee for service appears to be the least risky because you as the provider, get paid for what you do. The risk sits with the funder. Although as we've seen over this year, if demand drops for services, then of course that the providers business can be at risk. The main unintended or sometimes it's intended. The main incentive effect for fee for service is to drive volume. Fee for service payment systems result in higher volumes of care being provided. Capitation we've seen and the main issue there is twofold. It's around cream skimming, which is clearly a way of reducing your risk if you can get a better class or a better risk category of patient and also skimping. Trying to do more with less. But when we talk about skimping, we're usually talking about unintended compromises on quality.

  Jane Hall:

  Global budget, again, have similar issues. They reduce the risk for the funder though, because there is a state ceiling of payment. But again, for the provider who holds that budget, the incentive is for skimping and to ignore access outside once you've reached your global budget. To worry about the payments within the institution, rather than the broader equity considerations. So what we see when we start to look at these, is that no one payment mechanism achieves the right sort of balance of incentives. So not surprisingly, we rarely see a pure single payment method, they're often blended in some way to try and overcome unintended consequences. And I want to just take a step back and remind you that financing and funding of healthcare are very tied up together. Although we think of financing as how we raise money from people and funding, how we pay for the health care that's provided. And that we have multiple objectives in setting up our financing and funding system. It's important to make sure that people do not get necessary treatment simply because they cannot afford it. And indeed to reduce the financial risk even for those who can afford it. So that they're not pushed into serious financial difficulty because of the cost of health care.

  Jane Hall:

  We look for equity and financing, access to services and health outcomes. We want to ensure value for money of what is bought and provided and in the long term, we need a sustainable system. Sustainable financially, but sustainable in other ways as well. And the current emphasis and indeed our emphasis at the moment is on how do we ensure value for money, but let's not forget that those other objectives are also important. So let's have a look at National Health Reform over the last 40 years. I'm going to start with the introduction of Medicare, because it was our major sort of Big Bang reform. And Scotton and Deeble in their researching and talking about the need for universal tax financed health insurance. Really made the point that the primary purpose of all of that change was around the financing of health care. Although they recognize that financing and how you provide care and the use of resources are all interwoven.

  Jane Hall:

  So in terms of the major National Health Reform steps and of course, there are many and I'm sorry, if I've ignored your favorite step. But this, I think is where I think the major steps are, we had the introduction of Medicare, which was all about the financing of care. We had the National Health Reform Commission, which was wide ranging produced a large number of recommendations, but didn't gain a lot of traction immediately. But in 2011, we had the National Health Reform Agreement, which was primarily around hospital funding, but did have some other clauses in it. And according to the agreement itself, this was major structural reform, and it was going to develop deliver sustainable funding for the Australian healthcare system. The National Health Reform Agreement has been renewed now on a number of occasions. And increasingly, what we see coming through that is this issue of needing to pay for value, needing to have incentives for the improvement of outcomes and that new and flexible payment methods will be needed. So that's the sort of why reform and why around payments.

  Jane Hall:

  Now if we look at what's been done, I talk about the payment architecture of the Australian health care system. And we really got three major funding channels, the Pharmaceutical Benefits Scheme, what is now called Medicare Benefits Scheme and public hospitals and State Commonwealth Agreements. And what you can see from there, I'm going to ignore the financing but look more at the paying for outcome side. Is that the changes to Medicare did not change that payment architecture. That payment architecture has remained in place. What has happened within the PBS is not a major structural change. But we have attempted to improve the value delivered through incorporating the need for health Technology Assessment before New pharmaceuticals are listed. We see the similar approach on the MBS, with the introduction of that requirement for cost effectiveness analysis. We've seen the introduction of practice incentives and so this incentive payments which are a move away from fee for service. But they remained small and they remained not universally taken up.

  Jane Hall:

  We've seen the MBS review which has been targeted at removing low value items from the schedule. But more interestingly, brought on by the need to make changes in the face of the pandemic. We've seen the expansion of telehealth and the move away from quite such a rigid definition of inputs to funding services. In the hospital sector, the biggest change has been the introduction of activity based funding which we can See as a move to trying to move to value by paying for an episode of care rather than paying for individual inputs. And again, I think one of the major changes that came through that was changing the Commonwealth's contribution to state budgets. To pay on the levels of activity and pay at efficient prices. So we've seen some innovations that are trying to move us down the value based path. Outside of that, of course, they've been the coordinated and integrated care trials and programs that have been introduced. Now, the first Australian coordinated care trials began in the mid 1990s. And at that stage, they really attracted a lot of worldwide attention as pioneering and ambitious efforts to improve the integration of care.

  Jane Hall:

  And the features of those trials with were funds pooling, though, often it was notional rather than real. Active coordination of services, something of a purchasing function in commissioning or purchasing services required and very targeted at complex patient. Similarly, integrated care programs. But if you look at the various evaluations that have been done, the overall verdict is that these trials were difficult to implement and failed to deliver promised benefits. And in fact, it's worth noting, I think that particularly in the early trials, if the funding had in fact been held by the service, the service would have failed financially. They were unable to cover their costs. The current approach to this is through health care homes and that initial phase of that program has not been completed yet. Health care homes have an aim to provide enhanced access for patients flexibility in the use of resources and an emphasis on team and coordinated care.

  Jane Hall:

  Now, these barriers come from the initial evaluation report and barriers to implementation have been the voluntary participation of patients, but more so of practices and of individuals within general practices. The need for complex administrative changes within practices, to deliver the services as intended and the sheer level of inexperience in managing these bundles within the general practice environment. So again, great aspirational goals, but really difficult to achieve. Okay, so what would we expect to see, if we're looking for Value Based Payment? I think everybody knows this, but we put it there anyway. The value... But the idea of value is that is simply outcomes of a cost. And so what we want with Value Based Payment is something that will drive improved health outcomes and at the same time, better use of resources. Now looking at the various schemes that now go by the name of Value Based Payment around the world, we've come up with these features of what distinguishes Value Based Payment.

  Jane Hall:

  That it will drive better outcomes. That care is coordinated around the person and across sectors. That there is a flexible use of inputs and that there is a reward for providers for improved performance. So those features are essential parts of saying we have a Value Based Payment System. And we see three approaches that are being explored and implemented, episode-based payments, Condition Specific Payments and population based payments. So the episode-base payment puts the funding together in a bundle around an episode of care for the patient. So it has a distinct beginning and has a distinct end. And outcomes can generally be measured at the end of the episode. Joint replacement is a good example of this. Inputs can be used flexibly and the provider can be rewarded for efficiency If they're allowed to keep what they don't spend in that episode... On that episode. So that's what that would look like.

  Jane Hall:

  Condition Specific Payments would be bundles around patients with a chronic condition. Now, this is much less likely to have a clear beginning and certainly not likely to have a clear end on the basis of the very definition of chronic conditions is that they're ongoing. And so the payment often has to be time based, often annual. The outcomes that you're looking for a much less clear cut. Because chronic conditions often have a tendency to deterioration, they'll certainly be flexible use of inputs and the provider can be rewarded for the better management of inputs. But again, we can see some issues around potentials for cream skimming and skimping. Population-based payments, the bundle is for the comprehensive care for a defined cohort. And indeed, part of the incentive. The theory that drives this is that it should be all care, not just components of care.

  Jane Hall:

  The care should meet individual needs, because that'll provide better outcomes and reduce the demand for the services or the need for services. Flexible again, because the inputs aren't specified, its a population payment and the provider again, will be rewarded for better management of costs. That's clear, we'll talk a bit more about the management of outcomes. So these are the challenges for Value Based Health care. So, yes, a key feature is measuring outcomes. But there is a real challenge in deciding what should be measured, but also when it should be measured. The episode based bundle, if it has a clear end, has a neat point at which you would want to measure the outcomes. But for population-based payments or chronic condition based payments, when is the right time to measure the outcome and what should it be? It's also important that those outcomes are outcomes that are attributable to the intervention or to the management of there condition. What we want to reward is effective intervention, not just luck and certainly not cream skimming of lower risk patients.

  Jane Hall:

  We need to see coordination if we're going to achieve patient centered care. And so the provider has to have expertise as a purchaser or commissioner of services, to ensure that, that coordination happens. And the reason I put purchasing in there is because the services are not going to be provided by one person, team based approaches to care are intended to be a feature of this. And that means you need the ability to coordinate the providers, not just the services around the patient. And of course, that means that that coordinator or coordinating agency has to be able to bear the risk. As we saw with the early coordinated care trials, pools that are too small, cannot bear that risk and could be sent broke by a small number of very expensive, very needy patients.

  Jane Hall:

  We need to see flexibility in the use of resources and but we want to avoid unintended consequences. So by taking our focus away from the use of resources, we can allow for cream skimming for risk selection. So there's going to be a big need for appropriate and robust risk adjustment to avoid that unintended consequence of cream skimming. And similarly with skimping. It means there's going to need to be substantial monitoring of quality to ensure that the appropriate clinical care is given. And remember if we go back to the writings on this particularly National Health Reform Agreement, it's got to be, clinically sound evidence based care. And then the rewards; Well the rewards are going to be very much in the detail of the design of the payment mechanism. But we need to consider both upside and downside risk. The rewards provide a strong incentive for people to do better. And so there must be that positive reward for providers.

  Jane Hall:

  But what about the downside risk? How much are we going to expect providers to bear? And will it go to the point at which poor performing providers are allowed to go broke and go out of business. So bundle holding, what we need is an agency that is able to hold that bundle of funds. And these are the requirements, I think, for the agency. They have to be big enough to develop expertise in coordination, they have to be big enough to develop and manage contracting relationships across teams of providers. They have to be big enough to sustain risk exposure bearing that financial risk. But at the same time, a key feature of this is expected to be this centeredness around the patient or the person. And obviously, the bigger they get the further they get away from the individuals whom they're supposed to be serving. So What experience do we have in Australia, of anybody holding bundles? Well, I'm going to talk about LHDs in hospitals. So hospitals, particularly with the move to ABF, clearly a managing bundles. Bundles for patients, although they're restricted to within hospital care, by and large. In fact, the organizational structure rests with the local health districts. And indeed, the Commonwealth and state payments for public hospitals, go through those LHDs in a common pool. So we have some experience there.

  Jane Hall:

  Primary Health Networks have not had such a strong role. But recently, we're aware of the need for PHNs, to develop in the commissioning of Mental Health Services. And so there's some experience there and if people in the audience have experience of that commissioning we would be really interested to hear about that. And of course, we've seen general practices as the focus for this with health care homes, but various other Integrated Care Pilots as well. And I think what we've seen with general practice, is that it can be particularly difficult if they're at a small size to develop the administrative changes, processes that are needed to handle this. And that often, the incentive is not strong enough to push a change in the complete business model.

  Jane Hall:

  So conclusions. The implementation of Value Based Health Care payment is challenging. If you read Porter, it sounds really easy, but actually it's not. And the experience from Australia and the experience from around the world shows just how difficult it is. Very often what we've seen in schemes is that the incentives are too small and the change that required is too big to make it even worth bothering about. And we see that with the claims of various pay for performance efforts. It sounds like if we could... So data collection, it sounds like if we could just focus on outcomes, we don't have to worry about looking at inputs and looking at processes. But in fact, moving to Value Based Healthcare doesn't remove the need to look at processes and inputs. Simply because measuring outcomes and being able to attribute them to the care that's provided is quite challenging. And the longer term the outcome that's relevant, the more other factors can intervene.

  Jane Hall:

  So whilst it sounds like our data collection requirements might be reduced, in fact, data collection requirements are really going to be increased. If we're going to not just pay for outcomes, but pay fairly for outcomes that have been achieved and ensure that there isn't skimping. The skill and expertise requirements will increase across all sorts of provider groups, in terms of coordination and in terms of financial risk management. And I think the sort of final message is that we're going to need to change our funding architecture. If we look at what's been done with the various and attempts to move closer to Value Based Payment In Australia, it's being done within the funding stream. Within the silos that we have of PBS, MBs and public or public hospital funding. And we need to gain to change those from silos that are rigid, to streams that can be much more flexible and joined together.

  Jane Hall:

  So that's really the talk. But the questions that I still have from the thinking I've done is, where should the bundles sit? Who's the best agency to hold it? Who will monitor the bundle holders? And who will make sure they're providing the quality that's required? And will there be one model of payment suitable for Australia? Or should there be several models that deal with different problems in the system? Though, thanks for listening.

  Kees van Gool:

  Thank you, Jane. It's a great history trip and some great thoughts there and lots of very interesting discussion points. So, as promised, we now have a Q and A session. We have some time and I can see that there is one person who has a question. So I'll try, I'll do the reading for the person. And once I read out that question, feel free to raise your hands and we can go to you or pop other questions in the chatroom if you'd like. But just to get us started. This is from Lynn Perry. Given the recent concerns at the drop off from NDIS H care funding, the recommendation to move to streams seems apposite. Any indication this is in train. Is that question clear or Lynn did you want to expand on that?

  Jane Hall:

  Yeah, no, the question is clear, the answer isn't clear. Look, I think the intent is clear. Because if you go and look at the Health Care Reform Agreement, it talks about flexible payment models. And I think there's an implication that it will happen on some sort of regional basis. Because of the commitment to exploring joint planning and funding at the local level. So this is my opinion, listening to people talk about it, rather than a research based evidence... Strongly evidence based answer, it's a bit more anecdote. But when I hear people talking about it, they seem to think it's going to be quite easy. And I think a really important message is that it isn't. So I think it will take longer than the aspirational goals that we're seeing.

  Kees van Gool:

  Thank you. Another question that's popped into the chat room. Is there any impact of accountability of provider on technical efficiency of local health districts or public district hospitals in Australia?

  Jane Hall:

  Is there any accountability?

  Kees van Gool:

  Yeah.

  Jane Hall:

  I think, yes, one of the features of ABF in this country, compared to a lot of other countries, is that it's based on a comprehensive collection of cost data. And that's been valuable, not just in stating the efficient price, but in providing benchmarking. And we know that data is used by hospitals, we know it's used by state treasuries, we know that it's looked at. So once... If you can develop that level of transparency, then you get accountability. I'm not sure if that's what the question is about. There's another bit though, that is about consumer responsiveness. And I took that section actually out of the talk because I was worried about going too long. But one of the views of needing to have competition in any system is that it gives consumers choice. And if they don't like what they get from one provider, they can move to another provider. But does that create sufficient accountability? And if it doesn't, how do you get that accountability within a system that doesn't give you the the leave option?

  Kees van Gool:

  Yeah, I think that's a great segue to the next question. Isn't one of the greatest challenges to any option involving primary care still, that we do not have mandatory registration of patients with practices in Australia. Until we do, you can never really belong to a practice for funding purposes. And without that, can GPs really make this work?

  Jane Hall:

  Yes, I think there are lots of challenges for GPs. And many of them are because of the size. I mean General Practices, I think are mostly small and medium enterprises. And so you've got to recognize that as a business model and work around that. Our research has shown that if you ask patients, most of them think they have a regular GP, but the GP doesn't know that. Because they do not know where else they go for care. I think to remove choice altogether, by having that compulsory registration, would not be popular in the Australian system. And indeed, what we've seen is targeting of enrollment. So we see that through the health care homes and of course, it was due to have a broader rollout this year, but has been delayed due to COVID. A registration of older patients or patients with complex conditions, voluntary enrollment with General Practices.

  Kees van Gool:

  Next, at a clinician level... Sorry its kind of... it's just not a question.

  Jane Hall:

  Yeah.

  Kees van Gool:

  At a clinician level, ABF is used as an internal stick. Would you like to comment on it? Or Will you let me fulfill my dream and say, we'll take that as a comment.

  Jane Hall:

  Yes, I think we can take that as a comment.

  Kees van Gool:

  Okay, great.

  Jane Hall:

  I do think what bundling payments does, it's going to change the relationships between the person who holds the purse strings and and some of those service providers.

  Kees van Gool:

  Correct. What about care delivered via the private versus public sectors? The MBS funds the former, the NHRA, the latter. How would payment bundles across the two sectors work? Sorry, Sarah, I'm misquoting you. How would payment bundles be managed to cross the two sectors?

  Jane Hall:

  Well, if you've got someone holding a bundle, they can commission or purchase the service from a public or a private provider. And you would hope that there would be that option, so that again, they're driven by high quality, best outcomes, good price.

  Kees van Gool:

  Next question, we regularly hear that we need more funding in health or is that we have enough resources we just don't apply them efficiently enough? What's your view?

  Jane Hall:

  Yes, I think it's very difficult to have a firm answer on that question of how much is enough? I think there is increasing concern that a lot of what is... There is over provision in some areas and under provision in others. So it's about getting the the use of resources, better managed, rather than just pull more money into it. I think there is a knee jerk reaction to pull more money into something where it looks like a problem without thinking about how it's spent. And I think psychological counseling is a really good example. I mean, ever since I've been working, which is a long time as Casey has told you, mental health has been an area of need and governments have poured progressively more and more money into it. And it still seems to be a need. We have an increasing burden of mental health. And what we've seen in the sort of COVID support for counseling services is that it's gone from 10 sessions to 20 sessions. So it's more money, but not necessarily better use. And I think it's that distribution that's the problem.

  Jane Hall:

  We see the same problem in education, where we keep being told that more and more money is being poured into education, but Australian school students are performing worse and worse on world levels. But of course, it's not the total that matters here. It's how it's distributed across schools and according to need.

  Kees van Gool:

  Okay, sorry, I do have another question. How do you think linked health data can support Value Based Care, bundled payments?

  Jane Hall:

  Oh, linked health data is absolutely essential. If you can't follow people around the health system, you can't manage a bundle of any sort. And having a universal collection of individual health identifiers, so that we can follow people through the system and see what's happening to them is absolutely essential.

  Kees van Gool:

  And can I give myself a plug there for that question that the previous webinar addressed that issue in some detail. So it is on the CHERE website if you'd like to see a recording of that webinar. Any other questions in the chatroom? No one. Well, let me give you an opportunity for anybody who wants to raise their hands to ask any further questions. In the meantime, if people still are formulating their questions

• Overview of Australia’s data infrastructure in cancer care

  25 November, Professor Kees van Gool

  Play video

  Read transcript

  Transcript

  Kees van Gool:

  I go. I'm sharing the screen, which I hope I am doing. Hopefully everyone can see that

  Phil Haywood:

  Yeah.

  Kees van Gool:

  Great. Thanks for confirming. Okay. Well, thank you very much for joining us. It's a great privilege to kick this seminar series off and maybe be part of the next phase of implementing the Center for Research Excellence, which is almost coming up to a year. Now, of course, the CRE is embedded in the center for the Health Economics Research Evaluation, which has been around for coming up to two 30 years, and I've been part of, well, at least talking about coming up to 20 years of that, so two thirds of that 20 years. And in that time, it's fair to say that, we've evaluated a fair number of the payment reforms that have taken place in this country, Medicare Safety Net, bulk-billing incentives, integrated care, Coordinated Care, changes to copayments, et cetera.

  Kees van Gool:

  And if I were to summarize that body of work in one sentence, it's that the reforms were mostly expensive and had some impact. Some of it was good, some of it was bad. And that mirrors much of the international experience as well. There's an idea that we need to change the way we pay for healthcare, then there's the actual idea for reform. It gets implemented then evaluated. And we discover that the impact of the reform was probably not as great as we'd hoped or worse, that there were many unanticipated consequences that led to adverse events.

  Kees van Gool:

  Now I was quite happily producing that type of evidence and chugging along with that and telling governments what they'd gotten wrong. And then I had one of those mind-blowing conversations a few years ago with a fairly senior public servant who gave me this quote, which is, don't tell me what I've done wrong tell me how to get it right. So this senior public servant presented us with a challenge, what contribution can a team of applied health economists makes it a paying reform agenda, aside from telling the government that they got it wrong.

  Kees van Gool:

  And the challenge is that, by necessity, our data and our evidence always looks to the past, and how can the past inform how we should change payments for the future? So this similar provides an overview of how we're addressing that challenge. We start by discussing a current state of play of healthcare payments and where we think there is room for improvement. And second, we examine how retrospective data can inform the future design of payment reform ideas. And finally, we investigate whether our current data infrastructure, Australia's data infrastructure, is ready to help us meet that challenge.

  Kees van Gool:

  So let's look at payment reform and why we still focus on that as a research topic. So, ultimately we want our healthcare dollars to produce health outcomes. Think of those as more life years or better quality of life. And some instances, we do pay for health outcomes in the healthcare system. In the case of Australia's Pharmaceutical Benefits Scheme, government decides to pay for drugs on the basis of their anticipated cost per outcome. And I say anticipated because it's not actual outcome, it's promised outcomes based on the evidence from clinical trials. In other instances, we pay for output. Over the last 30 years, Australia has reformed the way we pay for hospital care. Nationally, we now pay for an episode of care. Hospitals get paid for the number of separations they perform, and each payment is calculated on the basis of the complexity of that separation.

  Kees van Gool:

  The hospital receives a payment for the output, and it decides how to combine all the inputs such as doctor time, nursing time, pharmaceuticals, equipment, capital, et cetera. Now, outside of drugs and hospitals, we still tend to pay for inputs of care. This is particularly true where we consider the management and treatment of diseases that require care over extended periods of time. Think about diabetes, a complex disease that requires multidisciplinary care over extended periods of time to manage the disease well and avoid short and long-term complications. They, in Australia, would pay for single services of GPs, allied health, diagnostics, and we hope that all these inputs combined together to produce the outcomes that we want.

  Kees van Gool:

  The same is true for cancer care. We know that a complex disease like cancer, requires a wide variety of healthcare inputs that need to come together in a coordinated and integrated fashion to deliver optimum care pathway, from screening to diagnosis, to staging, to planning, to surgery, to chemotherapy, radiotherapy, and then ongoing monitoring and evaluation.

  Kees van Gool:

  Now, does it matter what we pay for? Whether we pay for inputs or outputs or outcomes. Well, it does, if you believe the old notion that, you get what you pay for. Let's look at paying for inputs, a little bit more closely. The recent experience of introducing telehealth services into Medicare provides a very good case study. Tell me health is not new. Medical providers have had the ability to pick up the phone for quite some time and talk to patients. But there's been a reluctance to do so. Why? Because GPs, up to very recently, were not paid for this activity. Before COVID hit a GP telephone call was not remunerated and came at the expense of losing revenue from a face-to-face consultation. So, despite the fact that a telephone call is often the most practical, convenient and efficient form of communication between a doctor and a patient, the payment system was at odds with the delivery of that service.

  Kees van Gool:

  It's the flip side of, you pay of what you get for. In other word that is, you don't get what you don't pay for. Then the COVID pandemic struck. And in March 2020, in-person GP consultations dropped by around 20% to 30%. The government quickly introduced telehealth items into the MBS, at the end of March, for the entire population, and between April and June, GPs provided nearly four million telehealth consultations per month, representing about a $50 million increase in GP expenditure for each month, after March.

  Kees van Gool:

  Now, this all makes sense when dealing with a pandemic. The value of the telehealth consultation is not just the ability of GPs to provide services to their patients, it's also an essential part of the response to reduce the risk of COVID spread. However, once the COVID situation is under control, government will start to ask the question about whether to retain the telehealth items in the MBS. Now, without COVID, the value will be diminished because it doesn't make... But it doesn't mean that it's without value. The question of course is, does the additional $50 million in monthly GP benefits deliver greater health outcomes and convenience, or could we have spent the 50 million on other parts of the healthcare system and get even more better health outcomes?

  Kees van Gool:

  So the question is whether paying for the inputs gets us to where we want to be. By paying for inputs, do we get the best evidence-based pathway of care? Our argument is that, that's not necessarily the case. Paying for inputs will deliver outputs and outcomes, but these may not always be the optimum and most efficient and make a good use of our resources.

  Kees van Gool:

  So, here's a bit of an example of some work that we did using the Medicare data linked to the 45 and Up Study. So, as part of the 45 and Up Study, it asked questions about whether people were diagnosed with diabetes, and importantly, it also asks when they were diagnosed. And so when we use that information from the survey, we can link that to the administrative data in Medicare and examine how the onset of diagnosis... sorry, the diagnosis of diabetes leads to a change in healthcare utilization. And we did this by looking at GP visits and specialist visits. So, on the horizontal axis, we're measuring time, where zero is the year in which the person was diagnosed with diabetes, minus two is obviously the two years prior to diagnosis, and one and two, three and four are the years following diagnosis.

  Kees van Gool:

  And what you can see is that people tend to increase their utilization in years, leading up to the point of diagnosis, and in before both GPs and specialists. But if you look at more closely, you can also see that there are quite differences depending on the income of those people. So it's clear that everybody increases their use of GP visits at the point of diagnosis or the lead up to the diagnosis. But only the lower income people actually stay above the initial periods of GP visits. Whereas, the higher income people revert back to the level of consumption of GP services prior to the two years of diagnosis. Whereas, on the specialist side, you can see a little bit of opposite effect. Lower income initially increase their use of specialist services, then go back down to the level where they were, whereas higher income people increase their level and stay there for at least the next three to four years.

  Kees van Gool:

  So you can see that people navigate the healthcare system following a diagnosis of diabetes. And they navigate that system, presumably, on the basis of, for example, the cost. It's well known that GP visits come at a lower copayment generally than specialist visits, and so it's perhaps not surprising that we find that lower income people try and find the lowest cost healthcare, compared to the relatively well off.

  Kees van Gool:

  Now, we don't know what the optimum bundle of inputs is when delivering diabetes care. We do not know whether, for example, GP visits and specialist visits can be substituted for one another. That you can, you know, the two types of services are pretty much the same, and you can switch between one another. Or in fact, we do not know whether they are compliments, and that in fact, a series of GP visits complemented by a specialist visits provides the optimum care. But what we do know is that, people choose their own pathways and that there is variation in those pathways. In this case, depending on the out-of-pocket costs that may lead to... depending on the evidence and how the inputs come together on future outcomes.

  Kees van Gool:

  So what is important there is how the system interacts and how payments come into play when choosing the level of services. So one of the things that is a challenge for payments is that, of course, we want the incentives to be aligned to good practice. So in the previous case, there is no incentive for GPs or specialist to work together, there is an incentive to just financial incentive to increase the volume of the work in this fee for service system that we have.

  Kees van Gool:

  Where payment reform comes in, is to try and change those incentives and to align the financial incentives for the provider with the delivery of optimum care. So all the way we've been looking at this is, through a conceptual framework that looks at provider utility. And the provider utility can be seen, and this is based on the work from others, into two parts. The first part, I hope you can see my mouse, is the part of the utility where is motivated by selfish motivations. Essentially it's the income that doctors can obtain for their services. And the second part is the altruistic notion of the doctor caring about the welfare and the well-being of their patients.

  Kees van Gool:

  Now, each individual provider can be different. Some will care more about their altruistic part then their selfish parts, and that is represented by AI in this equation. So in other words, if AI equals one, then their own utility to provide is completely driven by the altruistic part. But if AI equals zero, then their motivation is completely driven by selfish motivations. So their utilities is driven by selfish motivations. AI can be anything in between zero and one for any individual provider. Now, the selfish part is driven by the revenue that they deliver. And so, X amount of healthcare at peak prices minus the cost of delivering that health care for each patient, and of course they see a number of patients J to one. That gets them their income.

  Kees van Gool:

  On the altruistic side, they care about the health that the healthcare provides. So they care about how the healthcare is converted into health for the patient. But they also care about the financial consequences of their patients, because that is a part of their well-being. So well-being is not just the health effects, but is also the financial effect of patients having to pay for some of the healthcare.

  Kees van Gool:

  And that, which is again, derived by the prices that the patient pays and the amount of healthcare that they do. But this is a very simplistic model and can be extended by including, for example, insurance arrangements. If insurance were introduced in this model, then obviously that has two effects. The first effect is that patients no longer pay the full price of the healthcare, but they also have to pay healthcare premiums. So there would be another parameter here, which is the healthcare premium, which can either be actual private health insurance premium, or in fact, a portion of the tax that they have to pay. Now, the challenge for payment reform is to align both the selfish and the altruistic motivation. And so, that in fact, whenever a decision is made to provide healthcare, both of the selfish motivation and the altruistic motivation move in the same direction, which in the case of fee for service, of course, is not always the case. They can move in opposite direction. And that's where value based payments reform comes in.

  Kees van Gool:

  So, the challenge for us is to use the conceptual model like that to establish the evidence for payment reform design and estimate the impact, include the behavioral change prior to implementation. We want to analyze variations in the patterns of care for patients with similar health care needs. We want to understand the role that payment systems can have in explaining variations. And we want to design use as payment systems around outputs rather than inputs of care. And finally, we want to simulate that impact of alternative payment mechanisms on the efficiency, funding sources, and provider institutional behaviors and equity.

  Kees van Gool:

  That's really what this graph is about... is that what are the conditions for payment reform? We recognize that not all parts of the healthcare system requires reform and the whole kind of reform, but there are some areas where that might be the case. And the way to identify this is, is there evidence of suboptimal care? And what we mean by suboptimal care is patients with specific conditions not receiving care according to the best available evidence, such as guidelines, systematic variation in the healthcare use or even a systematic variation in health outcomes. And we can look at such variation with respect to regions or socioeconomic groups or cultural groups.

  Kees van Gool:

  So if the answer is... to the question is there evidence of suboptimal care, and is there evidence that that payment mechanisms are a driver of that suboptimal care, that in fact, becomes a reason to look at where the value based payments, may be beneficial.

  Kees van Gool:

  So one form of payment reform that's receiving considerable attention are bundled payments. Around the bundled payments is a series of healthcare services or healthcare inputs, as we previously described them, are bundled together. Instead of paying a fee for a service for each individual service across an episode of care, it brings those fees together under a lump sum payments. The bundle is usually specific to a particular treatment phase of a particular disease. Internationally, bundled payments have been piloted in US, Netherlands, Sweden Portugal, and have shown promise in improving quality financial savings and increased satisfaction.

  Kees van Gool:

  Now, because we are buying an episode of care, it becomes plausible to think about the quality of care, which in turn can be incorporated into payment mechanisms. In practice, quality can be linked to the financial incentives, or at the very least, be part of a routine performance measure of benchmarking exercises. In such models, providers and institutions have a financial incentive to address the complex care needs of patients, including those in our oncology population, while systematically decreasing the use of services that do not improve health outcomes. And compared to fee for service, it gives providers more flexibility in how inputs are combined to deliver an episode of care.

  Kees van Gool:

  Now, there are numerous evaluations of bundled payments. And whilst they show promise, one fair criticism of this body of work is that bundled payment reforms are routinely treated as though they are a homogeneous intervention, when clearly each bundle payment reform is very different, whether they are implemented in different countries, different health systems or for different diseases or different parts of the treatment pathways. A bundled payment evaluation is not like evaluating the effectiveness of a drug. Each bundle payment is different and therefore its success or failure depends very much on the design of the bundle payment and the context within which it is implemented.

  Kees van Gool:

  So this is where I think we are meeting the challenge post to us at the start of this presentation by a senior public servants. Looking retrospectively will help us identify where we think the biggest problems are occurring with respect to unwarranted variation in care. It will help us design an evidence-based bundle of care that brings together the right inputs at the right price, with all the necessary safeguards to ensure high quality care. And that goes down to answering the set of questions that you see before you. Many of which can be answered through the existing retrospective datasets.

  Kees van Gool:

  Then finally, we use that evidence to inform microsimulation models, to examine the input that such [inaudible 00:20:49] may have on costs, funding sources and outcomes.

  Kees van Gool:

  So is Australia's data infrastructure ready and able to help us answer some of these questions about [inaudible 00:21:03]? So, the promise of linked data is vast. In the right hands it has got the capacity to answer fundamental questions about the origins of disease and examine the success of health care in treating disease. And by link data I mean bringing together various data sets that can be administered through link like Medicare claims or hospital admission data, as well as survey data. Importantly, they are linked at the individual patient. And in doing so it creates a profile of an individual's health status, their healthcare use over the course of their lifetime, or at least a significant portion of their lifetime.

  Kees van Gool:

  Linked data has the potential to observe both modifiable and non-modifiable risk factors. We often use surveys to obtain information on the person's lifestyles. Some studies even have bio-banks that can provide genetic information. Through both surveys and administrative data, we're able to ascertain information on places of residence or other factors such as occupation that provide valuable information on the person's environmental factors that may affect their life course. Linked administrative data is becoming incredibly powerful when it comes to specifying the time and diagnosis of some of the major diseases. For example, cancer registries provide information on the type of cancer diagnosed, the stage of diagnosis, and the date of diagnosis.

  Kees van Gool:

  More recent advances in linking GP health records, will fill an important gap in being able to identify a broader range of diagnostic information, although linkage of these records are still active in their infancy. Moving on to treatments, it means the data sets like emergency department presentations, hospital admissions, drug prescriptions, Medicare consultations are all very good at allowing us to identify what treatment a patient receives. And finally, through linkage with the cause of death registries, we're able to make sure that mortality and survival and administrative data can also look at some aspects of quality of care through, for example, readmissions or hospital acquired infections or complications. Many survey-based cohorts include, patient reported outcomes including quality of life measurements, and patient experiences. But these are limited to surveys, and as a result are limited [inaudible 00:23:17] there are immediate biases in very sick people who may leave the study, and also the timing of the survey is unlikely to align with a major illness or treatment.

  Kees van Gool:

  For the last 25 years, I've seen amazing progress and substantial investment by government and academia in our technical ability to link the data. University of Western Australia and WA departments, taking out and not to all our WA friends who are joining us on this call, were really innovators in this area and set the scene for many other jurisdictions. Establishment of the New South Wales Center for Health Records Linkage to state based administrative data in 2006, was a major advancement. And it was quickly followed by a similar centers in Queensland, Victoria, South Australia and Tasmania. More nationally and more recently BHW has established a data integration services center.

  Kees van Gool:

  And of course we now have longstanding cohort studies of specific population groups, such as women, children, and those aged 45 and over. And of course, many researchers have established their own cohort studies, many in fact. Now, obtaining this data is rightfully not easy, but there are clear pathways by which bonafide researchers can get permission to access the data. Investments in secure computing environments have made access more feasible for a wider research group community. But be warned, it all takes time, effort, and money, and a bit of know-how. It doesn't go straight on.

  Kees van Gool:

  There's, why the amazing progress? As users of the data, we still face some significant barriers. Linked data is usually cobbled together from various sources and are owned by various agencies and jurisdictions. There are processes in place to seek permission to use, but it all takes time and success is not always guaranteed. There are many known unknowns. The data sets most commonly linked, we know, have some blind spots, private health insurance claims, palliative care, some drugs, community care and public hospital, outpatient care, to name, but a few.

  Kees van Gool:

  And there are unknown unknowns. Despite 10 years of experience in using this data, every time we embark on a new project, there seem to be new set of surprises. Now, one of the few advantages of getting older is that your lived experience dates back to when Medicare was first introduced. And I think I have a reasonable knowledge of all the policy changes that have occurred over that time that may affect the way we interpret the Medicare data. But every time, like I said, there are surprises and all the certainty we have is that we know that there will be surprises. And of course, every data set comes with its own limitations. The data set is never clean, and this is a given, but it's particularly true for administrative data remembering that most of the commonly used data sets are not designed for research. Research is a secondary purpose.

  Kees van Gool:

  There are duplicate records, negative values where there shouldn't be, extreme and preposterous values, and over time we've developed our own rules about what to do when there are these potential data errors. But there's nothing particularly scientific about this. We make judgments that we think are defensible and good practices to test if these judgements matter in the analysis. There's no guarantee that these judgments are correct. Another challenge is the different formats that linked data arrive in. Longitudinal surveys that follow the same participants over time comes in waves every year or every three years, depending on the study. The linked administered data, however like MBS and PBS comes by claims. Every claim an individual participant makes represents a row of data. The question is, how do you bring all that data together in a suitable format so that all the time periods and the observations make sense, so the purpose of analysis?

  Kees van Gool:

  The answer is there but it takes skills and time. And that's why new degrees in data science and data analytics are so important to help us design... that are designed to develop these skills. Now, despite the deluge of data, the ever present challenge of undertaking high quality policy relevant research persists, the challenge has always been to use observational data to make causal inferences. It's so difficult because observational data is rife with selection issues and endogeneity implying that the results we get maybe bias and any correlation found does not imply causation. To make stronger causal inferences, it requires a sound theoretical framework, good data, and a research design that can address potential selection biases and endogeneity, like using quasi-experimental techniques for applied econometric techniques. And finally, some reflections, someone is increasingly coming up against some significant analytical roadblocks through the way we derived our samples for our linked studies. We are intensive users of the 45 and Up Study, and we are very appreciative of the great work that the Sax Institute has done there.

  Kees van Gool:

  But a large number of our studies we hit a few roadblocks. For example, despite the fact that there are 260,000 people originally in the 45 and Up Study, we hit sample size problems because it doesn't take long before you run into a sample size when you're dealing with, luckily, rare diseases. Federal creates issues around generalizability. If we select the sample on the basis of where we live, we won't know if the results apply to other areas. And most importantly, selecting it on the basis of patients, stops us from having high quality data on the other most important decision maker in the healthcare system, and that is providers. We've been able to use cohort studies to examine how providers respond to policy change, but only in a limited fashion. We can only observe provider behaviour if the patient is enrolled in the study. So we get a rich picture of a patient, but not necessarily everything that the doctor does, particularly if the patient is not enrolled in the study, like 45 and Up.

  Kees van Gool:

  So one of the final steps... Sorry, a summary of.... And make steps of where we think we are at. So, first of all, I think change to cycle of payment reform fail. I spoke about the fact that we tend to implement, then evaluating and work out [inaudible 00:29:50]. I think we can do better than that by understanding that we have the potential behavioural responses to payment change and recognise the context and the design of payment reforms and that impact on its potential success. And then finally, there's still a lot of work to do in the data infrastructure. Even though there's been amazing progress over the last two decades there are some issues around the way we do develop our cohorts and the cohort studies, which have limitations particularly when it comes to looking at provider issues.

  Kees van Gool:

  And in addition, there is a need to develop the community of researchers that look at these data sets. Even though we are all looking at different cohorts of different studies, at the end of the day, we're all looking at MBS data, we're all looking at BBS data, we're all looking at hospital admissions data, and there are many common issues. And so, that community of linked data research is something that I think is very close to my heart, to come up with some standards and finding some common solutions to the same problems that we keep on facing.

  Kees van Gool:

  That's it for [inaudible 00:31:06], and so I'm happy to take questions or in fact, it'd be delightful to open discussions.

  Phil Haywood:

  Thank you Kees. So there's been a sum total of zero questions so far. So, I'll take the opportunity to ask a couple that I've been meaning to ask you. So given you've taken up the challenge from the unnamed public service for over a year now, what do you think might be the low hanging fruit for payment reform in Australia and what, after an experience of a year, do you think might be the most difficult ones?

  Kees van Gool:

  Okay. So, well, low hanging fruit is relative. So, the way we've structured the work so far is straightforward. One was around, where can we find the variation. Second is, where can we use previous reforms to look at behavioural change. And the third one is to bring together the work around the design of different payments, making the design of bundled payments and implementing microsimulation exercise. Go to the first one, we're looking at several parts of variation, we're looking at radiotherapy, we're looking at palliative care, and also looking now at MBS where there are differences in out of pocket costs, which may impact on people's access.

  Kees van Gool:

  So we're basically just trying to tease out parts of the pathway by which patients enter the system and go through the system, and looking at variation where, for example, there's in palliative care or radiotherapy where there's variation for a fairly homogeneous group of patients, they all have the same cancer, they were diagnosed at the same stage, and where there is variation in the use of those services for say radiotherapy or where there's differences in the costs. That gives us some notion of where we might be thinking about where bundle payments, where we if we bring together the services, radiotherapy, the consultations, and any other services that might come combine, to do that.

  Kees van Gool:

  The other aspect is actually looking beyond cancer, even though we are focused obviously on cancer payment reform, but other areas of changes to Medicare provides us with some important insights in how doctors change their behaviour when, for example, we take something away from them. So the work, I know that Phil is working on closely, that's why it's a good question because he asked himself, is around joint injections. And so, there's some work going on where we look at the fact that a joint injection Medicare item was removed and we're