Chris Sampson’s journal round-up for 2nd December 2019

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

The treatment decision under uncertainty: the effects of health, wealth and the probability of death. Journal of Health Economics Published 16th November 2019

It’s important to understand how people make decisions about treatment. At the end of life, the question can become a matter of whether to have treatment or to let things take their course such that you end up dead. In order to consider this scenario, the author of this paper introduces the probability of death to some existing theoretical models of decision-making under uncertainty.

The diagnostic risk model and the therapeutic risk model can be used to identify risk thresholds that determine decisions about treatment. The diagnostic model relates to the probability that disease is present and the therapeutic model relates to the probability that treatment is successful. The new model described in this paper builds on these models to consider the impact on the decision thresholds of i) initial health state, ii) probability of death, and iii) wealth. The model includes wealth after death, in the form of a bequest. Limited versions of the model are also considered, excluding the bequest and excluding wealth (described as a ‘QALY model’). Both an individual perspective and an aggregate perspective are considered by excluding and including the monetary cost of diagnosis and treatment, to allow for a social insurance type setting.

The comparative statics show a lot of ambiguity, but there are a few things that the model can tell us. The author identifies treatment as having an ‘insurance effect’, by reducing diagnostic risk, a ‘protective effect’, by lowering the probability of death, and a risk-increasing effect associated with therapeutic risk. A higher probability of death increases the propensity for treatment in both the no-bequest model and the QALY model, because of the protective effect of treatment. In the bequest model, the impact is ambiguous, because treatment costs reduce the bequest. In the full model, wealthier individuals will choose to undergo treatment at a lower probability of success because of a higher marginal utility for survival, but the effect becomes ambiguous if the marginal utility of wealth depends on health (which it obviously does).

I am no theoretician, so it can take me a long time to figure these things out in my head. For now, I’m not convinced that it is meaningful to consider death in this way using a one-period life model. In my view, the very definition of death is a loss of time, which plays little or no part in this model. But I think my main bugbear is the idea that anybody’s decision about life saving treatment is partly determined by the amount of money they will leave behind. I find this hard to believe. The author links the finding that a higher probability of death increases treatment propensity to NICE’s end of life premium. Though I’m not convinced that the model has anything to do with NICE’s reasoning on this matter.

Moving toward evidence-based policy: the value of randomization for program and policy implementation. JAMA [PubMed] Published 15th November 2019

Evidence-based policy is a nice idea. We should figure out whether something works before rolling it out. But decision-makers (especially politicians) tend not to think in this way, because doing something is usually seen to be better than doing nothing. The authors of this paper argue that randomisation is the key to understanding whether a particular policy creates value.

Without evidence based on random allocation, it’s difficult to know whether a policy works. This, the authors argue, can undermine the success of effective interventions and allow harmful policies to persist. A variety of positive examples are provided from US healthcare, including trials of Medicare bundled payments. Apparently, such trials increased confidence in the programmes’ effects in a way that post hoc evaluations cannot, though no evidence of this increased confidence is actually provided. Policy evaluation is not always easy, so the authors describe four preconditions for the success of such studies: i) early engagement with policymakers, ii) willingness from policy leaders to support randomisation, iii) timing the evaluation in line with policymakers’ objectives, and iv) designing the evaluation in line with the realities of policy implementation.

These are sensible suggestions, but it is not clear why the authors focus on randomisation. The paper doesn’t do what it says on the tin, i.e. describe the value of randomisation. Rather, it explains the value of pre-specified policy evaluations. Randomisation may or may not deserve special treatment compared with other analytical tools, but this paper provides no explanation for why it should. The authors also suggest that people are becoming more comfortable with randomisation, as large companies employ experimental methods, particularly on the Internet with A/B testing. I think this perception is way off and that most people feel creeped out knowing that the likes of Facebook are experimenting on them without any informed consent. In the authors’ view, it being possible to randomise is a sufficient basis on which to randomise. But, considering the ethics, as well as possible methodological contraindications, it isn’t clear that randomisation should become the default.

A new tool for creating personal and social EQ-5D-5L value sets, including valuing ‘dead’. Social Science & Medicine Published 30th November 2019

Nobody can agree on the best methods for health state valuation. Or, at least, some people have disagreed loud enough to make it seem that way. Novel approaches to health state valuation are therefore welcome. Even more welcome is the development and testing of methods that you can try at home.

This paper describes the PAPRIKA method (Potentially All Pairwise RanKings of all possible Alternatives) of discrete choice experiment, implemented using 1000Minds software. Participants are presented with two health states that are defined in terms of just two dimensions, each lasting for 10 years, and asked to choose between them. Using the magical power of computers, an adaptive process identifies further choices, automatically ranking states using transitivity so that people don’t need to complete unnecessary tasks. In order to identify where ‘dead’ sits on the scale, a binary search procedure asks participants to compare EQ-5D states with being dead. What’s especially cool about this process is that everybody who completes it is able to view their own personal value set. These personal value sets can then be averaged to identify a social value set.

The authors used their tool to develop an EQ-5D-5L value set for New Zealand (which is where the researchers are based). They recruited 5,112 people in an online panel, such that the sample was representative of the general public. Participants answered 20 DCE questions each, on average, and almost half of them said that they found the questions difficult to answer. The NZ value set showed that anxiety/depression was associated with the greatest disutility, though each dimension has a notably similar level of impact at each level. The value set correlates well with numerous existing value sets.

The main limitation of this research seems to be that only levels 1, 3, and 5 of each EQ-5D-5L domain were included. Including levels 2 and 4 would more than double the number of questions that would need to be answered. It is also concerning that more than half of the sample was excluded due to low data quality. But the authors do a pretty good job of convincing us that this is for the best. Adaptive designs of this kind could be the future of health state valuation, especially if they can be implemented online, at low cost. I expect we’ll be seeing plenty more from PAPRIKA.

Credits

Rita Faria’s journal round-up for 13th August 2018

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Analysis of clinical benefit, harms, and cost-effectiveness of screening women for abdominal aortic aneurysm. The Lancet [PubMed] Published 26th July 2018

This study is an excellent example of the power and flexibility of decision models to help inform decisions on screening policies.

In many countries, screening for abdominal aortic aneurysm is offered to older men but not to women. This is because screening was found to be beneficial and cost-effective, based on evidence from RCTs in older men. In contrast, there is no direct evidence for women. To inform this question, the study team developed a decision model to simulate the benefits and costs of screening women.

This study has many fascinating features. Not only does it simulate the outcomes of expanding the current UK screening policy for men to include women, but also of other policies with different age parameters, diagnostic thresholds and treatment thresholds.

Curiously, the most cost-effective policy for women is not the current UK policy for men. This shows the importance of including the full range of options in the evaluation, rather than just what is done now. Unfortunately, the paper is sparse on detail on how the various policies were devised and if other more cost-effective policies may have been left out.

The key cost-effectiveness driver is the probability of having the disease and its presentation (i.e. the distribution of the aortic diameter), which is quite frequent in cost-effectiveness analysis of diagnostic tests. Neither of these parameters requires an RCT to be estimated. This means that, in principle, we could reduce the uncertainty on which policy to fund by conducting a study on the prevalence of the disease, rather than an RCT on whether a specific policy works.

An exciting aspect is that treatment itself could be better targeted, in particular, that lowering the threshold for treatment could reduce non-intervention rates and operative mortality. The implication is that there may be scope to improve the cost-effectiveness of management, which in turn will leave greater scope for investment in screening. Could this be the next question to be tackled by this remarkable model?

Establishing the value of diagnostic and prognostic tests in health technology assessment. Medical Decision Making [PubMed] Published 13th March 2018

Keeping on the topic of the cost-effectiveness of screening and diagnostic tests, this is a paper on how to evaluate tests in a manner consistent with health technology assessment principles. This paper has been around for a few months, but it’s only now that I’ve had the chance to give it the careful read that such a well thought out paper deserves.

Marta Soares and colleagues lay out an approach to determine the most cost-effective way to use diagnostic and prognostic tests. They start by explaining that the value of the test is mostly in informing better management decisions. This means that the cost-effectiveness of testing necessarily depends on the cost-effectiveness of management.

The paper also spells out that the cost-effectiveness of testing depends on the prevalence of the disease, as we saw in the paper above on screening for abdominal aortic aneurysm. Clearly, the cost-effectiveness of testing depends on the accuracy of the test.

Importantly, the paper highlights that the evaluation should compare all possible ways of using the test. A decision problem with 1 test and 1 treatment yields 6 strategies, of which 3 are relevant: no test and treat all; no test and treat none; test and treat if positive. If the reference test is added, another 3 strategies need to be considered. This shows how complex a cost-effectiveness analysis of a test can quickly become! In my paper with Marta and others, for example, we ended up with 383 testing strategies.

The discussion is excellent, particularly about the limitations of end-to-end studies (which compare testing strategies in terms of their end outcomes e.g. health). End-to-end studies can only compare a limited subset of testing strategies and may not allow for the modelling of the outcomes of strategies beyond those compared in the study. Furthermore, end-to-end studies are likely to be inefficient given the large sample sizes and long follow-up required to detect differences in outcomes. I wholeheartedly agree that primary studies should focus on the prevalence of the disease and the accuracy of the test, leaving the evaluation of the best way to use the test to decision modelling.

Reasonable patient care under uncertainty. Health Economics [PubMed] Published 22nd August 2018

And for my third paper for the week, something completely different. But so worth reading! Charles Manski provides an overview of his work on how to use the available evidence to make decisions under uncertainty. It is accompanied by comments from Karl Claxton, Emma McIntosh, and Anirban Basu, together with Manski’s response. The set is a superb read and great food for thought.

Manski starts with the premise that we make decisions about which course of action to take without having full information about what is best; i.e. under uncertainty. This is uncontroversial and well accepted, ever since Arrow’s seminal paper.

Less consensual is Manski’s view that clinicians’ decisions for individual patients may be better than the recommendations of guidelines to the ‘average’ patient because clinicians can take into account more information about the specific individual patient. I would contend that it is unrealistic to expect that clinicians keep pace with new knowledge in medicine given how fast and how much it is generated. Furthermore, clinicians, like all other people, are unlikely to be fully rational in their decision-making process.

Most fascinating was Section 6 on decision theory under uncertainty. Manski focussed on the minimax-regret criterion. I had not heard about these approaches before, so Manski’s explanations were quite the eye-opener.

Manksi concludes by recommending that central health care planners take a portfolio approach to their guidelines (adaptive diversification), coupled with the minimax criterion to update the guidelines as more information emerges (adaptive minimax-regret). Whether the minimax-regret criterion is the best is a question that I will leave to better brains than mine. A more immediate question is how feasible it is to implement this adaptive diversification, particularly in instituting a process in that data are systematically collected and analysed to update the guideline. In his response, Manski suggests that specialists in decision analysis should become members of the multidisciplinary clinical team and to teach decision analysis in Medicine courses. This resonates with my own view that we need to do better in helping people using information to make better decisions.

Credits

Thesis Thursday: Till Seuring

On the third Thursday of every month, we speak to a recent graduate about their thesis and their studies. This month’s guest is Dr Till Seuring who graduated with a PhD from the University of East Anglia. If you would like to suggest a candidate for an upcoming Thesis Thursday, get in touch.

Title
The economics of type 2 diabetes in middle-income countries
Supervisors
Marc Suhrcke, Max Bachmann, Pieter Serneels
Repository link
https://ueaeprints.uea.ac.uk/63278/

What made you want to study the economics of diabetes?

I was diagnosed with type 1 diabetes when I was 18. So while looking for a topic for my master’s thesis in development economics, I was wondering about how big of a problem diabetes – in particular, type 2 diabetes – would be in low- and middle-income countries (LMICs), because I had never heard about it during my studies. Looking for data I found some on Mexico, where, as it turned out, diabetes was a huge problem and ended up writing my master’s thesis on the labour market effects of diabetes in Mexico. After that, I worked at the International Diabetes Federation as a health economist in a junior position for about a year and a half and at one of their conferences met Prof Marc Suhrcke, who is doing a lot of global health and non-communicable disease related work. We stayed in contact and in the end he offered me the possibility to pursue a PhD on diabetes in LMICs. So this is how I ended up at the University of East Anglia in Norwich studying the economics of diabetes.

Which sources of data did you use for your analyses, and how was your experience of using them?

I exclusively used household survey data that was publicly available. In my master’s thesis, I had already worked with the Mexican Family Life Survey, which is quite an extensive household survey covering many socioeconomic as well as health-related topics. I ended up using it for two of my thesis chapters. The nice thing about it is that it has a panel structure now with three waves, and the last waves also included information on HbA1c levels – a biomarker used to infer on blood glucose levels over the last three months – that I could use to detect people with undiagnosed diabetes in the survey. The second source of data was the China Health and Nutrition Survey, which has many of the same qualities, with even more waves of data. There are more and more surveys with high-quality data coming out so it will be exciting to explore them further in the future.

How did you try to identify the effects of diabetes as separate from other influences?

As in many other fields, there is great worry that diabetes might be endogenous when trying to investigate its relationship with economic outcomes. For example, personal characteristics (such as ambition) could affect your likelihood to be employed or your wage, but maybe also your exercise levels and consequently your risk to develop diabetes. Unfortunately, such things are very difficult to measure so that they often remain unobserved. Similarly, changes in income or job status could affect lifestyles that in turn could change the risk to develop diabetes, making estimates prone to selection biases and reverse causality. To deal with this, I used several strategies. In my first paper on Mexico, I used a commonly used instrumental variable strategy. My instrument was parental diabetes and we argued that, given our control variables, it was unrelated to employment status but predicted diabetes in the children due to the genetic component of diabetes. In the second paper on Mexico, I used fixed effects estimation to control for any time-invariant confounding. This strategy does not need an instrument, however, unobserved time-variant confounding or reverse causality may still be a problem. I tackled the latter in my last paper on the effect of diabetes on employment and behavioural outcomes in China, using a methodology mainly used in epidemiology called marginal structural models, which uses inverse probability weighting to account for the selection into diabetes on previous values of the outcomes of interest, e.g. changes in employment status or weight. Of course, in the absence of a true experiment, it still remains difficult to truly establish causality using observational data, so one still needs to be careful to not over-interpret these findings.

The focus of your PhD was on middle-income countries. Does diabetes present particular economic challenges in this setting?

Well, over the last 30 years many middle-income countries, especially in Asia but also Latin America, have gone from diabetes rates much below high-income countries to surpassing them. China today has about 100 million people with diabetes, sporting the largest diabetes population worldwide. While, as countries become richer, first the economically better-off populations tend to have a higher diabetes prevalence, in many middle-income countries diabetes is now affecting, in particular, the middle class and the poor, who often lack the financial resources to access treatment or to even be diagnosed. Consequently, many remain poorly treated and develop diabetes complications that can lead to amputations, loss of vision and cardiovascular problems. Once these complications appear, the associated medical expenditures can represent a very large economic burden, and as I have shown in this thesis, can also lead to income losses because people lose their jobs.

What advice would you give to policymakers looking to minimise the economic burden of diabetes?

The policy question is always the most difficult one, but I’ll try to give some answers. The results of the thesis suggest that there is a considerable economic burden of diabetes which disproportionately affects the poor, the uninsured and women. Further, many people remain undiagnosed and some of the results of the biomarker analysis I conducted in one of my papers suggest that diagnosis likely often happens too late to prevent adverse health outcomes. Therefore, earlier diagnosis may help to reduce the burden, the problem is that once people are diagnosed they will also need treatment, and it appears that even now many do not receive appropriate treatment. Therefore, simply aiming to diagnose more people will not be sufficient. Policymakers in these countries will need to make sure that they will also be able to offer treatment to everybody, in particular the disadvantaged groups. Otherwise, inequities will likely become even greater and healthcare systems even more overburdened. How this can be achieved is another question and more research will be needed. Promising areas could be a greater integration of diabetes treatment into the existing health care systems specialised in treating communicable diseases such as tuberculosis, which often are related to diabetes. This would both improve treatment and likely limit the amount of additional costs. Of course, investments in early life health, nutrition and education will also help to reduce the burden by improving health and thereby economic possibilities, so that people may never become diabetic or at least have better possibilities to cope with the disease.