Alastair Canaway’s journal round-up for 10th June 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.

Analytic considerations in applying a general economic evaluation reference case to gene therapy. Value in Health Published 17th May 2019

For fledgling health economists starting in the world of economic evaluation, the NICE reference case is somewhat of a holy text. If in doubt, check the reference case. The concept of a reference case for economic evaluation has been around since the first US Panel on Cost-Effectiveness in Health and Medicine in 1996 and NICE has routinely used its own reference case for well over a decade. The primary purpose of the reference case is to improve the quality and comparability of economic evaluations by standardising methodological practices. There have been arguments made that the same methods are not appropriate for all medical technologies, particularly those in rare diseases or where no treatment currently exists. The focus of this paper is on gene therapy: a novel method that inserts genetic material into cells (as opposed to a drug/surgery) to treat or prevent disease. In this area there has been significant debate as to the appropriateness of the reference case and whether a new reference case is required in this transformative but expensive area. The purpose of the article was to examine the characteristics of gene therapy and make recommendations on changes to the reference case accordingly.

The paper does an excellent job of unpicking the key components of economic evaluation in relation to gene therapy to examine where weaknesses in current reference cases may lie. Rather than recommend that a new reference case be created, they identify specific areas that should be paid special attention when evaluating gene therapy. Additionally, they produce a three part checklist to help analysts to consider what aspects of their economic evaluation they should consider further. For those about to embark on an economic evaluation of a gene therapy intervention, this paper represents an excellent starting point to guide your methodological choices.

Heterogeneous effects of obesity on mental health: evidence from Mexico. Health Economics [PubMed] [RePEc] Published April 2019

The first line of the ‘summary’ section of this paper caught my eye: “Obesity can spread more easily if it is not perceived negatively”. This stirred up contradictory thoughts. From a public health standpoint we should be doing our utmost to prevent increasing levels of obesity and their related co-morbidities, whilst simultaneously we should be promoting body positivity and well-being for mental health. Is there a tension here? Might promoting body positivity and well-being enable the spread of obesity? This paper doesn’t really answer that question, instead it sought to investigate whether overweight and obesity had differing effects on mental health within different populations groups.

The study is set in Mexico which has the highest rate of obesity in the world with 70% of the population being overweight or obese. Previous research suggests that obesity spreads more easily if not perceived negatively. This paper hypothesises that this effect will be more acute among the poor and middle classes where obesity is more prevalent. The study aimed to reveal the extent of the impact of obesity on well-being whilst controlling for common determinants of well-being by examining the impact of measures of fatness on subjective well-being, allowing for heterogeneous effects across differing groups. The paper focused only on women, who tend to be more affected by excess weight than men (in Mexico at least).

To assess subjective well-being (SWB) the General Health Questionnaire (GHQ) was used whilst weight status was measured using waist to height ratio and additionally an obesity dummy. Data was sourced from the Mexican Family and Life Survey and the baseline sample included over 13,000 women. Various econometric models were employed ranging from OLS to instrumental variable estimations, details of which can be found within the paper.

The results supported the hypothesis. They found that there was a negative effect of fatness on well-being for the rich, whilst there was a positive effect for the poor. This has interesting policy implications: policy attempt to reduce obesity may not work if excess weight is not perceived to be an issue. The findings in this study imply that different policy measures are likely necessary for intervening in the wealthy and the poor in Mexico. The paper offers several explanations as to why this relationship may exist, ranging from the poor having lower returns from healthy time (nod to the Grossman model), to differing labour market penalties from fatness due to different job types for the rich and the poor.

Obviously there are limits to the generalisability of these findings, however it does raise interesting questions about how we should seek to prevent obesity within different elements of society, and the unintended consequences that shifts in attitudes may have.

ICECAP-O, the current state of play: a systematic review of studies reporting the psychometric properties and use of the instrument over the decade since its publication. Quality of Life Research [PubMed] Published June 2019

Those who follow the methodological side of outcome measurement will be familiar with the capability approach, operationalised by the ICECAP suite of measures amongst others. These measures focus on what people are able to do, rather than what they do. It is now 12-13 years since the first ICECAP measure was developed: the ICECAP-O designed for use in older adults. Given the ICECAP measures are now included within the NICE reference case for the economic evaluation of social care, it is a pertinent time to look back over the past decade to assess whether the ICECAP measures are being used and, if so, to what degree and how. This systematic review focusses on the oldest of the ICECAP measures, the ICECAP-O, and examines whether it has been used, and for what purpose as well as summarising the results from psychometric papers.

An appropriate search strategy was deployed within the usual health economic databases, and the PRISMA checklist was used to guide the review. In total 663 papers were identified, of which 51 papers made it through the screening process.

The first 8 years of the ICECAP-O’s life is characterised by an increasing amount of psychometric studies, however in 2014 a reversal occurred. Simultaneously, the number of studies using the ICECAP-O within economic evaluations has slowly increased, surmounting the number examining the psychometric properties, and has increased year-on-year in the three years up to 2018. Overall, the psychometric literature found the ICECAP-O to have good construct validity and generally good content validity with the occasional exception in groups of people with specific medical needs. Although the capability approach has gained prominence, the studies within the review suggest it is still very much seen as a secondary instrument to the EQ-5D and QALY framework, with results typically being brief with little to no discussion or interpretation of the ICECAP-O results.

One of the key limitations to the ICECAP framework to date relates to how economists and decision makers should use the results from the ICECAP instruments. Should capabilities be combined with time (e.g. years in full capability), or should some minimum (sufficient) capability threshold be used? The paper concludes that in the short term, presenting results in terms of ‘years of full capability’ is the best bet, however future research should focus on identifying sufficient capability and establishing monetary thresholds for a year with sufficient capability. Given this, whilst the ICECAP-O has seen increased use over the years, there is still significant work to be done to facilitate decision making and for it to routinely be used as a primary outcome for economic evaluation.

Credits

Alastair Canaway’s journal round-up for 10th July 2017

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.

Use-of-time and health-related quality of life in 10- to 13-year-old children: not all screen time or physical activity minutes are the same. Quality of life Research [PubMedPublished 3rd July 2017

“If you watch too much TV, it’ll make your eyes square” – something I heard a lot as a child. This first paper explores whether this is true (sort of) by examining associations between aspects of time use and HRQL in children aged 10-13 (disclaimer: I peer reviewed it and was pleased to see them incorporate my views). This paper aims to examine how different types of time use are linked to HRQL. Time use was examined by the Multimedia Activity Recall for Children and Adolescents (MARCA) which separates out time into physical activity (sport, active transport, and play), screen time (TV, videogames, computer use), and sleep. The PedsQL was used to assess HRQL, whilst dual x-ray absorptiometry was used to accurately assess fatness. There were a couple of novel aspects to this study, first, the use of absorptiometry to accurately measure body fat percentage rather than the problematic BMI/skin folds in children; second, separating time out into specific components rather than just treating physical activity or screen time as homogeneous components. The primary findings were that for both genders, fatness (negative), sport (positive) and development stage (negative) were associated with HRQL. For boys, the most important other predictor of HRQL was videogames (negative) whilst predictors for girls included television (negative), active transport (negative) and household income (positive). With the exception of ‘active travel’ for girls, I don’t think any of these findings are particularly surprising. As with all cross-sectional studies of this nature, the authors give caution to the results: inability to demonstrate causality. Despite this, it opens the door for various possibilities for future research, and ideas for shaping future interventions in children this age.

Raise the bar, not the threshold value: meeting patient preferences for palliative and end-of-life care. PharmacoEconomics – Open Published 27th June 2017

Health care ≠ end of life care. Whilst health care seeks to maximise health, can the same be said for end of life care? Probably not. This June saw an editorial elaborating on this issue. Health is an important facet of end of life care. However, there are other substantial objects of value in this context e.g. preferences for place of care, preparedness, reducing family burdens etc. Evidence suggests that people at end of life can value these ‘other’ objects more than health status or life extension. Thus there is value beyond that captured by health. This is an issue for the QALY framework where health and length of life are the sole indicators of benefit. The editorial highlights that this is not people wishing for higher cost-per-QALY thresholds at end of life, instead, it is supporting the valuation of key elements of palliative care within the end of life context. It argues that palliative care interventions often are not amenable to integration with survival time in a QALY framework, this effectively implies that end of life care interventions should be evaluated in a separate framework to health care interventions altogether. The editorial discusses the ICECAP-Supportive Care Measure (designed for economic evaluation of end of life measures) as progress within this research context. An issue with this approach is that it doesn’t address allocative efficiency issues (and comparability) with ‘normal’ health care interventions. However, if end of life care is evaluated separately to regular healthcare, it will lead to better decisions within the EoL context. There is merit to this justification, after all, end of life care is often funded via third parties and arguments could, therefore, be made for adopting a separate framework. This, however, is a contentious area with lots of ongoing interest. For balance, it’s probably worth pointing out Chris’s (he did not ask me to put this in!) book chapter which debates many of these issues, specifically in relation to defining objects of value at end of life and whether the QALY should be altogether abandoned at EoL.

Investigating the relationship between costs and outcomes for English mental health providers: a bi-variate multi-level regression analysis. European Journal of Health Economics [PubMedPublished 24th June 2017

Payment systems that incentivise cost control and quality improvements are increasingly used. In England, until recently, mental health services have been funded via block contracts that do not necessarily incentivise cost control and payment has not been linked to outcomes. The National Tariff Payment System for reimbursement has now been introduced to mental health care. This paper harnesses the MHMDS (now called MHSDS) using multi-level bivariate regression to investigate whether it is possible to control costs without negatively affecting outcomes. It does this by examining the relationship between costs and outcomes for mental health providers. Due to the nature of the data, an appropriate instrumental variable was not available, and so it is important to note that the results do not imply causality. The primary results found that after controlling for key variables (demographics, need, social and treatment) there was a minuscule negative correlation between residual costs and outcomes with little evidence of a meaningful relationship. That is, the data suggest that outcome improvements could be made without incurring a lot more cost. This implies that cost-containment efforts by providers should not undermine outcome-improving efforts under the new payment systems. Something to bear in mind when interpreting the results is that there was a rather large list of limitations associated with the analysis, most notably that the analysis was conducted at a provider level. Although it’s continually improving, there still remain issues with the MHMDS data: poor diagnosis coding, missing outcome data, and poor quality of cost data. As somebody who is yet to use MHMDS data, but plans to in the future, this was a useful paper for generating ideas regarding what is possible and the associated limitations.

Credits

Thesis Thursday: Ayesha Ali

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 Ayesha Ali who graduated with a PhD from Lancaster University. If you would like to suggest a candidate for an upcoming Thesis Thursday, get in touch.

Title
Essays on health economics: trans fat policies, commuting, physical activity, and body mass index in the US
Supervisors
Colin Green, Bruce Hollingsworth
Repository link
http://www.research.lancs.ac.uk/portal/en/publications/-(182eadeb-a873-4d45-93bc-a987899c6587).html

What drew you to this particular topic and made you want to dedicate your PhD to it?

I’ve always been very fascinated about how people make decisions when it comes to health-related behaviors. My Mom was a doctor, so health has always been a main driver in what our family ate and what sort of values my parents emphasized. Growing up, I think we were always kind of “the weird family” in the neighborhood and in our extended family, because we often put health before cultural norms (i.e. changing traditional recipes to be healthier, avoiding all of those colorful and fun children’s cereals and candies, etc.), so I’ve been very aware and very curious about what drives health-related choices for different people. I’ve also lived in a number of very different communities, where norms about behaviors related to health varied significantly from place to place, and that variation has always been something that I’ve sort of observed and wondered about as well. So, I think these observations are what drew me to economics, as a way of understanding choices. My thesis work was somewhat of an attempt to crack the surface.

I don’t think I got the chance in my thesis to dig as deeply into the issue where health behaviors and culture meet as I had wanted, but I was lucky to get to use some of the types of data and some of the econometric tools that will come in handy as I continue to explore this area.

Your study focussed on US data, but you studied at a UK university. Was this a help or a hindrance?

I think it was definitely a challenge to work with non-UK data while studying at a UK University; for example, I didn’t find certain types of support (in terms of local workshops or having other colleagues using the same data) or familiarity that was often easily available to someone using UK data. However, my supervisors were very supportive of me and there were students in my cohort using other non-UK data so I was not alone in that regard.

One challenge that I had was that I couldn’t just ask someone who uses the data, but instead had to spend a bit of time trying to understand how other researchers in the literature used the data and then try to figure out whether or not their assumptions and ways of using the data were applicable to my work. I spent a lot of time reading data documentation files — both of my datasets (NHANES and ATUS) have really good online resources that I’m now very familiar with! I think that a lot of this being “on my own” with the data helped me to develop a feeling of confidence that I may not have had otherwise. I’m somewhat prone to second-guessing myself, so being able to learn to have confidence in my work was really valuable.

So, although challenging, overall I would say it was definitely more of a help than a hindrance.

Methodologically, what was the most challenging aspect of the research?

In my third chapter I use time use data, a two-part model, and a recursive bivariate probit model to estimate physical activity participation and duration decisions given an individual’s commuting time, with an instrument for commuting. There are some conflicting ideas on how best to use time-use data in the literature (see: Franzis and Stewart, 2010; Gershuny, 2012; Stewart, 2013) and few examples of instruments for commuting (see: Baum-Snow, 2007; Gimenez-Nadal and Molina, 2011).

I received a lot of different feedback on the best estimation approach to use. I wanted to estimate participation and intensity elasticities for individuals who do physical activity on a given day. There are a number of ways to deal with these two questions separately or simultaneously and also to deal with the large number of zeros in the data (i.e. individuals who did not participate in physical activity). At one point, I remember that I thought I had everything figured out with this chapter; and then following a conference presentation of this work, one commenter was dead set that I should be using a switching model. I hadn’t considered that approach and wasn’t familiar with it at all, so I had to go look it up, figure out what it was and whether or not it worked with what I was doing. So, just figuring out the best way to deal with my data and with the questions I was asking in this third chapter were probably the most challenging part of the thesis for me.

If you could have a decision maker implement one policy change supported by your work, what would it be?

If I could ideally have policy makers do one thing that is supported by my work, I think it would actually be a rather general thing, not related to one specific policy, but sort of related to the entire approach of policy-making. I would want policy makers to consider the groups they are targeting with more care. For example, my third chapter looks at time use among obese individuals and healthier-weight individuals and finds that many decisions, such as the decision of where to live and work, are often driven by different factors in these different groups. In general, my work suggests that different groups may be driven by different motivating factors and if we don’t understand what these are, policies might not successfully reach those who could most benefit. That being said, this probably isn’t as easy as it sounds, as there are a lot of political influences on how policy decisions are made.

If you had to do it all again (perish the thought), is there anything you would have done differently?

I think overall, I’m really happy with my experience; I had some good resources and Lancaster was a great environment for me. My PhD cohort was close-knit and faculty in my department were very approachable and supportive. If anyone is interested, I also found McCloskey’s Economical Writing and Thomson’s Guide for the Young Economist to be really helpful at various stages of the PhD.

If I could have changed one thing though, it was how I dealt with insecurity. Even in such a supportive environment, the competitive nature of academia can contribute to feelings of insecurity; I worked hard to recognize my own insecurities and fight through them, but I didn’t always succeed. So, if I could do something differently, I would like to have been less afraid to own up to not knowing something and to just keep asking questions until I understood.

I really enjoyed having the chance to talk with you about my experiences and motivation. And I hope that if anyone can find my experiences useful to them at their stage of the process, they do. Thanks again for inviting me.