Every Monday our authors provide a round-up of the latest peer-reviewed journal publications. We cover all issues of major health economics journals as well as some other notable releases. 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.
Volume 21, Issue 7
The latest issue of EJHE manages to cover many of my research interests, with papers on mental health, valuing ‘dead’, and screening for diabetic retinopathy. There’s also plenty beyond my areas of expertise, which I won’t cover here.
But first, leading the way is the obligatory COVID-19 article. This one is on Europeans’ willingness to be vaccinated. The paper has garnered quite a bit of attention because it suggests that around 1 in every 14 people in Europe wouldn’t want a vaccine.
As regular readers will know, I recently finished my PhD, for which I conducted research on screening for diabetic retinopathy. The PhD was inspired in part by a growing body of literature showing that extended screening intervals are safe for many people and, therefore, cost-effective. This issue includes another article to add to that pile. The authors estimated the cost-effectiveness of biennial (rather than annual) screening in Wales using a Markov-style decision model. They estimated decremental cost-effectiveness ratios, which is a term I don’t think I’ve seen written down before, and found that the magnitude of cost savings is worth the very small increase in risk for screening attenders. But various sensitivity analyses show that extending screening intervals is not likely to be cost-effective for people with certain risk factors. This is why a risk-based screening approach is most certainly the way forward.
Covering another of my main research interests – mental health – is a study from Austria. The authors report on findings from a survey with a representative sample of people, who were asked a battery of questions in face-to-face interviews. Predictably, the study finds much higher resource use and costs for people with diagnosed mental disorders, and even more so for those with severe illness. The researchers also tried to control for physical comorbidities in their modelling and found mixed results for different categories of costs. The reported numbers will be valuable to researchers and decision-makers alike. Wisely, the authors seem to avoid coming up with a cost-of-illness style estimate for Austria. Unwisely, they include a pie chart in their results.
Bringing us up to date with my topic du jour (the unbearable lightness of being dead) is a paper on anchoring for the EQ-5D-Y. The prospect of a dead child, even in a hypothetical scenario, is troubling. This makes the process of valuing health states, relative to being dead, very difficult. If you ask me, we shouldn’t be presenting people with such trade-offs in the first place. But, until everyone agrees that we should drop dead, a solution is needed to estimate QALYs. In this study, the researchers compared four different approaches in the context of discrete choice experiments (DCEs): i) using a visual analogue scale, ii) using a duration attribute in the DCE, iii) using the lag-time time trade-off method, and iv) something called the location-of-dead approach, which is based on a series of paired comparisons between health states and being dead. These alternatives were tested in interviews in which adult participants were asked to adopt an adult’s perspective as well as that of a 10-year-old child. In general, the four approaches were pretty similar, so the authors provide some guidance on how to choose your methodology.
On a similar topic, this issue also includes a systematic review of 38 studies that used discrete choice experiments for health state valuation. The included studies exhibited a lot of methodological variation, which has been shown before in a review that I discussed in a previous blog post.
Finally, I was also drawn to a couple of papers that are – nominally – about costing.
The first is about measuring intangible costs associated with substance dependence, though it seems to have nothing to do with costs. Rather, the study compares the extent to which alternative generic measures (EQ-5D, SF-6D) reflect the quality of life burden of substance dependence when a sample is compared to the general population.
The second is about the burden of caregiving for Duchenne muscular dystrophy. Rare genetic disorders in childhood (such as DMD) are associated with a lot of costly spillovers and data are often lacking. This study provides estimates of financial burden (from an inevitably small sample) according to various clinical indicators.
Volume 20, Issue 3
The latest issue of IJHEM includes five articles on a diverse range of topics.
Of most interest to me is an article on economic fluctuations and mental health. The authors used variation in economic conditions across Catalonia’s regions as their testbed. Overall, there was no significant impact of local economic fluctuations on self-reported mental health or on the use of anti-depressants or anti-anxiety drugs. By chopping up the sample, the researchers do find some significant effects around reduced sleep, but this seems a bit speculative given how much testing was conducted on the sample.
There are a couple of articles on the demand for care in the US.
One study asks whether uninsured people demand less. The author homes in on readmissions for people whose insurance status changes between visits. The headline result is that people who become uninsured use 6% fewer services. This is likely to be a lower estimate because the study is conducted in Maryland where uninsured people pay much lower prices.
Another study looks at how evidence on the ineffectiveness of treatments influences demand. Dishearteningly for those of us who like evidence, the two case studies – fenofibrate and dronedarone – exhibited a slow change in practice. But the key contribution of the paper is in exploring the characteristics that were associated with a faster or slower change in practice. The Medicaid and Medicare enrolment status of the patient, and the gender and specialty of the provider, were correlated with the rate at which prescription of the drugs decreased. The findings also suggest that a drug that is shown to be unsafe will be de-adopted more quickly than a drug that is simply shown to be ineffective.