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 ethics, politics, and economics of medically assisted death is tricky, to say the least. This thought provoking paper sets out six perspectives that seem to contradict economic theory (traditional or behavioural). The six notions that are contradicted are: i) utility maximisation, ii) discounting future losses, iii) coding risk as probability, iv) acting with self-interest, v) valuing outcomes not processes, and vi) ignoring post-mortem states. Plausible scenarios are outlined in each case that challenge basic economic theory. For example, people may sacrifice future time in sub-optimal (but acceptable) health in a way that does not maximise utility. Indeed, it is generally a legal requirement that an individual is competent to make the choice of assisted death, at which moment they may not be in an extremely bad state. The examples are related to the policy context of Canada, where medical assistance in dying has been legal since 2016. The authors conclude that economic theory needs to be strengthened in order to have validity in this context. I’d say we’ve got a long way to go.
Smoking to cope: addictive behavior as a response to mental distress. Journal of Health Economics Published 4th May 2020
People with mental health problems tend to smoke more. Various reasons have been posited for this. In economics, many researchers still turn to the mostly unsatisfactory ‘rational addiction model’ to explain smoking behaviour. But that can’t do much to explain the association between mental health and smoking. This paper builds on such a model in a sensible way, by focussing on a specific context and identifying a plausible mechanism.
The theory is that smoking is a coping mechanism. It’s quite similar to psychological theories of self-medication, differing in that coping implies a response that could be observed in people without ‘clinical’ characteristics. In the short term, smoking can offer neurochemical ‘rewards’ that relieve stress. The utility of smoking might therefore be greater for people experiencing higher levels of distress, such as people with depression. What’s more, adolescents might be expected to be more likely to make the decision to smoke as a coping response, because they tend to underestimate the harms of smoking.
The author analyses data from a US general population cohort study with 3,746 people under the age of 19. The data included multiple waves with information on smoking behaviour and the occurrence of two traumatic life events: experiencing the death of a (non-family) person that they were close to and being a victim of a violent crime. Regression models were used to test for the impact of a traumatic life event on first cigarette use and smoking in the last 30 days. Mental health was captured using the Center for Epidemiologic Studies Depression (CESD) scale.
A recent shock was associated with increased first cigarette use (8.2% points), smoking in the past 30 days (9.1% points), and daily use in the past 30 days (5.9% points). Even though a minority of people reported a traumatic event, the author estimates that the impact of these effects is substantial, with around 5% of first cigarette use in the whole sample being explained by the traumatic events. In support of the author’s proposed coping mechanism, the effect is greater in people with higher levels of ex ante mental distress and smaller for people facing higher cigarette taxes.
I don’t really see how this is any different from the self-medication hypothesis. Despite the author’s desire for this novelty, the paper doesn’t need it for the findings to be useful. The implication is that young people with depression are at high risk of becoming smokers. Reducing their baseline level of distress (say, through access to mental health services) could prevent them from taking up smoking following a traumatic life event.
Efficiency ratio and rocketing drug prices: old concerns and new possibilities. The European Journal of Health Economics [PubMed] [RePEc] Published 5th May 2020
Recent years have seen rising drug prices. This has got lots of people worried about the sustainability of health care. And the benefits achieved from the use of these drugs does not seem to be adequately correlated with their relative prices. In the context of oncology, evidence suggests that very high prices are associated with small benefits. This is the context set out by the authors of this paper who, after a somewhat rambling opening, propose a new basis for setting prices.
In essence, the proposed framework establishes a minimum QALY gain required if the manufacturer is to be allowed to charge a higher price than the comparator. If the QALY gain is less than this minimum, they can charge the same price as the comparator. The higher price that they can charge for a sufficiently effective new medicine is determined by a willingness to pay threshold multiplied by the QALY gains over and above the minimum.
This framework doesn’t make sense to me. It seems to be a solution to a problem that doesn’t exist, because the problem it solves is not the problem that the authors describe. If we think that rising prices are a problem per se, then maybe this is a sensible solution. But they aren’t. Rising prices are only a problem if they are not associated with sufficient gain. A standard cost-effectiveness threshold approach can ensure that gains are sufficient to maintain efficiency. If prices get too high, we don’t buy the drugs. If all the drugs are too expensive it’s because we can achieve greater health outcomes by opting for non-pharmacological treatments. For this framework to work, we would need to be able to identify a minimally important QALY gain. No such thing exists. All QALYs matter, from the first to the last.
Does the EQ-5D usual activities dimension measure what it intends to measure? The relative importance of work, study, housework, family or leisure activities. Quality of Life Research [PubMed] Published 23rd April 2020
Arguably, we don’t know a whole lot about what the EQ-5D is measuring. Partly, that’s because we don’t know what it should be measuring. The selection of the five dimensions of the EQ-5D was pretty unscientific by today’s standards. So this study piqued my interest. What “usual activities” are being captured by that dimension?
The authors analyse data from the Multi-Instrument Comparison study, with 7,933 participants with (or without) various health conditions. Using correlation and regression analyses, the association between the usual activities dimension of the EQ-5D-5L and various other items was measured. The other items were selected from the SF-36 and the AQoL to represent i) work/study activities, ii) housework activities, iii) family activities, and iv) leisure activities. These are all explicitly mentioned on the EQ-5D questionnaire. For reasons I don’t understand (and which aren’t explained) the EQ-5D dimension is dichotomised into no problems (i.e. 1) and any problems (i.e. 2-5). A high correlation was found between each of the indicators of the different types of activities and the EQ-5D dimension. Controlling for sociodemographics and disease status, housework was the most important predictor and family the least important.
Though I do think it’s an interesting research question, the study seems to be conceptually flawed. The items from the SF-36 and the AQoL are not predictors of health-related quality of life as measured by the EQ-5D. They’re items that are attempting to capture (roughly) the same aspects of health-related quality of life as the usual activities dimension of the EQ-5D. This implies quite a different (and less interesting) conclusion to that presented by the authors. Rather than being able to say that the usual activities dimension is mostly predicted by housework, all we can say is that whatever determines an answer to a specific question about housework also determines a question about usual activities. We have no real idea about the importance of housework per se because differences in the strength of association are likely to arise from differences in the framing of the different questions. But even with more appropriate indicators of different types of activities, the authors’ conclusion that the usual activities dimension of the EQ-5D measures what it intends to measure would still be untenable, because we don’t have a clear idea of what it intends to measure.