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. If you’d like to write one of our weekly journal round-ups, get in touch.
As far as I can tell, this issue is not a ‘special’ on COVID-19, but it looks a lot like one. Of the 17 articles that are included, eight focus on issues relating to COVID-19.
One article addresses a point previously discussed on this blog; the relative importance of lockdown and fear in explaining the economic decline. It provides ammunition in the battle against the naive perspective that pandemic response involves a straight trade-off between health and the economy. The study uses mobile phone records to identify people’s visits to places of business across time and different locations in the US, modelling the average number of visits to each business per week. The researchers had access to data from March to May 2020 for about 10% of mobile devices in the US and more than 2 million business locations. The headline finding is that consumer traffic fell by 60 percentage points and that local lockdown policies explain only 7 of those percentage points. Within industries, the analysis found that people switched to visiting smaller and less busy stores. The findings are by no means conclusive – the data can’t reveal changes in expenditure per visit – but they are broadly supportive of the hypothesis that ‘it’s the virus, stupid‘.
The determinants of the success of policy is a key area where learnings from the pandemic could be of vital importance in the future. Another article using mobile phone data examines the importance of civic duty in adherence to social distancing. The analysis covers the US and European countries and civic duty is measured by indicators such as voter participation and recognised indices of social capital. Areas with greater civic duty exhibited stronger and more prolonged adherence to social distancing. For instance, survey data from the US showed that people in counties with higher civic duty reported being more likely to wear a mask. There is also an article on public trust in science, with the sinister but satisfying title ‘revenge of the experts’. The researchers used data from a large survey to show that people exposed to epidemics between the ages of 18-25 have less trust in science. This distrust is concentrated among people with less science education and translates into poorer compliance with health policies.
Something that is now becoming clear in the research is the ways in which the impact of COVID-19 and policy responses have had hetergeneous – and inequitable – impacts. A study in this issue looks at inequalities in the impact of COVID-19 in terms of labour market outcomes. The researchers used the UK’s Understanding Society household panel survey, which includes additional COVID-related questions in 2020. The analysis shows that people in the bottom income quintiles before the pandemic experienced the biggest income shocks, and that a common mitigation strategy was to borrow money from friends and family. Another study highlights inequalities in the demand for online learning in the US, based on Internet search data. Areas with higher incomes and better Internet access saw greater increases in activity, suggesting the potential for the reinforcement of inequalities in educational attainment along these lines.
Using Internet search data is a popular strategy for timely research outputs nowadays. Another study in this issue used Google Trends data to identify the impact on well-being of lockdown policies. The results are a mixed bag, with searches relating to loneliness, sadness, and worry increasing in prevalence while stress and suicide declined.
And there is yet more on COVID-19, with a study on trends in government employment during the pandemic and a study on the relationship between COVID-19 and migration in South Asia. Beyond COVID, there is a study that looks at intergenerational mobility in health. Non-health topics covered include four-day school weeks, discrimination in student evaluations of teaching, responses to inheritance, family and government insurance, populism, political responses to violence, and social assistance and employment in Germany and Uruguay.
Volume 39, Issue 1
The latest issue of PharmacoEconomics opens with a pair of papers on sensitivity analysis in cost-effectiveness modelling. The first is a review of recent advances in deterministic sensitivity analysis (DSA). The authors summarise the key limitations of the ‘classic’ approach and outline three developments that address these, namely i) stepwise DSA, ii) probabilistic DSA, and iii) distributional DSA. This paper is followed by another (which I peer-reviewed) that introduces the conditional net benefit frontier, a tool to use alongside probabilistic one-way sensitivity analysis (what the authors of the first paper call probabilistic DSA). The cNBF is a familiar-looking visualisation, which plots the net benefit for an intervention at each centile of a probabilistically sampled parameter value.
This issue includes a couple of quite big reviews of economic evaluations. One sought to identify all economic evaluations conducted alongside stepped-wedge trials. For the benefit of the uninitiated, these are trials where the intervention is administered in waves rather than a simple 1:1 simulataneous randomisation. The authors of the review sought to identify methodological practices in this context. The review included 69 studies, the vast majority of which were protocols. Of the 15 articles reporting results, it was common for the reporting to fail to mention key methods, such as accounting for correlation between costs and outcomes and adjusting for time in the analysis. Given the high number of protocols identified, there is a good opportunity for this study to influence the analysis and reporting of studies currently in progress.
Another review in this issue considers the inclusion of productivity effects in economic evaluations of drugs, which identified 208 studies. Almost all studies included absenteeism, about 15% included presenteeism, and a few captured other aspects of productivity loss. Some studies (31) did not state what aspects of productivity were included, but costed them anyway. Most studies (75%) did not state whether they used a human capital approach or a friction cost approach. As well as extracting information about the methods used to include productivity estimates, the authors summarise associations between disease context and the influence of productivity estimates. In most cases, including productivity costs was favourable to the reported ICER, as you might expect.
A couple of papers in this issue report on choice tasks. One study explores drug coverage preferences in New Zealand. The researchers used an analytical hierarchy process with a survey of 500 people, in which participants were asked to make trade-offs between treatments for rare and common conditions. Overall, the findings support the notion that rarity is not important in itself; coverage should be determined on the basis of treatment effectiveness, quality of life, and equity. Another study reports on a best-worst scaling study of the relative importance of education and crime outcomes in economic evaluation. The online survey was with 39 health economists and HTA experts, who were asked to choose the most important and least important from a set of five intersectoral costs and benefits. These were things like prison expenditures and absenteeism from school. The study reports a ranking of the different items according to importance, to inform future studies. But it also reveals that most of the respondents don’t practice what they preach.
This issue also includes a study on synthesising estimates of relative effects for sequential treatments and just one applied analysis; a budget impact analysis in the context of acute myeloid leukemia. Finally, there is a comment and a reply for a previously published study on the Core Obesity Model.