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.
Placing values on statistical life years has important public policy implications in measuring who benefits and how much they benefit from interventions. The authors of this study provide what they describe as the most comprehensive evidence to date against a constant value for a statistical life year, an assumption they argue is also applicable when calculating QALYs. Using a Spanish household survey collected over a large sample size (approximately 6,000 individuals), the authors study the relationship between willingness to pay (WTP) and age, by estimating individual WTP for reduction in risk of mortality due to acute myocardial infarction. Three different WTP elicitation procedures were performed. Parametric, semi-nonparametric and non-parametric models using marginal and total approaches were applied to understand the relationship using many alternative methods. Binary variables for income (proxied on a measure of self-perceived social status), education (>lower secondary level) and gender were also included as controls for the models. The results of the linear model show that WTP is lower as age increases. Those with higher income (i.e. social status) and education have higher WTP, while gender is not significant in any model. Sensitivity tests were as hypothesised. The non-parametric model produces similar results to the others, albeit with a higher senior discount. The senior discount is not independent of the income variable. From this, the authors estimate the value of a statistical life year for an 85 year old to be 3.5 times higher than that of a 20 year old. The authors are keen to highlight the strengths of their findings with a large sample size allowing for the robustness of results to be tested across a number of different model types. However, they do flag up the lack of comparability with previous studies that have focused on risk reductions with a lower probability of mortality. The assumption that the authors make that their findings for life years have direct applicability for QALYs is somewhat questionable, particularly for non-acute conditions and QALYs calculated for them. The rationale behind the three types of preference elicitation methods and how/why they were chosen is not apparent in the paper itself. The social status measure they use as a proxy for income is also questionable, and appeared to be applied to maximise sample size. If data for real income was used or imputation of income was included for missing data, it would be interesting to see what impact this may have had on their study findings.
Preferences for public involvement in health service decisions: a comparison between best-worst scaling and trio-wise stated preference elicitation techniques. European Journal of Health Economics [PubMed] Published 10th December 2016
Public involvement in health care is something that has become increasingly recognised as important to do and to be informed by public perspectives when making important decisions for their community. How and where that public involvement should feed into decision making is less well understood. In this study, the authors compare two methods, best worst scaling (BWS) case 2, and a new method the authors call ‘trio-wise’ where the choice task is presented in an equilateral triangle. Using ‘trio-wise’, respondents are able to click in any part of the triangle; this the authors argue gives additional insight on the strength of a respondent’s preferences and also accommodates indifferent preferences. Public preferences are sought using these two methods to understand what aspects of public involvement are most important. Eight general characteristics are included in the exercises. Respondents completed either BWS or the ‘trio-wise’ task (not both) using web based surveys. Approximately 1,700 individuals per arm were sampled. Only three of the eight general characteristics could be completed at any one time due to the trio-wise triangle approach. There was some evidence of position bias for both exercises. The authors say that weak preferences were observed using the trio-wise approach but this could be due to difficulty participants faced in choosing which generic characteristic was more important without further information. Impact and focus of public involvement are found to be the most important characteristics across both BWS and trio-wise. The authors find preference intensity has no bearing on choice probabilities, but this could be an artefact of the weak preferences observed in the sample. Although I can see the appeal of using the trio-wise approach when there are only three characteristics, BWS is advantageous in tasks with more characteristics. Indeed it feels that the findings from this experiment were impeded by the use of the trio-wise approach when much more useful information on guiding future public involvement practice could have been gathered using either BWS or a discrete choice experiment (DCE) across all eight characteristics and the options of public involvement within each characteristic.
How do individuals value health states? A qualitative investigation. Social Science & Medicine [PubMed] Published 22nd November 2016
The valuation tasks of health states used to generate QALYs have been previously found to be complex tasks for members of the general public to complete, who have little experience of such health states. This qualitative study seeks to gain a better understanding as to how the general public complete such tasks. Using a purposive sample, 21 individuals were asked to complete eight DCEs and three TTO tasks, based on the EQ-5D-5L valuation protocol. Participants were asked to complete the valuation tasks using think aloud, followed by semi-structured interviews. Three main themes emerged from the framework analysis undertaken on the interview transcripts. Firstly, individuals had to interpret a health state, using their imagination and experience to help visualise a realistic health state with those problems. Knowledge, understanding of descriptive system, additional information for a health state, re-writing of health states and problems with EQ-5D labels all impacted this process. The second theme was called conversion factors, which the authors took to mean in this context the personal and social factors that affected how participants valued health states. Personal interests, values and circumstances were said to have an effect on the interpretation of a health state. The final theme was based on the consequences of health states, that tended to focus on non-health effects caused by health problems, such as activities, enjoyment, independence, relationships, dignity and avoiding being a burden. The authors subsequently developed a three-stage explanatory account as to how people valued health states based on the interview findings. Although I would have some concerns about the generalisability of these findings to general public valuation studies, given the highly educated sample, it does highlight some issues about what health economists might implicitly think individuals are doing when completing such tasks compared to what they actually are doing. There are clearly problems for individuals completing such hypothetical health states, with the authors suggesting a more reflective and deliberative approach to overcome such problems. The authors also raise an interesting comment as to whether participants actually do weigh the consequences of health states and follow compensatory decision-making or instead are using simplifying heuristics based on one attribute, which I agree is an area that requires further investigation.