Economic evaluation, screening for prostate cancer, and 'value for money'?

Economic evaluation has now been applied to health care for over 50 years, sometimes to good effect, sometimes for ill. This study seeks to give an understanding of what ‘economic evaluation’ can offer decision-makers, but also sets out to acknowledge its problems and pitfalls. In addition, this thesis applies one approach to economic evaluation, utilising data from the Finnish Randomised Study of Screening for Prostate Cancer (FinRSPC) after 20 years of the trial. Started in 1996, the FinRSPC is a pragmatic population-based trial investigating invitation to prostate-specific antigen (PSA) testing as a basis for mass screening to detect prostate cancer. Although the last invitations to screening were sent in 2007, health-care registers provide almost complete and almost continuous follow-up of some of the subsequent costs and effects, for up to 20 years in total.

This thesis includes four peer-reviewed articles related to economic evaluation or the FinRSPC, as well as a synthesis of those articles in a summary, which includes an integrated assessment of the development and current state of economic evaluation in the field of health care. The main rationale for much of this study is that care should be taken when interpreting economic evaluation. This is mainly because the language and terminology surrounding economic evaluation is often ambiguous and it often oversells what research can credibly offer. One main finding from this study is that there should be humility about our ability to adequately evaluate prostatecancer screening. Data on costs and effects looks likely to remain deficient in some respects, due to problems in measuring outcomes, as well as due to problems in conducting trials on this topic. A second main finding is that economic approaches to evaluation inescapably include their own judgments about value; far from being a value-neutral form of assessment, these values determine its scope, content and appropriateness, to an important extent. Together with other findings, this study suggests a more classical approach to valuation would be useful: deliberative analysis. Such deliberative processes would include comprehensive quality assessment of information offered by any economic evaluation, integrating sources of information as and when necessary.

The articles which constitute this thesis each provide some evidence to support the suggestion of greater use of deliberative analysis. The measurement of health-related quality-of-life effects related to prostate-cancer screening in Article I faces challenges due to the timing and mode of data collection, due to the ‘measures’ used, as well as that much of the analysis is prone to biases. Article II mainly focuses on differences in average costs between the arms, as estimated from the accessible healthcare registers, for the men in the FinRSPC after 20 years. However, it should be noted that those estimates cannot account for, e.g., the potential impact of overdiagnosis or variations in treatment practices. Article III surveys a wide range of methodological literature and concludes that language and terminology related to health-economic evaluation should be used with sufficient humility to prevent it being oversold. It strongly recommends that the qualities of any ‘economic’ evidence be thoroughly checked for quality and relevance, especially if that evidence is to be utilised as part of priority setting processes. Article IV uses cost-effectiveness analysis to combine cost information from registers with analysis of data on effectiveness, in terms of different indicators of mortality, from the FinRSPC. The approach to economic evaluation used here revealed a (non-statistically significant) increase in mortality in the screening arm, but perhaps more importantly, it highlighted the uncertainties surrounding the evaluation of overdiagnosis associated with prostate-cancer mass screening.

In summary, although it is often claimed that economic evaluation can be useful in informing decision-making, in practice such claims are conditional on the qualities of the content of each economic evaluation. This thesis provides a rationale for assessing the qualities of all economic evaluations: their credibility should always be checked. Such quality assessment should look carefully at the uncertainties surrounding the evidence base, including any potential for medico-industrial influence. Every stakeholder should bear responsibility for appropriate interpretation of the available evidence, especially if interventions have consequences which are complicated, complex, or are simply not amenable to quantification or robust evaluation. Judging the appropriateness of the information from economic evaluation to any policy question should be approached with humility about what is, or can be, known.

This thesis concludes with three propositions. Firstly, that economic evaluation should continue almost unchanged, with an important exception being that the current overselling of economic evaluation, both in principle and in practice, should be curbed. Secondly, attempts to improve the estimation of the likely costs and effects of interventions should continue, but with increased humility about the extent to which costs and effects can, in the foreseeable future, ever be ‘measured’ comprehensively. Finally, and perhaps most importantly, given that health-economic evaluation easily neglects many important impacts and, given the difficulties in changing language, terminology or research practices, those wishing to use information from economic evaluation should always engage in thorough critical assessment of its qualities.