Setting the 2nd Panel Straight on Discounting
This post summarizes a recent critique published by Paulden et al. (2016) of recommendations on discounting costs and health effects made by the 2nd Panel on Cost-Effectiveness in Health and Medicine. Discounting is used in cost-effectiveness analysis (CEA) to apply less weight to future costs and effects relative to those occurring in the present. Although not immediately intuitive, discounting is important for achieving consistency both with the observed human preference to enjoy good things sooner rather than later and the economic reality that positive interest rates mean there is a return to delaying costs until the future.
Discounting matters because it can influence whether an intervention is cost-effective or not. The classic example is a preventive intervention, such as vaccination, that typically has up-front costs and results in health gains in the future. A vaccine might appear highly cost-effective based upon undiscounted results, but then appear very poor value for money once discounting is applied. Naturally this makes discounting a contentious subject.
Discounting has been hotly debated within health economics. Differential discounting, whereby a lower discount rate is applied to health effects than to costs, has sparked particularly heated exchanges. Nevertheless, a remarkable degree of consensus was achieved by Claxton et al. (2011) in a paper describing the conditions under which differential discounting is justified. That work showed that the appropriate approach depends on whether it is ‘health’ or ‘social welfare’ that is to be maximized and whether the healthcare budget is fixed or flexible.
Fast forward to November 2016 and the publication of the updated recommendations for conducting CEAs made by the 2nd Panel. These recommendations are an update of the landmark 1996 textbook edited by Martha Gold and colleagues. Updating this influential text is important, since it provides an opportunity to reflect important technical advances over the previous 20 years and set out clear guidance to CEA practitioners for the future. The revised guidelines were put together by a panel of CEA experts primarily drawn from the United States. Draft chapters were released for a brief public consultation in 2015. The final text was launched in December 2016 at an event at the National Academy of Sciences in Washington DC.
The new discounting guidelines recommend equal discounting of costs and effect at a rate of 3% per annum. This recommendation is essentially the same as that in the 1996 text, although with a different theoretical justification. Surprisingly, the 2nd Panel’s recommendations explicitly rule out differential discounting. Indeed, what the Paulden critique regards as particularly surprising is that the 2nd panel started out from the same theoretical basis as Claxton et al. (2011), but somehow reached a different conclusion: that differential discounting is not justified. The 2nd Panel did not clearly explain how they reached a contradictory conclusion.
The Paulden critique delves into the 2nd Panel’s workings to show how muddled assumptions and algebraic errors led the 2nd Panel to an erroneous set of conclusions. The critique proposes how the recommendations may be corrected, which leads back to conclusions consistent with Claxton et al. (2011).
In addition to the mistaken algebra, the second panel did not make best use of relevant evidence on what discount rate to recommend. Real government borrowing costs in recent years would suggest discounting at a rate of 1.5% per annum or lower, but the 2nd Panel seems to have ignored relevant sources and decided to stick with the existing rate of 3% per annum, partly out of concern for ‘comparability’ with analyses conducting in accordance with previous guidelines. Paulden et al. argue that comparability with prior analyses is not a justifiable reason for ignoring recent theoretical and empirical evidence and for holding back methods development.
Paulden et al. conclude that the 2nd Panel have missed an important opportunity to move CEA methods forward. The 2nd Panel’s mistakes could have been avoided if there was greater engagement during the drafting stage, including a greater use of peer review. Hopefully the critique will stimulate fresh debate on the topic of discounting. In any case, it would have been better to have had this debate before the new textbook was printed, bound and shipped.
Claxton K, Paulden M, Gravelle H, Brouwer W, Culyer AJ. Discounting and decision making in the economic evaluation of health‐care technologies. Health Economics. 2011 Jan 1;20(1):2-15.
Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG, editors. Cost-Effectiveness in Health and Medicine. Oxford University Press; 2016 Oct 3.
Paulden M, O’Mahony JF, McCabe C. Discounting the Recommendations of the Second Panel on Cost-Effectiveness in Health and Medicine. PharmacoEconomics. 2016. doi:10.1007/s40273-016-0482-0.