Elsevier

Journal of Health Economics

Volume 27, Issue 6, December 2008, Pages 1645-1649
Journal of Health Economics

Reply
Future costs in economic evaluation: A comment on Lee

https://doi.org/10.1016/j.jhealeco.2008.07.007Get rights and content

Abstract

In a recent article in this journal Lee argued that indirect medical costs should be ignored in economic evaluations. To reach this conclusion, Lee uses an unrealistic and uncommon budget constraint. This comment highlights a number of methodological problems in Lee's analysis. Moreover, it highlights that looking at current practice of economic evaluation, Lee's model implies the inclusion rather than the exclusion of indirect medical costs.

Introduction

The topic of whether to include future (medical) costs in economic evaluations remains controversial. If an intervention increases the longevity of some patient group, it is reasonable to assume that these survivors will also consume resources during these added life years. This consumption may entail both medical as well as non-medical resources (housing, clothing, etc.). Given the focus on health in economic evaluations of health technologies, the debate in the health economic literature has mainly considered what costs for medical consumption in life years gained need to be included in economic evaluations.1 Normally, two types of future medical costs are distinguished: (i) related and (ii) unrelated future medical costs. Related medical costs are a direct consequence of the intervention under study. Examples would be treatment costs for wound infections after an open-heart surgery or life long anti-rejection medication after an organ transplant. These costs are denoted as being conditionally dependent, because, conditional on survival, the size of these costs is influenced by the initial intervention. There is no controversy in the literature about the inclusion of these costs (e.g. Garber and Phelps, 1997, Meltzer, 1997, Brouwer et al., 2001, Nyman, 2004, Van Baal et al., 2007, Lee, 2008). The inclusion of unrelated future medical costs is, however, a matter of much debate. These costs are related to the intervention only to the extent that the intervention alters survival. For instance, if someone surviving acute heart failure thanks to heart surgery at age 60 will receive a total hip replacement at age 70 after a traffic accident, the costs of the hip replacement are considered unrelated. However, obviously, they would never have occurred if the treatment of the heart attack had not taken place. Therefore with Weinstein (1980) one may argue that these costs “…that would not otherwise have arisen must be considered” in economic evaluations. However, Russell (1986) strongly opposed this proposition. Later, Garber and Phelps (1997) showed that inclusion of unrelated medical costs was irrelevant. This was subsequently convincingly disputed by Meltzer (1997) and Meltzer and Johannesson (1999) who showed inclusion to be consistent with welfare maximization. Nyman (2004) recently argued that consistency requires future medical costs to be included on the cost-side of an economic evaluation, since the associated health gains are usually included on the effect-side. Van Baal et al. (2007) concluded that including future medical costs as well as the health effects of that future medical spending resulted in ratios that were both consistent and optimally informative for decision makers. The recent literature therefore seems to converge to a consensus on inclusion of these unrelated future medical costs.

However, recently Lee (2008) argued in favor of excluding these costs (as well as other unrelated future costs). In a paper which claims to resolve several controversies in CEA, he re-addressed the work of Garber and Phelps (1997) and Meltzer (1997), and demonstrated that Meltzer's findings depend on his budget constraint. Combining the budget constraint used by Garber and Phelps with the objective function used by Meltzer, Lee derives that unrelated future costs should be excluded from economic evaluations. He continues to argue that Meltzer's budget constraint is inappropriate.

This comment responds to the paper of Lee. Our objections primarily focus on the relation between the model assumptions and current practice of cost-effectiveness analysis, considering the models in Lee's sections 2 and 6. In this way, the current reaction complements a recent comment by Meltzer (2008), which focused on the formal modeling part of the Lee paper, as well as on Lee's discussion in section 4 of conditional and annuity models for budget constraints. Lee's results crucially depend on the choice of budget constraint. This will be shortly addressed in section 2 below, since this point is clarified in the recent comment by Meltzer (2008). Section 3 of the current comment then relates the model results to the reality of cost-effectiveness and cost-utility analysis.2 We explain how the current practice of CEA warrants inclusion of costs even given Lee's assumptions on the budget constraint.

Section snippets

The budget constraint: not a trivial matter

Using Lee's notation, his basic model is as follows: Choose the level of medical consumption mt = m0, …, mt, so as toMaxExpt=0TβtSt(Mt)Ut(ct,Ht)s.t.Ht=H(m0,...,mt),fort=0,...,Tyt=ct+pmtwith t being a time indexe, β a discount factor, S the probability of survival that depends on health, H, with health in turn depending on all previous health care, U the utility of a representative agent that depends on current consumption and current health and hence also on past and current health care, yt

From modeling reality to muddling in reality

An important criterion to judge an economic evaluation is the applicability of its results to the problem addressed. Apart from more theoretical arguments, only if that requirement is met, economic evaluations serve their purpose (Brouwer et al., 2008). We argued that a cost utility ratio that contains all substantial effects on future costs as well as future benefits is the most informative ratio for a decision maker (Van Baal et al., 2007). This is normally the viewpoint in a cost-benefit

Acknowledgement

We are grateful for the useful comments made by David Meltzer and John Nyman on an earlier version of this comment. The usual disclaimer applies.

References (23)

  • A. Deaton

    Franco Modigliani and the Life Cycle Theory of Consumption

    Banca Nazionale del Lavoro Quarterly Review

    (2005)
  • Cited by (30)

    • Estimating the costs of non-medical consumption in life-years gained for economic evaluations

      2021, Social Science and Medicine
      Citation Excerpt :

      As a result, the inclusion of both future medical and future non-medical costs, is still uncommon (de Vries et al., 2018). Disagreement is strongest concerning the inclusion of non-medical costs, sometimes also referred to as survivor consumption (Feenstra et al., 2008; Gandjour, 2006; Garber and Phelps, 1997, 2008; Lee, 2008; Lundin and Ramsberg, 2008; D Meltzer, 1997; Meltzer, 2008; Nyman, 2004, 2011; Richardson and Olsen, 2006). While the debate on whether to include these costs in life years gained is ongoing, only a few studies have concentrated on the empirical estimation of future non-medical costs and their influence on outcomes of economic evaluations (Kruse et al., 2012; Manns et al., 2003; D Meltzer, 1997; Meltzer et al., 2000).

    • Catalog of Age- and Medical Condition—Specific Healthcare Costs in the United States to Inform Future Costs Calculations in Cost-Effectiveness Analysis

      2021, Value in Health
      Citation Excerpt :

      For example, the Second Panel on Cost-Effectiveness in Health and Medicine recommends that the reference case CEA be performed from both the healthcare sector and the societal perspectives.1 From either perspective, there has been a theoretical debate about whether and how future costs (medical and nonmedical consumptions and production during added life years owing to medical intervention) should be included in the cost part.2-7 The Second Panel also recommends that all future costs be included in CEA.1

    • Broadening the Concept of Value: A Scoping Review on the Option Value of Medical Technologies

      2021, Value in Health
      Citation Excerpt :

      According to the third definition, the claimed value does not originate from the uncertainty around a decision and the flexibility of deferring it, as in definition B, but rather it stems from the consideration that the value of a life-extending technology should also include the benefits of future treatments that otherwise would be precluded to patients if they did not benefit from improved survival. This definition of value pertains to the broader discussion on whether future costs and benefits not directly linked to the intervention being assessed should be accounted for when evaluating a technology.73-79 Therefore, we recommend that research related to this definition adopt the term “option value of survival.”

    • A welfare-theoretic model consistent with the practice of cost-effectiveness analysis and its implications

      2020, Journal of Health Economics
      Citation Excerpt :

      The first is whether future costs (both non-healthcare consumption and healthcare costs) should be included in an additional year of life produced by an intervention. ( Garber and Phelps, 1997; Meltzer, 1997; Feenstra et al., 2008; Meltzer, 2006, 2008; Nyman, 2004) Second, there is the debate on whether costs and benefits should be discounted at the same rate in the future years. ( Gravelle et al., 2007; Claxton et al., 2011; Paulden and Claxton, 2012) Unfortunately, the CEA literature looking at these three debates has approached these subjects in isolation and, consequently, generated lingering questions when an assumption in one debate ignores the requirements of the other.

    • Development of an Official Guideline for the Economic Evaluation of Drugs/Medical Devices in Japan

      2017, Value in Health
      Citation Excerpt :

      When costs such as adverse events and future events are difficult to calculate on a fee-for-service basis, average prospective payment may be applied, but not for the target technology. There is some controversy regarding how to treat unrelated medical costs in economic evaluation [31–37]. Treatment of high blood pressure reduces mortality from cardiovascular events and stroke and leads to prolonged life and increased unrelated medical costs (e.g., dementia, diabetes, and dialysis).

    • Costs and health effects of adding functional foods containing phytosterols/-stanols to statin therapy in the prevention of cardiovascular disease

      2011, European Journal of Pharmacology
      Citation Excerpt :

      Finally, neither previous study considered costs caused by diseases other than cardiovascular disease, acquired later in the life-years saved. Yet, it is more and more recommended that these indirectly related health care costs should be included in economic evaluations (van Baal et al., 2007; Feenstra et al., 2008; Nyman, 2004; Meltzer, 2008). Nevertheless, with respect to the latter two aspects, sensitivity analyses which disregarded discontinuation of phytosterols/-stanols or indirect health care costs did not substantially alter the results.

    View all citing articles on Scopus
    View full text