Use of Number Needed to Treat in Cost-effectiveness Analyses

Vishvas Garg MBA BPharm, PhD; Xian Shen MS, PhD; Yan Cheng MS, PhD; James J Nawarskas PharmD BCPS; Dennis W Raisch PhD MS RPh


The Annals of Pharmacotherapy. 2013;47(3):380-387. 

In This Article


The flow diagram of the literature search is shown in Figure 1. Our initial PubMed and EconLit searches yielded 169 and 21 CEA articles mentioning NNT and NNH, respectively. Articles that did not meet our inclusion criteria (n = 121) were reviews (n = 33), editorials (n = 4), non-English-language (n = 11), and those that did not report NNT/NNH values (n = 73). Furthermore, 4 QALY articles including NNT/NNH were excluded because associated costs were not described. Therefore, after executing inclusion and exclusion criteria, 69 articles remained for full-text review and analysis.

Figure 1.

Literature search flowchart. CEA = cost-effectiveness analysis, NNH = number needed to harm, NNT = number needed to treat.

From full-text review, we did not find any CEA studies incorporating NNH. This finding is based on the terminology used in the reviewed articles to represent NNT and NNH. None of the NNT values reported used an implied meaning of NNT to represent NNH/NNTH. The selected articles were summarized by treatment/disease condition, focus of study, type of CEA, and measure of NNT in Appendix I (available at We included information about journal title, year of publication, and journal impact factor and provided additional comments. Key features, including journal type, journal impact factor, focus of studies, the way NNT was reported in CEA, and data source for each article are displayed in Table 1.

Among the 69 articles that incorporated NNT into CEA, 80% were published in clinician-practice-focused journals.[8–11,14–16,18–67] Health-economics-focused journals published 11 (15.9%) articles.[68–78] Overall, the majority (72.4%) of the articles incorporating NNT/NNH in CEA were published in the highest quartile of journals based on impact factor (>3.0).[18,19,21–35,38,39,41,42,44,45,49,53,55–57,59,60,63,64,66,69–86] Moreover, 8.7% of all articles were published in journals with impact factors greater than 10, which were all clinician-practice-focused journals.[23,24,26–28,74]

A large percentage of articles focused on disease treatment (40.5%) or disease prevention (40.5%) (Table 1).The cost-effectiveness of patient education was assessed in 4.3% of the articles.[42,64,87] NNT separated from CEA based on QALY was reported in 24.6% of the studies,[23,27,34,37,41,43,45,46,48,54,56,77–80,82,83] whereas 27.5% of the studies reported NNT separated from CEA based on events.[18,19,24–26,28,33,39,40,59,62,67,71,72,74,76,84–86] Literature was used as the data source in 53.6% of the articles.[18,19,21–25,27,32,34–37,40,41,43,45,47,53,56,57,61,62,64,68–71,73,74,76–78,81–86] Randomized controlled trials (O-RCT-CEA and RCT-CEA) comprised 24.6% of the studies'data sources.[26,28–30,33,38,39,42,46,48,49,58,72,75,79,80,87]