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


We searched PubMed and EconLit from 1966 through September 2012, using the Boolean indicator "and," the MeSH term cost-benefit analysis, or the key words cost-effectiveness analysis, cost-effectiveness, and cost-utility analysis. These terms were separately combined with the following: number needed to treat, number needed to harm, NNT, or NNH. We excluded reviews, editorials, non-English-language articles, and articles that did not present NNT, NNH, or cost-effectiveness ratios. We included CEA studies reporting cost per life-year gained, per QALY, or other cost per effectiveness measure (eg, cure, complication avoided). Of all included studies, full-text articles reviewed to extract study information by type of journal, impact factor of the journal, focus of study, data source, publication year, how NNT/NNH was reported, and outcome measures.

Based on the description and statement of each journal on the publisher's website, we classified journals into 3 categories: clinician-focused practice journals (eg, Journal of Infectious Diseases), policy-focused journals (eg, Journal of Managed Care Pharmacy), and health-economics-focused journals (eg, Pharmacoeconomics). We further reviewed the 2010 impact factor of each journal in which the articles were published. We identified greater than 3.0 as the top quartile of journals by impact factor as follows: (1) we consulted the Journal Citations Reports by Thompson Reuters, (2) we compiled a list of all of the medical, disease specialty, health policy, health services research, and health economic journals, and (3) we determined the total number of journals and divided them into quartiles by impact factor. The health care focus of each included article was defined as disease treatment, disease prevention, disease management, or patient education. We also summarized the studies according to data sources: (1) randomized controlled trials (RCTs) conducted primarily for the purpose of CEA (RCT-CEA), (2) RCTs with primary purpose other than CEAs (O-RCT-CEA), (3) observational studies based on a large database of health care observational data (OBS-CEA), and (4) CEA based on data obtained from a review of published literature (LIT-CEA). The way NNT/NNH in the CEA was reported was also summarized into 3 groups. We distinguished studies that stated NNT/NNH separately from CEA versus those that reported NNT/NNH as part of the CEA. Among those that reported NNT/NNH separately from CEA, we differentiated the studies using QALYs as the effectiveness measure.

Two authors (DWR and VG) conducted the literature search and extracted full text of the included studies. Four authors (DWR, VG, XS, and YC) reviewed the included studies. One author (VG) reviewed all of the articles. Three authors (DWR, XS, and YC) reviewed one third of the articles each. Any conflicts were resolved by discussion among all 4 authors.