Sample size calculation in economic evaluations

• Published in: Health economics Vol. 7; no. 4; pp. 327 - 335
• Main Authors: Al, Maiwenn J., Van Hout, Ben A., Michel, Bowine C., Rutten, Frans F.H.
• Format: Journal Article
• Language: English
• Published: Chichester Wiley Subscription Services, Inc., A Wiley Company 01.06.1998; John Wiley & Sons, Ltd
• Series: Health Economics
• Subjects: 2-Pyridinylmethylsulfinylbenzimidazoles; Angioplasty, Balloon, Coronary - economics; Anti-Ulcer Agents - economics; Anti-Ulcer Agents - therapeutic use; Calculation; Cost-Benefit Analysis; Costs and Cost Analysis; economic evaluation; Esophagitis, Peptic - drug therapy; Esophagitis, Peptic - economics; Fibrinolytic Agents - economics; Fibrinolytic Agents - therapeutic use; Health economics; Health Services Research - methods; Humans; Lansoprazole; Models, Statistical; Myocardial Infarction - drug therapy; Myocardial Infarction - economics; Myocardial Infarction - therapy; Omeprazole - analogs & derivatives; Omeprazole - economics; Omeprazole - therapeutic use; Probability; Sample Size; sample size calculation; Streptokinase - economics; Streptokinase - therapeutic use
• ISSN: 1057-9230; 1099-1050
• DOI: 10.1002/(SICI)1099-1050(199806)7:4<327::AID-HEC342>3.0.CO;2-U

A simulation method is presented for sample size calculation in economic evaluations. As input the method requires: the expected difference and variance of costs and effects, their correlation, the significance level (α) and the power of the testing method and the maximum acceptable ratio of incremental effectiveness to incremental costs. The method is illustrated with data from two trials. The first compares primary coronary angioplasty with streptokinase in the treatment of acute myocardial infarction, in the second trial, lansoprazole is compared with omeprazole in the treatment of reflux oesophagitis. These case studies show how the various parameters influence the sample size. Given the large number of parameters that have to be specified in advance, the lack of knowledge about costs and their standard deviation, and the difficulty of specifying the maximum acceptable ratio of incremental effectiveness to incremental costs, the conclusion of the study is that from a technical point of view it is possible to perform a sample size calculation for an economic evaluation, but one should wonder how useful it is. © 1998 John Wiley & Sons, Ltd.