Analysis as‐randomized and the problem of non‐adherence: An example from the veterans affairs randomized trial of coronary artery bypass surgery
Authors:
- Janet WittesJWJanet Wittes
- Katherine DetreKDKatherine Detre
- Theodore HolfordTHTheodore Holford
- 1 January 1993
- journal article
- clinical trial
- Published by Wiley in Statistics in Medicine
- Vol. 12 (13) , 1185-1195
Abstract
In most randomized clinical trials not all patients adhere to the therapy to which they were randomly assigned. Instead, they may receive the therapy assigned to another treatment group, or a therapy different from any prescribed in the protocol. When non-adherence occurs, problems occur with the analysis comparing the treatments under study. Rigorous statistical principles require attributing outcome events to the original random treatment assignment ('intent-to-treat' analysis). Using data from the Veterans Administration Cooperative Study of Coronary Artery Bypass Surgery, we report the intent-to-treat analysis and apply four other methods of analysis for analysing non-adherers: 1. exclude non-adherers from analysis; 2. transfer them to the alternative treatment group at the time of randomization; 3. censor them at the time of treatment change, and 4. transfer them to the alternative treatment group at the time of treatment change. Inherent problems and biases of these four other methods are discussed.Keywords
This publication has 14 references indexed in Scilit:
- Compliance as an Explanatory Variable in Clinical TrialsJournal of the American Statistical Association, 1991
- Characteristics and outcome of medical nonadherers in the veterans administration cooperative study of coronary artery surgeryThe American Journal of Cardiology, 1984
- Coronary artery surgery study (CASS): a randomized trial of coronary artery bypass surgery. Survival data.Circulation, 1983
- A computer program for transition life tablesComputer Programs in Biomedicine, 1981
- Effect of bypass surgery on survival in patients in low- and high-risk subgroups delineated by the use of simple clinical variables.Circulation, 1981
- Influence of Adherence to Treatment and Response of Cholesterol on Mortality in the Coronary Drug ProjectNew England Journal of Medicine, 1980
- Veterans administration cooperative study of surgery for coronary arterial occlusive diseaseThe American Journal of Cardiology, 1977
- Covariance Analysis of Heart Transplant Survival DataJournal of the American Statistical Association, 1977
- Evaluation of Response-Time Data Involving Transient States: An Illustration Using Heart-Transplant DataJournal of the American Statistical Association, 1974
- Survivorship Analysis of Heart Transplant DataJournal of the American Statistical Association, 1974