Factors associated with loss to follow-up in a large tuberculosis treatment trial (TBTC Study 22)

• Published in: Contemporary clinical trials Vol. 28; no. 3; pp. 288 - 294
• Main Authors: Conwell, Donna Sepulveda, Mosher, Ann, Khan, Awal, Tapy, Jan, Sandman, Laurie, Vernon, Andrew, Horsburgh, C. Robert
• Format: Journal Article
• Language: English
• Published: New York, NY Elsevier Inc 01.05.2007; Elsevier
• Subjects: Antibiotics, Antitubercular - therapeutic use; Attrition in clinical trials; Bacterial diseases; Biological and medical sciences; Cardiovascular; Clinical trial. Drug monitoring; Clinical Trials, Phase III as Topic; Ethnic Groups; Follow-Up Studies; General pharmacology; Hematology, Oncology and Palliative Medicine; Homeless Persons; Human bacterial diseases; Humans; Infectious diseases; Loss to follow-up; Medical sciences; Multicenter Studies as Topic; Multivariate Analysis; Patient Dropouts; Pharmacology. Drug treatments; Randomized Controlled Trials as Topic; Retention in clinical trials; Retrospective Studies; Reward; Surveys and Questionnaires; Tuberculosis; Tuberculosis and atypical mycobacterial infections; Tuberculosis, Pulmonary - drug therapy; Loss to follow-up; Tuberculosis; Attrition in clinical trials; Retention in clinical trials; Infection; Human; Physician patient relation; Treatment; Bacteriosis; Clinical trial; Mycobacterial infection; Predictive factor
• ISSN: 1551-7144; 1559-2030
• DOI: 10.1016/j.cct.2006.09.003

Abstract Introduction Loss to follow-up in clinical trials compromises achievement of study goals. We evaluated factors associated with loss to follow-up after completion of treatment phase in a large tuberculosis treatment trial (TBTC/USPHS Study 22) in the U.S. and Canada. Methods Patients who were lost to follow-up were compared to those who reached a study end-point or successfully completed follow-up. A generalized estimating equation model was used to combine patient-specific and site-specific factors. Results Of 1075 patients enrolled, 965 (89.8%) reached a study end-point, died, or completed the 2 year post-treatment follow-up phase, and 110 (10.2%) did not. Multivariate analysis showed the following factors to be independently associated with loss to follow-up: birth outside USA/Canada (OR 2.07, 95% CI 1.25–3.40, p = 0.005), history of homelessness (OR 1.94, 95% CI 1.00–3.80, p = 0.05), enrollment at a health department (OR 2.71, 95% CI 1.27–5.79, p = 0.010), and use of any kind of incentive (cash/cash equivalent) during treatment phase (OR 3.04, 95% CI 1.73–5.33 p = 0.0001). Conclusions Cultural or linguistic factors and lack of stable housing contribute to loss to follow-up. Attention to these factors could improve long-term retention in clinical trials. Enrollment at a health department and use of incentives during treatment phase may be markers for other factors leading to loss to follow-up.