Propensity score methods in health technology assessment: principles, extended applications, and recent advances
- 1London School of Hygiene and Tropical Medicine (LSHTM), United Kingdom
- 2Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Medical Sciences Division, University of Oxford, United Kingdom
- 3Parc de Salut Mar, Spain
- 4Universidade de Sorocaba, Brazil
- 5Center for Data Integration and Knowledge for Health (CIDACS), Brazil
- 6Institute of Public Health, Federal University of Bahia, Brazil
- 7Department of Medical Statistics, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, United Kingdom
- 8Department of Statistics, Federal University of Bahia (UFBA), Brazil
- 9Department of Non-Communicable Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, United Kingdom
Randomized clinical trials (RCTs) are considered the gold-standard approach to estimate effects of treatment on outcomes. They are also the designs of choice for health technology assessment (HTA). Randomization ensures comparability, in both measured and unmeasured pre-treatment characteristics, of patients assigned to treatment and control or comparator. However, even adequately powered RCTs are not always feasible for reasons such as cost, time, ethical, and practical constraints. RCTs rely on data collected on selected, homogeneous population under highly controlled conditions; hence, they provide evidence on efficacy of interventions rather than on effectiveness. Alternatively, observational studies can can provide evidence on the relative effectiveness or safety of a health technology compared to one or more alternatives when provided under the routine setting of health care practice. In observational studies, however, treatment assignment is a non-random process hence treatment groups may not be comparable in their pre-treatment characteristics. As a result, direct comparison of outcomes between treatment groups lead to biased estimate of treatment effect. Propensity score methods have been used to achieve comparability of treatment groups in terms of their measured pre-treatment covariates and thereby controlling for confounding bias in estimating treatment effects. Despite the popularity of propensity scores methods and recent important advances, misunderstandings on their applications and limitations are too common. In this article, we provide a review of the methods, extended applications, recent advances, strengths and limitations.
Keywords: Bias, Confounding, effectiveness, Health Technology Assessment, Propensity Score, Safety, Secondary data, Observational study
Received: 18 Apr 2019;
Accepted: 31 Jul 2019.
Edited by:Ileana Mardare, Carol Davila University of Medicine and Pharmacy, Romania
Reviewed by:Bogdan ILEANU, Bucharest Academy of Economic Studies, Romania
Douglas Faries, Eli Lilly (United States), United States
Copyright: © 2019 Ali, Prieto-Alhambra, Lopes, Ramos, Bispo, Ichihara, Pescarini, Williamson, Fiaccone, Barreto and Smeeth. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Mx. M Sanni Ali, London School of Hygiene and Tropical Medicine (LSHTM), London, United Kingdom, firstname.lastname@example.org
Prof. Luciane Lopes, Universidade de Sorocaba, Sorocaba, Brazil, email@example.com