- Search for Research Summaries, Reviews, and Reports
- DEcIDE Project
- Development and Validation of Instrumental Variable Methods for Effectiveness Research
- Research Report May 14, 2010
Related Links for this Topic
- Brookhart MA, Rassen JA, Schneeweiss S. Instrumental variable methods in comparative safety and effectiveness research. Pharmacoepidemiol Drug Saf 2010:19:537-54.
- Brookhart MA, Schneeweiss S. Preference-based instrumental variable methods for the estimation of treatment effects: assessing validity and interpreting results. Int J Biostat 2007;3:14.
- Rassen JA, Solomon DH, Glynn RJ, et al. Simultaneously assessing intended and unintended treatment effects of multiple treatment options: a pragmatic “matrix design.” Pharmacoepidemiol Drug Saf 2011 Jul;20(7):675-83. PMID: 21626604.
Abstract - Draft
Development and Validation of Instrumental Variable Methods for Effectiveness Research
Observational studies of drugs and medical procedures based on administrative data are increasingly used to inform regulatory and clinical decisions. However, the validity of such studies is often questioned because available data do not contain measurements of many important prognostic variables that guide treatment decisions. Recently, various approaches to this problem have been proposed that use instrumental variables that are defined at the level of a health care provider or group of providers. Implicitly, these approaches attempt to estimate causal effects by using differences in medical practice patterns as a quasi-experiment. Despite the early promise of these approaches, many questions persist about the validity of the methods and the appropriate interpretation of their results. As part of this task order, we will develop new methodology that can be used to explore how violations of instrumental variable assumptions and treatment effect heterogeneity may bias the standard IV estimator with respect to the average treatment effect in the population. Using these methods, we will describe how one can anticipate the likely direction of bias by using both empirical data and commonly available subject matter knowledge, such as whether medications or procedures tend to be overused, underused, or often misused. These methods will be described in the context of a study comparing the GI bleeding risk attributable to different NSAIDs. A second example will apply these methods using the hospital as the basis of an instrumental variable in a study of direct thrombin inhibitors in patients undergoing percutaneous coronary interventions and in-hospital hemorrhagic events. The results will be validated against existing RCT data. Lastly, to facilitate broader and appropriate use of instrumental variable methods in comparative effectiveness research this task-order will develop a practical guide on the implementation and evaluation of such methods.
*To receive a print copy of this report, send a request to firstname.lastname@example.org. Please include the report title(s) and a complete mailing address. For electronic copies of the report(s), please visit the appropriate journal Web site.