A most stubborn bias: no adjustment method fully resolves confounding by indication in observational studies

Abstract

Objective-To evaluate the effectiveness of methods that control for confounding by indication, we compared breast cancer recurrence rates among women receiving adjuvant chemotherapy versus those who did not.Study Design and Setting-In a medical record review-based study of breast cancer treatment in older women (n=1798) diagnosed 1990-1994, our crude analysis suggested adjuvant Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Conclusion-Conventional methods do not control for unmeasured factors, which often remain important when addressing confounding by indication. PS and IV analysis methods can be useful under specific situations, but neither method adequately controlled confounding by indication in this study. NIH Public AccessAuthor Manuscript J Clin Epidemiol. Author manuscript; available in PMC 2011 January 1.

Interpretation of the evidence for the efficacy and safety of statin therapy

“…For example, there have been many claims about the benefits of various vitamin supplements based on observational studies 128,129 that have been reliably refuted by large-scale randomized trials. [130][131][132] Similarly, when compared with the results from randomized trials of the effects of treatments for several different cancers, observational studies have generated improbable results despite controlling for comorbidity, extent of disease and many other characteristics that were recorded in detailed databases [133][134][135] (as is also the case for reported associations of statins with lower rates of cancer 90,[136][137][138] ). These findings are consistent with empirical studies in which biases in observational studies were shown to be large enough to conclude falsely that treatment produced benefit or harm, with none of a range of statistical strategies (such as regression analysis or propensity matching) capable of adjusting adequately or predictably for bias.…”

Impact of intensity of conditioning therapy in patients aged 40–60 years with AML/myelodysplastic syndrome undergoing allogeneic transplantation

“…The results of such retrospective design can be significantly influenced by differences in patient characteristics related to prognosis that drive the decision of whether to use RIC or CIC. 33 Patients with co-morbidities could have been assigned to RIC regimens, whereas those with higher-risk disease were assigned to the CIC regimens according to the discretion of treating physicians. This should have resulted in a higher TRM in the RIC group unless the RIC regimens truly lowered it, which is likely given the higher co-morbidity score difference between the two groups.…”

Instrumental Variable Analysis in Epidemiologic Studies: An Overview of the Estimation Methods

“…Also in case of these outcome variables, IV analysis has been applied with two-stage method. In that case, the second-stage model could be a Cox proportional hazards model [78][79][80]. However, Brookhart et al [3] stated that this approach for IV analysis is not motivated by a theoretical model and, therefore, parameters that are obtained from this approach may not be causally interpretable.…”

“…However, Brookhart et al [3] stated that this approach for IV analysis is not motivated by a theoretical model and, therefore, parameters that are obtained from this approach may not be causally interpretable. Examples of this approach are a study of the effect of rosiglitazone on (time to) cardiovascular hospitalization and all-cause mortality using facility-prescribing patterns as an IV [78], and a study of the effect of adjuvant chemotherapy on (time to) breast cancer recurrence using physician preference as an IV [79].…”