Sensitivity Analysis and External Adjustment for Unmeasured Confounders in Epidemiologic Database Studies of Therapeutics

Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA
Pharmacoepidemiology and Drug Safety (Impact Factor: 2.94). 05/2006; 15(5):291 - 303. DOI: 10.1002/pds.1200
Source: PubMed


Large health care utilization databases are frequently used to analyze unintended effects of prescription drugs and biologics. Confounders that require detailed information on clinical parameters, lifestyle, or over-the-counter medications are often not measured in such datasets, causing residual confounding bias.Objective
This paper provides a systematic approach to sensitivity analyses to investigate the impact of residual confounding in pharmacoepidemiologic studies that use health care utilization databases.Methods
Four basic approaches to sensitivity analysis were identified: (1) sensitivity analyses based on an array of informed assumptions; (2) analyses to identify the strength of residual confounding that would be necessary to explain an observed drug-outcome association; (3) external adjustment of a drug-outcome association given additional information on single binary confounders from survey data using algebraic solutions; (4) external adjustment considering the joint distribution of multiple confounders of any distribution from external sources of information using propensity score calibration.Conclusion
Sensitivity analyses and external adjustments can improve our understanding of the effects of drugs and biologics in epidemiologic database studies. With the availability of easy-to-apply techniques, sensitivity analyses should be used more frequently, substituting qualitative discussions of residual confounding. Copyright © 2006 John Wiley & Sons, Ltd.

Full-text preview

Available from:
  • Source
    • "This level of sunlight exposure has been estimated to pertain to B25% of a Western population (English et al, 1998). Using these figures combined with the exposure prevalence of 16% for statins in our study and the 'rule-out' approach (Schneeweiss, 2006), we estimated that even a statin–sunlight exposure association as low as 1.38 would fully explain an observed OR of 1.1 for statins and SCC. The slight increase in risk of NMSC associated with statin use observed in our study can therefore easily be explained by residual confounding by sunlight exposure or other unmeasured confounding. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background: Evidence is conflicting regarding statin use and risk of basal cell (BCC) and squamous cell skin cancer (SCC). Methods: Using Danish nationwide registries, we identified all patients with incident BCC/SCC during 2005-2009 and matched them to population controls. We computed odds ratios (ORs) for BCC and SCC associated with statin use. Results: We identified 38,484 cases of BCC and 3724 cases of SCC. Statin ever use was associated with ORs of 1.09 (CI: 1.06-1.13) for BCC and 1.01 (CI: 0.91-1.11) for SCC. Conclusions: Statin use was not associated with risk of SCC. Residual confounding plausibly explains the marginally increased risk of BCC.
    British Journal of Cancer 10/2014; 112(1). DOI:10.1038/bjc.2014.527 · 4.84 Impact Factor
  • Source
    • "The conduct of more validation studies and sensitivity analyses is also recommended to better understand methodological issues and sources of uncertainty (Chatterjee and Wacholder 2002), (e.g. Greenland 1996; Rosenbaum 2005; Schneeweiss 2006; VanderWeele and Arah 2011). 3) Improve communication about epidemiologic uncertainty Full disclosure of uncertainty in epidemiology is encouraged as a matter of transparency. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background: There is a recognized need to improve the application of epidemiologic data in human health risk assessment especially for understanding and characterizing risks from environmental and occupational exposures. Although there is uncertainty associated with the results of most epidemiologic studies, techniques exist to characterize uncertainty that can be applied to improve weight-of-evidence evaluations and risk characterization efforts. Methods: This report derives from a Health and Environmental Sciences Institute (HESI) workshop held in Research Triangle Park, North Carolina, to discuss the utility of using epidemiologic data in risk assessments, including the use of advanced analytic methods to address sources of uncertainty. Epidemiologists, toxicologists, and risk assessors from academia, government, and industry convened to discuss uncertainty, exposure assessment, and application of analytic methods to address these challenges. Synthesis: Several recommendations emerged to help improve the utility of epidemiologic data in risk assessment. For example, improved characterization of uncertainty is needed to allow risk assessors to quantitatively assess potential sources of bias. Data are needed to facilitate this quantitative analysis, and interdisciplinary approaches will help ensure that sufficient information is collected for a thorough uncertainty evaluation. Advanced analytic methods and tools such as directed acyclic graphs (DAGs) and Bayesian statistical techniques can provide important insights and support interpretation of epidemiologic data. Conclusions: The discussions and recommendations from this workshop demonstrate that there are practical steps that the scientific community can adopt to strengthen epidemiologic data for decision making.
    Environmental Health Perspectives 07/2014; 122(11). DOI:10.1289/ehp.1308062 · 7.98 Impact Factor
  • Source
    • "To further assess the robustness of our results, we performed a subgroup analysis to evaluate the risk of death in patients who had been hospitalized because of TBI to see if there is a ‘dose-response’ effect in the relationship between TBI and death. We also performed sensitivity analyses [17,18] to evaluate how large the effect of an unmeasured confounder would be to account for the results. A two-tailed P value of <0.05 was considered significant. "
    [Show abstract] [Hide abstract]
    ABSTRACT: It is known that the risk of death in elderly patients with moderate to severe traumatic brain injury is increased. However, the relationship between mild traumatic brain injury and death has never been established. We investigated the mortality rates of older patients with mild traumatic brain injury in Taiwan to evaluate if there is a higher risk of death compared with the general population. We utilized a sampled National Health Insurance claims database containing one million beneficiaries. We followed all adult beneficiaries older than 65 years from January 1, 2005 till December 31, 2009 to see if they died. We further identified patients with mild traumatic brain injury and compared their risk of death with the general population. We identified 5997 patients with mild traumatic brain injury and 84,117 patients without mild traumatic brain injury. After controlling for age, gender, urbanization level, socioeconomic status, diabetes, hypertension, history of alcohol intoxication, history of ischemic stroke, history of intracranial hemorrhage, malignancies, dementia and Charlson Comorbidity Index score, the adjusted hazard ratio was 1.25 (95% confidence interval, 1.16--1.34). Mild traumatic brain injury is an independent significant risk factor for death in the elderly.
    Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 01/2014; 22(1):7. DOI:10.1186/1757-7241-22-7 · 2.03 Impact Factor
Show more