Search

PURPOSE: Supplementing investigator-specified variables with large numbers of empirically identified features that collectively serve as 'proxies' for unspecified or unmeasured factors can often improve confounding control in studies utilizing administrative healthcare databases. Consequently, there has been a recent focus on the development of data-driven methods for high-dimensional proxy confounder adjustment in pharmacoepidemiologic research. In this paper, we survey current approaches and recent advancements for high-dimensional proxy confounder adjustment in healthcare database studies. METHODS: We discuss considerations underpinning three areas for high-dimensional proxy confounder adjustment: (1) feature generation-transforming raw data into covariates (or features) to be used for proxy adjustment; (2) covariate prioritization, selection, and adjustment; and (3) diagnostic assessment. We discuss challenges and avenues of future development within each area. RESULTS: There is a large literature on methods for high-dimensional confounder prioritization/selection, but relatively little has been written on best practices for feature generation and diagnostic assessment. Consequently, these areas have particular limitations and challenges. CONCLUSIONS: There is a growing body of evidence showing that machine-learning algorithms for high-dimensional proxy-confounder adjustment can supplement investigator-specified variables to improve confounding control compared to adjustment based on investigator-specified variables alone. However, more research is needed on best practices for feature generation and diagnostic assessment when applying methods for high-dimensional proxy confounder adjustment in pharmacoepidemiologic studies. ispartof: PHARMACOEPIDEMIOLOGY AND DRUG SAFETY vol:31 issue:9 pages:932-943 ispartof: location:England status: published

Background and aim Postoperative infection (POI) is a major source of morbidity and prolongation of hospitalization in inflammatory bowel disease (IBD) patients. This large observational study was conducted to further describe risk factors and to quantify the proportion of POIs that are preventable. Methods We conducted a retrospective cohort analysis of the Optum US health insurance claims database. The study population included adults with ulcerative colitis (UC) or Crohn's disease (CD) who underwent lower gastrointestinal (GI) surgery of small intestine, colon, rectum, or anus during September 2014 to September 2016. Multiple logistic regression was used to identify and quantify risk factors and determine the proportion of infections that are preventable. Results A total of 3360 adult IBD patients with lower GI surgery were included in the study. Their mean age was 51 years, 52.5% were women, and 59.5% had CD. The 30-day POI incidence was 15.1% (95% confidence interval: 14.0-16.4%). We identified the following nonmodifiable or procedural risk factors: history of POI, open procedure, red blood cell transfusion within 6 months, preoperative hospital stay of at least 4 days, lower GI ostomy surgery, lower GI resection surgery, and a history of chronic obstructive pulmonary disease. Modifiable risk factors included corticosteroid use and anemia prior to surgery, but few infections were attributable to these modifiable factors. Conclusions This large, observational, real-world evidence study from the US found that the majority of the observed risk factors were nonmodifiable or procedure-related. Corticosteroid use and anemia before surgery were identified as modifiable risk factors.