Methodological considerations when analysing and interpreting real-world data.

In the absence of relevant data from randomized trials, nonexperimental studies are needed to estimate treatment effects on clinically meaningful outcomes. State-of-the-art study design is imperative for minimizing the potential for bias when using large healthcare databases (e.g. claims data, electronic health records, and product/disease registries). Critical design elements include new-users (begin follow-up at treatment initiation) reflecting hypothetical interventions and clear timelines, active-comparators (comparing treatment alternatives for the same indication), and consideration of induction and latent periods. Propensity scores can be used to balance measured covariates between treatment regimens and thus control for measured confounding. Immortal-time bias can be avoided by defining initiation of therapy and follow-up consistently between treatment groups. The aim of this manuscript is to provide a non-technical overview of study design issues and solutions and to highlight the importance of study design to minimize bias in nonexperimental studies using real-world data.

Employment characteristics and cause-specific mortality at automotive electronics manufacturing plants in Huntsville, Alabama.

BACKGROUND: This study was carried out in response to worker concerns over their exposure to lead solder and chlorinated solvents at automotive electronics manufacturing plants in Huntsville, Alabama. METHODS: A study of 4396 United Autoworkers members ever-employed at the plants between 1972 and 1993 was conducted with mortality follow-up through 2016. Poisson regression was used to estimate mortality rate ratios (RR) according to employment characteristics, including calendar period of employment. RESULTS: Pre-1977 hires exhibited elevated adjusted rates of all-cause (RR, 1.29; 95% confidence interval [CI], 1.09-1.52), cardiovascular (RR, 1.38; 95% CI, 1.03-1.86), and digestive system (RR, 2.31; 95% CI, 1.04-5.10) disease mortality relative to the most recent hire group (1984-1993). Never- versus ever-employment in a skilled trade job was associated with elevated adjusted rates of all-cause, all-cancer, and cardiovascular disease mortality. Nervous system disorder mortality was greatest among 1977-1983 hires. CONCLUSIONS: Elevated mortality among pre-1977 hires is consistent with worker concerns over greater exposure to hazards at the original plant building.

Mortality among autoworkers manufacturing electronics in Huntsville, Alabama.

BACKGROUND: Workers raised concerns over suspected excesses of mortality at automotive electronics manufacturing facilities in Huntsville, Alabama. METHODS: A study of 4396 UAW members ever-employed at Huntsville facilities between 1972 and 1993 was conducted with mortality follow-up through 2016. Standardized Mortality Ratios (SMRs) were estimated using U.S. and Alabama reference rates. RESULTS: Relative to U.S. rates, there was a modest excess of all-cause mortality among White female workers (SMR 1.08, 95%CI: 0.99-1.18) and among all workers hired <1977 at the original plant building (SMR 1.10, 95%CI: 0.99-1.22). There was excess nervous system disorder (SMR 1.24, 95%CI: 0.91-1.65) and brain and nervous system cancer (SMR 1.31, 95%CI: 0.67-2.28) mortality. Estimates for several causes of interest were imprecise. CONCLUSIONS: All-cause mortality estimates were greater than anticipated based on results from other UAW cohorts. The excess of nervous system disease mortality is consistent with other studies of electronics workers exposed to lead-solder and chlorinated solvents.

Odors from sewage sludge and livestock: associations with self-reported health.

OBJECTIVES: Class B treated sewage sludge (TSS) contains microbes and toxicants and is applied to land in areas where livestock wastes may be present. We evaluated relationships of reports of TSS and livestock odors with acute symptoms and excessive flies. METHODS: A total of 158 adults living near liquid TSS application sites, 85 living near cake TSS application sites, and 188 living in comparison areas responded to a household survey regarding odors, health, and demographics. We identified symptom groups using factor analysis. We used generalized estimating equations to fit linear models for associations between factor scores and odors, and Poisson models for associations with specific symptoms. RESULTS: Most factor scores were similar between exposure groups. Covariate-adjusted z-scores for lower respiratory symptoms were 0.28 (95% confidence interval [CI] -0.10, 0.65) higher among residents who reported moderate to very strong liquid TSS odor than among residents in comparison areas, and 0.28 (95% CI 0.05, 0.50) higher among residents who reported moderate to very strong livestock odor compared with residents reporting no or faint livestock odor. The factor score for dermatologic conditions was higher among residents who reported higher liquid sludge odor (0.27, 95% CI -0.13, 0.68), primarily due to skin rash (prevalence ratio = 2.21, 95% CI 1.13, 4.32). Excessive flies were reported twice as commonly among respondents who reported moderate to very strong TSS odor than among other residents. CONCLUSIONS: Reported odors from TSS and livestock were associated with some acute symptoms. Health departments should monitor land applications of human and animal wastes and conduct surveillance of health problems reported by neighbors.