Good practices for quantitative bias analysis LASH, Timothy L; FOX, Matthew P; MACLEHOSE, Richard F ...
International journal of epidemiology,
12/2014, Letnik:
43, Številka:
6
Journal Article
Recenzirano
Odprti dostop
Quantitative bias analysis serves several objectives in epidemiological research. First, it provides a quantitative estimate of the direction, magnitude and uncertainty arising from systematic ...errors. Second, the acts of identifying sources of systematic error, writing down models to quantify them, assigning values to the bias parameters and interpreting the results combat the human tendency towards overconfidence in research results, syntheses and critiques and the inferences that rest upon them. Finally, by suggesting aspects that dominate uncertainty in a particular research result or topic area, bias analysis can guide efficient allocation of sparse research resources. The fundamental methods of bias analyses have been known for decades, and there have been calls for more widespread use for nearly as long. There was a time when some believed that bias analyses were rarely undertaken because the methods were not widely known and because automated computing tools were not readily available to implement the methods. These shortcomings have been largely resolved. We must, therefore, contemplate other barriers to implementation. One possibility is that practitioners avoid the analyses because they lack confidence in the practice of bias analysis. The purpose of this paper is therefore to describe what we view as good practices for applying quantitative bias analysis to epidemiological data, directed towards those familiar with the methods. We focus on answering questions often posed to those of us who advocate incorporation of bias analysis methods into teaching and research. These include the following. When is bias analysis practical and productive? How does one select the biases that ought to be addressed? How does one select a method to model biases? How does one assign values to the parameters of a bias model? How does one present and interpret a bias analysis?. We hope that our guide to good practices for conducting and presenting bias analyses will encourage more widespread use of bias analysis to estimate the potential magnitude and direction of biases, as well as the uncertainty in estimates potentially influenced by the biases.
Pregnant women are exposed to a mixture of endocrine disrupting chemicals (EDCs). Gestational EDC exposures may be associated with changes in fetal growth that elevates the risk for poor health later ...in life, but few studies have examined the health effects of simultaneous exposure to multiple chemicals. This study aimed to examine the association of gestational exposure to five chemical classes of potential EDCs: phthalates and bisphenol A, perfluoroalkyl substances (PFAS), polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), and organochlorine pesticides (OCPs) with infant birth weight.
Using data from the Health Outcomes and Measures of Environment (HOME) Study, we examined 272 pregnant women enrolled between 2003-2006. EDC concentrations were quantified in blood and urine samples collected at 16 and 26 weeks gestation. We used Bayesian Hierarchical Linear Models (BHLM) to examine the associations between newborn birth weight and 53 EDCs, 2 organochlorine pesticides (OPPs) and 2 heavy metals.
For a 10-fold increase in chemical concentration, the mean differences in birth weights (95% credible intervals (CI)) were 1 g (-20, 23) for phthalates, -11 g (-52, 34) for PFAS, 0.2 g (-9, 10) for PCBs, -4 g (-30, 22) for PBDEs, and 7 g (-25, 40) for OCPs.
Gestational exposure to phthalates, PFAS, PCBs, PBDEs, OCPs or OPPs had null or small associations with birth weight. Gestational OPP, Pb, and PFAS exposures were most strongly associated with lower birth weight.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Aims
To estimate and test the difference in rates of violent and non‐violent crime during medicated and non‐medicated methadone treatment episodes.
Design, Setting and Participants
The study involved ...linkage of population level administrative data (health and justice) for all individuals (n = 14 530) in British Columbia, Canada with a history of conviction and who filled a methadone prescription between 1 January 1998 and 31 March 2015. Methadone maintenance treatment was the primary independent variable and was treated as a time‐varying exposure. Each participant's follow‐up (mean: 8 years) was divided into medicated (methadone was dispensed) and non‐medicated (methadone was not dispensed) periods with mean durations of 3.3 and 4.6 years, respectively.
Measurements
Socio‐demographics of participants were examined along with the main outcomes of violent and non‐violent offences.
Findings
During the first 2 years of treatment (≤ 2.0 years), periods in which methadone was dispensed were associated with a 33% lower rate of violent crime 0.67 adjusted hazard ratio (AHR), 95% confidence intervals (CI) = 0.59, 0.76 and a 35% lower rate of non‐violent crime (0.65 AHR, 95% CI = 0.62, 0.69) compared with non‐medicated periods. This equates to a risk difference of 3.6 (95% CI = 2.6, 4.4) and 37.2 (95% CI = 33.0, 40.4) fewer violent and non‐violent offences per 100 person‐years, respectively. Significant but smaller protective effects of dispensed methadone were observed across longer treatment intervals (2.0 to ≤ 5.0 years, 5.0 to ≤ 10.0 years).
Conclusions
Among a cohort of Canadian offenders, rates of violent and non‐violent offending were lower during periods when individuals were dispensed methadone compared with periods in which they were not dispensed methadone.
Indoor and outdoor fine particulate matter (PM2.5) are both leading risk factors for death and disease, but making indoor measurements is often infeasible for large study populations.
We developed ...models to predict indoor PM2.5 concentrations for pregnant women who were part of a randomized controlled trial of portable air cleaners in Ulaanbaatar, Mongolia. We used multiple linear regression (MLR) and random forest regression (RFR) to model indoor PM2.5 concentrations with 447 independent 7-day PM2.5 measurements and 87 potential predictor variables obtained from outdoor monitoring data, questionnaires, home assessments, and geographic data sets. We also developed blended models that combined the MLR and RFR approaches. All models were evaluated in a 10-fold cross-validation.
The predictors in the MLR model were season, outdoor PM2.5 concentration, the number of air cleaners deployed, and the density of gers (traditional felt-lined yurts) surrounding the apartments. MLR and RFR had similar performance in cross-validation (R2 = 50.2%, R2 = 48.9% respectively). The blended MLR model that included RFR predictions had the best performance (cross validation R2 = 81.5%). Intervention status alone explained only 6.0% of the variation in indoor PM2.5 concentrations.
We predicted a moderate amount of variation in indoor PM2.5 concentrations using easily obtained predictor variables and the models explained substantially more variation than intervention status alone. While RFR shows promise for modelling indoor concentrations, our results highlight the importance of out-of-sample validation when evaluating model performance. We also demonstrate the improved performance of blended MLR/RFR models in predicting indoor air pollution.
Display omitted
•Indoor air pollution is an important determinant of personal exposure.•We used multiple linear regression (MLR) and random forest regression (RFR) to model indoor PM2.5 concentrations.•Blended models combining MLR and RFR approaches outperformed stand-alone models.•RFR and blended MLR/RFR approaches show promise for modelling indoor pollution.
To our knowledge, this is the first application and evaluation of random forest regression and blended models for modelling indoor air pollution concentrations, and these techniques show promise for estimating exposure in health risk assessment and epidemiology.
Developmental exposure to airborne particulate matter (PM) may increase children’s risk of developing autism spectrum disorder. We quantified the impact of reducing PM exposure during pregnancy on ...the development of autistic traits in children. We also assessed associations between indoor fine PM (PM2.5) concentrations during pregnancy and autistic traits.
In this parallel-group randomized controlled trial, we randomized 540 non-smoking pregnant women to receive HEPA filter air cleaners or to a control group, which did not receive air cleaners. We administered the Social Responsiveness Scale (SRS-2) to caregivers when children were a median of 48 months (range: 48 to 51 months). Our primary outcome was the SRS-2 total T-score. We imputed missing data using multiple imputation with chained equations and our primary analysis was by intention to treat. In secondary analyses, we estimated associations between full pregnancy and trimester-specific indoor PM2.5 concentrations and T-scores.
We enrolled participants at a median of 11 weeks’ gestation. Our analysis included 478 children (233 control, 245 intervention). The intervention reduced average indoor PM2.5 concentrations by 29 % (95 % CI: 21, 37 %). The mean SRS-2 total T-score was 0.5 units lower (95 % CI: −2.5, 1.5) among intervention participants, with evidence of larger benefits for children at the high end of the T-score distribution. An interquartile range (9.6 µg/m3) increase in indoor PM2.5 during pregnancy was associated with 1.8-unit (95 % CI: 0.3, 3.2) increase in mean SRS-2 total T-score. Effect estimates for PM2.5 concentrations by trimester were smaller and confidence intervals spanned no effect.
Reducing indoor PM during pregnancy had little impact on mean autism-related behavior scores in children. However, indoor PM2.5 concentrations during pregnancy were associated with higher scores. Exposure to particulate matter during pregnancy may influence the development of autistic traits in childhood.
Trial registration: ClinicalTrials.gov: NCT01741051.
Sensitivity analysis is used widely in statistical work. Yet the notion and properties of sensitivity parameters are often left quite vague and intuitive. Working in the Bayesian paradigm, we present ...a definition of when a sensitivity parameter is "pure," and we discuss the implications of a parameter meeting or not meeting this definition. We also present a diagnostic with which the extent of violations of purity can be visualized.
•Gestational air pollution exposure may impact cardiometabolic health in childhood.•We evaluated air cleaner use during pregnancy and obesity-related outcomes in young children.•The intervention was ...associated with effects in the expected directions.•Reducing air pollution during pregnancy may improve children’s cardiometabolic health.
Gestational exposure to particulate matter (PM) air pollution may increase the risk of childhood obesity, but the impact of reducing air pollution during pregnancy on obesity-related outcomes in childhood has not been examined.
To assess the impact of reducing gestational PM exposure on body mass index (BMI) at two years of age.
In this single-blind, parallel group randomized controlled trial in Ulaanbaatar Mongolia, we randomly assigned 540 pregnant women to receive 1–2 portable high efficiency particulate air (HEPA) filter air cleaners or no air cleaners. We measured height and weight when children were a mean age of 23.8 months. Our primary outcome was age- and sex-specific BMI z-score based on the World Health Organization 2007 Growth Charts. Secondary outcomes included age- and sex-specific weight z score, overweight/obesity (defined as BMI z-score > 2.00), and catch-up growth (defined using various cut-offs to identify children with relatively low birth weight for sex and gestational age and relatively high age- and sex-specific weight in childhood). We imputed missing outcome data using multiple imputation with chained equations and our primary analysis was by intention to treat (ITT). We estimated intervention effects on continuous and binary outcomes using linear and logistic regression, respectively.
After excluding known miscarriages, still births, and neonatal deaths our analysis included 480 children (235 control and 245 intervention). The mean (SD) child BMI z score was 0.79 (1.0); 9.8% of children were overweight or obese. The mean BMI z score of children who were randomly assigned to the intervention group was 0.16-units lower (95% CI: −0.35, 0.04) than children in the control group. The intervention was also associated with reductions in overweight/obesity (odds ratio = 0.59; 95% CI: 0.31, 1.12). Catch-up growth occurred less frequently in the intervention group, but effect estimates varied depending on the specific definition of catch-up growth and confidence intervals consistently spanned no effect.
We found that the use of portable air cleaners during pregnancy was associated with improvements in obesity-related outcomes, although some effect estimates lacked precision. Reducing PM exposure during pregnancy may lead to improvements in cardiometabolic health in childhood.
Background: Numerous observational studies show that statin use is associated with lower risk of osteoporotic fractures. However, a causal relationship is not supported by data from randomized ...trials. Unmeasured confounding is implicated as a likely culprit for the controversy because of failure to measure and adjust for patient-level tendencies to engage in healthy behaviors. However, an alternative explanation is selection bias because of the inclusion of prevalent users of statins in the analysis. The relative importance of either bias has not been investigated in a quantitative sensitivity analysis. Methods: We conducted a systematic review to summarize the pattern of association between statin use and fracture risk in observational studies. Our objective was to quantify the magnitude of unmeasured confounding and selection bias in a sensitivity analysis. Results: In 17 published studies, the pooled relative risk for the association between current use of statins and fracture risk was 0.75 (95% confidence interval = 0.66—0.85). Upon adjustment for individual-level use of preventative health services, the pooled relative risk shifted by less than 5% on the log scale. However, a sensitivity analysis for selection bias revealed that moderate levels of bias could eliminate the association between statins and fracture risk. Conclusions: It appears that confounding from unmeasured variables cannot explain the protective association between statins and fractures that has been observed in the literature.