Session 2 - Opportunities for Improving Health through Big Data and Data Science: Propensity Score Methods for Reducing Bias in Observational Studies: An Example from the PRAMS Study
Presentation Type
Invited
Track
Health Care Application
Abstract
Background: Conventional cigarette (CC) smoking is one of the most preventable causes of adverse birth outcomes. Although electronic cigarettes (ECs) are considered to be safer than CCs during pregnancy, the evidence is yet to be presented. This study examines the effects of EC use on birth outcomes compared to those of CC smokers using multiple analytic methods to adjust for risk factors of tobacco use. Methods: Data were extracted from 3,727 exclusive CC smokers and exclusive EC users who participated in the Phase 8 survey of the Pregnancy Risk Assessment Monitoring System (PRAMS). Adverse outcomes included small-for-gestational-age (SGA), low birthweight (LBW), and prematurity. Analyses were conducted using three different methodologies to account for nine covariates: 1) weighted, unadjusted logistic regression, 2) weighted, adjusted logistic regression, 3) inverse propensity weighted (IPW) logistic regression. Results: While unadjusted logistic regression analyses showed variable favorable effects of EC use on SGA (OR 0.64; 95% CI 0.39–1.04) and LBW (OR 0.58; 95% CI 0.40–0.82) compared to CC smoking, accounting for nine covariates using IPW rendered the results statistically insignificant. Using ECs does not significantly decrease the odds of neonates being SGA (OR 0.64, CI 0.28–1.45), LBW (OR 0.65, CI 0.35–1.20) or born prematurely (OR 0.80, CI 0.37–1.75) compared to smoking CCs. Conclusion: The study fails to support the common perception that ECs are safer than CCs during pregnancy. Until the safety of prenatal EC use is demonstrated, all pregnant women should be encouraged to abstain from all tobacco products including ECs. The use of advanced methodologies is recommended for future research to more robustly account for shared risk factors.
Start Date
2-11-2020 11:00 AM
End Date
2-11-2020 12:00 PM
Session 2 - Opportunities for Improving Health through Big Data and Data Science: Propensity Score Methods for Reducing Bias in Observational Studies: An Example from the PRAMS Study
Pheasant Room 253 A/B
Background: Conventional cigarette (CC) smoking is one of the most preventable causes of adverse birth outcomes. Although electronic cigarettes (ECs) are considered to be safer than CCs during pregnancy, the evidence is yet to be presented. This study examines the effects of EC use on birth outcomes compared to those of CC smokers using multiple analytic methods to adjust for risk factors of tobacco use. Methods: Data were extracted from 3,727 exclusive CC smokers and exclusive EC users who participated in the Phase 8 survey of the Pregnancy Risk Assessment Monitoring System (PRAMS). Adverse outcomes included small-for-gestational-age (SGA), low birthweight (LBW), and prematurity. Analyses were conducted using three different methodologies to account for nine covariates: 1) weighted, unadjusted logistic regression, 2) weighted, adjusted logistic regression, 3) inverse propensity weighted (IPW) logistic regression. Results: While unadjusted logistic regression analyses showed variable favorable effects of EC use on SGA (OR 0.64; 95% CI 0.39–1.04) and LBW (OR 0.58; 95% CI 0.40–0.82) compared to CC smoking, accounting for nine covariates using IPW rendered the results statistically insignificant. Using ECs does not significantly decrease the odds of neonates being SGA (OR 0.64, CI 0.28–1.45), LBW (OR 0.65, CI 0.35–1.20) or born prematurely (OR 0.80, CI 0.37–1.75) compared to smoking CCs. Conclusion: The study fails to support the common perception that ECs are safer than CCs during pregnancy. Until the safety of prenatal EC use is demonstrated, all pregnant women should be encouraged to abstain from all tobacco products including ECs. The use of advanced methodologies is recommended for future research to more robustly account for shared risk factors.