|
Abstract:
|
After the Affordable Care Act was passed, many states in US expanded their Medicaid program and created marketplace to offer affordable health insurance options to low/middle income families in 2014. Stakeholders are interested in the effectiveness of such policy, particularly, the potential impact on individuals' health and healthcare utilization. To gauge the policy impact using available data, we propose a statistical framework for estimating the average population causal effect using cross-sectional survey data. Large population health surveys are major sources of data for policy planning and program evaluation. However, literature on how to appropriately conduct causal inference using complex survey data is relatively scarce. Propensity score based adjustments are popular in analyzing observational data. Ad-hoc propensity score adjustment incorporating survey weights have been used with complex survey data, without rigorous justification. We propose a new super population framework, which includes a pair of potential outcomes for every unit in the population, to streamline the propensity score analysis for complex survey data. Based on the proposed framework, we explore diffe
|
