Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper analyzes 30-day "treatment spells"—fixed-length periods that commence with a service after a gap in provider contact—to examine how health-care utilization and spending of insured employees at large firms are influenced by health plan types. We focus on differences between preferred provider organizations (PPOs) and two recent innovations: plans that feature a narrow panel of providers, and plans that allow free choice of providers but increase demand-side cost sharing: consumer-driven/high-deductible health plans. Health plan effect estimates change dramatically after controlling for endogenous plan type choice, and individual fixed effects. With these controls, narrow-panel plans reduce the probability of new treatment spells relative to PPOs by 34 percent with little effect on chronic, repeat visit spells. Visit reductions are more concentrated in less severe conditions in narrow-network plans, hence diagnostic coding on the remaining patients increases. We find no evidence that either narrow panel or higher-cost-sharing plans pay lower prices per procedure or have less intensive treatment given initiation of treatment. With controls, consumer-driven/high-deductible health plans are associated with higher total spending on procedures than PPO plans.