This study examines how well risk assessment models identify the medical needs of children in managed care plans and will build a full information pediatric specific risk assessment model that uses all of the automated data available from health plans. Inadequate attention has been paid to the unique challenges of pediatric risk assessment. Our hypothesis is that current risk models are not sensitive enough to correctly identify medical need among children as a group and among children with special needs. Moreover, because these models are used to adjust managed care payments, we expect health plans will have financial incentives to avoid enrolling children and families with children, particularly children with special needs. Using a dataset comprised of the largest number of people involved in a pediatric risk assessment study, 1.9 million people, including 515,729 children, the specific aims of this research are to: 1) Determine whether current risk assessment models are sensitive enough to identify expected healthcare cost and utilization for children in managed health care plans. We will study how well risk assessment models predict medical need among children and families with children by comparing existing risk assessment models focused on the general population to the few pediatric specific risk assessment models. Specific attention will be paid to the needs of children with high cost chronic medical conditions. 2) Evaluate if there exist financial disincentives for health plans to enroll children and families with children. 3) Assess how additional health status information improves the predictive performance of risk assessment instruments. This study is unique because no comprehensive research has examined the sensitivity of risk assessment methods to identify medical need among children in managed care regardless of their insurance source. Pediatric risk assessment research has also not explored opportunities to use all information available in managed care information systems.
