Socioeconomic (SE) factors strongly affect child health and healthcare. On average, children from low SE backgrounds begin life with more health problems, and face more health risks from their homes and neighborhoods throughout childhood than do children from high SE circumstances. Physicians and insurers seeking to deliver high quality care must tailor medical services to children's SE background. However, they are not compensated for the added complexity of caring for children from low SE backgrounds. Consequently, they face incentives to avoid these patients, and under-recognize or under-treat SE-related health issues. Payment methods that recognize the added effort that low SE children require may improve the willingness of physicians and insurers to care for and insure these populations and improve the quality of care delivered to them. Risk adjustment (RA) is the statistically based approach used by the healthcare system to allocate resources so that those who care for complex patients are given more resources than those who do not. Existing pediatric RA methods are underdeveloped and perform poorly, predicting pediatric health care spending only half as well as adult spending. Further, we are not aware of any pediatric studies evaluating whether improving RA algorithms reduces incentives to "cream-skim" healthier patients or "dump" sicker ones (though many adult studies exist).
Our Specific Aims are (1) to evaluate whether geographically based SE information improves the ability of RA algorithms to predict pediatric spending;(2) to simulate the extent to which including geographically based SE information in RA algorithms reduces incentives for providers to avoid children from low SE backgrounds. We will conduct a series of 3 cross-sectional panel studies using claims data from Blue Cross Blue Shield of Massachusetts (BCBSMA) 2007-2010 and census tract data from the U.S. Census'American Community Survey (ACS) 2006-2010. Our study population will be the ~25,000 children insured by BCBSMA and cared for by primary care physicians affiliated with Boston Children's Hospital (BCH) through 72 privately-owned community-based offices located across Massachusetts and 2 BCH-owned hospital-based practices. These datasets contain all the information needed to "geocode" SE variables from patient addresses and to calculate total annual medical spending (i.e., services provided at all institutions, not just BCH). Preliminary analyses show that the patient population is geographically and socioeconomically diverse. Knowing whether geographically based SE information can improve pediatric RA models will fill a critical knowledge gap in the pediatric RA field, and advance our understanding of how to better align healthcare payments with pediatric patient complexity. The study will facilitate ongoing health system and policy efforts aimed at ensuring that children from low SE backgrounds have access to insurance and medical care and narrowing disparities in child health and healthcare.
This study examines the degree to which incorporating geographically based socioeconomic (SE) information advances our understanding of the risk adjustment methods needed to better match healthcare payments with pediatric patient complexity. The study will facilitate ongoing health system and policy efforts aimed at ensuring that children from low SE background have access to insurance and medical care and narrowing disparities in child health and healthcare.