Integrating Social Determinants of Health Into Risk Adjustment Models: An Opportunity to Improve Care Coordination Strategies For Medicaid Beneficiaries

点击翻译按钮获取中文摘要

PROJECT SUMMARY The purpose of this project is to provide evidence of the value of measuring social determinants of health (SDH) in the Medicaid population. This project will demonstrate that risk stratification using traditional risk adjustment models (RAMs) can be improved significantly with the inclusion of SDH information in addition to the information these models typically rely on: age, gender, and medical diagnoses. In addition, this project will quantify incident and cumulative risk of various SDH factors on the onset of incident disease within a one year period, highlighting the need to address social and environmental disadvantages that drive health inequities so that Medicaid can achieve the triple aim: improved care and population health at lower cost. To accomplish our objectives, we will conduct a prospective cohort study that will enroll 9,600 adult Medicaid beneficiaries treated at two medical facilities located in Washington, DC. The study sample will be predominately black (i.e. 90%) with a slightly higher percentage of females (60%). Medicaid beneficiaries who participate will complete a comprehensive SDH assessment during their initial medical encounter that will include validated SDH questions that measure housing stability, food availability, financial strain, health behaviors, social support, etc. Six and 12 months after enrollment, we will conduct a telephone follow-up interview with subjects to determine if there have been any major changes to their social and environmental circumstances. We will merge the interview data with DC Medicaid claims data (two years pre and one year post enrollment date). We will use the prior claims data to characterize health care utilization and expenditures by: (1) different types of service (i.e., inpatient care, outpatient care, prescription drugs, etc); (2) acute versus chronic treatments; and (3) preventable versus non preventable. We will risk stratify the study sample using three RAMs (i.e., Chronic Disability and Payment System, The Johns Hopkins Adjusted Clinical Groups (ACG) System, and 3M Clinical Risk Groups (CRGs). We will compare predictions of next year's health care utilization and expenditures for each RAM using generalized linear models that include or exclude SDH variables. We will also quantify the incident and cumulative risk associated with different SDH factors on disease onset during a one year follow-up period using generalized estimating equation models. To reduce the large health inequities that exist in the Medicaid population, this project aims to demonstrate the critical need to measure and address SDH in our care delivery strategies.