Association of a Targeted Population Health Management Intervention with Hospital Admissions and Bed-Days for Medicaid-Enrolled Children

IMPORTANCE: As the proportion of children with Medicaid coverage increases, many pediatric health systems are searching for effective strategies to improve management of this high-risk population and reduce the need for inpatient resources. OBJECTIVE: To estimate the association of a targeted population health management intervention for children eligible for Medicaid with changes in monthly hospital admissions and bed-days. DESIGN, SETTING, AND PARTICIPANTS: This quality improvement study, using difference-in-differences analysis, deployed integrated team interventions in an academic pediatric health system with 31 in-network primary care practices among children enrolled in Medicaid who received care at the health system’s hospital and primary care practices. Data were collected from January 2014 to June 2017. Data analysis took place from January 2018 to June 2019. EXPOSURES: Targeted deployment of integrated team interventions, each including electronic medical record registry development and reporting alongside a common longitudinal quality improvement framework to distribute workflow among interdisciplinary clinicians and community health workers. MAIN OUTCOMES AND MEASURES: Trends in monthly inpatient admissions and bed-days (per 1000 beneficiaries) during the preimplementation period (ie, January 1, 2014, to June 30, 2015) compared with the postimplementation period (ie, July 1, 2015, to June 30, 2017). RESULTS: Of 25 460 children admitted to the hospital’s health system during the study period, 8418 (33.1%) (3869 [46.0%] girls; 3308 [39.3%] aged ≤1 year; 5694 [67.6%] black) were from in-network practices, and 17 042 (67.9%) (7779 [45.7%] girls; 6031 [35.4%] aged ≤1 year; 7167 [41.2%] black) were from out-of-network practices. Compared with out-of-network patients, in-network patients experienced a decrease of 0.39 (95% CI, 0.10-0.68) monthly admissions per 1000 beneficiaries (P = .009) and 2.20 (95% CI, 0.90-3.49) monthly bed-days per 1000 beneficiaries (P = .001). Accounting for disproportionate growth in the number of children with medical complexity who were in-network to the health system, this group experienced a monthly decrease in admissions of 0.54 (95% CI, 0.13-0.95) per 1000 beneficiaries (P = .01) and in bed-days of 3.25 (95% CI, 1.46-5.04) per 1000 beneficiaries (P = .001) compared with out-of-network patients. Annualized, these differences could translate to a reduction of 3600 bed-days for a population of 93 000 children eligible for Medicaid. CONCLUSIONS AND RELEVANCE: In this quality improvement study, a population health management approach providing targeted integrated care team interventions for children with medical and social complexity being cared for in a primary care network was associated with a reduction in service utilization compared with an out-of-network comparison group. Standardizing the work of care teams with quality improvement methods and integrated information technology tools may provide a scalable strategy for health systems to mitigate risk from a growing population of children who are eligible for Medicaid.