Grant Recipient: Mathematica Policy Research
Term: 2019 –2024
Principal Investigators: Randall Brown, Ph.D., Mathematica
Arkadipta Ghosh, Ph.D., Mathematica
Summary: TCM is a nurse-led hospital discharge and home follow-up program for chronically ill older adults, aimed at preventing health complications and re-hospitalizations and improving patients’ experience with care. TCM is delivered by a Master’s level transitional care nurse who works with the patient, his/her family, and the patient’s doctors while the patient is hospitalized to develop an individualized plan of care and, following discharge, accompanies the patient to his/her first physician visit and conducts an average of 12 home visits over three months to monitor symptoms and ensure the patient is taking medications. The nurses also will help patients address depression and social needs that can increase their risk of readmission.
TCM is backed by uniquely promising evidence of sizable reductions in rehospitalizations and net healthcare costs. Two well-conducted RCTs in Philadelphia, published in 1999 and 2004, found reductions in rehospitalizations of 30 – 50%, and net healthcare cost savings of approximately $4,500 per patient, within 5 – 12 months after patient discharge.
Under this project, the developer of TCM at UPenn will partner with four health systems (including nine hospitals) that serve diverse populations in five states – Washington, Michigan, California, Missouri, and Ohio. UPenn and the hospitals will participate in an RCT to determine whether the impacts found in the prior RCTs can be reproduced under expanded implementation conditions in health systems that are diverse in terms of geographic location, patient demographics, post-acute services offered, payer mix, and hospital size. Mathematica will enroll a total sample of 1,600 patients (800 treatment and 800 control), with each hospital system contributing 400 patients to the sample. The study will estimate program effects on hospital readmissions, and healthcare costs and utilization, over the 12 months post-random assignment using data from administrative records (obtained from Medicare claims, Medicare Advantage plans, and hospitals). The study will also estimate program effects on health-related quality of life measured in a one-time survey administered 3 months post- random assignment. The study will also investigate how program impacts vary across health system, patient characteristics (including primary diagnosis), insurer, and fidelity to the intervention.