Tracy L. Johnson, Ph.D., is Director of Health Care Reform Initiatives at Denver Health and Hospital Authority (DH) and Assistant Professor at the Colorado School of Public Health. She has more than 20 years of experience in health policy development, research, and evaluation. Her research spans the population health spectrum and includes health insurance coverage, population segmentation/risk stratification, delivery system design, health system financing and payment models, safety net provider issues, and health equity. In her current position, she advises DH on state and federal delivery system and payment reforms, designing and implementing its population segmentation strategy, and directing the evaluation of its $20 million Centers for Medicare and Medicaid Services (CMS) Health Care Innovation Award. The evaluation quantified the inpatient utilization and cost avoidance associated with a large-scale practice transformation effort to integrate population health approaches into the delivery of primary care services. Previously, Johnson served as owner and principal of Health Policy Solutions, Inc., a consulting firm specializing in strategic planning, health policy analysis, and health services research and evaluation. In that capacity, she advised the state agency that administers Medicaid and Child Health Plan Plus (CHP+) on programmatic and data-related issues. Johnson holds a Ph.D. in Health Policy from the Johns Hopkins Bloomberg School of Public Health.
Project: Risk Prediction for New South Wales Planning and Health Systems Evaluation: Proof of Concept
With an increasingly aging population and simultaneous growth in health expenditures, there is broad international interest in improving quality of care and overall population health while reigning in the total cost of care. Because a small percentage of the population accounts for a large share of overall health expenditures, use of statistical methods to predict which patients will be costly in the near future (predictive risk modeling) has emerged as an important strategy for delivery system design efforts focused on reducing avoidable utilization and for evaluating those efforts. Drawing on similar research conducted at DH, Johnson will seek to adapt a validated tool for predicting hospital patients’ likelihood of being rehospitalized and examine whether predictive performance may be improved by incorporating non-hospital information, including social and behavioral factors. This adapted tool will be used to identify cohorts of New South Wales (NSW) public hospital patients at high-risk of readmission in 2010 and 2015, for which a detailed descriptive analysis will be completed.
Additionally, a comparison of 30-day readmission rates for both cohorts will assess health system performance in NSW (and in individual hospital catchment areas), pre- and post-implementation of Australian health care reforms since 2010. The project will use NSW administrative datasets such as the NSW Admitted Patient Data Collection (APCD) and the NSW Emergency Department Data Collection, which include services provided to public patients in public hospitals. It will also employ the 45andUp Study, a large, prospective dataset that links population survey data with primary care, hospital, and pharmacy claims. The proposed research will have important policy implications in both the U.S. and Australia, while advancing predictive risk techniques relevant to administrators, clinicians, policymakers, and health services researchers internationally.