Other Co-Principal Investigator studies

R01MH136658

Horvitz-Lennon/Normand (multi-PI), Role: Co-PI

04/2024-03/2029

Serious mental illness (SMI), namely schizophrenia, bipolar disorder, and severe major depressive disorder (MDD), is associated with a heavy burden of disease, with racial/ethnic minorities disproportionately affected. This burden stems from enduring disability and significantly reduced life expectancy caused by a high risk for serious medical comorbidity and suicidality. Providing healthcare to adults living with SMI is costly, particularly for Medicaid, a key payer for this population. Despite high costs, adults diagnosed with SMI receive low quality of mental and physical health care, with quality generally lowest for racial/ethnic minorities.

State Medicaid programs have implemented policies, such as quality reporting and value-based payment, to improve quality of care or value of care (i.e., cost-effectiveness) for high-need beneficiaries including those with SMI. However, policymakers lack the necessary information to develop, target, implement, and evaluate these policies. For example, the associations between person- and area-level characteristics and quality of SMI care are not well understood, precluding the targeting of corrective interventions. Moreover, concerns have been raised about SMI quality measurement, with some arguing that available process-based indicators do not cover key areas of SMI care and may not be associated with key patient outcomes including social outcomes. Crucially, the quality and cost effects of the broad adoption of telehealth to deliver mental health care following the COVID-19 pandemic are not well understood, nor are its effects on disparities. Last, because little is known about the relationships between quality, health and social outcomes, and Medicaid and other costs or their interplay with race/ethnicity, policymakers lack a full understanding of the state budgetary and societal impacts of interventions to improve the quality and equity of SMI care.

We propose a research program to improve the quality, value, and equity of Medicaid-funded care received by adults with SMI. To achieve this goal, we will leverage an ongoing partnership with the New York State Medicaid program’s mental health authority, availability of several patient-level datasets, and expertise in quality measurement, Medicaid policy, policy analysis, racial/ethnic disparities and social determinants of health, and causal inference approaches using machine learning. Specifically, we seek to fill evidence gaps hindering policymakers’ efforts to improve quality, equity, and value of SMI care, including: What are key predictors of quality of care? Do “whole person” quality measures constructed with process-based indicators impact health and social outcomes? What are the effects of telehealth on quality and costs of mental health care? What are the costs (and cost offsets) associated with providing higher quality of care, from the perspective of Medicaid and from a broader state budgetary perspective? Do these associations vary by race/ethnicity? To address these questions, we will link Medicaid and other person-level datasets (e.g., New York City homelessness data) to create a 2017-2023 SMI cohort and construct health (e.g., suicidality), social (e.g., homelessness), and healthcare cost outcomes. Predictors of quality will be obtained from this multi-domain dataset (e.g., social risk, provider characteristics) and from public domain data (area-level social risk). The literature will provide information on other costs. Our Specific Aims are to:

Aim 1. Identify person- and area-level predictors of quality of care and determine if the associations vary by race/ethnicity. Using composite measures of quality of mental health and physical health care and machine learning algorithms to identify predictors of quality, we will characterize subgroups of patients receiving lower versus higher quality and assess the modifying effect of race/ethnicity.

Aim 2. Estimate the effects of receipt of high-quality care on health and social outcomes and determine if the effects vary by race/ethnicity. We will employ data adaptive algorithms for causal inference to quantify effects of quality of mental health and physical health care on up to 6-year outcomes and assess the modifying effect of race/ethnicity. 

Aim 3. Estimate the effects of telehealth on quality of mental health care and costs to Medicaid and determine if the effects vary by race/ethnicity. We will use a quasi-experimental approach to assess if quality of mental health care and costs to Medicaid (mental health care and total) differ for patients accessing care through telehealth versus in-person visits and assess the modifying effect of race/ethnicity. 

Aim 4. Develop alternative quality-improving interventions and compare their effects on health and social outcomes, Medicaid and broader state spending, and racial/ethnic equity. Using results from Aims 1-3 and existing evidence, we will develop ≥3 alternative interventions and estimate and compare their effects on outcomes of high significance to patients, state policymakers, and society.

            Our proposed research is responsive to PAR-23-095 and aligned with the NIMH Strategic Plan’s Goal 4 (Advance Mental Health Services to Strengthen Public Health) because among other tasks and as called by Goal 4, we will (a) identify mutable factors that are likely to influence disparities in quality and outcomes for underserved groups, (b) use large representative data sets and novel computational approaches to improve mental health care and its outcomes, and (c) evaluate the impacts of an innovation (telehealth). 

Project Number: 1RF1AG083033-01
Name of PD/PI: Hsu, John/Normand, Sharon-Lise
09/2023-08/2026

This application proposes to examine potential disparities in skilled nursing/therapy delivered in
beneficiary homes resulting from new Medicare policies.