Assessing performance of ZCTA-level and Census Tract-level social and environmental risk factors in a model predicting hospital events.

Predictive analytics are used in primary care to efficiently direct health care resources to high-risk patients to prevent unnecessary health care utilization and improve health. Social determinants of health (SDOH) are important features in these models, but they are poorly measured in administrative claims data. Area-level SDOH can be proxies for unavailable individual-level indicators, but the extent to which the granularity of risk factors impacts predictive models is unclear. We examined whether increasing the granularity of area-based SDOH features from ZIP code tabulation area (ZCTA) to Census Tract strengthened an existing clinical prediction model for avoidable hospitalizations (AH events) in Maryland Medicare fee-for-service beneficiaries. We created a person-month dataset for 465,749 beneficiaries (59.4% female; 69.8% White; 22.7% Black) with 144 features indexing medical history and demographics using Medicare claims (September 2018 through July 2021). Claims data were linked with 37 SDOH features associated with AH events from 11 publicly-available sources (e.g., American Community Survey) based on the beneficiaries' ZCTA and Census Tract of residence. Individual AH risk was estimated using six discrete time survival models with different combinations of demographic, condition/utilization, and SDOH features. Each model used stepwise variable selection to retain only meaningful predictors. We compared model fit, predictive performance, and interpretation across models. Results showed that increasing the granularity of area-based risk factors did not dramatically improve model fit or predictive performance. However, it did affect model interpretation by altering which SDOH features were retained during variable selection. Further, the inclusion of SDOH at either granularity level meaningfully reduced the risk that was attributed to demographic predictors (e.g., race, dual-eligibility for Medicaid). Differences in interpretation are critical given that this model is used by primary care staff to inform the allocation of care management resources, including those available to address drivers of health beyond the bounds of traditional health care.

Association of Participation in the Maryland Primary Care Program With COVID-19 Outcomes Among Medicare Beneficiaries.

Importance: Advanced primary care is a team-based approach to providing higher-quality primary care. The association of advanced primary care and COVID-19 outcomes is unknown. Objective: To evaluate the association of advanced primary care with COVID-19 outcomes, including vaccination, case, hospitalization, and death rates during the first 2 years of the COVID-19 pandemic. Design, Setting, and Participants: This retrospective cohort study used Medicare claims data from January 1, 2020, through January 31, 2022, and Maryland state vaccination data. All Part A and B Medicare claims for Maryland Medicare beneficiaries were included. The study population was divided into beneficiaries attributed to Maryland Primary Care Program (MDPCP) practices and a matched cohort of beneficiaries not attributed to MDPCP practices but who met the eligibility criteria for study participation from January 1, 2020, through December 31, 2021. Eligibility criteria for both groups included fee-for-service Medicare beneficiaries who were eligible for attribution to the MDPCP. A forced-match design was used to match both groups in the study population by age category, sex, race and ethnicity, Medicare-Medicaid dual eligibility status, COVID-19 Vulnerability Index score, Maryland county of residence, and primary care practice participation. Exposures: Primary care practice participation in the MDPCP. Main Outcomes and Measures: Primary outcome variables included rate of vaccination, monoclonal antibody infusion uptake, and telehealth claims. Secondary outcomes included rates of COVID-19 diagnosis, COVID-19 inpatient claims, COVID-19 emergency department claims, COVID-19 deaths, and median COVID-19 inpatient admission length of stay. Claims measures were assessed from January 1, 2020, through October 31, 2021. Vaccination measures were assessed from January 1, 2020, through March 31, 2022. Results: After matching, a total of 208 146 beneficiaries in the MDPCP group and 37 203 beneficiaries in the non-MDPCP group were included in this study, comprising 60.10% women and 39.90% men with a median age of 76 (IQR, 71-82) years. Most participants (78.40% and 78.38%, respectively) were White. There were no significant demographic nor risk measure baseline differences between the 2 groups. The MDPCP beneficiaries had more favorable primary COVID-related outcomes than non-MDPCP beneficiaries: 84.47% of MDPCP beneficiaries were fully vaccinated, compared with 77.93% of nonparticipating beneficiaries (P < .001). COVID-19-positive beneficiaries in MDPCP also received monoclonal antibody treatment more often (8.45% vs 6.11%; P < .001) and received more care via telehealth (62.95% vs 54.53%; P < .001) compared with nonparticipating counterparts. In terms of secondary outcomes, beneficiaries in the MDPCP had lower rates of COVID-19 cases (6.55% vs 7.09%; P < .001), lower rates of COVID-19 inpatient admissions (1.81% vs 2.06%; P = .001), and lower rates of death due to COVID-19 (0.56% vs 0.77%; P < .001) compared with nonparticipating beneficiaries. Conclusions and Relevance: These findings suggest that participation in the MDPCP was associated with lower COVID-19 case, hospitalization, and death rates, and advanced primary care and COVID-19 response strategies within the MDPCP were associated with improved COVID-19 outcomes for attributed beneficiaries.

Predicting avoidable hospital events in Maryland.

OBJECTIVE: To develop and validate a prediction model of avoidable hospital events among Medicare fee-for-service (FFS) beneficiaries in Maryland. DATA SOURCES: Medicare FFS claims from Maryland from 2017 to 2020 and other publicly available ZIP code-level data sets. STUDY DESIGN: Multivariable logistic regression models were used to estimate the relationship between a variety of risk factors and future avoidable hospital events. The predictive power of the resulting risk scores was gauged using a concentration curve. DATA COLLECTION/EXTRACTION METHODS: One hundred and ninety-eight individual- and ZIP code-level risk factors were used to create an analytic person-month data set of over 11.6 million person-month observations. PRINCIPAL FINDINGS: We included 198 risk factors for the model based on the results of a targeted literature review, both at the individual and neighborhood levels. These risk factors span six domains as follows: diagnoses, pharmacy utilization, procedure history, prior utilization, social determinants of health, and demographic information. Feature selection retained 73 highly statistically significant risk factors (p < 0.0012) in the primary model. Risk scores were estimated for each individual in the cohort, and, for scores released in April 2020, the top 10% riskiest individuals in the cohort account for 48.7% of avoidable hospital events in the following month. These scores significantly outperform the Centers for Medicare & Medicaid Services hierarchical condition category risk scores in terms of predictive power. CONCLUSIONS: A risk prediction model based on standard administrative claims data can identify individuals at risk of incurring a future avoidable hospital event with good accuracy.

A Population Health Measurement Framework: Evidence-Based Metrics for Assessing Community-Level Population Health in the Global Budget Context.

Population health is one of the pillars of the Triple Aim to improve US health care. The authors developed a framework for population health measurement and a proposed set of measures for further exploration to guide the population health efforts in Maryland. The authors searched peer-reviewed, expert-authored literature and current public health measures. Using a semi-structured analysis, a framework was proposed, which consisted of a conceptual model of several domains and identified population health measures addressing them. Stakeholders were convened to review the framework and identified the most feasible population health measures considering the underlying health information technology (IT) infrastructure in Maryland. The framework was organized based on health system factors, determinants of health, and population-based and clinical outcomes. Measurement specifications were developed that addressed different aspects of selected measures and assessed various national and local data sources for selected measures. Data sources were identified based on their key characteristics, challenges, opportunities, and potential applicability to the proposed measures, as well as the issue of interoperability of data sources among different organizations. The proposed framework and measures can act as a platform to quantify the determinants of health and the state overall population health goals. Key considerations for developing a population health measures framework include health IT infrastructure, data denominators, feasibility, health system environment, and policy factors. Measurement development and progression using the framework will largely depend on the users' focus areas and availability of data. The authors believe that the proposed framework and road map can serve as a model for communities elsewhere.

A State-wide Health IT Infrastructure for Population Health: Building a Community-wide Electronic Platform for Maryland's All-Payer Global Budget.

Maryland Department of Health (MDH) has been preparing for alignment of its population health initiatives with Maryland's unique All-Payer hospital global budget program. In order to operationalize population health initiatives, it is required to identify a starter set of measures addressing community level health interventions and to collect interoperable data for those measures. The broad adoption of electronic health records (EHRs) with ongoing data collection on almost all patients in the state, combined with hospital participation in health information exchange (HIE) initiatives, provides an unprecedented opportunity for near real-time assessment of the health of the communities. MDH's EHR-based monitoring complements, and perhaps replaces, ad-hoc assessments based on limited surveys, billing, and other administrative data. This article explores the potential expansion of health IT capacity as a method to improve population health across Maryland. First, we propose a progression plan for four selected community-wide population health measures: body mass index, blood pressure, smoking status, and falls-related injuries. We then present an assessment of the current and near real-time availability of digital data in Maryland including the geographic granularity on which each measure can be assessed statewide. Finally, we provide general recommendations to improve interoperable data collection for selected measures over time via the Maryland HIE. This paper is intended to serve as a high level guiding framework for communities across the US that are undergoing healthcare transformation toward integrated models of care using universal interoperable EHRs.

Innovative Payment Mechanisms in Maryland Hospitals: An Empirical Analysis of Readmissions under Total Patient Revenue.

BACKGROUND: The state of Maryland implemented innovative budgeting of outpatient and inpatient services in eight rural hospitals under the Total Patient Revenue (TPR) system in July, 2010. METHODS: This paper uses data on Maryland discharges from the 2009-2011 Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID). Individual inpatient discharges from eight treatment hospitals and three rural control hospitals (n=374,353) are analyzed. To get robust estimates and control for trends in the state, we also compare treatment hospitals to all hospitals in Maryland that report readmissions (n=1,997,164). Linear probability models using the difference-in-differences approach with hospital fixed effects are estimated to determine the effect of the innovative payment mechanisms on hospital readmissions, controlling for patient demographics and characteristics. RESULTS: Difference-in-differences estimates show that after implementation of TPR in the treatment hospitals, there were no statistically significant changes in the predicted probability of readmissions. CONCLUSIONS: Early evidence from the TPR program shows that readmissions were not affected in the 18 months after implementation. IMPLICATIONS: : As the health care system innovates, it is important to evaluate the success of these innovations. One of the goals of TPR was to lower readmission rates, however these rates did not show consistent downward trends after implementation. Our results suggest that payment innovations that provide financial incentives to ensure patients receive care in the most appropriate setting while maintaining quality of care may not have immediate effects on commonly used measures of hospital quality, particularly for rural hospitals that may lack coordinated care delivery infrastructure.