Pittsburgh, NYC Public Health Departments Report Progress on Multi-Sector Data Efforts

Feb. 21, 2018
Executives from four big-city public health departments gathered online Dec. 14 to share insights from their work to use data to more closely align their work with housing, education, social service, economic development, and safety.

Executives from four big-city public health departments gathered online Dec. 14 to share insights from their work to use data to more closely align their work with housing, education, social service, economic development, and safety.

The officials from New York City, Baltimore, Pittsburgh and Seattle are all grantees of Data Across Sectors for Health (DASH), a program of the Robert Wood Johnson Foundation, which aims to identify barriers, opportunities, promising practices and indicators of progress for multi-sector collaborations to connect information systems and share data for community health improvement.

New York City

Kevin Konty, director of the methods unit in the division of epidemiology at the New York City Department of Health, described an effort to bring data from several city agencies and Medicaid data down to the neighborhood level rather than the larger census tracks.  “The motivation for the project stemmed from a focus on social determinants of health,” he said. The city created “Neighborhood Tabulation Areas (NTAs),” which hold about 15,000 people each and better align with actual neighborhood boundaries than the census’ public use microdata areas (PUMAs), which are about 140,000 people each and may be at a sale that is courser than the actual notion of a neighborhood, Konty said.

At the NTA level, the city can overlay on a map more than 100 indicators, including data from Medicaid, social services, corrections, education, and emergency department visits and hospitalizations.

Konty shared maps of Staten Island, and pointed out that premature mortality rates there vary quite a bit by NTA within that district. Public health officials can respond by working with community partners to respond to social determinants at that neighborhood level.

The data is allowing the city to identify health concerns at a neighborhood scale. It can monitor for pockets of high burden areas and uncover social determinants. “That can help drive community prevention planning and investments,” he said.

Allegheny County, PA

Pennsylvania’s Allegheny County, which includes Pittsburgh, has worked the University of Pittsburgh to create a data warehouse and done predictive modeling around social determinants and cardiovascular disease, according to Karen Hacker, director of the county’s health department.

The county, which as 1.2 million residents and 130 municipal governments, sought to build atop a regional health improvement plan already in place, Hacker said.

With data from three large managed care organizations, the project had de-identified information on approximately 60 percent of the area’s residents. In areas of high concentration of cardiovascular disease and risk, they could overlay a series of health issues and social determinants such as being on food stamps, obesity, diabetes, hypertension, and percent of housing in poor condition. The researchers looked at things like cardiovascular disease deaths overlayed with vacant properties. They tried to model what impact changes to those social determinants might have on cardiovascular health.

These datasets are being used to create a predictive model of cardiovascular disease in Allegheny County. The Public Health Dynamics Laboratory from the University of Pittsburgh has developed the Framework for Reconstructing Epidemic Dynamics (FRED) predictive model to predict infectious disease. FRED has now been modified to model chronic disease throughout the county and predict the effectiveness of possible interventions through the DASH project.

Hacker said the team has taken the information to all regional stakeholders, including an advisory coalition. Groups like the American Heart Association are using it to target their efforts. “The next step is considering how to use the data to refocus on other outcomes such as asthma and opioid overdoses,” she added.

Baltimore City

The Baltimore City Health Department is working with the Maryland state health information exchange, CRISP, to target fall prevention interventions. In 2015, more than 3 million older adults in the United States were treated for falls in the emergency department, said Darcy Phelan-Emrick, the city health department’s chief epidemiologist. “That is about the population of Utah, so it is a large public health burden,” she said. (Approximately 4,000 of those falls are in Baltimore.) “We are leveraging the HIE for a public health use case for surveillance of falls,” she said. The project is called B’Friend (Baltimore Falls Reduction Initiative Engaging Neighborhoods and Data).

Partners in the effort besides CRISP include Baltimore City Housing, the 311 system, social service providers, hospitals and academic institutions. “The goal is to decrease the rate of falls leading to emergency department visits or hospitalizations among older adults by one-third in three years in Baltimore City,” she said.

By pulling ICD-10 codes, the HIE can help Public Health do epidemiological analyses or “hot-spotting.” Public health is partnering with groups such as emergency medical services and the transportation department, which is responsible for repairing sidewalks. “We can alert community partners about ways they could focus services in subpopulations,” Phelan-Emrick explained. Working across sectors is more difficult than she expected, she added, but partners do get excited when they get new data that is local and meaningful.

Seattle and King County, Washington

The effort in Seattle and King County is to better align public health and housing efforts by linking data. Amy Laurent, an epidemiologist for the county, noted that 75 percent of adults who live in King County public housing also receive Medicaid and yet data between these two sectors isn’t linked.

The project links Medicaid claims data with public housing authority resident data. “It provides the public housing authorities a de-identified dataset and visualizations with coded health conditions to assess and evaluate, and allows them to take a deeper dive into the data,” she said. They can generate maps to help identify Medicaid enrollment opportunities. They are working to create longitudinal data sets rather than just snapshots of certain points in time. In the short-term, it will give them a better understanding of housing and health, and will lead to cross-sector partnerships.