Analyze

The index produces an overall score derived from 12 indicators distributed across four domains (current infection rates, baseline population concentration, current interventions, and health and environmental vulnerabilities. Each vulnerability factor is represented as a slide of a radar chart (see below).

The dashboard can also be used to visualize changes over time in cases, deaths, PVI, and PVI rank (with a line chart and a bar chart), as well as predicted changes in cases and deaths (with a line chart), see below.

Additional visual layers can be added to the PVI map (e.g. number of cases and deaths).

The PVI offers a visual synthesis of information to monitor disease trajectories, identify local vulnerabilities, forecast outcomes, and guide an informed response (e.g. allocating resources). This includes short-term, local predictions of cases and deaths. The PVI dashboard creates profiles (called PVI scorecards) for every county in the United States.

The PVI dashboard can be customized to specific needs by adding or removing layers of information, filtering by region, or clustering by profile similarity. The Predictions panel connects historical tracking to local forecasts of cases and deaths. The dashboard applies an integrated concept of vulnerability composed of both dynamic (infection rate and interventions) and static (community population and health care access) factors.

The statistical modeling supporting the PVI dashboard (generalized linear models of cumulative outcome data) has indicated that following population size, the most significant predictors of cases and deaths were the proportion of Black residents, mean fine particulate matter [particulate matter ≤2.5μm in diameter (PM2.5)], percentage of population with insurance coverage, and proportion of Hispanic residents.

The ToxPi*GIS framework, from which the PVI was built, is a free tool that integrates data streams from different sources into interactive profiles that overlay geographic information systems (GIS) data. This enables people using the tool to compare, cluster, and evaluate the sensitivity of a statistical framework to component data streams. In other words, this enables the integration of data that are not normally compared (data are combined into a matrix comprised of various domains or categories, varying weights and represented by color schemes).

County

Data is drawn from the Social Vulnerability Index (SVI) of the Centers for Disease Control and Prevention (CDC), testing rates from the COVID tracking project (produced by the Atlantic Monthly Group), social distancing metrics from mobile device data, and USA Facts’ measures of disease spread and case numbers.

The creation of the Covid-19 Pandemic Vulnerability Index (PVI) Dashboard was a collaborative effort between scientists from the National Institute of Environmental Health Sciences, North Carolina State University, and Texas @&M University. Their mission was to provide a resource to support dynamic, community-level decision-making in response to the Covid-19 Pandemic.

Each PVI county profile is calculated using Toxicological Prioritization Index (ToxiPi) software, which integrates data within a geospatial context. ToxPi*GIS is meant to promote the development of targeted, effective community policies. ToxPi*GIS was created by the Reif Lab at North Carolina State University. The overarching goal of the lab is to understand the interactions between human health and the environment through the application of analytical/visual methods and experimental design. Data sources include epidemiological studies of human health, high-throughput screening (HTS) of environmental chemicals and model organism data. The lab is run by Dr. David Reif (Professor in the Department of Biological Sciences) and members include students from several degree programs as well as post-doctoral and senior scientists.

The HPI draws data about 25 community characteristics into a single indexed HPI score. The includes sub-scores for 8 “Policy Action Areas”: Economic, Education, Housing, Health Care Access, Neighborhood, Clean Environment, Transportation, and Social Factors. These scores are meant to be used to evaluate health geographically. Each policy action area includes the following individual indicators and weights:

ECONOMIC (0.32)

• Poverty
• Employment
• Income

EDUCATION (0.19)

• Pre-school enrollment
• High school enrollment
• Bachelors attainment

HEALTHCARE (0.05)

• Insured adults

HOUSING (0.05)

• Severe cost burden low income
• Homeownership
• Kitchen and plumbing
• Crowding

NEIGHBORHOOD (0.08)

• Retail jobs
• Supermarket access
• Parks
• Tree canopy
• Alcohol establishments

CLEAN ENVIRONMENT (0.05)

• Diesel PM
• Ozone
• PM2.5
• Drinking water

SOCIAL (0.10)

• Two parent household
• Voting

TRANSPORTATION (0.16)

• Healthy community
• Automobile access

*The steering committee for the HPI sought to include race/ethnicity as a 9th policy action area, but they were prohibited from doing so by state law which does not allow California state agencies to use race as a basis for public contracting.

The primary HPI Index is designed to align with life expectancy at birth as a predictive measure of community health status. However, the Healthy Places mapping tool can also be used to create custom scores using different indicators. The mapping tool includes detailed definitions of each indicator.

Each indicator is linked to a policy guide, which outlines concrete actions (e.g. best practices, emerging policy options) that local jurisdictions can take to improve HPI indicators. These actions are sometimes aimed at addressing direct links between policy and an action area, and other times aimed at addressed the root causes of an action area. The mapping tool also enables filtering results by “Decision support layers” like health outcomes, health risk behaviors, race/ethnicity, climate change effects, and other layers that the alliance identifies as important for advancing “resilient, equitable communities in California”. Geographies (e.g. census tracts) can also be compared by indicator using a ranking tool. The pool function can be used to create customized aggregations of data to map (e.g. adding several census-tracts together).

Data is available at several different scales: census-tracts, congressional districts, state assembly districts, state senate districts, cities, core based statistical areas, elementary school districts, metropolitan planning organization and medical service study areas.  

Creating maps by different combinations of indicators or geographic aggregations could be tinkered with to produce provocative data visualizations. Ranking scores can be used to draw distinction between different census tracts. However, clear inequities are evident even without these adjustments, with the HPI index score clearly demonstrating noticeable differences across geographies. 

The California Healthy Places Index is made available by the Public Health Alliance of Southern California. Their mission is to “make health equity and racial justice a reality” through collaboration and data (https://www.thepublichealthalliance.org/). They engage in advocacy and mobilization to generate this change. They are composed of a coalition of executives representing 10 local health jurisdictions in Southern California (including Long Beach, Los Angeles, Orange, and Riverside, among others), an area they highlight as representing 60% of California’s population (with which they blur the boundaries between “California” and “Southern California”).

The alliance emphasizes pursuing equity using publicly available data and collaboration (with government agencies, legislators, hospitals, health plans, philanthropy, and community advocates). They present the Healthy Places Index (HPI) as a tool for exploring how life expectancy is impacted by community conditions.

More specifically, the HPI was created by a steering committee made up of epidemiologists and 3 public health coalitions led by the alliance.

This data has been used for assessments, decision making, and planning on a state, regional and local level in California by a wide range of actors, which include:

·  California Department of Public Health

·  Governor’s Office of Planning and Research

·  California Environmental Justice Alliance

·  the Hospital Association of Southern California

·  County Public Health Departments

·  Local/regional healthcare providers

For instance, Kaiser Permanente used the HPI in conducting a community health needs assessments for several areas in southern California (to comply with federal tax law requiring them to conduct a health needs assessment at least once every three years). They used the tool to identify the most under-resourced geographic communities and identify the factors that are most predictive of negative health outcomes. 

For this community health assessment, researchers also consulted residents, community leaders, government and public health department representatives through surveys, stakeholder interviews, and focus groups. The assessment identified several health needs that needed to be prioritized: access to healthcare, economic security, mental health, stroke, and suicide. This was used to guide implementation strategies in partnership with community-based organizations, hospitals and groups (e.g. identifying reducing food insecurity as a strategic priority and designing/implementing food benefits programs).

Other reports using the HPI include the Solano County Public Health Departments’ report entitled “Maternal and Child Verification of Cumulative Health Impacts from Social Factors,” the Los Angeles County Department of Public Health city and community health profiles, and the California Environmental Justice Alliance’s SB 1000 Toolkit.