Established Data Infrastructure Enables Local Governments to Meet Need During Natural Disasters & Health Emergencies
By: Emily Berkowitz & Matthew Katz
In 2020, the United States has seen devastating health and economic impacts from the COVID-19 pandemic, a historically destructive wildfire season on the West Coast, and one of the worst hurricane seasons on record. Crises like these historically cause the greatest harm to the most under-resourced groups in society. Without ample government support directed intentionally to those experiencing the most need, increased inequality resulting from natural disasters or health emergencies will continue to worsen.
Use of cross-sector administrative data can be a key tool for identifying effective crisis responses that meet the needs of those most affected. In a previous post, we discussed examples of data use specifically related to COVID-19. In this post, we broaden our lens to consider how state and local governments can leverage their data assets to respond more effectively to future natural disasters and health emergencies. Importantly, we find that communities that invest in developing strong data-sharing relationships and data infrastructure prior to a crisis have a leg up, and provide examples of ways in which several have rapidly adapted to meet emerging needs and improve equity.
Lessons from previous crises
In the United States, both the health and economic effects of COVID-19 disproportionately impact low-income communities and Black, Indigenous, and People of Color (BIPOC)1. This phenomenon isn’t new; during past crises and natural disasters, BIPOC have consistently faced greater rates of displacement and loss of material possessions. These disparities are consistent across natural disasters and regions. For example, during Hurricane Katrina in 2005, Superstorm Sandy in 2012, and Hurricane Harvey in 2017, residents living below the poverty line and identifying as BIPOC were more likely to live in neighborhoods that experienced flooding.
During wildfire season in California, older adults and low-income communities are the most likely to be displaced or remain unhoused as a result of fires. Though fire-prone areas in the U.S. are more likely to be populated by high-income households, the impacts experienced by individuals in rural areas, low-income neighborhoods, and immigrant communities are more long-lasting. These communities are less likely to have access to the necessary resources, including insurance and previous investment in fire safety, to withstand destruction and loss. As a result, disparities in preparedness and recovery compound existing inequities.
With a more fine-grained geographic view of risk and impact, governments are better able to respond equitably during emergencies. In 2019, the University of Michigan’s Poverty Solutions Initiatives and Princeton University’s Center for Research on Child Well-Being linked Census and administrative data in three domains: 1) income, using poverty and deep poverty rates; 2) health, using life expectancy and low birth weight measures; and 3) social mobility, using new social mobility estimates for counties and cities. The resulting dataset was used to construct an “index of deep disadvantage” in order to map the communities with the most disparate economic and social conditions. University researchers conducting field interviews in Nichols, South Carolina — a community identified by the index as deeply disadvantaged — found most Black families were still displaced almost three years after Hurricane Matthew (2016) and one year after Hurricane Florence (2018), despite $400 million in hurricane recovery aid allocated to the region2. The mixed-methods approach to studying disadvantage helped lift up stories from residents who either didn’t know about or were unable to complete the paperwork necessary to receive aid.
As this example demonstrates, integrated data can provide researchers with the ability to identify vulnerable communities and assess the root drivers of disadvantage. In addition, integrated data has the potential to take us a step further, providing local and state leaders with actionable information to assist in allocating resources before, during, and after disasters. Actionable Intelligence for Social Policy (AISP) helps state and local governments collaborate to build such approaches, which we refer to as Integrated Data Systems (IDS). IDS provide the rules of the road, as well as the human and technical capacity for routine data sharing and integration.
Unfamiliar with administrative data reuse and IDS? New to data sharing and integration? Check out AISP’s Introduction to Data Sharing and Integration document.
Balancing risks & benefits of data sharing
Like any tool, IDS can be misused and are not without risks. Legacies of racism in policy and planning have locked many low-income communities and BIPOC out of opportunities. BIPOC are also overrepresented in government datasets on public services. Broader, more expansive records, like birth records and DMV records, can be difficult to access, and are less frequently employed by data integration efforts in the U.S.
For this reason, communities building data linkage infrastructure should balance the risks and benefits of sharing and engage diverse stakeholders to grapple with three key questions:
- Is this linkage legal?
- Is it ethical?
- Is it a good idea?
An important note: it may be difficult to engage in a meaningful risk/benefit analysis with stakeholders about starting to build an IDS during times of crisis when fast action is required, and when the risk of exacerbating inequities is already heightened. Communities that have already done the hard work of building both infrastructure and public trust will be much better positioned to pivot their existing data work to address urgent needs while still maintaining open dialogue with the public about how any new risks will be mitigated.
See A Toolkit for Centering Racial Equity Throughout Data Integration for a nuanced discussion of risk vs. benefit in data sharing and for actions your site can take to center racial equity in your work.
The value of integrated data in times of crisis
In order to mitigate rather than compound inequities, approaches to crisis planning and recovery will need to use an interdisciplinary approach to holistically respond to need. Typically, efforts will do this in one of the following ways:
1. Using datasets with personal identifiers removed to measure and map vulnerability and need.
2. Using datasets that include identifiable information to perform individual risk assessment and to allocate resources.
Using datasets without personal identifiers can involve a number of activities — from basic dashboarding at the community level to more in-depth analysis of individual trajectories at the aggregate level. Because personal information about individuals is removed or sufficiently safeguarded, this type of data integration and analysis is less sensitive and more common. The linking and use of datasets that include individual-level, identifiable information requires a more complex legal and technical approach and therefore is generally only undertaken by well-established and well-resourced partnerships.
For more on this differentiation, please see the footnote below3 and visit the following resource: https://fpf.org/2016/04/25/a-visual-guide-to-practical-data-de-identification/
Building a strong foundation for data use before a crisis
Data integration efforts continue to refine the core elements of IDS as they mature. A practice of continuous improvement and engagement pays off in times of crisis when priorities rapidly change, and data access is necessary to inform life-saving decisions.
As an example, in 2012, during Hurricane Sandy, New York City’s Center for Innovation through Data Intelligence (CIDI) collaborated with the Office of Emergency Management, National Guard, and other emergency response teams to identify areas of greatest need. When establishing their linked data infrastructure in the early 2000s, CIDI prioritized developing strong governance standards and supportive relationships with city agencies, allowing project needs to dictate IT infrastructure rather than the other way around. Doing so meant that when Super Storm Sandy hit, CIDI was ready to serve residents of the city in a way that was responsive to the needs of people and departments as well as technically and legally feasible. CIDI staff produced maps of residential units that were known to be without heat, electricity, or telephone service in the impacted areas of the city, and the data visualization allowed city workers to better plan how many responders to send to a specific area. Additionally, CIDI staff analyzed survey data collected on a daily basis by the National Guard to create multi-agency reports used to assess whether residents’ needs were being met by services from the appropriate agencies.
During less critical periods in NYC, CIDI has continued to build partner trust through responsible use of administrative data to produce actionable insights about service utilization and vulnerability. The following examples demonstrate how CIDI and other established IDS have leveraged their pre-crisis capacity to adapt in moments of crisis, using both identifiable datasets and those with personal identifiers removed.
Using datasets with personal identifiers removed to map vulnerability, need & dispossession
CIDI Maps Vulnerability During COVID-19
CIDI’s experience during Hurricane Sandy prepared them as a reliable, community-engaged organization able to help guide equitable response at the time when NYC was the epicenter of the COVID-19 pandemic. In 2020, CIDI rapidly repurposed their existing data capacity and vulnerability mapping tools to help other City agencies and community organizations provide emergency assistance as new priorities emerged amdist the health crisis. The Identification and Mapping System for Vulnerable Populations was developed by integrating data from NYC Health and Human Services (HHS) using indicators collaboratively identified by community-engaged workgroup members from City agencies. Given the segregation of NYC neighborhoods and schools, “vulnerability” clusters are inherently an issue of racial equity. For that reason, this practice of using diverse voices to determine measures of vulnerability is notable. Other efforts have embedded racial equity considerations in their response and recovery work through methods that capture lived experienced and other forms of public engagement.
In NYC, the system showcases geographical clusters of vulnerable people as well as each neighborhood’s concentration of shelter sites, housing authority buildings, and retirement communities. With CIDI at the helm, NYC was able to utilize this aggregated individual-level data to more accurately and equitably distribute aid.
LA County Data Capacity Used to Assess Pandemic Impacts on Unhoused Populations
In Los Angeles County, California, long established infrastructure has been leveraged to support people who are both experiencing homelessness and at a greater risk of contracting COVID-19. Specifically, researchers from LA County, the University of Pennsylvania, and UCLA worked with the Enterprise Linkages Project (ELP) — an IDS hosted by the LA Chief Executive Office of Research and Evaluation — to identify to sub-groups of the population at high risk for severe health complications from coronavirus exposure. By linking information from health care and homeless management information systems (HMIS) datasets, researchers were able to assess discrete levels of vulnerability among the aging homeless population, propose housing and service models that matched those levels of vulnerability. They were also able to estimate potential cost offsets to Medicaid and the County that would help recapture funds needed to help stabilize people in housing.
Using datasets that include identifiable information for risk assessment & resource allocation
King County, Washington Uses Data to Aid in Isolation & Quarantine Decision-Making
In King County, Washington, the Integrated Data Hub routinely uses their infrastructure and client-level data resources for program evaluation, internal quality improvement, and care coordination. At the start of COVID-19 lockdowns, officials in King County established a robust system for isolation and quarantine (I/Q) sites to assist people unable to safely isolate in their homes, including the unhoused population and congregate living residents. The I/Q facilities are maintained by the county’s Department of Community and Human Services (DCHS) in partnership with Public Health Seattle King County. Placement determinations are made by a team of experts who are able to assess the unique needs of clients. I/Q coordinators, however, were concerned about the sustainability of this process in the face of rapid spread of COVID-19. In response, King County DCHS and Public Health staff facilitated an extension to their existing data sharing agreement (DSA) with the Washington State Health Care Authority to allow for the use of identified Medicaid data to make treatment decisions about individuals seeking I/Q support during the COVID-19 response.
Using the Data Hub, DCHS staff were able to rapidly build out a decision support tool that provided a client-level view of potential I/Q guests. The tool brought together behavioral and physical health information and homeless response system engagement to prioritize assessments otherwise performed during coordinated calls between clinical staff. Luckily, the county has been able to stem the spread of COVID-19 to date and did not need to fully deploy the Data Hub’s tool. Nonetheless, having the capacity to build the resource was essential to crisis response preparation. DCHS’ existing infrastructure and well-established data-sharing relationship with the State allowed the team to quickly get approval for this new COVID-19 response tool.
Camden, NJ Uses Both Health Information Exchange & IDS to Identify Risks
Camden Coalition of Healthcare Providers — a multidisciplinary non-profit focused on the overlapping health and social needs of people in Camden, New Jersey — is currently running their Health Information Exchange (HIE) in conjunction with their IDS (Camden ARISE) in order to identify at-risk populations based on clinical markers from the CDC (e..g, respiratory conditions, cardiovascular conditions) with social vulnerability risk factors (e.g., food and/or housing insecurity, utility needs) to guide outreach. Community-based HIEs facilitate information and data exchange between healthcare providers and partner agencies in order to improve coordination of care, often with attention to social determinants of health. As such, this HIE routinely holds (and protects) identifiable information. By establishing authority to use information in connection with their IDS, the Camden Coalition is able to rapidly and safely merge existing information for a limited but clear purpose: to help the most at-risk populations during an unprecedented health crisis. The HIE and IDS are also being used to conduct population health surveillance with internal dashboards that track daily emergency department and inpatient encounters, as well as to better understand demographic and geographic trends in COVID-19 outcomes. As with other sites, these tools could not have been scaled so quickly without prior authority, relationships, and demonstrated use cases.
Curious about how states have used data integration capacity to respond to COVID-19? Check out the following resources: AISP’s previous blog on integrated data’s role in COVID-19 response and recovery, a report from the Beeck Center at Georgetown University on the role of state CDOs in supporting equitable recovery with data, and AISP’s informal list of COVID-19 data sharing examples from the Network.
The future of IDS & crisis response
As the climate crisis and a multitude of health emergencies proliferate, jurisdictions across the U.S. with established integrated data capacity and strong vertical and horizontal communication pathways will continue to benefit from more nimble crisis response. Capacity is not measured soley in technical sophistication of data infrastructure; rather, it requires strategic partnerships across agencies and collaboratively-created processes to securely share data and turn findings into action. Investment in these key elements of data-driven collaboration prior to an emergency is essential to an effective response that mitigates rather than exacerbates inequalities.
Actionable Intelligence for Social Policy (AISP) receives grant funding from the Bill & Melinda Gates Foundation to study the use of integrated data for economic mobility. This blog was produced in anticipation of an expanded series of briefs and case studies on this topic expected February 2021.