Pandemic Demonstrates Need for Improved SDOH Data Management

By Rahul Sharma 

Roughly $4 trillion is spent on healthcare annually in the U.S. Despite this massive expenditure, only 20% of health outcomes can be attributed to medical care. The vast majority of health outcomes are rooted in social determinants of health (SDOH), according to County Health Rankings. 

The deadly pandemic has made it painfully clear that unaddressed social determinants and health inequities can result in poor health outcomes. 

“Long-standing systemic health and social inequities have put many people from racial and ethnic minority groups at increased risk of getting sick and dying from COVID-19,” the Centers for Disease Control and Prevention (CDC) wrote in February 2021.

It’s equally clear that if payers and providers could leverage a comprehensive medical and social longitudinal health record (LHR) that includes SDOH data and enables them to connect patients with community support resources, they could reduce healthcare costs while improving lives. Such an approach would make SDOH data securely available to clinicians at the point of care, giving them a holistic view of the patient.

Sadly, healthcare stakeholders have yet to efficiently capture and integrate SDOH data, in large part due to persistent interoperability problems and lack of effort. Only 24% of hospitals and 16% of physician practices in the U.S. screened for SDOH factors such as housing instability, food insecurity, utility needs, transportation needs, and interpersonal violence, a 2019 Dartmouth study published on JAMA Network showed. Those screening rates could be much higher if the healthcare industry were able to integrate and share SDOH data across the healthcare/wellness continuum.

Fortunately, technology exists that allows data from different sources and in different forms (structured and unstructured) to be collected and synthesized, including SDOH. This technology can integrate external data sets (such as the 12 Dimensions of the Social Environment) containing SDOH information to augment standard screening tools, providing potentially valuable insights at the point of care. Clinicians can incorporate this data into LHRs for individual patients. 

SDOH data can be difficult to access because much of it is unstructured. Leveraging Artificial Intelligence (AI) algorithms to digitize unstructured SDOH data can provide critical data for patient LHRs. And by applying machine learning (ML) to that data, providers and payers can create models for patient medical and social risk scores and predict costs and patient behavior. 

One of the biggest challenges health officials and providers faced in early 2020 as they sought to contain the spread of a communicable disease was lack of high-quality, verified data about patients, the community, and the COVID virus itself. But technology exists that would allow payers, providers, and other relevant healthcare stakeholders to synthesize data in different forms and from multiple sources, and to identify and correct errors. Real-time access and analysis of SDOH helps inform mitigation strategies and management of resources while empowering payers and providers with actionable information on the social factors most impacting a patient’s health.

Obstacles to SDOH 

Before the healthcare industry and society at large can fully benefit from a deep understanding of SDOH, barriers to collecting and sharing SDOH data must be eliminated. 

One such barrier to SDOH in healthcare is siloed data.  Many hospitals, payers, health organizations, government agencies, and community service providers have highly valuable social and health data trapped in their databases, with no way to easily share it.  A lot of this data is also stored in proprietary non-standard formats.  Lack of standardization plus lack of adoption of existing standards are issues that need a resolution.

The data’s form is another barrier. The vast majority (80%) of healthcare data is unstructured and can include clinician notes, patient-reported text describing a physical condition, images, Audio/Video recordings, surveys, IoMT (Internet of Medical Things) data or transcripts from telehealth visits. Unlocking value from this unstructured data will require healthcare organizations to apply technology like Natural Language Processing (NLP) to digitize the data.  In addition, for addressing the needs of non-English speaking populations, technology solutions need to have multilingual support and translation service capabilities. 

Distributed ledger technology (DLT) can be used by payers and providers to enable authorized disclosure in a private permissioned manner in near real-time. DLT ensures the reliability of the data and its source(s) and provides all parties with full transparency and eliminates the massive needs of re-conciliation of data. 

It is essential that stakeholders sharing patient data make a priority of securing protected health information (PHI) as well as personal information. Data (both in transit and at rest) should be encrypted, while biometrics data should be considered for patient identity verification and matching.

A third barrier to SDOH data sharing in healthcare involves regulatory exposure. Since some SDOH stakeholders (such as schools, municipal offices, and law enforcement agencies) aren’t bound by HIPAA rules, it becomes necessary to obtain consent from patients to share data. This would require a means to secure consent and an immutable audit trail of disclosures. Further, the technology used should support multi-channel communication – including audio, video, secure text, and mobile apps – for obtaining and managing consent. 

In addition to implementing technology that allows different systems to share SDOH data, healthcare stakeholders need ways to securely exchange information. One option is a next-generation communications engine with end-to-end secure capabilities for B2B and B2C that powers email, SMS, and MMS with extensions supporting ubiquitous instant messaging software.

To fully support SDOH networks, the technologies used by stakeholders must offer a simple user interface and easy-to-navigate workflow from which data can be accessed and shared with authorized parties. Key patient data categories and information sources – including administrative claims data, clinical data, demographics, disease registries, employment, immunity tests, prescriptions, sensor data, SDOH data, and others – are crucial to the creation of LHRs. 

Finally, while COVID-19 laid bare the inequities in healthcare caused by social factors, it also highlighted the importance of the physical supply chain, particularly during a crisis. Imagine a healthcare system able to leverage near real-time integration of supply chain data from different medical/healthcare vendors, as well as carriers like FedEx and UPS, for resource and emergency response planning.

Conclusion 

Patients, providers, communities, and payers all have a stake in improving individual and population health while reducing healthcare costs. To accomplish these goals, healthcare stakeholders must capture and leverage SDOH data, which research shows accounts for four out of five patient outcome and population health trends. By accessing relevant SDOH data from verified sources, healthcare professionals can identify and mitigate health risks that overwhelm our system. This can be accomplished through deployment of technologies enabling multiple stakeholders to build and share LHRs in a private permissioned manner while optimizing holistic care coordination and data-driven population health initiatives. 

Rahul Sharma is the CEO of HSBlox, which enables SDOH risk-stratification, care coordination and permissioned data sharing through its digital health platform.