Home  >  Interoperability   >   SDoH, Data Standards, and Innovation for Unstructured SDoH Data

SDoH, Data Standards, and Innovation for Unstructured SDoH Data

by Jody Ranck | August 30, 2021

Key Takeaways

The ONC is advancing the creation of standards for social determinants of health (SDoH) data as a priority in their efforts to improve outcomes and population health. USCDI (United States Core Date for Interoperability) version 2 guidance was recently released and is including SDoH, gender identity and sexual orientation in the forthcoming standards.

Unstructured data contains a great deal of SDoH data that needs to be normalized and standardized to improve risk stratification and patient profiles. With nearly 80% of medical records as unstructured data, the use of natural language processing (NLP) for SDoH data extraction and creation of longitudinal records is of growing importance.

Standards, coupled with public goods approaches to financing, offer a way to reduce the costs of SDoH interventions and improve ROI. Unlocking SDoH data for driving social service partnerships is one pillar of a broader public goods approach to investing in SDoH interventions.

Introduction

Making productive use of SDoH data is a difficult challenge for many organizations on a number of levels. From the lack of standards to the cost of some datasets to the messiness of unstructured data in EHRs, organizations find a number of challenges in accessing the data to create an accurate social risk profile of patients and data on social services.

For the past two years, we’ve been following the SDoH space as vendors offer more mature solutions and investment activity grows. For example, the stunning $150m investment in Unite Us in March 2021 indicates the level of investor interest. One study covering 2017-19 found that across 917 hospitals, systems were spending over $2.5B annually, over half of which went to addressing housing needs.

A fundamental component of making SDoH strategies more effective rests on the development and use of standards for SDoH data that can facilitate data computability and sharing. Improvements on these fronts can lead to better interventions and greater ROI from SDoH investments. Recent efforts by the ONC on the standards front, as well as some innovative approaches utilizing NLP to unlock value from unstructured data in EHRs relevant to SDoH interventions, are on the horizon.

ONC SDoH Standards Activities

The ONC has made the development of standards for SDoH data a core pillar of its strategy for mobilizing technology to improve outcomes and population health. The most important focal point of the current strategy in this regard is the cooperative agreement with HL7 to develop standards for SDoH data. Other components include:

  • Development of a toolkit for states’ payers and community-based services for implementing SDoH-based data exchange that will include components on the SDoH standards
  • Exploration of electronic data tagging capabilities and clinical guidelines with SDoH for electronic decision support

Recently the ONC released USCDI Version 2, a standardized set of data classes and constituent data elements for interoperable health exchange.  What is notable about version 2 of USCDI is the addition of SDoH elements for assessments, goals, interventions, and problems, as well as gender identity and sexual orientation data elements.

Figure 1: ONC Focal Areas

The process of adding the new SDoH standards include submissions from collaborators such as HL7’s Gravity Project, which has been developing a FHIR-based implementation guide (IG) for SDOH. HL7 also runs connectathons so developers can test and validate Gravity’s IGs.

The Gravity Project stems from the Siren program at UCSF. Siren has attempted to frame SDoH infrastructure as a public good in order to overcome the free rider problem where organizations refrain from investing if the gains go to a third party. The public effort to foster the development of standards can be seen as one of the pillars of creating this third space that can facilitate HCO’s investments in SDoH, with the risk shared across a diverse ecosystem of public and private sector actors.

Unstructured Data and SDoH

Some additional challenges around SDoH data beyond the standards includes the substantial amount of data in EHRs that contains SDoH-relevant information, but is unusable because it is unstructured. Furthermore, many stakeholders in SDoH interventions are outside of the health sector and beyond HIPAA’s regulatory reach, which does create privacy risks for HCOs partnering with community-based organizations. Once standards for SDoH are created, there is still the challenge of making the unstructured SDoH data accessible, while also maintaining patient privacy.

New Approaches Can Help

One approach that we have encountered in recent months involves the use of blockchain and AI to create a longitudinal record of patients that includes unstructured data and a wide range of SDoH data. HSBlox is an SDoH-focused platform that integrates technologies from the digital supply chain, fintech, and health IT to create longitudinal records from disparate data sources, with the added component of patient consent controls. This method uses blockchain, an important capability for proper governance of SDoH data and programs.

A core capability of the HSBlox platform, CureAlign, is the ability to standardize audio and video files, including those in Spanish, in a manner that creates a secure, traceable data trail across social care and medical care boundaries. Unlocking the value of unstructured data in the EHR can create more reliable risk stratification and surface social factors attributing to outcomes that are valuable for referral networks with community-based organizations.

The HSBlox standard is applied when data enters their interface. A longitudinal record of a patient can be generated from the digitized, unstructured data to create a picture of the patient that includes approximately 12 data elements associated with SDoH.

Figure 2: HSBlox CureAlign platform example

The blockchain component of the platform enables proof of provenance of the data captured and is part of the consent management protocol. The consent management issue is an important one for sharing of SDoH data; just because there are standards does not mean patients will always share behavioral and SDoH data with providers. Building trust through control over data has the potential to improve data sharing as well, the HSBlox team asserts. Also included is a fintech component that enables invoicing for social services.

Conclusion

The social determinants of health market has been attracting a great deal of investor attention over the past year. However, the sector is still immature, and a number of barriers make leveraging SDoH in interventions that are cost-effective and responsive to patient needs challenging. One issue is the access to data that provides a 360-degree view of the patient and their social needs. The ONC’s leadership role in developing USCDI v2 that includes SDoH data elements is an important step to addressing this issue. However, standards alone are insufficient to bridge the gaps in the current system.

A growing amount of research on SDoH programs is pointing to the need for more use of public goods strategies. This would be in order to fund more sustainable interventions, given the timelines to find an ROI on social interventions, such as early childhood education. Standards and technologies that can improve access, and the quality of data that informs partnerships, can also be viewed as elements of enabling more sustainable SDoH programs. The dominant approach to SDoH platforms at the moment is focused on creating repositories of social service providers that can be matched to social needs identified at the point-of-care. Standards, unlocking unstructured data, and innovative financing mechanisms could enable far more robust analytics and AI/ML approaches to SDoH, with more granularity and specificity to geographical regions.

One response to “SDoH, Data Standards, and Innovation for Unstructured SDoH Data”

  1. Mary-Sara Jones says:

    The real value of SDOH will be achieved when we stop depending on healthcare as the entry point. High need individuals/families will typically appear to Human Services/Social Services programs before healthcare identifies SDOH needs. By more holistically addressing needs when individuals/families first present, we have the opportunity to avoid negative health outcomes instead of simply managing how far they progress.

Leave a Reply

Your email address will not be published. Required fields are marked *

Stay up to the minute.

[bestwebsoft_contact_form]