Leveraging SDOH Data to Improve Healthcare Outcomes

by | Aug 18, 2020

A Focus on Social Determinants of Health (SDOH) – Improving Outcomes and Gaining Insights
For information on our SDOH tools or to gain access, please contact us.

Social Determinants of Health (SDOH) are the conditions in which people are born, grow, live, work, and age. Given the disparate and complex range of variables in conditions across geographic populations, deriving actionable results from SDOH data is a challenge. That is where technology can step in to lend a hand. With guidance from health experts shaping the approach, SDOH data can be harnessed to gain insights that can inform care delivery, policy, and quality initiatives, thereby improving outcomes.

CVP is actively developing solutions that use SDOH data to navigate changes in healthcare. With our direct experience in Agency-level statistical tools and datasets, including Census, Centers for Medicare and Medicaid Services (CMS), and Agency for Healthcare Research and Quality (AHRQ), we are driven to create unique strategies, platforms, and tools linking datasets and enabling users to leverage SDOH and help change the future of healthcare.

CVP’s SDOH Data, Analytics, and Visualization Platform to Inform Healthcare Quality: Measuring, Discussing & Addressing Disparities Within Social Determinants of Health at the Local Level

CVP’s initial foray into SDOH data-driven solutions was for the AHRQ’s Visualization Resources of Community-Level Social Determinants of Health (SDOH) Challenge. In that challenge, AHRQ sought tools that support visualizing social determinants of health and community service digital data sets to enhance research and analysis of community-level health services. The end goal was to have solutions that allow the agency to better predict and plan for optimal use of limited resources and ultimately improve health outcomes.

When developing our submission, CVP looked to an underused resource, hospital Community Health Needs Assessments (CHNA), for guidance in selecting SDOH factors that are most relevant at the community level. For this platform, CVP created a data lake comprising publicly available data sets from the Environmental Protection Agency, Census Bureau, Departments of Transportation, Labor, Education, and Health and Human Services (i.e., CMS, CDC), and privately generated data (e.g., Dartmouth Atlas and Kaiser Family foundation). Topics included income and poverty, education, housing quality, housing cost, environment, employment, transportation, and food access. Going beyond the requirements of the AHRQ Challenge, we included health outcomes, access to primary care, behavioral healthcare, dentists, and hospitals, and care quality, such as diabetes care and hospital quality, to enable us to explain relationships among social determinants, care quality, and care outcomes.

CVP also created a repository of CHNAs from hospitals across the country and applied natural language processing (NLP) and other advanced technologies to enable qualitative analysis within and across CHNAs to identify trends and policy priorities. CVP’s tool addresses the following questions:

  1. Which social determinants of health and health outcomes face the greatest disparities in counties?
  2. How can CHNAs highlight the SDOH factors that most need attention to improve health in specific communities?

Based on AHRQ’s desire for a visualization product that is easy to maintain, inexpensive, and open-source, CVP built its solution on a suite of open source technology to promote transparency and accessibility. It is built with Metabase for Visualization, PostgreSQL for databases and querying, Python for data loading, text analytics, and NLP, AWS for cloud hosting, using servers from their free-tier and GitHub for code and setup.

The resulting platform displays meaningful relationships among SDOH, health outcomes, and health quality indicators. For example, we observed that communities with higher rates of persistent poverty had fewer hospitals with high readmission rates (Pearson’s Chi-squared test: X-squared = 31.62, df = 2, p-value = 1.361e-07). Communities with higher rates of low education had fewer hospitals with high readmission rates (Pearson’s Chi-squared test: X-squared = 16.781, df = 2, p-value = 0.000227).

CVP’s Phase 1 submission was recognized by AHRQ as a Visualization Resources of Community-Level Social Determinants of Health (SDOH) Challenge semifinalist.

Because the platform can be used to understand how social determinants interact with care access and quality, CVP submitted an abstract to the National Quality Forum (NQF) for consideration as a Next-Generation Innovator for the 2020 Annual Conference.  

CVP’s NQF submission was recognized as a 2020 Next-Generation Innovator Abstract winner and can be used by measure developers, policymakers, and researchers to explore SDOH, healthcare access, and quality. Little did we know how helpful our platform could be in the coming weeks.

CVP: COVID-19 SDOH Visualization Resources

When the COVID-19 crisis arose, CVP immediately recognized the importance of SDOH data analysis and our unique ability to leverage it in our SDOH tools. We developed a COVID-19 Visualization Tool focusing on SDOH COVID-19 risk factors for serious diseases, including diabetes prevalence, smoking rates, age, and socioeconomic status, in areas without Federally Qualified Health Centers (FQHC) or hospitals. There are four visualizations in total:

  • Underserved counties by hypertension risk
  • Underserved counties by prevalence of diabetes
  • Underserved counties with no ICUs and population over 65
  • Physician staffing by county

The tool is available for everyone to access on our website here.

CVP continues to develop solutions leveraging SDOH to empower our healthcare customers and end-users to engage with and make smart decisions to improve outcomes for all. If you would like to learn more about how CVP can help you leverage SDOH data to improve care, please contact us.

Pin It on Pinterest

Share This