Fixing our public health data infrastructure and governance

The COVID-19 pandemic exposed longstanding problems in public health. IMO’s Chief Strategy Officer, Dale Sanders, suggests ways to fix an ailing system.
Share on facebook
Share on twitter
Share on linkedin
Share on email
Improving public health with higher data quality in healthcare

Let’s start the discussion here: The median ROI for public health interventions is 4:1, and median cost benefit ratio is 10:1.[1] The best way to reduce the never-ending upward curve of US healthcare costs is to invest upstream in public health.

If you’ve ever wished for a great overview of our nation’s public health system, I highly recommend reading The Patchwork US Public Health System, recently published in the New England Journal of Medicine (NEJM). In the article, Megan Wallace, DrPH, and Joshua M. Sharfstein, MD, outline key components of the system’s architecture, including its 21 federal agencies; 50 state agencies; and 2,459 local agencies.

“[T]here is no clear administrative structure that organizes the many federal agencies involved in public health,” explain the authors. “Twenty-one major federal agencies have a role in pandemic preparedness and response, for example, and more than 100 federal offices have been engaged in work during the COVID-19 pandemic. At the state and local levels, variation is the rule, not the exception.” In addition, the authors note that, “although information technology and data capacity are key to public health capacity, much of state and local public health work remains based on paper, with large gaps in the ability of health departments to obtain, analyze, and share information expeditiously.”[2]

COVID-19 put a spotlight on public health problems that have existed for years, yet public health accounts for only 3% of healthcare spending in the US. Clearly, our public health system needs a decade-long improvement plan, including funding. In addition to the advice in this article, I offer a few suggestions and advice of my own, below:

  1. Public health data operating system: The Centers for Disease Control and Prevention (CDC) needs funding to host a multi-tenant, single source of data truth — call it something like the “Public Health Data Operating System” — that the State, Tribal, Local and Territorial (STLT) agencies can leverage while maintaining a degree of local data privacy, security, autonomy, and agility. It can be done. Healthcare.gov for the state insurance exchanges is a great example and role model.
  2. Governance and economic sustainability: The Cloud makes these data operating systems a near utility from a technical perspective — and that’s coming from a person whose early career was based on having the technical expertise to build these platforms. That skill base is largely commoditized now, and I couldn’t be happier. Today, the harder data issues are found in sustainable economics and governance — data standards, taxonomies, and formats; and legislative, policy, and cultural barriers surrounding the collection and utilization of the data. But all those issues can be solved, too.
  3. CDC skills modernization: The CDC will need to staff the right kind of leadership, skills, and organizational structure to build and maintain this sort of platform. Historically, that has not been a strong point for the agency, as evidenced by the current state of our public health data infrastructure — but it can be. The skills exist in the market.
  4. Data acquisition and quality: Within its data governance structure, the CDC needs to lead the development of a formal data acquisition, standards, and quality strategy, based on Data ROI = (Value to the mission) ÷ (Cost to collect and maintain high quality data). As a guiding principle, the CDC must minimize its dependence on manual, human data entry and acquisition, and build a strategy which maximizes passive, automated, and sensor-based data acquisition. Examples of high-value, high-ROI data which need attention include: lab orders and results, diagnoses, social determinants of health (SDOH), mortality/cause of death, overdoses. Note that I omitted contact tracing, which would be an engaging topic for debate. In my opinion in today’s world, it’s too costly to collect high-quality contact tracing data, and if you can’t collect high-quality data, you’ll end up with a misdirected and false sense of situational awareness and hypothesis generation. Bad data that you believe is good is worse than no data.
  5. Healthcare delivery = public health: Public health and population health — that is, infectious disease and chronic disease management — must come together. Every major healthcare delivery system in the community must have a formal public health liaison on its leadership team, and they must jointly develop and sign strategic plans to include overlaps in the joint data strategies. For example, the number of CMS and other payer quality measures for providers must be reduced to make room for the data needs of public health. Said otherwise, public health priorities must be reflected in the data we ask clinicians to document in EHRs, and we can’t burden them with more data entry.

According to the NEJM article, in the American Rescue Plan $47B is earmarked for COVID-19 mitigation, $7.7B for the public health workforce, and $500 million for the CDC to update the public health information technology infrastructure. Let’s make sure we spend our tax dollars judiciously, for the long term.

For more on the challenges of managing COVID-19 data, click here to listen to Dale’s recent HIMSS podcast.

[1] Masters R, Anwar E, Collins B, et al. Return on investment of public health interventions: a systematic review. J Epidemiol Community Health 2017;71:827-834.

[2] Wallace M and Sharfstein J. The Patchwork US Public Health System. New England Journal of Medicine. 6 January 2022. Accessed 7 Jan 2022 via nejm.org.

Share on facebook
Share on twitter
Share on linkedin
Share on email

Ideas are meant for sharing.

Sign up today and have Ideas delivered straight to your inbox.

Related Ideas