Can contact tracing afford to be wrong 30% of the time?

Looking forward to a post COVID-19 peak world, the Centers for Disease Control and Prevention (CDC) is focusing on containment. “A critical part of opening up America again is to make sure we have the public health capacity to stay in containment mode: early case identification, isolation, and contact tracing,” explains CDC director Dr. Robert Redfield in this interview with NPR.

CDC is providing states with $45 million to expand public health teams to lead containment efforts. Hiring these teams is vital, but they will also need the right tools. They will manage the most extensive contract tracing endeavor in our nation’s history.

As a first order of business, these contact tracers will need systems to collect and manage the data about who has been tested, where they’ve been, and who they’ve been in contact with. Public health agencies across the country are racing to implement systems to collect and connect this data.

Six million duplicate records?

I’m worried though, because as an industry and as country, we have been remiss in building a strong data management foundation for systems in healthcare and public health.

Like Kim Bellard said in his recent blog post, we’ve buried our heads in the sand.  

We know that at least 30% of demographic information about patients, gathered into electronic health record systems throughout the country, is inaccurate or missing data.

Infrastructure that solves this problem has a low adoption rate. This low adoption has long been justified by complexity. Infrastructure has been complicated to deploy, expensive to maintain, requires intense collaboration, often relies in-demand data scientists, and it’s much harder to visualize.

Instead, health systems have relied on the tireless efforts of health information management professionals to fix the tip of the iceberg—fixing duplicate patient record rates of 8-12 percent that are caused by an underlying, widespread mass of bad information.

While individual hospitals and health systems have been able to tolerate this situation, it simply doesn’t scale when we need accurate data to fight a nationwide pandemic.

Let’s assume, for simplicity’s sake, a low duplication rate of 2 percent and extrapolate it to the U.S. population of 328.2 million. That’s 6 million duplicate citizen records.

Duplicate patient records create confusion and prevent a positive lab test result from making its way to the person tested. Duplicates will also prevent contact tracers from identifying and reaching each infected person’s numerous contacts.  

We can fix this now

This is meant to be a shining moment for health IT. Even though the HHS rule  on interoperability is delayed, we can act now. Learn more in this short video.

photo credit: partners in health