Based on discussions with public health experts and policy makers, we are responding to feedback on the way we assess contact tracing. We have removed the critical (red) range for contact tracing and it will no longer be possible to be marked critical (red) due to low numbers of contact tracing staff alone. While low numbers of contact tracing staff should be acknowledged, it doesn’t indicate an active or imminent outbreak. In addition, contact tracing is one our least refined datasets, largely because states themselves are highly variable in their reporting of contact tracing data.
Furthermore, the best metric to measure the success of a contact tracing program is the percentage of new cases not traced to a known case. However, to our knowledge, the only state releasing this data is Oregon. Therefore, as a proxy, we currently look at the number of contact tracers on staff for each state to estimate the number of contacts that could be traced within 48 hours if the contact tracing program was otherwise working perfectly. The number of contact tracers hired is a very unreliable proxy, because we do not know how effective the contact tracing programs are, even if they have enough staff. The number of contact tracing staff reported can also be incorrect. There are other unknowns, such as whether the number of contact tracing staff reported is merely the number of planned hires, or the number of actual hires. We also do not know if contact tracers have been given sufficient training or whether they are following up in person with people who do not pick up their calls.
For all these reasons, we are updating the way we categorize contact tracing. First, we are removing the critical range. It will no longer be possible to be marked red for low numbers of contact tracing staff alone.
The 0-10% range will now be High Risk.
- The 10-90% range will remain Medium Risk.
- The 90-100% will remain Low Risk.
This change will impact one state: Missouri will become orange (High Risk) instead of red (Critical). With 94 contact tracers and an average of 819 new daily cases, we estimate Missouri needs 4,095 contact tracing staff to trace all new cases in 48 hours, assuming their contact tracing program is otherwise 100% effective. Even with a completely effective program, Missouri’s staffing means that they would likely be able to trace only 2% of new COVID infections. While this number is alarming, we believe that based on infection growth rate, positive test rate, and ICU headroom, Missouri is not in active or imminent outbreak. Florida and Alabama also would have changed to orange with the adjustment in our contact tracing indicator, except the addition of our Incidence metric keeps them in the red, due to their high incidence, 53.4/100K population and 39.1/100K population, respectively.
Second, we are also updating the language around our contact tracing metric, to make it more clear that it measures only the contact tracers hired rather than the overall success of the program. We are working to develop a more accurate metric to evaluate whether a contact tracing program is effective. More on that coming soon!
In the meantime, you can read about our new metric, daily new cases per 100K population, here.