Data needed to manage the monkeypox response
An update on using metrics to end the monkeypox outbreak
As I wrote last month, there is a major piece missing from our monkeypox response: metrics and targets. In order to end the outbreak, we should identify a small number of simple metrics with which to benchmark our progress. If the data are heading in the right direction, we can be confident that we are making progress toward containment. If the data are unfavorable, we will know where to focus efforts and whether any changes we make to the response are yielding the desired result.
Ideally, there would be a discussion among public health officials about which metrics and targets are useful and feasible. As a first volley, I recommend three based on those used during the Ebola response: number of tests (and test positivity); proportion of cases arising from known contacts; and time from rash onset to isolation. I think reasonable targets for these metrics could be 20%, 80% and 2 days, respectively.
In the last few weeks, some information on each of those has become available.
Testing: Nearly 8,000 monkeypox tests were run the week of Jul 29, a four-fold increase from a month prior. In that time, test positivity has held in the 40-60% range, which is well above the 20% that I propose as a target. There are several possible reasons why test positivity is elevated. The country is only using 10% of available capacity, so the bottleneck does not seem to be on the lab side.
Rather, patients may not recognize their symptoms or have access to care, and providers may not be skilled at spotting monkeypox cases or accessing testing. An alternate hypothesis is that providers are very good at filtering based on clinical presentation and that only specimens that are likely positives get submitted for testing. This seems unlikely to me because monkeypox is a new illness both for patients and providers in the U.S., so I would not expect widespread familiarity, but it is a possibility worth investigating.
CDC has been conducting extensive outreach to at-risk groups and providers, so I hope test positivity is lower by the next update.
Contact tracing: Ideally, all new cases would arise from known contacts. This would imply that new cases are already in quarantine when they become infectious and therefore did not pass the virus on. It also implies that epidemiologists have identified all chains of transmission and so containment is close at hand.
On this metric, there is very little information available. The UK Health Security Agency put it at 8% in their July 18 technical report. Higher is better, so 8% is not reassuring. Early in the outbreak, UK epidemiologists reported that contact tracing is difficult for this outbreak because some cases are unwilling or unable to identify their contacts, given the sensitive and sometimes anonymous nature of those interactions. UKHSA did not publish an update in their August 1 report and I have not seen any information from the U.S.
Time to isolation: Time from rash onset to isolation is detailed in the CDC technical report. (The technical report details time to test result, not isolation. Here I am assuming they are the same.) As I reviewed in a recent summary, the interval has been steady at 6-7 days for several weeks now. This is too long and reinforces that timely case ascertainment and diagnosis is an issue. In addition to furthering patient and provider education, test turnaround time could be an area to explore for improvement.
It would be useful to see more data available on each of these metrics, particularly broken out by jurisdiction and demographic group. This information would highlight where our response is succeeding and where we still need to focus our efforts.