Can we flatten the curve in North Carolina?

Our paper on North Carolina is now available here. My last few posts essentially give a summary of the paper but all the details including the mathematical model are in the paper.

We need more contact tracers in North Carolina... Way more...

And of course we also need a public that supports (or at least responsive to) contact tracing efforts. As we discussed in the previous posts for NC, with a large testing capacity and effective contact tracing, there might be a way to keep the state open without letting the cases explode. One thing we have not addressed is the state's contact tracing capacity. One of the implicit assumptions of our model is that regardless of how effective contract tracing is there are at least enough resources to try to track people who test positive and identify their close contacts. There are mainly two parts to contact tracing. One is interviewing individuals who test positive and identifying their close contacts, and the other is making follow-up calls to the close contacts in an effort to make sure that they are socially isolating, get tested, and are doing generally fine given their high-risk status. The following figure shows our projections for how many new COVID-positive cases will be identified each day until the end of the year assuming that the state will remain at Phase 2 level social distancing levels. (Note that with proper face covering measures and other precautions that state might be able to keep current Phase 2 level social distancing while continuing with its reopening plan.) These projections give us a sense of how many individuals contact tracers will need to interview each day and roughly how many close contacts they will identify. There are three curves each corresponding to a different level of contact tracing effectiveness.

 As we can see from the figure, number of positive cases will peak some time in Fall and there will be more cases and thus more work for contact tracers if contact tracing is less effective.The following figure shows how many people will be in self-isolation due to contact tracing (and will need follow-up calls) each day until the end of the year.

As we can see from the figure depending on how effective contact tracing is our projections for how many individuals will need to be kept in self-isolation will differ significantly. In any case, they will peak some time in Fall. But, here one obvious question is the following: How many contact tracers does the state need? According to our estimates, even under the best-case scenario of CTEP  = 10 considered above North Carolina would need more than 2,800 contact tracers, which is significantly higher than the 1,500 the state reportedly has. In conclusion, the only way to keep the economy open is through increased testing and effective contact tracing capacity. One without the other would be pointless. The state needs to urgently work on increasing its capacity and capability for both. 

How efficient is contact tracing?

One of the challenges in predicting future evolution of the COVID-19 pandemic is figuring out what the impact of contact tracing will be. Contact tracing at such a large scale will be new for most states and to our knowledge there are no studies that can inform us about the effectiveness of the practice in the US. In our analysis, we made some assumptions on contact tracing effectiveness based on data from China. (What we mean by contact tracing effectiveness is the degree with which infected individuals become more likely to be detected through contact tracing.) Admittedly, however, there is significant uncertainty around this assumption. Therefore, we investigated how our results would change with changes in contact tracing effectiveness. The following figure illustrates our main observation. In the figure, CTEP stands for Contact Tracing Effectiveness Parameter. In the figures we posted so far for North Carolina, CTEP was set to 5. The figure shows how things would look like if contact tracing is less effective (CTEP = 2) and if contact tracing is more effective (CTEP =10) under the assumption the state stays in Phase 2 until the end of the year.

As we can see from the figure, the effectiveness of contact tracing will have a significant impact on how bad things will get and so it is possible that (if for instance contact tracing is so effective that CTEP would be even larger than 10) future evolution of the pandemic may not be as disastrous as we predicted. However, this is really a big if not only because, as suggested by some news articles, contact tracing so far is not working as well as hoped but also contact tracing capacity is far below the levels it needs to be to meet the anticipated workload. We clearly see this from our analysis. I will explain this in more detail in my next post.