Review 1: "Contact Tracing Efficiency, Transmission Heterogeneity, and Accelerating COVID-19 Epidemics"

RR:C19 Evidence Scale rating by reviewer:

• Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.

***************************************

Review:

This is a helpful modelling study taking a careful look at the impact of contact tracing on Covid-19 transmission dynamics, specifically the reproductive number. The authors have made use of an SEIR model that allows the researchers to explore various dimensions of contact tracing (e.g. timing, caseload per tracer, percent of infectious contacts traced). The model also incorporates social distancing measures. Overall this is a useful paper that appears to be well-done. The code is provided so others can modify to their particular scenarios. The model appears to be appropriately set up. A few notable limitations exist:

1. There are assumed exponential distributions in many places due to the nature of the model. This is not consistent with what is known about Covid-19 epidemiology. More sensitivity analyses would improve the robustness of the results.

2. The paper really highlights the need to have wide testing availability to make sure cases are identified and traced early. This could be emphasized more. In fact universities with strong testing and tracing protocols are demonstrating this.

3. More demonstration of the stochastic results would be helpful. There is clearly substantial stochastic variability in infectious disease dynamics and running the model more stochastically would be helpful.

4. A bit more detail on how calibration was done would be helpful. It is not clear exactly how that was done and f it is rigorous enough that the actual model outputs can be trusted.

5. Related to the previous comment, it is hard to know how much to trust the actual values produced from this exercise as opposed to relative values. Given the wide variability of communities and stochasticity of covid, relative values might be more meaningful.

6. It should be noted that this modelling strategy assumes homogenous mixing, which is most likely violated-it is not clear what impact this would have. It would be nice to see this analysis replicated in a network type model that can explore these heterogeneities further.