How are Health Authorities actually going to use Safe Paths?

To evaluate the Safe Paths + Safe Places solution in Testing, and ensure it is fit for purpose, it would be extremely useful to understand how Health Authoroties are actually going to use the solution.

I’ve not seen much documentation on this - I have some ideas of what might be sensible ways to use the solution, but no idea whether these match up with HAs' thinking, or the specific local conditions they are operating under.

DM: Kyle says that documentation of this will be provided as part of our development of training materials for HAs.

I also don’t know how much we aim to provide guidance to HAs vs. how much we intend to let them use the system however they want.

• I’d expect that initially we will be very led by HAs

• Once we have experience with several HAs in production, new HAs may well look to us for guidance on best practices.

• In either case, it would be useful for us to now be articulating how we expect the system to be used for - i.e. what range of uses we are designing for - to help inform testing, and to help identify any misaligned expectations that HAs may have.

Given current status, this is mostly a list of questions to be answered, rather than prescriptions.

1/ Who is the HA inviting to share their JSON data & begin a Redaction Interview that might eventually lead to published data?

• Just patients with a positive test? (covid-tracing)

  • KT-> At this point, it’s only being considered as a follow-up from a COVID19 positive test

• Patients with symptoms? (symptom-tracing)

• Patients who have received an exposure notification via the app, but may not yet be showing symptoms, and may not yet have a test? (2nd-tier-tracing)?

• Anyone else?

2/ What will be the HA’s procedures in moving from a Contact Tracing interview to a set of data to publish?

• What types of data will they redact for privacy reasons?

  • KT->We are providing all of the training on this currently. Personal residences.

• What types of data will they redact for efficacy reasons?

  • KT->Travel via private transportation, bad data points due to GPS wonkiness.

• What other modifications will they make to data? (e..g edit points that appear to be out of place; extend points to cover the entirety of an establishment visited, add points that the patient recalls, which weren’t logged by the phone etc.)

  • KT-> MVP1 is looking to include manual entry of GPS data points / time based on Patient memory

• Any other considerations?

3/ Under what circumstances might a patient be invited to share their data with an HA?

• Only patients entering the full process of Redaction and Publication?

  • KT->Right, after a COVID19 positive test

• Also in other circumstances? For example following an exposure notificatiion, this might be used as a means of better determining the risk level associated with an exposure notification.

  • KT-> Not currently planned

4/ What does the HA advise patients do following an exposure notification?

• Assume they are positive & quaranting themselves & their family for 2 weeks?

  • KT->They want to control this and make it highly relative to the current stage of outbreak.

  • KT->Because the feature where they control this is new, we haven’t had a chance to collect much of their feedback on this. Needs more engagement / planning as part of go-live.

• Engage in a discussion with the HA about their exact movements, and interactions with points of concern to better assess overall risk?

• Get a COVID test?

• Enter a 2nd tier contact tracing process?

• Some combination of the above? Or something else?

5/ How quickly can COVID tests typically be accessed, and how long do results take to come back?

• KT->I haven’t collected feedback on this one.

• Answers to this question are likely to vary a lot in different places. And the different answers are likely to feed into different positions on questions 1-3 above.

• See some discussion here: How fast does tracing need to be?

6/ Meta-question: How flexible are we in accommodating all or any of the above vs. providing specific recommendations for protocols to be followed? (factoring in local conditions that can’t be easily changed such as COVID prevalance, testing availability & speed etc.).

• KT->I would say very flexible within the constraints of ethical use of the tools. We are all inventing / learning how to do digital contact tracing.

7/ For Health Authorities who want guidance on best practices, are we able to provide recommendations on the above, given our experience with Health Authorities, our design intent of the app, and/or epidemiological modeling of how the Safe Paths solution complements Contact Tracing?

• In what form will this be? Written guidance? A consultation?

• KT->To some degree. In several cases, relaying how other HA’s plan to use the tool has been helpful guidance. We need to get serious about the M&E side of this with a real epi expertise in the mix.

8/ In assessing the efficacy of the solution, are we able to track which of the above approaches are taken by different Heath Authorities, so that we can evaluate which approaches tend to work best?

• KT->This is one of the MVP1 features. We need to measure what we’re doing here.

• To do so, we would probably want to define some sort of basic questionnaire structure for capturing the data from HAs in a structured way.

• This article might provide the right structure for such a questionnaire.