ClinicArrivals - helping Australian GPs with Covid-19

Apr 1, 2020

A couple of weeks ago, Nathan Pinskier reached out to me on behalf of the RACGP expressing concern about the impact of COvid-19 on Australian GPs.

Talking to Nathan, it was clear that there’s 2 acute challenges facing GPs:

  • sudden real concern about infection control and therefore keeping patients out of the waiting room
  • Rapid changes in telehealth arrangements.

As a gross generalisation, over a 2 week period, GPs have moved from being not allowed to use telehealth at all to being required to do it for a rapidly growing percentage of their population. You just can’t make changes to your system and workflow that quickly. And though there are good tele-health solutions out there, they are not integrated into the existing systems that GPs are using. As a consequence, GPs are trying to use WhatsApp or Skype, but these are a workflow disaster in this context (see, for instance, how to get your skype id).

What is needed is something that is zero impact for patients, ubiquitously available, and deeply integrated into a GPs workflow, since that’s increasing at this time. Further, it has to be available NOW, without requiring system change for every GP in the country.

The only way I could think that we could possibly solve that was if I partnered with Brett Esler from Oridashi, and we used his FHIR access library to access the GPs Practice Management System (PMS) Appointment diary, and then message using SMS the patients to find out whether a telehealth consultation was possible/appropriate, and then, if it is, set up a video session, and if it’s not, let them SMS that they are waiting from the car park, and then released that as an open source application.

I’m happy to say that as of today the ClinicArrivals application is now available for testing in limited production settings.

I didn’t do this by myself. In fact, I only did a fraction of the work. Other people contributed - in fact, dropped everything to contribute, all in their own time:

  • Brett Esler gave us free use of his FHIR GP Access library to access the appointment diary for Best Practice, Medical Director, and ZedMed and then worked to improve it for us.
  • Brian Postlethwaite wrote the actual guts of the application over a 96 hour period
  • Shovan Roy worked with me to figure out how to do the video-conferencing
  • James Berry set up the video conferencing server on AWS
  • Vadim Peretokin organised the build/release framework
  • Mel Grieve helped with the documentation for the video usage

All of those people contributed to the design, and it wouldn’t have happened with out them. Also, I need to give credit to Josh Mandel for sage early advice, Best Practice for waiving the commercial fees to be a partner so this is above board (and doing so in a timely fashion at a time of immense pressure), and most of all Nathan Pinskier for advice, encouragement, and volunteering to be the initial prototype site.

For video conferencing, we looked at a number of different approaches. Our criteria was that it be zero-install and near zero impact for the patient, not involve enrollment of the patient in anything, and not cause any delay for the GP, and have a RESTful API that we could use to orchestrate the service.

In the end, the only choices we found were two open source video conferencing solutions:

  • Jitsi - a very excellent video conferencing service that my family is now using while we’re staying at home
  • OpenVidu- another very flexible video-conferencing solution.

We chose Openvidu because it is near zero install for patients (doesn’t work on IE or Edge) and does have a simple to use RESTful API. Unfortunately, it requires a server which is not simple to install (though James is working on that). For now, then, I’m running that server as a free service while we figure out what to do about that.

There’s still plenty to be done, but the application seems to have reached MVP and so 10 days after it was conceived, we are trialing it with real patients today.

My deep thanks to everyone who helped with this process. Hopefully it will make a small difference in our grand fight against this virus.

P.S. where does this go long term? I can’t see it being a product. Actually, I’m hoping that since it’s open source, the PMS vendors will integrate the basic approach and methods into their own products, and this will naturally be superceded.

P.P.S: What this does underscore is why APIs are so important - what we wrote is just a mash-up between a bunch of RESTful APIs for the PMS, SMS, and videoconferencing. That’s why APIs matter: they create a resilient responsive eco-system that’s able to respond like this. Long term, the focus of governments should be clear: force your system to adopt APIs (that, of course, is not today’s problem)

Translations: