Building Video Consultations - Part 2, Delivery
Topics: Building a solution
At the end of week three we already had Nottingham University Hospital testing our video consultations, and the following week they began trialling it with real patients.
Reflecting at the end of that week, we were quite chuffed at how quickly we’d delivered. We also held a retrospective - a team meeting to reflect on what had gone well, and what could have been better, and talked about some of the decisions that had helped us.
First off, we built our Video solution as a standalone microservice and a new web app. There was early talk of having the video room embedded into our existing patient & staff apps, but that would have slowed us down to no great benefit. Architecturally at DrDoctor we are trying to evolve away from our existing monolith (the “classic method” of service development). If you don’t build your new solutions as stand-alone services, then frankly you’re never going to make the change! Still, it was a brave decision when trying to move fast.
So we stood up a new service and database. It actually took our developers Ferenc, Liam and Cristi less than a week. And it had immediate payback. We were now able to put two development teams on this new product. While the first team focused on the Video service & app experience, the other team led by Andy and Anup focused on integrating it with our appointment management workflows in the monolith. Now there were no merge or release conflicts; both teams went as fast as they could.
We also switched from our usual two week sprints to one week sprints. This emphasised the focus on “what will we be showing at the end of this week?” and helped us make the toughest possible decisions about reducing scope and accepting more technical debt than usual. We left our standard scrum-style rituals behind and just picked up stories that were ready to start, working them out by doing more collaboration ad-hoc.
We already had tacit agreement that we’d keep iterating on our video solution until it had proper market fit, and that gave us more confidence in cutting corners knowing we’d be able to come back and fix them up.
After six weeks our first Trust was using video consultations on a daily basis, and two more were starting testing and piloting.
As we took a breath and deliberately agreed to slow the hectic pace, we could see some of the quid-pro-quo from the way we’d been working. Throwing out our normal process has made us faster, but now our backlog was in a state, we’d started developing some bad habits, and there were more bugs and loose ends to tidy up than we’d usually like. Conclusion? Moving fast works, but it’s not sustainable over more than about six weeks.
Get in touch here to have a chat and see how DrDoctor Video Consultations can help your patients.