My time at Kry (Europe’s largest digital healthcare provider) is approaching the end and I’ve decided to leave on a high and do a quick summary report to see what impact our team had. I have spoken about some wins before but here is a deeper dive. We were a small gang who were often responsible for video consultations, chat consultations, voice calls and messaging patients & this is what we achieved.

Care Portal 1.0 & the early days

When I arrived the company was still letting clinicians use a product called Care Portal 1.0. Some still use it to this day but over the last 14 months or so we’ve been moving people over to a new product called Kry Pro.

Care Portal 1.0 had some downsides in that it was developed mainly to do Swedish video meetings but the company had grown. It had expanded to new markets, more features, more types of consultations and the old platform did not deal with these very well. It became hard to build on top of & scale any further. It was envisoned the new platform would give a much better user experience, give us more flexibility for new services/features but also be built in a way that we could be more nimble for changing demands in all of our markets & reduce development time.

Volume

Given some of this might be commercially sensitive I’ll give % changes rather than raw numbers (sorry reader!).

From Feb 2022 until April 2023 we saw quite the jump in the number of completed consultations per month. The company aims to transform how “people access high-quality healthcare” so naturally aims to increase the amount of patients that get to speak with clinician, which can be a frustrating experience when seeking primary care via traditional routes. Between those two dates we saw a 26.85% increase in net total meetings. A sign of healthy demand and we have a good chunk of returning users to our services. Our record month was actually March which would alter the stats to a 35.12% increase from when I joined. Basically the short answer is we see a positive trend from when we started to now. Which is good!

What type of consultations?

As I said earlier we only had video and physical consultations when I joined but since then we’ve released some new ones! These were entirely new services, created with other colleagues across teams that required a chunk of service design and a fair amount of iteration along the way.

First we helped introduce:

• nurse chat consultations
• doctor chat consultations
• voice consultations

Volume wise the vast proportion of meetings are still video. Roughly 9/10, some of these types of consultations are only available in half of the markets we are in too so that deflates potential volumes for some of the newer services. However, a nice little chunk are meetings that did not exist when we first landed. Not every patient wishes to be seen via video - whether it be a patient with lower digital literacy so perfers a phone call to a patient that does not want to be “seen” and finds a chat more comfortable.

Completion Rates

Consultations are great but we also need to make sure we keep no shows down and try to avoid late consultations. Both of these are bad for business plus bad for clinicians who wait around not helping people and patients having longer queues with people who don’t turn up clogging them up.

The good news is we hit a record high last month with 82.3% of meetings being completed (the rest being cancellations, no shows or technical issues). Given the industry averages for remote primary care being in the sub >80% range I’d say we were doing very well/extremely competitive.

We’ve been focusing a lot on improving this metric and this has meant a lot of time digging through potential frailties in our service (to reduce technical errors that could cause a meeting to not be completed) to thinking about how can we ease the pain of lateness/delays (often a big reason for no shows or late cancellations).

Drop-in Video Meetings

Our most frequent type of meeting is a drop-in video consultation. Back in February 2022 these types of meetings had an OK cancellation rate and no show rate but in July 2022 it was at record highs. Since we’ve made some tweaks, and as of April 2023 we’ve seen decreases of:

• 23.52% decrease in cancellation rates from Feb 2022
• but a whopping 62.63% from our record highs in July
• 19.78% decrease in no show rate from the month of Feb 2022
• but again a gargantuan fall of 52.25% in no show rate from our peaks.

At the scale we are talking this is many many many thousands of meetings saved from being no shows.

What did we do to aid this

A big contributor was our work into voice meetings as a fallback which increased completion rates of meetings. It effectively helped tackle issues that can arise before or during a call (e.g. a patient has a very bad internet connection). Stats show that 70% of clinicians that used the feature succesfully connected with a patient after failing to connect via video. This means that the feature was a great success in potentially saving us from a no show consultation, with the clinician able to communicate with the patient rather than be sat in silence.

We’ve also been somewhat obsessive in trying to make sure consultations don’t fail. This includes refactoring, tackling nitty gritty bugs, monitoring our services well and of course adding some extra features.

A sample of features we added:

• a way for clinicians to run diagnostic checks of if their setup is working (e.g. their audio/video setup is functioning well enough for a call to take place)
• guidance on how to fix common problems so clinicians can resolve issues without needing to contact support
• been involved with adding things like audio indicators within the video call to be able to diagnose sound issues. Other efforts from colleagues include having patients being able to add the consultation to their calendar of choice to help avoid no shows and patients being able to perform technical checks of their video and audio like a clinician can.

Service efficiency (completed meetings per hour)

As discussed earlier we’ve spent a good chunk of time migrating people over from an older platform called Care Portal 1.0 to Kry Pro. This required a lot of rethinking what services should be moved over, what should not be and how we could improve what was available given user feedback. We also invested time in making sure the performance of the services and product was great and made sure what we built was accessible.

The efforts have seemed to pay off as last month the difference in service efficiency (so how many meetings can a clinician complete in an hour) is much higher on the new platform, a 65.9% increase in efficiency. Effectively a clinician on the new platform can see a lot more patients per hour which is good for everyone. There are some mitigating factors for that number with some more complex meetings (that take longer) still running on the old platform however when we specifically look at just video consultation meeting efficiency the difference is still very large, with clinicians 64.96% more productive on Kry Pro.

For the markets that use Kry Pro we’ve had record efficiency numbers for April 2023. So our efforts in rethinking what clinicians need in a consultation appear to be paying dividends.

Meeting length

One way we’ve been trying to aid seeing more patients per hour is also to try and reduce time wasted within meetings. Ideally you don’t want to crush meeting time as that will eventually have a negative effect on patients and the quality of the consultation. Helping people can’t be like working in a factory. However, clinicians also aren’t always spending time directly with the patient. They are often reading patient notes, making prescriptions, waiting for calls to connect, messaging patients and other clinicians, writing medical notes, etc.

Video meeting breakdown of time

Comparing the total meeting length of video meeting from February 2022 to last month we see a 24.58% decrease in total time. Which if you think every meeting is almost ¼ quicker than before explains why you can increase your productivity. Good news though, the actual time with patients has barely dropped. So patients and clinicians still get a similar amount of time together.

In Feb 2022 this is the % of time spent:

• 21.7% was prep before the consultation started (e.g. reading notes)
• 2.6% was waiting for the patient to answer
• 35.3% was time with the patient
• 40.3% was post-meeting actions (e.g. writing notes)

Last month this changed to:

• 19.4% was prep before the consultation started - reduction
• 2.6% was waiting for the patient to answer - no change
• 44.1% was time with the patient - a healthy increase!
• 33.9% was post-meeting actions (e.g. writing notes) - reduction

Thus, as you can see. Time was being shaved from the call but not cutting into the time with people actually seeking help.

What we did to achieve this / Multi-tasking

On our mission to improve the consultation experience for clinicians we tried to make sure performance of the services was top notch, less technical issues but also added or tweaked some of the features to embrace clinicians being better able to multitask on a call. In theory, this would mean you could begin a call and still be reading notes or writing prescriptions while saying goodbye pleasantries to maximise time.

A sample of some features we added:

• Picture-in-picture mode for all video meetings so you could move the video around the page or to a second screen and change the size to suit your needs while you served the patient
• Added different size options of the video so they can minimise the patient video to maximise what else they can see on screen (e.g. go more in depth into the patients medical history)
• Make use of timers to indicate how long people had been on a call and when calls are expected to begin
• Make the call to action of contacting a patient very explicit instead of buried lower down in the page. Given we want clinicians to communicate with patients its good to give this the prominence it deserves

Satisfaction

Clearly one mission for the team was to have happy and eventually healthy users. We track patient ratings out of 5 and have a free text box for patients to give us comments. I also setup a little site to help me track reviews on app stores. One such recent patient review was:

Fabulous way to see a GP they are very helpful thank you (May 1st, 2023)

We have a 4.9/5 score on most appstores too.

We used to have a regular survey to clinicians but it’s became admittedly more ad-hoc in recent months. That being said, we’ve still been using feedback and iterating on many of the services we are responsible for towards the same goal of happier users.

Progress?

In terms of collecting feedback I was part of the effort to increase the % of responses. We worked on how to make it easier to let patients give us feedback and did a few content design changes. Back in February 2022 we only saw 8.5% of users give us feedback on meetings, last month that was up to 36.8%. Which is remarkable how much the jump has been! Peak feedback saw 42.1% of patients telling us how they feel about the services we provide.

Largely though top level patient satisfaction scores have remained static but lots of the year was focusing more on the clinician experience so this is unsurprising. As discussed in the earlier blog post, some of the services we helped introduce such as ‘Voice as Fallback’ did have the highest patient grades of all other meetings so some wins along the way.

One other service we created that saw a marked improvement in satisfaction scores was when we introduced video calls within a chat meeting. So effectively your chat consultation can switch to a video if needs be (e.g. you have coughing symptoms and the doctor wishes to see/hear what is happening rather than read about it). The meetings that combine both have seen a 20.55% increase in satisfaction scores from patients too.

For clinicians we’ve been iterating a bunch on our consultation pages. A recent quote from a French clinician after changes:

Love the new updated meeting view! (May 22nd, 2023)

Easing the pain when things go wrong

While uptime of core services remained great throughout the year, with very few incidents in memory and over 99.99% uptime, not everything was an overwhelming success. Mostly on the patient end we still saw that web users and Android users experienced more bugs than those on ioS devices (the majority of our users use an Apple product). We had also begun digging through some of these issues such as one more recently where the “stream was destroyed” and some odd behaviour about accidentally ending a call twice which then led to some funny business in our logs. We had more to do.

Costs

We massively reduced the cost of some of our services, especially in France, by switching suppliers that were being used. After some investigation we worked out that we could reduce the cost per minute of a video call by 54.6%. When we did it we also saw a slight increase in patient satisfaction and reduced no shows/technical faults by 41.2%.

In a time where the company is aggressively pursuing profitability then I am happy to say we were very much at the forefront of helping the company try and hit its targets.

Reflections

There was more on the roadmap & I probably oculd share many more stats but I hope by now you get the picture. The team was genuinely a total delight and we achieved a lot in a short-ish time together. One day I’d love to get the dream team back together but for now we’ll need to reflect on a job well done.