Healthy Nuggets #10: Chatbots, more chatbots, simple solutions and the best wellness apps
Apple updated its Watch range this week but simpler solutions can have major benefits. Photo by Oliur on Unsplash.

Healthy Nuggets #10: Chatbots, more chatbots, simple solutions and the best wellness apps

After last week's "no good news" hiatus from this newsletter, this week could have easily featured 20 good-news posts from around the digital health world. I'll try and compress it into a handful but it's good to see things are still moving along quickly.

This is also the second attempt at getting the newsletter done, after LinkedIn very kindly wiped an entire draft thanks to me pressing "refresh" by accident. One would have assumed the "Draft - saved" status might have saved me, but 500 words went down the drain. As a developer it was always faster to do something the second time, but no less annoying.

Onwards.

A lot of effort goes into these blogs, especially trying to make sure links out to interesting things are accurate and relevant, but if you want to know any more about anything written here, just drop me a DM.


With Large Language Models (LLMs) constantly in the news just now, and companies in all sectors wondering what to do with them, the hackneyed term "chatbot" has become ubiquitous once more.

Chatbots had their own hype cycle in 2016 as discussed in Nuggets #8 where I discussed early advances - and huge audiences - in the travel industry, where replacing forms with time-saving intelligent conversation was seen as a big deal. Sadly the industry moved onto the next thing and despite companies like Babylon - widely discussed on here for its troubles and sale - trying chatbot triage for GPs, chatbots did go on to develop something of a bad name.

But LLMs ought to breathe some life into the subject and a nice article from Alisa Moore in the Code Like A Girl blog, discusses some of the use cases made possible by careful use of chatbots; there isn't a single mention of LLMs in the piece but that's good in a way, as it resets the debate back to possibilities rather than technology.


The blog offers various use cases for chatbots, and is worth a longer read, but broadly they sit in four categories:

  • Information and assistance - helping people to find timely and accurate information, be it symptom checking, such as HealthHero 's Symptom Assessment service;
  • Monitoring and feedback - two-way channels to ensure patients are able to maintain awareness of their condition but also to improve the general quality of health, such as Doctify 's innovative feedback platform;
  • Automation and efficiency - filling out forms is painful and often unnecessary so chatbots - especially with some level of language intelligence that an LLM would improve - can definitely fill this gap. In some cases the forms might not be needed at all with RPA or another form of automated processing.
  • Reminder and notification - this could be a number of things but ultimately works best where bots are part of another platform; our chatbots at Skyscanner had huge numbers because they were present in Facebook, Skype, Cortana and Alexa with low barriers of adoption - but it meant they could also lean on those platforms' notification methods and not need separate permission. This could have been very useful during the early Covid-19 pandemic both for private and public messaging.

All of these cases save us our most precious asset - time - whilst offering varying purposes to make life easier. LLMs may make them simpler to implement but not entirely - they all require ultimately some form of personalisation or knowledge of a patient and ultimately their consent, as discussed in another article this week about the NHS's current AI programmes.

But ultimately the answer may well be chatbots that either communicate with each other or orchestrate a variety of disparate services, since no one company will be able to perform all these tasks. Is an open data ecosystem ultimately the answer, perhaps a health equivalent of Open Banking? Probably.


Image from "Top 10 Health Chatbots" by

Helpfully, The Medical Futurist also recently published its guide to the "top 10 health chatbots" showing the state of the art in this field just now - please do head to the article but given the comments about use cases "chatbots" above, a few of them are worth a mention:

OneRemission is something very close to home at the moment - aggregating important information for cancer sufferers and giving useful natural-language advice on day-to-day management and relief. This might prevent some of the 3am reassurance calls to stretched clinicians after chemotherapy, for example.

Florence by Pact Care is a more general health bot, but has a particular strength in helping people understand and remember to continue with their treatment or medication; with adherence a massive problem in any treatment course or trial.

Side note: just how many companies in health are called "Florence"? Dozens apparently! A quick reminder why that might be, but can't someone just use "Nightingale" instead, as the NHS did?

Sensely 's bot aims to "give personalisation new meaning" by offering patient engagement either on the personal level, as part of an employee health programme, or provided by public health authorities; it operates through chatbot channels in over 30 languages aiming to meet people where and how it suits them.


It's interesting to note that in an entire article about Apple's new WatchOS announced this past week, the closest a writer could get to health advances was "asking Siri how much you slept last night".

The new Apple Series 9 watch with OS 10 - photo borrowed from Apple, don't sue me!

Natural language may be in vogue but the watch does more than that, albeit incremental improvements in things like ECG, blood oxygen, menstrual cycle and mental health; Apple will learn more from the 53 million watches it sold last year than most health startups' data combined.


Elsewhere, an NHS practice in Slough has been trialing technology from Graphnet Health which allows for self-reporting of blood pressure recordings, and in return opens a two-way communication service between clinicians and patients.

The technology is straightforward rather than bleeding-edge but clearly two things are true here: firstly that age is no barrier to adopting some form of technology in self-managing one's health, given the age profile of blood pressure sufferers (and a high number of people want to 'age in place' rather than spend later years in some form of care, for which tech will be crucial). Secondly, Simple technology is just as good, if not better than more modern methods as long as the idea and pathway is sound.

In this case a wearable device gives patients readings they self-report to their practice where they are out of the ordinary; the surgery can then keep in touch via SMS message, guaranteed to work on any age of phone, if there are any next steps.

Dr Bharan Kumar, GP at Bharani Medical Centre in Slough, told Digital Health's news site this week:

“We wanted to develop a simple, more timely, cost-effective, efficient and sustainable model that would benefit the patients most in need of our help, and the use of data and technology has enabled us to do that.“
Patients can easily submit readings from home using blood pressure monitors, which are reviewed by a clinician. If results are normal, they are added to their records, and no further action is needed, reducing face-to-face appointments, and avoiding inconvenience for the patient.
“Those with higher readings are contacted again via text message to upload more readings, with medications then adjusted if needed to improve control.”

Since the project was implemented in December 2021, Bahrani Medical Centre has seen a 15% increase in patients aged 79 and under with controlled blood pressure. Well done to all involved.


After all the AI and chatbot chat, time to have a look at apps (which tend to deliver the best chatbot services anyway). Tipalti has been conducting analysis to come up with the "best wellness apps" in 2023 based on a variety of factors.

Flo, Calm and MyFitnessPal came out as the top three apps in a survey by

The top three: Flo Period Tracker ("Flo", not "Florence"!), MyFitnessPal and Calm , came in with over 2 million downloads between them in the month under analysis this year, with the growth particularly impressive for MyFitnessPal given they first launched in 2005 prior to apps really existing.

The survey is worth reading as it gets into other factors such as app quality, rating and revenue from a variety of sources but it's testament to the potential of health apps that the top two have 100 million downloads each.


Finally on health stuff, many thanks to everyone that took part in our meetup this past week in Edinburgh - featuring fantastic talks by Blair Walker , Stuart Dyer and the team from Lenus Health - and also to all who attended.

We're planning a third meetup this year in November and potentially a parallel one in London over the next couple of months, so watch that space for more!


Talking of which... On a personal note it was fantastic to see close friends (and meetup hosts) at Bellfield Brewery winning Scottish Brewery of the Year this past week. Maybe not health related, but the company was borne of the troubles coeliacs had in the simple pleasure of a good beer without making them ill, so making no apologies for featuring it here :)

Blurry photo of

Bellfield was started around 8 years ago to cater for that coeliac drinker, but has expanded way beyond that brief to making excellent beer everyone would choose; two more awards for particular brews proved that once more. But the national brewery award, ahead of some far better-resourced and long-serving competition, is validation for many years' persistence - particularly given the 2 years of Covid which ruined the going-out trade and crazy price pressures since that.

Well done to the founders, staff, customers and investors (full disclosure, I am one of those) who got it this far.


PS: LinkedIn, please stop messing around with the article format, it's REALLY annoying - especially burying stuff in hidden menus that used to be automatic. And fix your shitty interface so it doesn't lose 500 good words. How about some version control so it doesn't happen?


Small print: This newsletter goes out to subscribers and across LinkedIn every Sunday night around 7:30 pm. Feel free to contact me if you've seen or are creating something interesting in digital health. I work for?Waracle, but all opinions and content selections are my own. Anything in which I have a work or personal interest will be declared.

Cover photo by Oliur on Unsplash


Neil Gentleman-Hobbs

A giver and proven Tech Entrepreneur, NED, Polymath, AI, GPT, ML, Digital Healthcare, Circular Economy, community wealth building and vertical food & energy hubs.

1 年

Good piece David Low we should catch up at the end of the month after our round of shows

要查看或添加评论,请登录

David Low的更多文章

社区洞察

其他会员也浏览了