It won’t be an exaggeration to call 2020 as an epoch-making year for everyone and so it has been for me too. At the beginning of the year, there was the usual looking back at 2019 & making plans for the year to come; but within the first quarter, almost every single of those plans had to be looked at and reworked again, (and again and again). One thing that kept me going was my reading (old habit) and writing (newly found habit). When I look back at my 2020 in lockdowns, social distancing and overall pandemic anxieties, these habits helped me keep sane. Below are some notes containing a summary of my writing followed by notes on some readings that stand out for me.
Writing about healthcare-Long Form (sort of)
Through the year I researched and wrote eight articles, most of them reflection pieces on areas of digital health that was about the work/research I was doing at that time, including mHealth design, outcomes research and implementation in the real world. Of these eight, I especially loved the below three.
At the beginning of the year, I reviewed 50 original research articles on applications of machine learning published in 2019. There were some interesting findings-
- Most of the ML models were generated and evaluated on a single data set within a single hospital/healthcare system
- A surprising 57% of the studies did not use an independent validation set. Amongst these, some studies did not specify use of validation sets at all; some had randomly chosen validation sets or replaced by cross-validation (validating within training sets).
- On model performance, only 48% articles used additional performance metrics such as PPV, NPV and F1 scores.
In this two-part review of trying to measure ‘impact’ of mHealth apps, I reviewed research published in 2019 that used some kind of mHealth interventions-this resulted in 30 apps that reportedly used behavioral components and published their results in scientific journals. I reviewed the design and results of these studies for scientific validity and
Several observations followed-~65% of these trials were run at a single hospital/facility; ~60% trials were run for short duration and very few considered measuring app adherence/usage metrics. All of these factors would contribute to their limited applicability.
In the second part of the review, I provided some pointers about how to better design, conduct and report research for mHealth interventions-
- Who will the app really cater to? Where can we see the maximum impact?
- What does this patient group look like? How can we design the intervention to cater to this specific group?
- How can we show the effect of intervention is really connected to the app?
- After a successful study, where do we go next?
Oh, this was definitely the most fun to write about. I reviewed Form S-1s for a total of 13 firms in the digital healthcare world that went public in 2019 and 2020 to tease out some answers to these questions. Some fascinating insights ensued-The median age of CEOs for this cohort is 50, significantly lower than the median of 58 for CEOs at S&P 500; only two of the firms could be considered as B2C and the entire cohort was highly dependent on Big Tech.
Looking back at these writings, I think through all of these I tried showing how much hype exists behind the tall claims we hear about digital healthcare and the efforts we all need to take to start developing truly impactful innovations.
Writing about healthcare-Short Form
Around Q2 2020, I also realized that while my reading about healthcare has started getting more diverse, I haven’t yet found a good way to collate and curate this for future review/consumption. This was the time ‘Things & Thinks’ was born- it is a newsletter I publish every fortnight which has curated healthcare news, updates about digital health and my mini-musings on innovation, technology, governance and leadership in healthcare. I have published 16 editions so far and have started getting into a good rhythm. I will publish the ‘Best of 2020’ edition soon!
Reading during a pandemic
I ended up reading more books than usual this year. Some of my favorites through the year are-
- Fiction: So Much For That-Lionel Shriver: An incisive take on American healthcare, showing how messy the whole system is and how it impacts families and individuals. Favorite quote-
“And then all this talk, at the hospital, about ‘fighting,’ and ‘beating,’ and ‘winning.’ Of course you’d rise to that. Try to shine in the contest. But it’s not a contest. Cancer is not a ‘battle.’ Getting sicker is not a sign of weakness. And dying,” he said the word softly but distinctly, “is not defeat.”
- Non-Fiction: Hello World-Hannah Fry: is an interesting look into how what we call ‘algorithms’ have existed for some time now. Hannah Fry walks a fine line of balance when she considers both the ill effects as well as possible uses of data and its uses. She provides detailed examples of algorithmic experiments gone wrong and at the same time doesn’t shy away from pointing out how the so called ‘human touch’ is not far away from the same pitfalls that we blame the other side for. Favorite quote-
Whenever you see a story about an algorithm, see if you can swap out any of the buzzwords, like ‘machine learning’, ‘artificial intelligence’ and ‘neural network’, and swap in the word ‘magic’. Does everything still make grammatical sense? Is any of the meaning lost? If not, I’d be worried that something smells quite a lot like bullshit. Because I’m afraid — long into the foreseeable future — we’re not going to ‘solve world hunger with magic’ or ‘use magic to write the perfect screenplay’ any more than we are with AI
- Fiction: A Memory Called Empire-Arkady Martine: An unsual sci-fi set with a tight plot slowly progressing towards a closure that generates more questions than it answers, at the same time closing many loops to give you some satisfaction! Favorite quote-
A mind is a sort of star-chart in reverse: an assembly of memory, conditioned response, and past action held together in a network of electricity and endocrine signaling, rendered down to a single moving point of consciousness
- Non-Fiction: Gene- Siddhartha Mukherjee: This was a great read started back in 2019, then slowly savoured through the first half of 2020. The book as its most lucid, informative and exciting when the author writes about the interplay of various elements, the high point for me was reading about Human Genome Project, almost feeling like watching a movie! Favorite quote-
If we define “beauty” as having blue eyes (and only blue eyes), then we will, indeed, find a “gene for beauty.” If we define “intelligence” as the performance on only one kind of test, then we will, indeed, find a “gene for intelligence.” The genome is only a mirror for the breadth or narrowness of human imagination.
- Fiction: Infinite Detail- Tim Maughan: Reading this freakishly prescient sci-fi during the ongoing pandemic and US riots and tweets by @Anonymus made my head spin and added to all the disorientation that 2020 has been. Favorite quote-
The pinnacle of human effort had been to create a largely hidden, superefficient, globe-spanning infrastructure of vast ships and city-size container ports — and all to do nothing more than keep feeding capitalism’s hunger for the disposable. To move plastic trash made by the global poor into the hands of hapless, clueless consumers. A seemingly unstoppable beast built from parasitic tentacles, clenching the planet with an iron grip.
- Non-fiction: Wild Places- Robert Macfarlane: A mix of nature writing, travelogue and reflections, this is a fantastic lyrical exploration of the remaining ‘wild places’ in the UK. Favorite quote-
We cannot navigate and place ourselves only with maps that make the landscape dream-proof, impervious to the imagination. Such maps — and the road-map is first among them — encourage the elimination of wonder from our relationship with the world. And once wonder has been chased from our thinking about the land, then we are lost.