Archive

Happy December, and thanks and welcome to the 100+ folks who signed up since the last newsletter. This time around I've a bit of a holiday hodgepodge, something like our Thanksgiving dinner at Chief O'Neill's in Chicago (amazing and recommended)

The family that buffets together...

The past few weeks, I've been beavering away on commercial strategy for our NewCo. Am loving the beginning-ness of it and the chance to learn and explore new (to me) parts of healthcare and technology. Though we're a little ways from market, I keep hearing echoes of Roger Martin going on about strategy vs planning. Strategy "imagines a desirable future and makes a set of choices with the best chance of bringing it about. It is fundamentally not analytical." But everywhere and usually: "if you ask for strategy, you are likely to get planning." I love the concept of strategy as a theory about how to produce outputs you don't directly control. More to come soon on the new business and what we're up to.

With our NewCo, I've been paying more attention to AI/ML in medicine. Here's something: A new ML model can identify male from female human retinas. 👁️ Before this work, no one realized there was a difference! As Benedict Evans put it: "ML can give you infinite interns, but it can also give you one intern with infinite patience - that can look at 85k photos of retinas and spot a pattern no-one knew was there." Full-text

Here I am on another night of rain-snow mix, sitting in a cold car waiting for a kid's music lesson, and realizing it's been a full turn of the seasons since the last newsletter.

My big news is that I've started a new medical device company with two friends.

I've been working in respiratory disease for most of my career. For much of that time, I focused on digital health, from well before there was much interest or enthusiasm. Today the market is giant, but we allocate too much energy and capital to unimaginative, incremental projects and other that focus on solving/exploiting the weird problems of the US healthcare system. Not nearly enough is going to the diagnostics, devices and therapeutics needed to improve care and treatment.

How we design clinical trials for asthma and COPD - and how much of our research results from tightly-controlled studies that may not reflect everyday reality in the clinic and community - gets my vote for most underappreciated problem.

As I posted recently, in the case of asthma, a 2011 NEJM paper reported that eligibility criteria in most RCT treatment trials excluded an estimated 95% of patients with a current diagnosis of asthma.

Now we have a good analysis of the situation in COPD. A brand new paper in AJRCCM, titled Prognosis of Patients with Chronic Obstructive Pulmonary Disease Not Eligible for Major Clinical Trials, analyzed a large community database from Copenhagen and found that more than half of individuals with COPD would be excluded from major clinical trials when applying commonly-used eligibility criteria.

Randomized controlled trials primarily study whether an intervention works and therefore require the important outcomes (eg exacerbations) occur with some frequency. As a result, sponsors design trials to select individuals more likely to experience such outcomes in a reasonable time period. But this raises questions about whether results also apply to the less severe population commonly encountered in clinical practice.

The epidemiological data on asthma morbidity continues to puzzle and frustrate.

In a new study from Canada, among children, age-adjusted annual hospitalization rates per 100k decreased by 55% in females (152 to 69) and by 60% in males (270 to 108) from 2002 to 2017.

But a closer look at trends in the changes reveals that something stopped working in 2010.

The annual rate of decline all but evaporated: from -15.3 [females] and -25.8 [males] before 2010 to -0.6 [females] and -0.8 [males] after 2010.

Our approach to medication dosing is about to change.

Some quick thoughts about personalization of regimens and medication optimization:

Historically, we've mostly studied and prescribed medications for their average effects in a population, rather than how they perform in a given individual in daily life.

Even with clinical guidelines that encourage adjustment to achieve intended outcomes, few people ever see their regimen optimized once a medication is prescribed.