Discover more from DVS
Personal Risk and Preventive Medicine in Primary Care
Why I'm bullish on middlemen and optimistic about technology
If my recent exchange with my primary care provider is any indication, we’re not prepared for the emerging tools and strategies to reduce individual long-term risk.
I recently had a panel of labs done as part of an overdue annual visit. Two days later my results posted to MyChart, with the ordering physician’s name next to it (redacted).
I clicked on “Comments from Care Team” and saw this:
I took a quick look at the results. Everything seemed mostly normal except my LDL was “borderline high” and my non-HDL cholesterol was “above desirable.”
I thought about the risk estimate my physician had provided and concluded that while it might sound small I couldn’t get comfortable with his interpretation. I quickly tried to inventory where else in my life I was wittingly running similar odds of something that negative happening within that time frame.
Having recently read Outlive and some of the online discussion about emerging markers and diagnostics (cf Eric Topol’s review), I sent a MyChart message back to say that I wasn’t thrilled with my LDL level.
I asked if he had suggestions about what I should try to bring it down, or what additional testing might be valuable to help me better understand my accumulated risk and when other interventions would be warranted. I also asked for his interpretation of the rest of the studies we’d done during that session. What could he say about my inflammatory markers or my metabolic health or any of the other tests?
His response: “Fortunately your [sic] using the American Heart Association calculator, your 10-year risk of developing cardiovascular disease is estimated at 3.8%, which is considered ‘low risk.’ Starting cholesterol-lowering medication has been recommended when the estimated risk is over 7.5%.”
The conversation ended.
On social media this summer there’s an active debate about a new company that plans to market cash-pay whole-body scans.
On one side are a group who think the risks of unnecessary imaging and potentially harmful interventions driven by incidental findings outweigh any possible benefits. They warn that a wave of worried well seeking follow-up care will choke an already stressed delivery system.
And the costs, even if born by individuals, will compound as they travel through the system - when we should obviously want to spend less.
They highlight how these technologies distract from proven ways we have to reduce risk, and since these are expensive and privately paid, argue that they will increase health inequalities, though that unwittingly implies their beneficial effects.
On the other side is a group of enthusiasts who hold that the value of early detection (and intervention) outweighs its potential tradeoffs.
They have powerful anecdotes, set in a pervasive optimism about science and technology and its ability to enable individual consumers to do more for their health on their own than they’ve been able to achieve inside the healthcare system.
The same forces are structuring my conversation with my physician about my latest labs.
Historically, our approach to preventive medicine and screening has been dominated by our understanding of how risk behaves across a population. We try to balance the indeterminate utility and value of interventions against imperfectly-understood downside risks. Sometimes less medical care is more. Sometimes its absence means a delayed and catastrophic diagnosis.
A lot about the probability calculation depends on variables that are less well understood and capable of changing more quickly than the conversation assumes.
The odds of an inaccurate diagnosis are not static over time but shift as science accumulates clinical experience. The predictive probabilities of diagnostic tests change significantly with small increases in the prevalence of a disease, which we don’t often measure all that well. And the risks and benefits of interventions evolve as they take place in new care settings or with refined or different technologies, like they might when traffic patterns and roadways are better designed or a new passenger safety innovation is introduced into motor vehicles.
Finally, it’s not foolish to hold open the possibility that screening lots more people might materially change our understanding of a disease in some way (including in a way that establishes the value of the screening).
Yesterday, Sachin Jain published an article in Forbes arguing against the rise of “middlemen” in medicine, which he sees carving out and diminishing the patient-physician relationship. I agree that it’d be great to have a physician who’s able to work closely with me to make me healthier. But my experience is that the dominant conceptual model about individual risk and preventive medicine has mostly smothered this conversation.
While we encourage people to become active partners in improving their health, at some threshold of risk, derived mostly from effects averaged across a population, the valence of that aspiration suddenly changes. Medicine discourages and even shuns further curiosity and efforts. Those who continue are portrayed as though they’re taking part in some kind of health safari, characterized by medical care in such lavish proportions that only professional athletes and the worried wealthy can afford it.
In the case of cardiovascular disease, we have a well-described set of risk factors. But if someone has mostly addressed the gross determinants (smoking, weight, exercise, and so on), preventive medicine/primary care just stops talking to them about how to improve their cardiovascular health. The conversation ends. Traditional primary care is out of ideas and all but unable to engage. Of course, that silence is no guarantee that the individual situation and risk has been understood and covered, just that at that part of the distribution we’ve done everything we’re aware of that reduces risk on average.
We can, however, increasingly expect to have opportunities to use new tools and strategies to estimate and reduce individual risk. These approaches will work out where you are and what you should do, while considering (but underweighting) what worked for some average of people generally like you. Individual differences will inevitably start to emerge from behind the regression model estimates.
In my case, why wouldn’t I want to try to lower my LDL into the normal range, even when my risk, in aggregate, might be low? With more evidence about accumulated changes in my cardiovascular system I could yet be more aggressive in diet, in adopting other lifestyle modifications, or in starting medicines. And why would I want to take a 10 year view of CV risk rather than a 20 year one? If it’s only because that’s the time period on which we have solid population data, then we don’t have as much evidence against that approach as claimed.
My doctor just may not have the time or incentive to help me further. As patients, we’re stuck with traditional fee-for-service primary care while science and technology expand the possibilities to understand and improve an individual’s health. No surprise that we turn our attention elsewhere. To middlemen that create connection, to consumer products that allow us to act, to technology that promises to create an opening in the darkness (no matter how flawed). And when we can’t get anything out of their primary care doc, we’re not wrong or misguided in doing so. After all, we’re not sure how much time we have left, either.
My doctor may be right, empirically, that across the population it’s not worth my pursuing knowledge or risk reduction any further. He could take a stand against medicalization and suggest that I’d be better off focusing on music or gardening. I’d listen to that argument. But that’s not what he told me. There just was no physician engagement.
As a narrative strategy designed to encourage a person to improve his or her own health, and to develop a salutogenic relationship around that shared goal, his approach of benign neglect, rooted in some population estimate of my risk and therefore silent, is ignorant of its own limits and counterproductive to boot.