Reposted from January. The panel referred to is still meeting, and we are working on its final report. Stay tuned!

I went to a meeting of a panel this past week under the auspices of the Institute of Medicine. Called the Evidence Communication Innovation Collaborative [yes, I know], the group discusses a number of topics around the general ambit of communicating medical evidence to patients. We spent a lot of time, productively, on the topic of decision aids. A lot of people at the meeting really like them. (Here's a collection of decision aids, which includes a short primer on what they are; the Wikipedia article is informative.)

I like them, too, and it's not hard to understand why. Decision aids are the fuel of shared decision making. Information should not be concentrated in the hands of the doctor; rather, it should be presented to the decider - which we presume is the patient, the ordinary person - in a way relevant to them.

But they are not the panacea:

We forward thinking doctors know in our heart of hearts that decision-making should be shared with the patient; unfortunately, not all patients think that way. Some still rely on the physician to make their decisions.

Even the alternatives, and the risks and benefits attached to each one, are not so obvious without some thought. And that thought is not the view from nowhere, as a philosopher put it, but dependent on the point of view of the person thinking. The risks, benefits, and preferred alternatives depend on the kind of person doing the choosing. And who's to say that a patient from one race, say, or economic stratum, will react to alternatives the same way as another?

Count me encouraged but skeptical: there were a lot of people in that room ready to share decision making. But a decision aid is only as good as the decisions it includes. We need to know a lot more about how people make decisions, and how they talk to their doctors, before we can expect such aids to do more than reproduce our current health care system's inequalities and insensitivities.

Reposted from last year.

After a talk I gave about doctor-patient communication, a woman in her 30s raised her hand and said, "I work at a university, and whenever I see the doctor, he assumes I'm a student, and he talks down to me. This makes me very unhappy and I'm  not sure what to do about that." I assumed from her circumstances, and from the tone of her question, that she was not free to choose another doctor. What should be done by someone who has this problem? There are various strategies, but part of the answer is to understand why the doctor might be talking down to someone.

1. The doctor's personality is unsuited to an egalitarian interaction. If they were a chimneysweep there would be the same problem.

2. The doctor was not trained to appreciate the importance of doctor-patient communication.

3. The doctor does not know what to recommend, or recognizes, perhaps subconsciously, that there is no good evidence to prefer one route of treatment over another. This leads to a basic conflict: she wants to help the patient but also acknowledges to herself that there is no one convincing path. Since she has been conditioned to present a front of omniscience to patients, both through the system and the influence of colleagues, she is not allowed to present that ignorance, or to partner with the patient to address it. Thus, she appears abrupt or condescending.

Perhaps I am giving such a doctor too much slack? I would do the same for a patient. Systematic changes mean we should, on occasion, refrain from blaming people without considering the entire differential.

My neighbor, the famous pianist,
has gotten ever sicker the past two years,
his body has slowly shrunk
as if he were hiding himself from the world
and somewhere, with trembling hands,
holding onto a concealed
center of life.

And then one day
he entrusted me
with his apartment keys.
He had to go into the hospital
for an operation
and he asked me
to take care of his two vases
in the window.

"I can bring them to my place,"
I offered.

"No, no," he stammered.
"The piano has to live with somebody."
He hung his head
and added as if embarrassed,
"If you have time, it would be good
if once a week, in the evening,
you could just sit for a moment
by the piano."

(from Der Finfter Zman, CYCO Farlag, 2008. Original Yiddish here. Translation mine.)

Last week I had an absorbing conversation with Natasha Gajewski, someone who, besides being the daughter of a former patient, is also the CEO of Symple Health, a company that has produced a new app helping patients to keep track of their symptoms. That patient of mine was by training a statistician, and to each visit with me he would bring piles and piles of printouts, records of lab results, readings from dialysis. I felt it less than relevant to his care - at least most of the information was just superfluous. But, on the other hand, I didn't want to talk him out of something that was obviously important to him.

To hear now her daughter tell it, though, his data collecting was an obsessive thing, keeping his grandchildren away from him because  - by his own personal calculations - a visit from them might disturb his homeostasis.

Obviously, his daughter knew him much better than I did (he passed away a few years ago). But I can't help think that here we have several sides of an important phenomenon: that of the engaged patient who records their own data for the sake of their health. This medical version of the "quantified self" can lead - as this story hints - to reassurance and obsession both.

The quantified self, for those who don't know, refers to the practice of recording - through means both more and less technologically advanced - "readings" from daily life, and using that information for the benefit of health. Such an initiative has something in common with "big data": the effort to amass large-scale collections of data on a population level.

From a medical perspective, I don't know what we can expect from big data. Certainly many claims have been made for population-level data and its potential to improve health on a global scale. Who knows if that will pan out? But what about big data on an individual scale?

Data collection works if there is a question waiting for an answer. Otherwise it's like playing the stock market without any real investment, watching random numbers jump up and down. If you have high blood pressure, checking your blood pressure at home might help to control it. If you have diabetes and check your blood sugar, ditto. But to monitor yourself with thousands of data points, most of which will not answer any askable question, verges on what Abraham Verghese has termed the iPatient - the electronic version of the patient, the tests and results visible on the screen that doctors view to the exclusion of contact with the actual human being. Only, in the case of this hyper-self-analysis, the iPatient is analyzing themselves in a data-centric way to the exclusion of the bigger picture.

What we need is at least two things:

1. A sense of what data will actually be helpful for a given patient, and what "help" (i.e. clinical improvements, patient-relevant outcomes) can be expected

2. A realization that any new source of data will, in general, not be the game-breaking change that people are looking for. Health care problems, in general, are just hard, and data are only part of the solution, which involves patient preferences, the multitiude of possible ways to identify the problem (differential diagnosis), and the goal of patient and physician.

3. An appreciation that data collection can provide, to the individual, both very real reassurance and self-empowerment, as well as a source of obsession and paradoxical lack of control

As Natasha said about our conversation, there are no easy answers, neither unlimited data nor its lack.