Does Not Compute

What’s that nagging pain, you ask?  In today’s SF Chronicle, there’s an Open Forum piece on crowdsourcing medical diagnoses, a new and trendy way to find out what’s bothering you.  Patients submit their symptoms to an online forum of diagnostic enthusiasts, some of whom are medical professionals, and they respond with a list of potential diagnoses, the most popular one listed first.  Imagine Wikipediatrics.  Or Family Practice Feud (“our survey said…?”).  Any set of symptoms or data will generate a bell-curve of answers, and the theory is that the peak of the curve is most likely to be correct.

My first reaction on reading this was to choke on my coffee.  As a rheumatologist, I was impressed by how many diseases in my field present with almost exactly the same constellation of symptoms:  for example, fever, joint pain and rash.  Much more testing is usually needed, and even then, the power of every test is limited by its sensitivity and specificity.  For decades, I taught medical students and residents about the importance of subtle variations in the patient’s history and physical exam which could lead to the diagnosis and treatment, even in the face of contradictory and misleading test results.  To steer through a complicated landscape, it helps when the physician is dedicated not just to finding the answer, but to the larger goal of helping the patient.

Dr. Lisa Sanders, who teaches at my old medical school, has a column in the NY Times Sunday Magazine, in which she presents a challenging diagnostic case weekly.  Based on her description of history, physical and preliminary tests, readers suggest possible diagnoses.  And, predictably, the suggestions are all over the map:  remember, many diseases look remarkably similar!  Now this crowdsourcing of diagnoses brings the whole enterprise to a different level- for a fee, of course.

There is an analogy to how the practice of medicine is currently evolving:  diagnosis by computers.  Given a set of symptoms, a diagnostic algorithm can pop up a set of answers, with the most likely one on top.  But the old expression “Garbage in, garbage out” applies when subtle points of the history and physical are passed over, or when irrelevant data are swept into the equation.  Someone still has to decide what data to enter or leave out, and wouldn’t it be better for that person to have expertise, judgment and the goal of not just finding the answer, but the larger goal of helping the patient?  In other words, a good (non-cybernetic) doctor?

Here’s an excerpt from Fourth World, in which an intern, Kai, presents a puzzling case to his attending, Dr. Hol Chan:

Kai continued, “I have put W.P. through the Probot twice, and both times the results were identical:  signals of tissue injury or regeneration, inflammation, pre-mutagenesis and metabolic derangement are completely absent.  Epigenetic expression, including at the micro-RNA level, is normal.  Risk loci mapping and haplotype structure are unremarkable.  You can see on the next screen that the central and peripheral chi are not in any way obstructed.  I entered the patient’s history, systems review, family history, physical exam and lab data into the analyzer and found no matching diagnosis.  And so, without a suitable coding of his diagnosis, there is no way to initiate the billing process.”

Dr. Chan, studying the wall screen, nodded in agreement.

Kai looked up from his da-disc and shrugged.  “In fact, W.P. is perfectly healthy, even though obviously he is persisting in his illness behavior.”

W.P. stared fixedly downward at his legs, now pale and mottled in the cold room.  Unsure what “illness behavior” implied, at least he knew that his pain was very real.  It was excruciating, every minute of every day; the sleepless hours of each night passed exquisitely slowly…  His wife finally shattered the silence:  “Healthy? Perfectly healthy?  What are you talking about?  Can’t you see he’s in pain?  What’s perfectly healthy about that?  Can’t you just rearrange his genes and end this once and for all, instead of giving him all those… those therapeptides to control pain but only make him more lethargic than he already is?  You doctors and your damned machines:  scanning and probing here and there, coming up with nothing.  Epigenetic expression is normal.  Oh, so everything’s just fine then, is it?  Well it’s not, and I am at the end of the line, people!   I can’t stand it anymore- you find the problem, and you find the solution…you find it.”  She suddenly began to sob, her shoulders shaking, her arms folded tightly across her chest.

Kai had panic scrawled on his face.  “But the Probot is accurate within ten-to-the-minus-seven-percent!  As I said, there’s nothing wrong with your…”  Dr. Chan cut in abruptly:  “Kai.  Excuse me.”  Then, addressing everyone in the room, she said in her most calming voice, “I think we’d better break here, and collect our thoughts.  Why don’t we go to the conference room and review our findings?

“If you don’t mind,” she said to the patient’s wife, who had just as suddenly stopped her crying but shook her head slowly back and forth, unwilling to accept that she and her husband were going through this yet again.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s