Wednesday, June 17, 2015


and the FLY in the OINTMENT

Edward Kaplan and Paul Meyer came up with an elegant solution in discussing outcomes. They created the Kaplan-Meyer Estimator; a non-parametric estimator of survival function using timed data. Each entity logged in as present, absent or deceased. No gimmicks! A treatment for cancer would be graphed on an X-Y axis and the survival data would lend itself to an easy view. Each step-wise ladder would denote those that perished to the disease and those still surviving or censored as little tick marks on the steps. Comparisons with superimposed placebo survival curve identify the benefits or risks (as in the graph below) of the new treatment.

But here we are embarking on this lauded “Outcomes” as a metric to determine quality. I don’t know about you, but I have few misgivings about this form of measure. Actually I have two problems to be exact; a) How does one define Outcome and b) how does one define quality?

In a timed data if one were to treat pneumococcal pneumonia with Penicillin and the fever, chills and the chest XRay improved, then one could consider that as a good outcome. Easy peasy! However throw in the mix of fungus, immunological deficiency from age, disease, cancer and what you have putting it mildly is a major conundrum. No easy cure! Now let us bring in a patient with cancer into the equation. Is survival a good measure? If you say yes, then I have a bridge in Brooklyn available for sale. You see survival is not a simple measure in cancer care either. There are many tiny incalculable genes mutating as they play their game of life. The whole biology of the cancer and its rate of growth thing is a complex dance filled with many characters. There is no one character that struts on the stage momentarily to signify nothing. Every moving and non-moving genetic construct signifies something in the life of a human being. The many that are involved have each a bucketful of colorful lives and a plethora of epigenetic sauces to create the soup of life. So will the individual physician’s entire caseload of patients with a certain diagnosis be put through the K-M curve and compared to the national average? If we do that, even then we might be flawed in our thinking, especially if the physician has a disproportionate number of African American, or Hispanic or Asian patients or older or younger or more females than males, or right-handed than left-handed, or more artists than scientists, or movie-goers vs. internet-obsessed in that population (you get the drift).

Survival Curves 20-100 years & % of Survivors 
A: White women in US (1939-1941)
B: Women in India (1921-1938)
C: Theoretical Population with half-life of 18 yrs.

You see where I am getting at? It does grind out the cogs in the wheels of thought. But for simplicity’s sake, let us take a diabetic patient. If the measure Hemoglobin A1C is not below a certain limit then that would constitute a poor measure. Correct? Not so quick my dear Watson! The assumption here is that the world is a perfect place and that certain input leads to a perfect stream of output. But that is not the case, is it? The patient might like their case-load of Twinkies unbeknownst to the physician and never mention that they consume 3, 3-liter bottles of Cola daily and consider that as hydration. Or simply never take their medications on time due to forgetfulness, expense or laziness. How would that measure in the “Outcomes” category? Oh and while I have your attention, Is Outcome like a Warranty? You know like the many page document found with cars and electronic products? The Six Sigma approach might become necessary for root cause analysis in human outcomes as well. Eventually we as humans might need a MTBF (Mean Time Between Failure) for the population, drilled down to subsets of society and then individual level of existence for comparison, don't you think? And if doctors by virtue of intuition considered gods then so be it, let the Kaplan Meyer Curves roll!

Where exactly do Behavioral Modification and Behavioral Economics intersect? At the expense level or at the payer level? Or at the arbitrariness of the expert level, where most things today seem to live and breathe? Your guess is as good as any!

Now imagine given these simple scenarios that the elite experts who reside in those vaunted ivory towers consider outcomes as a measure of a good doctor and make payments predicated on such faulty thinking? Now before you go saying that cannot happen. No one would be so stupid to accept that? Guess what? That is exactly where we are today. “Outcome” is the new big buzzword. And CMS is trying mightily to convince the “Gruber’s Stupid” that, "this is the way it should be."

We are not done yet. Now add to it some “Satisfaction Scores” from the patients and we have a whole new enchilada in front of us. Say a patient perceived your smile as a cynical one. There goes your score. Or imagine you prescribed a medication that cost him or her more money than anticipated, or you did not recognize him at a department store, or your receptionist missed calling him or her on the specified time, or you as a physician had an emergency and the waiting room with five patients turned into one with ten patients sitting and filling out the “satisfaction Score” form? Any doubt that your “Reimbursement” from the Insurance would be returned “denied.” And you will have to jump through hoops to get a fraction of that expense 3 months later. That would constitute a certain Win-Win for the Insurer, wouldn’t it?

Ah the cleverness in it all!


The fly in the ointment continues to flutter its wings, trying mightily to get out, but the ossified mindset seems unlikely to remove it because that mindset believes in that  removing the fly means failure and more arduous hours of toil to rework a new flawed concept from inception. And guess what, according to them, the fly is part of the ointment and therefore makes for a perfect recipe.

No comments:

Post a Comment