Tuesday, March 4, 2014

"Transparency" games? Are heart surgeons cherry-picking patients to bump up their outcomes data?

An interesting post from The Incidental Economist today:

The unintended consequences of public reporting of mortality outcomes
  March 4th, 2014

New York tracks and reports the mortality rates of heart surgeons. And great news: Operative mortality is going down! But is this because surgeons are ‘cherry picking’ only the least sick patients?

The idea behind public reporting of mortality rates was to allow patients to select the best doctors and to pressure doctors to improve their practice. How well did it work?...


 Surgery is almost certainly getting safer. However, a surgeon can also reduce her operative mortality rate by choosing not to operate on the patients who are sickest and least likely to survive...

...it’s at least possible that surgeons are accurately judging whether patients would benefit from surgery. That is, perhaps they are declining to operate on those patients for whom the operative risk and cumulative trauma of surgery are such that the patient would gain no additional life from the procedure.

My bottom line is that it is premature to conclude that surgeons are letting patients die to advance their careers. The question is nevertheless important and I urge you to read Kolker’s article. For more TIE posts on the complexities of using outcomes to improve quality, see here, here, and here. For more on the importance of competition among hospitals for improving health care quality, see here.
"Data." "Big Data." 24/7/365, floods of HIT-enabled data. Yeah, it's gonna be Beneficent, Transformative Analytical Nirvana...


Not. Adroit empirical analyses will be more important than ever. Spurious correlations will be easier to produce than ever before. "Just gimme the p-value!" (pdf) What do you think?

From my 2008 Tranche Warfare post:
While at the credit card bank, I once interviewed a pleasant young hire prospect (playing house "liberal-arts-guy" dumb), a woman with a Master's degree in Statistics. Offhandedly, I asked her to explain to me, in plain plebian English, the concept of "Standard Deviation."

She couldn't do it. She haltingly gave me all the Stats textbook jargon: "Root Mean Squared (RMS) Deviation," the "Square Root of The Mean Squared Deviation, corrected for degrees of freedom" blah, blah, blah...

I dropped the line of questioning.

OK. The Standard Deviation is simply the "average" or "expected" variation around an "average." You calculate an arithmetic average. Unless each value is identical, there is variability. The Standard Deviation -- beneath the hood of all the Scary Greek Shit -- is simply the amount of variation to "expect," "on average."

We hired her anyway. It wasn't my call. She lasted about 3 months, did a few banal yet aesthetically pleasing Excel sheet graphs and Powerpoint assemblages, and then moved on to inflict her thoroughly academically pedigree'd ignorance elsewhere.
A quick trip over to Science Based Medicine for some thoughtful analytics:
Different Strokes for Different Folks: Assessing Risk in Women
  Harriet Hall

You may have noticed that men and women are different. I hope you have noticed. As the French say, vive la diff√©rence! It’s not just that one has dangly bits and the other has bumpy chests. Or that one has to shave a beard and doesn’t like to ask for directions while the other has menstrual periods and likes to discuss feelings. There are differences in physiology and in the incidence of various diseases. For instance, normal lab values for hemoglobin are higher for men than for women, and autism is more prevalent in males while multiple sclerosis is more prevalent in females.

In the past, women have been underrepresented in clinical studies; when the first studies of aspirin for cardiovascular prevention came out, we knew it was effective for men, but we didn’t have enough evidence to recommend it for women. This is changing; researchers today are more aware of the need to include women in their studies. Now the American Heart Association/American Stroke Association (AHA/ASA) has issued the first evidence-based guidelines for reducing the risk of stroke in women.

Previous guidelines were for all adults, without specifying any differences by sex. Interestingly, those guidelines said stroke was more prevalent in men than in women, directly contradicting what the new guidelines for women say. They say the lifetime risk of stroke in women is higher than in men (20% vs. 17%). In 2009, 60% of stroke-related deaths were in women. Women are different: genetic differences in immunity, coagulation, hormonal factors, reproductive factors, and social factors can influence the risk of stroke and impact stroke outcomes. The new guidelines were needed to reflect risk factors that are unique to women...


In the male-dominated fields of science and medicine, women have too often been given short shrift. The situation is radically different today from what it was when the women’s lib movement first started to raise our consciousness of gender inequities. As the Virginia Slims slogan said, “You’ve come a long way, baby” but we need to go still further. The new guidelines point out that risk assessment tables for cardiovascular disease are based largely on data from men; better data and new tables are needed to predict risk in women. I hope these new sex-specific guidelines will be the first of many such efforts, not just in cardiovascular disease but in every aspect of medicine.
__

IN OTHER NEWS:
EMR and Meaningful Use thoughts from MedPageToday

video
 __

WEDNESDAY UPDATE:
MEDICAL LICENSURE, 1512 EDITION

The Incidental Economist is on a tear of late.
Preventing sorceries, witchcrafts, and other inconveniences
March 5, 2014, Nicholas Bagley
The impending doc shortage has spurred a lot of discussion about whether scope-of-practice laws ought to be relaxed to allow nurse practitioners and other medical professionals to assume greater responsibilities for primary care. This is, of course, just the latest iteration of a never-ending legal debate over the proper role of the state in regulating the practice of medicine—a debate that stretches back to at least 1512 and the reign of King Henry VIII...

A NEW TWITTER CONNECTION


My Vegas friend John Lynn is everywhere, all the time in the Health IT space. Doing a great job.

HIT FUNDING REQUEST

From Healthcare IT news
Obama budget tags $1.8B for health IT
HHS budget trimmed from last year's total
WASHINGTON | March 5, 2014


In his $3.9 trillion fiscal year 2015 budget proposal released Tuesday, President Obama asked for $1.8 billion to support health information technology incentive payments — the same amount he requested last year. Actual spending for this category came to $1.07 billion in 2013.

The budget also included $77.1 billion in discretionary funding to support HHS’s mission, $800 million below the 2014-enacted level...
Pretty wimpy, if you ask me.
 ___

More to come...

No comments:

Post a Comment