Tuesday, August 28, 2012

On the Invisibility of Selection Bias and Regression to the Mean in Population Health Management

A DMCB Vacation....
While the Disease Management Care Blog is neglectfully vacationing at the New Jersey shore, the spouse is slathering it and the spawn with SPF 1003.  That's why the DMCB can't help but recall the recent advice of a media dermatologist: fear the sun.  Every melanoma patient he had ever cared for "recalled at least one severe sunburn!"

The population health management providers would be well advised to remember that skin doctor's silliness.  While every patient we recall "got better" also ask yourself the simple question that should have been posed by the radio-show host:

"What about the patients you didn't take care of?"

Talk to any telephonic-coach nurse or any medical home care manager, and he or she will likely be able to rattle off dozens of HIPAA-protected encounters that started with an expensive clueless patient and ended with a less-expensive empowered patient. They and their manager-supervisor-vice-presidents are sure they had an impact. They see little reason to believe otherwise and see no need for "proof."

They're wrong.

Unfortunately, for both dermatology and PHM, it cuts both ways.  Plenty of patients have had severe sunburns and never had to see a dermatologist.  And there are plenty of patients who go from high to low intensity without the benefit of care management.

Of course, don't credit the DMCB for this example of selection bias with regression to the mean.  Skeptic Francis Bacon commented on the invisibility of drowned worshippers centuries ago.

The DMCB won't mention that problem to the spouse when it enters the surf.

Image from Wikipedia

No comments: