Friday, January 30, 2009

Regional variations, costs - Practice Variations and Patients' Socioeconomic Status

Some, but not all, practice variation between various regions of the country stems from differences in the socioeconomic status of populations served – not necessarily to differences in physician-induced demand or hospital pricing. If you service a population of poor patients with low health literary and high chronic disease risks, your costs will be higher.

Flawed Last Blog

That’s the message I failed to convey in my last blog. Richard “Buz” Cooper, MD, Professor of Medicine at the University of Pennsylvania, corrected me with this comment.

I really don’t think its “unwarranted variation” vs. unbridled entrepreneurism. It’s whether the variation is due to suppliers or patients. The Dartmouth Group fails to adjust for risk (they say they do, but they really don’t) which leaves the variation” unexplained,” and since it’s not explained, it’s” unwarranted” and must be due to physicians and hospitals since no other reason is apparent. The Dartmouth crowd doesn’t want to find another reason. The one they have suits them just fine. But whenever anyone disaggregates the Dartmouth data, most of the variation is explained by clinical risk (high risk = higher use +poorer outcomes) and socioeconomic factors (poor patients used more service).

I don’t think anyone would argue about whether the health care system could function better or whether doctors could function better or whether doctors could be more efficient in their use of resources. Efforts to improve things are widespread, from clinical trials to decision analysis “hints” to system technologies and more – and there’s lots more to do. It’s a constant struggle just to keep up with the “inefficiencies ” that are inherent in caring for more complex patients, dealing with more system pressures (regulations, 80 hour weeks) and more. But this struggle has very little to do with why Mayo is different from hospitals in Newark and Chicago – the patient populations couldn’t be more different- and they will never be the same. It is disingenuous to tell policy makers that there’s 30% of health care spending out there just waiting to be saved by homogenizing the populace.

Lower Health Literary, Higher Risks, Higher Costs

In other words, differences with practice variation can often be traced to the poor socioeconomic status of the population. This view is reinforced in mu mind by a power presentation I was reading by Jerry Reeves, MD, one of the nation’s leading authorities on cost variation and author of “Report Cards, Incentives, and Reminders – Impacts on Health and Costs.” Reeves is president of the Los Vegas Operations for the Culinary Health Fund that provides health benefits for hotel and restaurant workers across the country.

Among other factors, explains Reeves, is that health risks differ in populations with low health literacy. These low health literacy rates are,

• White/Anglo 39%
• Hispanics/Mexicans 79%
• African Americans 75%
• Native Americans 64%
• Asian/Pacific Islanders 61%

These literacy rates could make a tremendous difference if you are at UCLA serving its poor population. Low literacy is the single best predictor of health cost, and leads to,

• More hospitalizations
• More emergency room visits
• Less screening
• Later stages of disease
• Lower treatment adherence
• Proper understanding of treatment
• Higher homicide and suicide
• Higher infant mortality of offspring

Poorer populations have a higher incidence for these measures,

Health Measure Added costs per year
• High blood sugar $1,150
• Overweight $690
• High tobacco use $447
• High cholesterol $428
• High Blood Pressure $390
• High sitting around $339
• No self-care $225

No Mystery

In short, there’s no mystery to why doctors and hospitals serving poor populations have higher costs and why health costs vary with the socioeconomic status of the population.

None of this is to say that variations do not exist in payment and practice patterns of physicians and hospitals. Whether these variations are “warranted” or “unwarranted” is another matter and depends on the judgment of the analyst.

But most fair-minded observers would concede that variability is part of the human and organizational condition. Individual physicians vary in their practices from day to day; physicians in the same practice vary from each other; physicians serving the same socioeconomic populations vary; hospitals vary in their payment policies.
Physicians and hospital administrators are not automats. They may not even be aware of these variations, so it is important to bring variations to their attention. But at the same time, you cannot homogenize humankind, and you cannot homogenize doctors and hospitals.

1 comment:

Dr. Bonis said...

Do you have the powerpoint with the data?