TUESDAY, March 15 (HealthDay News) -- The likelihood of Medicare patients being diagnosed with a chronic disease may depend on where they live, a disparity that makes it more difficult to assess the quality of care patients receive, a new study finds.
Certain groups of Medicare patients in regions with the most diagnoses also had a lower case-fatality rate for chronic conditions such as coronary artery disease and kidney failure, but the reasons for that are unclear, the researchers reported.
It would stand to reason that whether a person is diagnosed with a chronic disease has to do with how ill they are, the researchers said.
But instead, the findings suggest that chronic disease diagnosis is influenced by the "intensity of health care" in a particular region, which includes how many doctors and specialists are operating in a particular region, access to those doctors and the likelihood of doctors to send you to a specialist or to order lab and imaging tests.
"The study suggests disease diagnosis is not only a property of the patient, but associated with the intensity with which health care is delivered in a region," said senior study author Dr. John Wennberg, a professor emeritus and founder of the Dartmouth Institute. "For example, if in certain regions people see lots of doctors, have lots of visits to doctors and lots of lab tests, that could be because there is a perfect relationship between illness and the amount of care that's delivered. But it could be that the more doctors you see, the more diagnoses you get."
The study, conducted by Dr. H. Gilbert Welch of the Department of Veterans Affairs Medical Center, White River Junction, Vt., and colleagues, appears in the March 16 issue of the Journal of the American Medical Association. It was funded in part by the National Institute on Aging.
Researchers analyzed records on nearly 5.2 million Americans aged 65 and up who received fee-for-service Medicare benefits in 2007. In particular, they examined data on diagnoses of nine serious chronic conditions: cancer, chronic obstructive pulmonary disease, coronary artery disease, congestive heart failure, peripheral artery disease, severe liver disease, diabetes with end-organ disease, chronic renal failure and dementia.
The mean number of chronic health conditions diagnosed was 8.7 for every 10 people. But that varied significantly depending on which of the 306 U.S. regions patients sought care.
On the low end, patients in the Grand Junction, Colo., and Idaho Falls, Idaho, were diagnosed with 5.8 chronic illnesses for every 10 people, compared to more than 12 illnesses for every 10 people in Miami and McAllen, Texas.
Not surprisingly, the more chronic conditions a person had, the more likely they were to die. The fatality rate among people with no chronic health conditions was 16 per 1,000 annually; 45 per 1,000 for those with one condition; 93 per 1,000 for those with two conditions.
Yet paradoxically, among subgroups of Medicare patients in regions where patients tended to receive more diagnoses, the risk of death from a chronic condition falls.
Patients were divided in quintiles (or fifths) based on the frequency of diagnoses in that region.
Among patients with one chronic condition, 51 per 1,000 of those in the lowest quintile for diagnoses died, compared to only 38 per 1,000 in the highest quintile. For patients diagnosed with three conditions, 168 died in the lowest quintile compared to 137 per 1,000 in regions where chronic conditions were more readily diagnosed.
That could be because they receive far better care, Wennberg said. But more likely, the real reason is because in "high intensity" health-care regions, patients are more likely to be diagnosed with a chronic illness even if it's not particular serious or they're not actually ill.
The data show that as the number of diagnoses rise, so do the number of doctor visits, different doctors seen, imaging tests and lab tests done.
"Doctors make diagnoses," Wennberg said. "If you go to lots of doctors and doctors, they are going to make more diagnoses."
This complicates efforts to measure the quality and effectiveness of care, said Dr. Ashish Jha, an associate professor of health policy at the Harvard School of Public Health.
In an effort to improve health care, there's a growing movement to grade doctors and hospitals on performance, and to attach pay to how well doctors perform.
To make things fair, grading systems usually account for how sick patients are to avoid penalizing doctors who are taking care of a sicker group of people.
But research such as this study shows such methods may be flawed, because those who appear to have more chronic illnesses may not actually be any sicker than those with fewer chronic illnesses, and vice versa, he said.
Put simply: a patient diagnosed with heart disease and diabetes would be sicker than a patient with only one condition, but whether they are diagnosed with both may be more related to where they live and how many doctors they see.
Or, perhaps a patient gets an X-ray and is surprised to learn she has mild lung disease. If she were in another region were lower intensity healthcare, she might never get that X-ray and she'd feel fine and never learn she had lung disease.
But if those doctors in the high-intensity region "get credit" for having a sicker patient due to the chronic disease diagnosis, it may look like they're performing better when they're actually not.
"This is a really important study and a really important finding," Jha said. "If you are a doctor who is very good, but you don't order lots and lots of tests, it might look like your patients are not that sick and therefore your outcomes might look worse. Obviously, the last thing we want to do is penalize doctors for being cautious and prudent and not overspending on tests and imaging studies or being overly aggressive."
One potential solution is the increasing use of electronic health records, which will enable data from patients charts to be more easily and readily analyzed when doing risk adjustment, Jha said. Currently, with so many physicians still using paper records, risk adjustment is done using insurance claims data, which is incomplete and doesn't get into great detail about how sick patients actually are.
"We want to be able to differentiate between the one who is really sick with lung disease vs. someone who has something minor but is otherwise OK," Jha said. "You can't get it from claims or billing data, but you can get that from clinical records, and the broader use of electronic health records will help that."
For more on chronic illness care, go to the Agency for Healthcare Research.
SOURCES: Ashish Jha, M.D., MPH, associate professor, health policy, Harvard School of Public Health, Boston; John Wennberg, M.D, M.P.H., professor emeritus, Dartmouth College, Hanover, N.H.; March 16, 2011, Journal of the American Medical Association
All rights reserved