The most likely determining factors for healthcare spending and physicians' tendencies to order tests, treatments or referrals appear to be location, location and location, according to a report in the Oct. 24 issue of the Archives of Internal Medicine.
The report suggests that the geographic location--and not necessarily the health--of a patient is a key factor in predicting the level at which primary-care physicians recommend interventions.
The authors of the report analyzed survey data and compared regional spending to physician recommendations. For example, 32% of the physicians in high-spending regions said they would refer a hypothetical 60-year-old patient with symptoms of an enlarged prostate to a urologist while only 23% of the physicians in low-spending regions said they would do so. Another scenario involved a 35-year-old man with prolonged back pain and foot weakness. In high-spending regions, the patient was recommended for a magnetic resonance imaging scan 82% of the time. In low-spending regions, an MRI was recommended 69% of the time.
The analysis was performed by researchers with the VA Outcomes Group, a White River Junction, Vt., group trying to improve the presentation of medical information, and the Center for the Evaluative Clinical Sciences at Dartmouth, in Hanover, N.H. They compared how 5,490 family physicians and internists in different regions answered the Center for Studying Health System Change's 1998-99 Community Tracking Survey. The telephone survey included six vignettes describing different patient scenarios and, not surprisingly, it was found that physicians practicing in regions with high per capita Medicare spending recommended the most interventions.
"Physicians respond differently depending on where they practice, and the higher the spending in the area, the more likely they are to intervene," said the report's lead author, Brenda Sirovich, M.D., a staff physician with the VA Outcomes Group and an assistant professor of medicine at Dartmouth Medical School. "Healthcare spending does happen to reflect the decisions physicians make on individual interventions."
She said the survey's vignettes were designed so that they would have no obvious right or wrong answers. In fact, there was a general consensus on only one out of six vignettes. With the other five scenarios, it was calculated that the high-spending regions recommended interventions 10% more often than the low-spending regions.
Sirovich said that the study results were not surprising, but added that the strength of the study was that it isolated the role of physicians in explaining the wide variations seen in practice and spending across different regions. "There is a local culture about how medicine is practiced, but its origin--how it got to be that way--is less clear," Sirovich said, explaining that variations could be attributable to the local legal climate or to the number of specialists in the area who are available for referrals. "In areas of high spending, there tended to be a large supply of specialists, so that probably does influence how often physicians refer."
American Academy of Family Physicians President Larry Fields, an M.D. who practices in Ashland, Ky., said the results were interesting, but not surprising, noting that primary-care physicians are seeing patients with increasingly complex conditions, so ordering more tests is often prudent. Nevertheless, he added that in areas with intense legal climates there will be a tendency to order tests or refer patients to specialists even in cases where these actions may be seen as having a "low yield and high cost."
"It wouldn't surprise me if things haven't changed," Fields said. "But (another study) needs to be done with new data to see if the trends are different. You can't take 6- or 7-year-old data and draw conclusions about it today--no matter what you're measuring."
Sirovich said the 1998-99 survey was the last time the Center for Studying Health System Change used the vignettes in its tracking survey, so more current data isn't available. Still, she expressed confidence that newer data wouldn't yield significantly different results. Center spokeswoman Alwyn Cassil said, "It's unlikely that it has changed dramatically."
Cassil said that local supply may be a driver of this behavior. "Folks in Miami get more because they can," Cassil said. "If you build it, it will get used."