ACOs and the earlier pilot program, known as the Medicare Physician Group Practice Demonstration, emphasize shared savings for groups of health workers and award bonus payments based on their quality of care.

That model, called pay-for-performance, is one cost-saving strategy of President Barack Obama's health care reform law.

"A lot of people are talking about Accountable Care Organizations as a way to try to improve quality of care while reducing cost," said Dr. Craig Pollack from Johns Hopkins Medicine in Baltimore, who has studied ACOs and cost sharing but wasn't involved in the new study.

"There's increased effort (to understand) what makes a successful hospital system and what doesn't and how to optimize systems of care," Pollack told Reuters Health.

Lead author Carrie Colla said the new analysis of the group practice pilot program can give organizations a better idea of what kind of cost savings to expect, and on which patients, when ACOs are rolled out.

For example, the late-2000s program led to the biggest savings on treatment of people who qualified for both Medicare and Medicaid. Known as "dual beneficiaries," those patients have a disproportionate number of chronic health conditions and may suffer as a result of uncoordinated treatment.

"There was a lot of room for improvement in caring for this group," said Colla, from The Dartmouth Institute for Health Policy and Clinical Practice in Lebanon, New Hampshire.

"Some of the promise of these types of programs is in managing the most difficult patients."

Colla and her colleagues compared health care costs over time for Medicare patients seen at one of 10 systems with physician incentive groups and those treated at other facilities without the pilot program.

They found physician groups saved an average of $114 per patient each year.

Savings were more than $500 per year for patients covered by both Medicare and Medicaid. That was equal to a five-percent decrease in their cost of care - mostly due to fewer hospitalizations for heart attacks, strokes and other emergency conditions.

Pollack called that finding "exciting."

"The idea is to try to improve care for people who are the most vulnerable and the most expensive," he said. "That looks promising, based on this study."

The physician incentive programs weren't cost-effective across the board, however. Outcomes ranged from annual savings of $866 per patient at one organization to extra costs of $749 per patient at another.

Four of the 10 sites saved money overall, the researchers reported Tuesday in the Journal of the American Medical Association. Three had no change in their expenses and three spent more with the physician group program in place.

Colla told Reuters Health the variation in cost savings - or spending - by health system could be the result of different systems putting varying amounts of resources and effort into what was a temporary program, as well as the size and structure of the organizations.

Those and other organizations may see ACOs as more permanent, she added - which could lead to more significant reforms and thus more cost savings and quality improvement.

"The magnitude of savings that is possible is an important lesson," she said.

SOURCE: bit.ly/JjFzqx Journal of the American Medical Association, online September 11, 2012.

• A Permanent Economic Emergency