Very spread out places, such as Atlanta, had about 100 more health problems per 1,000 people than areas that were less so, such as the Greensboro-Winston Salem area of North Carolina. Washington had about 50 more health problems per 1,000 people than Baltimore, Sturm said.
Although some researchers have speculated that the social isolation that can occur in sprawling communities may also lead to more mental health problems, such as depression, the new study failed to find that link.
"We find a strong association on the physical health side, but surprisingly not on the mental health side," Sturm said in a telephone interview.
Other researchers praised the study, saying it adds to evidence that suggests the physical attributes of where a person lives can have a significant impact on their health.
"It's the first study to aggressively assess systematic relations between health outcomes and the built environment," said Lawrence D. Frank of the University of British Columbia, who studies sprawl.
"This is still a very new field of research, but every significant study that has come out so far has reached a similar conclusion," said Don Chen, executive director of Smart Growth America, a Washington-based advocacy group. "This may be a promising way to begin addressing some of these chronic health issues."
But critics dismissed the findings, saying the study was flawed and the link between sprawl and health was tenuous at best.
"I remain a skeptic of the research, in part because the results they find are weak," said Samuel R. Staley, a senior fellow at the Reason Foundation, a Los Angeles-based libertarian group. "This study seems particularly prone to spurious results -- results that are statistically related but really don't tell us much about causes."
Peter Gordon, a professor in the school of policy, planning and development at the University of Southern California in Los Angeles agreed, calling the study "junk science." The areas studied, for example, are so large they could not distinguish important neighborhood differences, he said.
"Describing places this large via a simple ad hoc 'sprawl' index is nuts," Gordon wrote in an e-mail. "People have been suburbanizing for a very long time. Yet, life expectancy keeps getting longer."