WASHINGTON (AP) – A report that expensive illnesses lead to nearly half of all personal bankruptcies is being challenged by researchers who looked at the same data and concluded that such costs lead to fewer than one in five bankruptcies.

The study looked at 1,771 bankruptcy filers in California, Illinois, Pennsylvania, Tennessee and Texas. It said illnesses and medical bills were cited as the cause, at least in part, of 46 percent of the bankruptcies.

That original study, by Dr. David Himmelstein of Harvard Medical School and colleagues, was published on the Internet by the journal Health Affairs last year. The challenge, by David Dranove and Michael L. Millenson of the Kellogg School of Management at Northwestern University, appears in Tuesday’s Internet edition of that journal.

“It is insufficient to show that medical problems are associated with bankruptcy; one must determine whether, and to what extent, medical spending causes bankruptcies,” Dranove and Millenson wrote.

A reanalysis, funded by a health insurance trade group, identified people who said illness or injury was the cause of bankruptcy and also said medical bills contributed to it. That led Dranove and Millenson to conclude that “medical expenditure bankruptcies” constituted 17 percent of all bankruptcies.

For that 17 percent, the researchers could not determine whether medical costs were the most important cause of the bankruptcy.

Himmelstein and colleagues responded that Dranove and Millenson “manipulate the data far beyond legitimate reinterpretation.”

For example, they said, the new interpretation excludes people who took out second mortgages to pay medical bills and then said their bankruptcy filing was “to save our home.”

The Himmelstein paper suggested a Canadian-style national health insurance plan as a way to prevent medical bankruptcies. Dranove and Millenson contended that since they found medical costs were only a part of the bankruptcy problem, any insurance system broad enough to give financial protection would be so costly it might undermine other safety net programs.

Himmelstein’s research was funded by the Robert Wood Johnson Foundation, Ford Foundation, Harvard Law School and New York University Law School. The Dranove and Millenson research was funded by the health insurance trade group America’s Health Insurance Plans.