Once upon a time it was believed that greedy insurance companies were bad for patients - now they are a sign of the have's and have nots. A new analysis in Neurosurgery finds that brain tumor patients with private insurance have fewer medical complications and thus are in the hospital less than those who are on Medicaid or were uninsured between 2002 and 2011.

Uninsured, who would all be under Medicaid today, and Medicaid patients were also at greater risk of developing a new medical condition in the hospital and 25 percent more likely to die during their stay. Those same patients ended up in a nursing home, rehabilitation center or hospice more frequently than people who had private insurance, according to nationwide data from 566,346 hospital admissions involving brain tumor cases.

People who are uninsured or use Medicaid also are less likely to benefit from early detection of brain tumors because they have less access to health care than those with private insurance, said Dr. Kristopher G. Hooten, a resident in the UF College of Medicine's department of neurosurgery and the study's lead author.

When brain tumor patients are hospitalized, much has already happened that affects their medical prognosis, Hooten said.

"When private-insurance patients start to have a problem, it gets picked up really fast. They go to a primary doctor, who makes a quick referral to a neurologist or neurosurgeon," he said.

People who use Medicaid don't always have that benefit, sometimes waiting and then going to an emergency room when their symptoms are more severe. That ultimately affects a patient's outcome.

"It's both an access-to-care and a quality-of-care issue before patients are admitted. (Uninsured or Medicaid patients) come in when their brain tumors are more advanced," Hooten said.

Once hospitalized, patients with private insurance and those on Medicaid also fared differently, the study found. The Medicaid patients were more prone to certain kinds of infections, postoperative respiratory issues and problems with blood sugar control.

They also were more at risk for so-called "hospital-acquired conditions," including pressure ulcers and vascular catheter infections. Medicaid and uninsured patients were almost twice as likely to have blood sugar problems compared with those with private insurance.

That isn't because hospitals treated individual patients differently based on their insurance. Instead, Hooten said, the Medicaid patients are more likely to have a broader set of medical problems.

All of those factors contribute to longer hospital stays, a higher death rate and a greater likelihood that Medicaid recipients and the uninsured will end up in a nursing home, rehabilitation center or hospice rather than going home, the study found. After adjusting for all of the hospital and patient factors, the outcomes and occurrence in hospital-acquired conditions were similar. However, Medicaid and uninsured patients still had an unexplained longer length of stay, which may be a direct result of their insurance status.

Researchers hope their findings will be used in several ways. Those include helping patients by identifying conditions like obesity and poor nutrition that put them at greater risk for other medical problems.

The findings also show how a federal agency's current method of comparing hospitals' quality could benefit from more precise information about patient populations. Hospitals that care for higher-risk Medicaid and uninsured patients should be judged differently than those that have more privately insured patients, said Dr. Maryam Rahman, an assistant professor in the UF department of neurosurgery and the senior author of the study. Some hospitals might publicize a negligible complication rate for a certain condition, but that alone doesn't tell the whole story.

"It's not due to the fact that they're amazing deliverers of health care, it's just that they take care of a low-risk population," she said.

Data used by the UF Health researchers came from the National Inpatient Sample, the largest health care database of its kind in the United States. The sampling did not identify patients or specify the hospitals where they were treated, Rahman said.

Hooten and Rahman hope that the findings will be used to affect health care policy and improve disparities in medical care.

"This type of research is important from a global standpoint to understand what goes into quality assessment, how hospitals are ranked based on quality and which patients are potentially high-risk. The true benefit is identifying areas of improvement and making things better for patients," Rahman said.