Tool may limit hospital admits from nursing homes

Tool may limit hospital admits from nursing homes

Helps predict risk of death from LRI

A new tool designed by researchers at the University of Missouri-Columbia School of Medicine and Washington University School of Medicine in St. Louis helps identify nursing home residents at relatively low risk for death from lower respiratory infection (LRI), which means some patients may be treated safely without transferring them to a hospital.

From a quality perspective, why might it be preferable not to transfer these patients to the hospital setting? "I think that hospitals are bad places for frail, old nursing home patients," asserts David R. Mehr, MD, MS, associate professor of family and community medicine at the University of Missouri-Columbia School of Medicine, and lead researcher.

The risk of complications is significantly greater in a hospital, Mehr asserts. "If you think about it, you take somebody who may be cognitively marginal, put them in a new environment with new people, subject them to a series of procedures, and they’re bound to be confused." (Eighty percent or more of nursing home patients have had some cognitive impairment, he notes.) "Most of them get IVs, which are at risk of being pulled out. You can get into a cascade of complications that lead to immobility, pressure sores and urinary tract infections, and the patient can end up coming back to the nursing home in worse shape than when he left. We often need the technology of the hospital, but if we have the technology to avoid hospitalization, you have fewer complications and people are happier. The fact is, they don’t like to go to the hospital."

Mehr’s earlier years of practice convinced him that such a tool was necessary. "I recognized that one of the key issues when I was seeing nursing home patients was what to do with pneumonia — what to treat it with and whether to hospitalize the patient or not," he recalls. "While doing fellowship and research work, it jumped out at me; it became clear that we couldn’t lump together people with a 2% chance of dying and a 60% chance of dying in the same study if we didn’t have some sort of basis for identifying that risk of dying."

The tool was derived from earlier work funded by the Agency for Healthcare Research and Quality, conducted by the Patient Outcomes Research Team (PORT) on community-acquired pneumonia. The PORT developed and validated the Pneumonia Severity Index (PSI), which is used to identify pneumonia patients living in a community who can be treated safely at home. However, because the PSI assigns higher risk based on age (1 point for every 10 years) and other variables common to elderly people, it predisposes most nursing home residents with respiratory conditions to hospitalization, whether or not their condition actually warrants it.

To make the new tool more sensitive to residents of nursing facilities, the researchers gave more weight to variables such as activities of daily living, mood decline, and markers of poor nutritional status. By assigning point values to each of eight variables, the researchers were able to develop a scale. A score of 0-4 was a low risk score, while a score of 5-6 indicated a relatively low 30-day mortality risk. (See table, below.)

Armed with such a predictive tool, to whom does the decision ultimately fall as to whether to transfer the patient to a hospital? "Nobody really knows the answer to that question," Mehr concedes. "In theory, you need a physician decision to admit, but in reality, the nursing home may make the determination and the doctor may not have much to say about it. The family can also say yes or no. You can have nursing homes that are concerned about how regulatory agencies look at them, and they may want to avoid deaths [bad outcomes]. Certainly, nursing homes and physicians might make use of this tool, although it’s probably not practical for it to be used by family members, who may not understand many of the terms we use."

The data the researchers have collected may fuel many more interesting studies, Mehr says. "We did not find a relationship between mortality and quality measures, but we see a relationship between hospitalization and some quality measures. We have collected a bunch of quality measure data, and we’re working on looking at how they relate to hospitalization."