System uses quality data to improve health care

Nursing homes can improve care and cut costs by using quality-of-care data, which they are required to collect. But that requires wading through a data swamp to find useful information.

Dr. Christie Teigland, a researcher at the New York Association of Homes and Services for the Aging in Albany, N.Y., is changing that. She has developed software that shows nursing homes which patients need special help to avoid falling or getting bed sores.

Using Teigland’s Equip for Quality system, one 300-bed New York state nursing home steadily reduced the number of falls among its patients, going from 93 in September 2002 to 53 in February 2003. Another New York nursing home using the system received a $30,000 reduction in its annual liability insurance premium.

Teigland is an econometric forecaster by training. She and her colleagues at the New York association began by building a model that crunches data to predicts with 83 percent accuracy which residents will fall during the next quarter. The systems uses the Minimum Data Set (MDS) to create the predictions.

The MDS is a standardized patient evaluation that nursing homes must complete on admission and at least every three months thereafter as a condition of participation in Medicare and Medicaid. The MDS assesses residents’ functional and cognitive status and other health conditions. It supports 34 quality-of-care measures and indicators determined by the Centers for Medicare and Medicaid Services. For example, the centers have standards not only for the percentage of residents who fall or have bed sores but also for the percentage of residents who are in pain or lose weight.

Although homes get their quality scores and can compare their ratings with others, administrators cannot easily discern patterns or determine why their ratings are improving or deteriorating, Teigland said. “We’re shifting the way the MDS data is used from retrospective reporting to a proactive focus,” she added.

Equip for Quality, which has a Web interface, connects the dots in the data and points to which residents are at high risk based on their history and risk factors. For example, Teigland said, incontinent patients and ones with behavior problems have a higher risk of falling. She said patients with behavior problems are more likely to fall because they annoy others, who are then more likely to shove them.

Equip for Equality has users in 400 nursing homes in 21 states. The system provides online charts that show which patients are likely to fall, and it describes preventive actions. After developing and validating the model for falls, Teigland and her team added models for bed sores and urinary tract infections. They are completing one that predicts who could sustain a bone fracture.

They are working on a model to predict weight loss, and “there will be more coming down the road,” Teigland says. Meanwhile, her team tweaks the models to improve their accuracy as more data becomes available.

Manual systems can analyze data, she said, but “most of the manual tools put way too many people at high risk.” It’s much more efficient and effective to focus on those residents who are at greatest risk, she said.

Development and testing of the original model was supported by a patient safety grant from the Agency for Healthcare Research and Quality, part of the Department of Health and Human Services. Now, nursing homes pay user fees that cover the costs of operating Equip for Quality, which has a staff of 15. “If it were a business, it would be doing really well,” Teigland said.

The success of Equip for Quality is a testament to the potential value of information technology for health care. And quality also cuts costs. Teigland said a nursing home with seven fewer cases of bed sores during a year will reduce its treatment costs by at least $25,200.