Skip to main content

Predictive Modeling May Reduce Hospitalizations

By HealthLeaders Media Staff  
   October 12, 2010

The Visiting Nurse Service of New York has adopted advanced information technology designed to improve patient care. The effort appears to be effective in reducing hospitalizations and re-hospitalizations, according to a paper published in the Journal for Healthcare Quality.

The VNSNY Center for Home Care Policy & Research launched the Outcomes Initiative to support research, evaluation, and informatics services. The system identifies patients at risk for hospitalization, identifies patients eligible for and in need of physical therapy, and assesses the performance of clinical staff and programs.

Together, these and other HIT initiatives have been instrumental in helping VNSNY achieve a 12% decrease in the overall patient rehospitalization rate between 2001 and 2009, according to the authors. They also report a reduction in patient episodes ending in hospitalization, from 37% to 27%, during that period.

Analysis of the VNSNY database revealed several factors that put home-care patients at increased risk of hospitalization: unhealed pressure or stasis ulcers, urinary incontinence, the presence of a catheter, respiratory symptoms, shortness of breath, and congestive heart failure. A computer model was devised to find the patients and place them in one of seven risk categories, ranging from very low to very high, so that appropriate clinical steps could be taken.

"[I]ntervening in the care of patients who have a moderate level of hospitalization risk was found to be an effective method of producing important and meaningful changes in the quality of care. There was also consensus among nurses and other clinicians on the usefulness of the intervention for improving medication reconciliation, communication with physicians, and awareness among clinicians about factors that predict hospitalization," the authors wrote.

The Quality Scorecard monitors goals and actual performance of VNSNY on such measures as staff oversight, scheduling of patients, discharge planning, treatment outcomes, and patient satisfaction. It updates monthly to show how well targets are being reached.

"Unless you have a feedback mechanism in place, you have no idea whether your ambitious programs are really working," Robert Rosati, PhD, vice president of clinical informatics, Center for Home Care Policy and Research, VNSNY, said in a prepared statement.

A predictive model identifies patients who should receive physical therapy based on a variety of baseline factors (e.g., demographic characteristics, functional limitations, symptom severity, diagnosis type, medications, comorbidities).

The approach has generated an average of 25 additional referrals to the rehabilitation department each week, according to the authors. They further point out that the efforts may reduce hospitalizations.

"Preliminary analyses from their investigation revealed that patients who received physical therapy were 37% less likely to be hospitalized compared with patients who did not receive physical therapy services," they write.  

Despite the reduction in hospitalizations, the authors stress that more research is needed. "While it is assumed that HIT initiatives such as the Quality Scorecard and Hospitalization Risk Score have contributed to better outcomes among VNSNY home healthcare patients, this needs to be explicitly tested in a controlled research study."

Tagged Under:


Get the latest on healthcare leadership in your inbox.