Translating numbers into useful information

Two hospitals use data to improve performance

Your rehab department has tracked outcomes data for years. In fact, you have two drawers full of quarterly reports to prove it. But have you done anything with the numbers after you’ve read the reports and filed them away?

Hospital Case Management spoke with two hospital administrators whose facilities are using outcomes data as performance measurement tools to help staff and department managers compare their performances with those of their peers and identify best practices and opportunities for improvement.

At Shepherd Spinal Center in Atlanta, the senior management team issues a monthly report that compares outcomes by teams based on 30 variables, says Gary Ulicny, PhD, the hospital’s president and CEO.

A performance improvement team, which includes a physician, developed the list. The physician-led teams are categorized by type of patient group, such as paraplegics or acquired brain injuries.

Shepherd created its own executive information system about two years ago that tracks the data, using the Functional Independence Mea sures as a starting point but adding variables looking at functional activities after discharge from the hospital. For example, what kind of ongoing medical care was needed? How much assistance did the patient need from others to perform daily living activities such as getting dressed?

Data are distributed monthly to physicians, although teams are identified by Physician A, Physician B, and other acronyms. "But everybody knows who’s who," Ulicny says. "They became pretty competitive. They see it as an opportunity to learn from colleagues."

Shepherd does not tie performance to compensation, he says, but uses the data in the strategic planning process. "What you see in so many hospitals is they get memos saying they need to submit three performance improvement ideas. Many people will grab on to easy or convenient things. We let the data in our report cards be the guide."

The next step for Shepherd is to dissect a couple of programs to determine what contributes to an outcome and what doesn’t, Ulicny says. One specific component includes a time-study analysis of all nurses in each of the programs. The analysis will determine how each nurse is using his or her time, whether there is work that technicians or assistants can do to free nurses for other duties, and whether other areas for improvement exist.

The hospital has made several improvements based on the outcomes data reported, Ulicny adds. "A couple of years ago, we found that discharge planning had gotten rushed, and [we] were able to retool that significantly," he says.

At National Rehab Hospital in Washington, DC, benchmark data are presented quarterly to administrators and program managers, says Jackie Ennis, director of outcomes management. Reports are reviewed by a leadership quality council — which includes the hospital’s president, administrator, vice president of nursing, head of clinical services, and head of outpatient services — and by medical directors and department managers.

Data are "cut" by impairment group, Ennis says. Measurements are grouped by these patient types: stroke recovery, orthopedic impairments and disabilities, spinal cord injury, and traumatic brain injury. National Rehab also tracks results within the MEDSTAR health system, the integrated delivery system it belongs to, she says.

"We try to determine how outcomes are affected by participating in other points of the system of care . . . not just by cost but by functional gains," Ennis says. For example, data tracked for a stroke patient might look not only at acute care rehab costs, but also outpatient costs and home health costs, she explains.

Both Ennis and Ulicny say their hospitals have shared outcomes data with payers and have sorted data by payer membership. Payers are particularly interested if you can isolate outcomes to their membership, they say. Their goal is to assign meaning to the numbers.

"Those of us in the field have the tendency to produce lots of numbers, drop the data, and run," Ennis says. "I think it’s really important to add value to the outcomes data by focusing on meaning and implication. What does the data mean in context of daily operations or in a strategic context, even if you’re simply posing the question with data?"

For example, in one presentation Ennis made to medical residents on staff, she presented data demonstrating that elderly spinal cord patients had shorter lengths of stay than non-elderly spinal cord patients.

"On the surface, this seemed counterintuitive," she says. But during the presentation, medical residents in the audience suggested that many of the non-elderly spinal cord patients were involved in accidents resulting from urban violence, which likely affects the severity of the injury and thus recovery time.

"We went away from the meeting with four requests to look at the data for spinal cord patients in a different way," she says.