Hospital slashes pneumonia rate with quality project for early intervention

Project helps identify at-risk patients, methods to prevent infection

When nosocomial pneumonia rates started going through the roof at St. Luke’s Episcopal Hospital in Houston, the quality improvement leaders knew it was time to break out the big tools and find a solution. After soliciting ideas from a multidisciplinary team, the hospital achieved an astounding 50% reduction in its nosocomial pneumonia rates without any major expenditures or complicated changes in clinical care.

Though the project itself was intensive, the actual solutions to the nosocomial pneumonia problem turned out to be as simple as hand washing and suctioning. In addition, the hospital developed a tool for assessing which patients are at high risk so they can receive preventive care as early as possible. After five years, the quality improvement project has been a major success, says Rosemary Luquire, PhD, RN, senior vice president for patient care and chief quality officer.

Luquire worked with Susan Houston, PhD, RN, CNAA, FAAN, assistant vice president of clinical management and outcomes research to develop the quality improvement project. Nosocomial pneumonia rates began to increase significantly in 1994, when the rate was 4.7 per 1,000 patient days per year. In 1996, the rate had reached 6.5.

"In 1996, we saw that we would top off the year at a high rate, and though we do a lot of work with infections, we had the greatest opportunity to reduce nosocomial pneumonia because it was increasing at a faster rate than the others," Houston says. "We got together a multidisciplinary practice collaborative team with nurses, physicians, infection control practitioners, pharmacists, administrators, and a lot of others."

The first task was to create a fishbone diagram with the different causes of nosocomial pneumonia. With brainstorming and educated guesses, many potential causes were identified, from hand-washing practices to reuse of disposables, patient location, and the retaping and rotating of endotracheal tubes. Then the team sought verification that those causes actually led to nosocomial pneumonia infections, but they found that there was no literature to support many of those supposed causes.

"We found that many of the things we think cause pneumonia are just gut thinking, hypothetical, not actually supported by any data," Houston says. "Our literature review also revealed that most of the research has been done on patients with emphysema, COPD, and asthma. But most of our cases are in cardiovascular surgery patients."

So to make the information more applicable to the patients at St. Luke’s, the team went back to the fishbone diagram. This time, team members conducted a case control study of 240 medical records and plotted the causes of nosocomial pneumonia on the fishbone. With a univariate statistical analysis, the team identified a number of factors associated with the infections, but then a multivariate analysis revealed that only four particular factors were most strongly associated with the patients who developed infections.

Those factors were the use of aortic balloon pumps, renal failure, reintubation, and the total intubation time. The analysis revealed that those four factors were strongly associated with infections, so the team hoped they could be used to identify patients at risk and also develop a protocol to address those issues.

"These are all factors that we know preoperatively or intraoperatively, so we can use them to predict and prevent instead of just waiting to see who would get an infection," Houston says.

Using those factors, the team developed a scoring tool to identify high-risk patients. The cardiovascular recovery room nurses scored the patients every 24 hours, and if the patient met a certain cutoff score, he or she was put on the nosocomial pneumonia prevention protocol.

These are the steps in the protocol:

  1. Obtain sputum sample.
  2. Order "pneumonia protocol" sputum culture on routine culture order screen.
  3. Order chest X-ray if not done within last 24 hours.
  4. Order CBC with machine differential if not done within last 24 hours.
  5. Ask respiratory care to institute use of in-line suction catheter.
  6. Repeat orders 1, 2, 3, and 4 every 48 hours until extubated or trached (whichever comes first).

Notify physician if:

  • Sputum gram stain is positive (greater than or equal to +3 WBC, 2+ GNR, or 2+GPC) or
  • One of the following organisms is identified in culture: P. aeruginosa, K. pneumoniae, enterobacter species, S. marcesoens, S. aureus.

Physicians should consider empiric therapy as follows:

  • If gram stain is positive for gram negative rods (GNR), consider ceftazidime 2g q 8h.
  • If gram stain is positive for gram positive cocci (GPC), consider vancomycin 1g q 12h (doses altered for renal function) or clindamycin 800 mg q 12 h, if reintubated and/or possibility of aspiration.
  • Reassess antibiotics at 72h or when culture and sensitivity results available.

If pulmonary disease and/or infectious disease physician is consulting, defer antibiotic decision to that service.

Protocol causes rates to drop immediately

"We implemented the nosocomial pneumonia protocol, and our rates dropped from 6.5 to 4.6 over a year. That’s huge," Houston says. "We knew the protocol was working, so we looked at some other factors, too."

The quality improvement team studied the hand-washing and suctioning practices at the hospital and found ample room for improvement. The team updated the policies and procedures for both, and then sent observers out periodically to monitor how well staff follow them. The hospital still conducts in-person monitoring every so often to keep people aware of the need for good hand-washing and suctioning techniques, and there is some discussion about implementing video monitoring.

"People always do it better when they know someone is watching," Houston says. "Then it drops off slowly as people become complacent, so we come back and stand there again, looking over their shoulders as they wash their hands. It raises the awareness level again."

The quality team also conducted a study comparing the use of a 0.12% chlorhexidine mouthwash (Peridex), comparing it to Listerine, in hopes that one or both would lower the infection rate. The study found lower colony counts in the respiratory tract with the chlorhexidine, but there was no associated reduction in pneumonia rates. Houston suspects a larger study might show a beneficial effect.

The quality initiatives have been in place for about five years now, and Houston and Luquire say the project is a major success. The nosocomial pneumonia infection rate declined from 6.5 per 1,000 patient days in 1996 to 2.8 in 2001, putting the hospital in about the 15th percentile of the infection rates collected by the Centers for Disease Control and Infection in Atlanta.

Those good results come with very little investment. Houston says the hospital spent roughly $20,000 on the project itself, and the pneumonia prevention protocol costs about $30 per patient.

A single nosocomial pneumonia infection costs the hospital about $8,000, so Houston says the project’s costs were recovered once it prevented only a few infections. With the lowered infection rates, she estimates the hospital avoids about 100 pneumonia infections per year.

Success leads to awards, opportunities

St. Luke’s recently was honored with a Premier Award for Quality for the pneumonia project, awarded by Premier Inc., a national alliance of 1,600 not-for-profit hospitals. Houston’s advice for other hospitals interested in emulating the project’s success is to "go into the project for the long haul."

She says the quality improvement team still meets regularly to assess pneumonia rates and look for ways to improve.

The success of the project was one factor that led St. Luke’s to joined 33 other hospitals that have achieved "Magnet" status from the American Nurses Credentialing Center in Washington, DC. The Magnet Nursing Services Recognition Program for Excellence in Nursing Services was developed by the American Nurses Credentialing Center in 1994 to recognize health care organizations that provide the best in nursing care and uphold the tradition within nursing that supports professional nursing practice.

Luquire says the hospital’s achievement is largely due to the fact that nurses are involved with making decisions about nursing and patient-related issues.

"At St. Luke’s, our nurses are empowered and encouraged to provide their input into decision making about issues related to their professional practice," Luquire says. "I’m convinced that our program of shared leadership — which is not just a concept, but a philosophy and organizational structure — had much to do with St. Luke’s earning Magnet status."

Luquire says that same approach was key to the success of the nosocomial pneumonia project because nurses and others close to the patient were encouraged to recommend improvements in care.

"We view shared leadership as first and foremost a philosophy that the caregiver closest to the patient, if given appropriate resources, can tell us what works best at the lowest cost with the best outcomes," she says.

"So we actively promote a structure in which professional nurses help make decisions and establish research-based policies that affect their clinical practice," Luquire says.