Keystone ICU project drives down mortality rates

The so-called Keystone ICU project was organized by the Michigan Health & Hospital Association's (MHA) Keystone Center for Patient Safety & Quality, which worked with patient safety experts from the Johns Hopkins University Quality & Safety Research Group.

According to MHA's estimates, the 127 ICUs where the program was put into effect saw 1,578 fewer deaths than would have otherwise been expected over the 15-month span from March 2004 to June 2005. The interventions also resulted in 80,000 fewer hospitals days, with a healthcare delivery cost savings of more than $165 million.

One key to the drop in spending and mortality was infection control; central intravenous line infections were halved over the course of the intervention, lowering Michigan's infection rate to among the nation's best. More than half of the participating ICUs have gone six months or more without bloodstream infections or ventilator-related pneumonias.

ICU care is among the most expensive health care in the United States, and ICU patients are responsible for about 30% of acute care cost-or $180 billion annually. Because patients in the ICU are at high risk for adverse events, the units have become a target for quality improvement efforts.

A growing body of research suggests several interventions can cut morbidity and mortality in ICU patients, though those interventions are not widely or systematically implemented. The Keystone effort sought to use that evidence base to fashion a program that could ensure best practices in the ICU. The project focused on six interventions:

As part of the process, all participating ICUs assembled a team that committed to the collection of data, as well as regular meetings and conference calls to compare results. The teams consisted of an executive, an ICU director, an ICU nurse manager, an ICU nurse, an ICU physician, a department administrator, and a pharmacist.

The pharmacists worked with the rest of the team on all projects, but efforts focused on better glycemic control of ICU patients were among the main areas of responsibility for the pharmacists, according to Michael Peters, R.Ph., a clinical ICU specialist at the Henry Ford Medical System's main campus.

Patients with normal blood sugar levels are significantly less likely to die in the ICU than are patients with poor glycemic control. Peters said that his facility boosted the number of patients with normal sugar number from 25% to 30% up to around 75%. "It's made a huge impact in these patients' outcomes," he said.

The presence of the pharmacist on the floor played a significant role in ensuring that the interventions were followed, Peters said. "If nurses are having difficulty with the protocol, they seek the pharmacist's advice. It's all hands-on. I'm part of a team that, three days a week, looks at every one of these initiatives."