Algorithm guides ICP infection investigation

Mining for sentinel events in compromised patients

Infection control professionals in a group of cancer centers have developed an algorithm to help meet new patient safety goals by the Joint Commission on Accreditation of Healthcare Organizations. (See algorithm.)

The Joint Commission requires health care organizations to manage as sentinel events all identified cases of unanticipated death or major loss of function due to nosocomial infections.

Because patients at cancer centers generally are at high risk for serious infections, such institutions face unusual difficulty in determining whether a nosocomial infection is unanticipated, says Lisa Roman, RN, BSN, OCN, infection control coordinator at Fox Chase Cancer Center in Philadelphia.

"We needed to address the fact that we have immunocompromised patients with multiple comorbidities," she says. "That makes it more difficult to determine if an infection is the cause of death or major loss of function."

The algorithm, which can be used to identify or rule out sentinel events caused by nosocomial infections, incorporates such data as reason for admission, cancer type and stage, cancer effect on immunity, and anatomic barriers to infection. The influence of these factors on infection likelihood and severity dictate the branch points in the algorithm, she explained.

For example, if a patient died of pneumonia but had advanced lung cancer that blocked lung drainage, the patient’s death would be deemed "anticipated." Similarly, when a patient comes in under palliative or "comfort" care then "you know then that they are in a point of their disease state where death or major permanent loss of function may be impending," Roman adds.

When death or morbidity is unanticipated, the incident is reported as a sentinel event. The case is referred to the risk management and quality improvement departments for a root-cause analysis. An action plan is implemented and monitored in accordance with Joint Commission requirements.

Investigation of suspicious cases is prompted by ICPs conducting routine surveillance and reviewing death reports to identify nosocomial infections that result in death or major loss of function.

"If something strikes us from any of those areas as a potential to be an infection that caused harm, then we would put this algorithm into place," Roman says. "If the nosocomial infection looks like it played a major role, then you are going to go right down that algorithm to consider it a potential sentinel event, go to a root-cause analysis, and develop an action plan."

Much more often than not, of course, the cases will be explained by the existing patient conditions and not lead to a sentinel event investigation. However, the overall review process is a good quality improvement exercise in itself, she says.

"Protecting patients has always been our goal, anyway," Roman notes. "This is just finding new ways to try and achieve that. Any kind of ongoing effort to look at your process — that constant, continuous review — is really quality improvement. You can always do better no matter how good you are, and this is just one more opportunity to pursue that."