Joint Commission: ICPs should report infection deaths as sentinel events if there is no other apparent cause

If the patients didn’t have the infection, would they have died?’

The Joint Commission on Accreditation of Healthcare Organizations has clarified its controversial request for data on fatal infections, emphasizing that such outcomes should be reported as sentinel events only if the infection was clearly the cause of death.

"People have been confused about this," says Paul Schyve, MD, Joint Commission senior vice president. "The first thing is, we’re not talking about all nosocomial infections. The second thing is, we are talking about voluntary reporting."

Indeed, the unprecedented request for data — first issued in an open letter by Dennis O’Leary, MD, Joint Commission president — raised immediate concerns in the infection control community. ICPs pointed out that it is difficult to ascribe a nosocomial infection as the sole reason for death in hospital patients with multiple underlying diagnoses across the broad spectrum of disease.

To clarify the issue, the Joint Commission recently issued a Sentinel Event Alert that states: "Manage as sentinel events all identified cases of death and major permanent loss of function attributed to a nosocomial infection (i.e., except for the infection, the patient would probably not have died or suffered loss of function)."1

The Joint Commission cited the Centers for Disease Control and Prevention’s (CDC) definition of a nosocomial infection as: "a localized or systemic condition: 1) that results from adverse reaction to the presence of an infectious agent(s) or its toxin(s), and 2) that was not present or incubating at the time of admission to the hospital."

Infections that lead to patient deaths or permanent loss of function should be regarded as sentinel events and subjected to a "root-cause analysis" with the idea of preventing such infections in the future.

"We are urging these sentinel events be reported to the database, so we can learn from multiple organizations what appear to be the underlying root causes," Schyve says. "What we suggested is that they look at it from the point of view of: If the patients didn’t have the infection, would they have died? There are still questions around that, and some organizations may define the parameters somewhat differently."

The issue of fatal infections as sentinel events was among several topics discussed at a Feb. 4, 2003, meeting of a special Joint Commission task force on infection control.

"Most people thought it was not epidemiologically sound, and it was not, in general, likely to result in a lot of improvement," says Bryan Simmons, MD, a member of the task force and health care epidemiologist at Methodist Health Systems in Memphis, TN. "Particularly, if you found yourself just looking at those [cases] and trying to determine whether the death of a patient who had an infection, which occurs all the time, was actually due to the infection," he explains.

Such an emphasis could result in a return to "total-house" surveillance (e.g., recording and analyzing every infection), an approach that has been largely abandoned in favor of targeted approaches such as surveillance by objective (e.g., reduce bloodstream infections).

"My impression was that the Joint Commission clarified that it was for events where there was no other explanation for the death," says Elaine Larson, RN, PhD, a member of the task force and professor of pharmaceutical and therapeutic re-search at Columbia University School of Nursing in New York City. "A lot of people die with infections, but they pointed out that this would only be if the infection were the primary cause of death. So in that regard, I think it is a sentinel event. I think if there are other explanations, then it is not considered a sentinel event," she says.

According to the CDC, some 2 million patients are infected annually in U.S. hospitals, and about 90,000 of these patients die as a result. Yet the Joint Commission database includes only 10 infection-related reports that have been reviewed under the sentinel event policy since its implementation in 1996. The reports include a total of 53 patients, 14 of whom died. While the age of the patients varied, the vast majority were infants (29) and seniors (19), and many of them were immunosuppressed, the Joint Commission reported. Settings included the newborn and pediatric intensive care units, long-term care facilities or units, general medical/surgical units, and endoscopy and obstetrics units. The infecting organisms included HIV, Pseudomonas aeruginosa, E. coli, methicillin-resistant Staphylococcus aureus, salmonella, and Clostridium sordellii.

"The number of reported infection-related sentinel event cases represents an insufficient sample from which to draw any generalizable conclusions and recommendations," the Joint Commission concluded, emphasizing that more reports are needed.

However, Clinicians have raised concerns about the confidentiality of the data, particularly due to possible legal action. "We specifically do not want to receive data that identify either patients or practitioners," Schyve says. "All we know is what organization it came from. When we place it in the larger, sentinel event database, we do not retain the name of the organization. The whole purpose of the database is to see whether there are patterns in the root causes that run across organizations."

Likewise, a report of a sentinel event will not necessarily prompt an inspection by the Joint Commission, he adds. "What we do expect — when an organization has voluntarily reported to us — is that they are doing a root-cause analysis and taking some action," he says. "Not because we are going to cite them in some way because of the event. The real question is: Are they responding to the event in a way that will help to improve safety in the future?"

As an example of the process, Schyve cites a recently published article by Julie Gerberding, MD, MPH, CDC director. All institutions could benefit from a careful internal review of how hospital infections are managed, she said.2 "Unless all institutions continuously strive to reduce their infection rates, there may be little motivation for better-than-average performers to improve, a situation that could lead to a culture of complacency,’" the CDC director warned. "A mindset that the current error rate can always be decreased almost certainly explains the fact that other complex industries routinely achieve error rates far below equivalent benchmarks in health care."

In addition to the issues of reporting infections, the Joint Commission task force discussed several other pressing issues at its recent meeting. "Issues that were considered included staffing and personnel; the use of evidence-based practices; the care of the environment, equipment, and supplies; and how [ICPs] collect and analyze data," Schyve says.

Also discussed was the Joint Commission’s continuing emphasis on compliance with the CDC’s new hand hygiene guidelines. Will the new guidelines, which call for the use of alcohol-based rubs, be officially added to the accreditation standards?

"We have recommended to people that they use those guidelines, and one of the issues that will be looked at by the infection control task force is whether those guidelines should be reflected in changes in the standards," Schyve says.

There is some speculation that the hand hygiene guidelines may surface as a patient safety goal for 2004, which means surveyors would be looking for compliance. In addition, upcoming standards changes may emphasize the need for professional staffing in infection control and a clear organizational presence for the program.

"There was pretty strong acknowledgement that one of the most important things that hospitals needed was well-educated, trained staff to do surveillance [and prevention]," Larson says. "I think they heard the message from people attending the task force that there seems to be a need to have — if not an infection control committee — then some centered [program] focus. Several people said when the Joint Commission took out the recommendation for a committee, they stopped their committee and then they had to restart it because things just got too diffused."

Though recent research supports the need for more infection control staffing than traditionally allotted (one ICP per 250 licensed beds), the Joint Commission is not expected to establish any specific formula. "Staffing was discussed, but I’m sure the standards are not going to talk about specific numbers and ratios," Larson says. "It’s going to be a more general [requirement] that the staffing be appropriate for the facility."

New standards may be issued by year’s end

Overall, the situation remains unresolved until the Joint Commission considers the insights provided by the task force and completes its revision of standards.

"Our expectation and hope is that we will have the actual standards — to be able to [distribute to] the field — later this year," Schyve says. "The second question is: When will the surveyors begin to survey for them, and then when will those results start to effect the accreditation of those institutions?"

The Joint Commission decided to form the task force after ICPs protested a proposal to consolidate and reduce the number of infection control standards in 2004, when the commission plans to implement its ambitious Shared Vision/New Pathways accreditation program. (See Hospital Infection Control, December 2002, under archives at But what began as a flap over standards has opened up into a broader reassessment of infection control and accreditation.

A series of critical, highly publicized articles by the Chicago Tribune last year added additional pressure on the Joint Commission to show it views infection control as a priority.

"They brought us together out of concern about whether or not their standards had enough punch in terms of getting the job done in infection control," Simmons says. "The main [questions] were: Are our standards adequate, and how should we strengthen our standards to make sure that infection control has all of the support that is needed?"

Though the task force did not express great concern that the Joint Commission was lax in the area of infection control, the panel decided to look at the situation in terms of "Can we do better?" Simmons says. "The group did think that there were some opportunities to strengthen the [ICP] position," he says. "This was an auspicious beginning. [The Joint Commission] seemed very concerned. We felt we were starting off right and that it was worth taking the time to do it because they were taking it seriously."


1. Joint Commission on Accreditation of Healthcare Organizations. Infection control-related sentinel events. Sentinel Event Alert Jan. 22, 2003; Issue 28.

2. Gerberding JL. Hospital-onset infections: A patient safety issue. Ann Intern Med 2002; 137:665-670.