A dose of predictive analytics could be part of the prescription to stop a deadly bacterium in its tracks, according to a staff member at a Burlington, Ont.-based hospital.
Joseph Brant Memorial Hospital recently experienced an outbreak of C. difficile, a common hospital bacterium that has become more deadly in recent years. The report has ramifications for hospitals province-wide.
The hospital released the results of a mortality review conducted by the Infection Control Unit of Toronto’s University Health Network on May 7. It revealed the outbreak was worse than previously thought.
From May 1, 2006 to Dec. 31, 2007, 91 patients diagnosed with the infection died while in hospital, with the bacteria contributing to 76 of those deaths. Ontario health minister George Smitherman reacted to the release by announcing that all Ontario hospitals will have to report cases of C. difficile to the province.
The lessons learned by Joseph Brant staff as a result of their outbreak are valuable to share with other hospitals that are now required to track the deadly bacteria. Staff in the decision support department has considered how the use of powerful predictive analytics software from Cary, N.C.-based SAS Institute Inc. could help frontline workers.
The software would complement existing tracking methods, not replace them, explains Mark Morreale, manager of decision support at the hospital.
The hospital recently adopted SAS software in a pilot program to measure other important indicators such as the Hospital Standardized Mortality Rate (HSMR), he says. The software combined data from across several different systems to allow for a close analytical look.
“What’s great about SAS is that we’re able to an in-depth analysis at the patient level,” Morreale says. “We’re able to calculate the amount of risk associated with a patient and report that to a nurse.”
Currently, the hospital tracks infection control information by using pencil-on-paper line sheets and Microsoft Excel spreadsheets, according to Barb Shea, the manager of infection prevention and control services. The number of patient days and incidents of infection are tracked and passed on to regional networks comparing the rates across multiple hospitals.
“Many of these statistical pool gathering software programs are being looked at also,” she adds. But the infection control team at Joseph Brant has not yet seriously considered such an option.
Collecting data about infections is laborious by nature, Shea says. A trained health care worker must go to the bedside and make a thorough assessment of the patient.
SAS software could help automate the tracking process, Morreale says. Nurses could link their lab reports directly to infection control information about C. difficile. This will boost “the ability to do rounds and their tracing ability as well.”
Joseph Brant staff has invested in improved infection control methods since the outbreak, according to their report. Improved reporting and treatment methods and better housekeeping has helped reduce the incidence of the infections, and no new cases of the bacteria have been reported since April 15.
Still, the report advises that hospitals should expect the harmful strain of bacteria will strike again and a plan is needed. C. difficile is an intestinal bacterium and victims are typically the old and frail who are inflicted with other sickness.
With SAS software being used by more than 50 hospitals across the province, the business intelligence software company hopes to play a role in the fight against infection. SAS software has the capacity to understand the root cause of a C. difficile outbreak, according to Pat Finerty, vice-president of alliance and business development for SAS Canada.
“The C. difficile issue is a bed-side care issue,” Finerty says. “We believe that understanding these causes and preventing these things before they occur is the way to maximize the resources in the system.”
SAS formed a health care advisory council in 2005, a collection of industry experts with the goal of swapping best practices between medical professionals. Now the company plans to reach out to hospitals and help them use their software to stop a C. difficile outbreak before it even starts.
“It’s wise if we can proactively prevent it from occurring,” Finerty says. “If hospitals have some best practices, and we can document them and predict how they’ll help, then we can share these with others in the system.”
Staff at Joseph Brant was a bit wary of the new software at first for tracking mortality rates, Morreale recalls. They felt a loss of control with the data being automatically crunched instead of their usual manual processing of the information.
But that anxiety was quickly cured when the software’s abilities became clear during the pilot project. He hopes for similar success when it comes to gauging the threat of an outbreak.
“A Web-based reporting tool with timely information about infections outbreaks would be very, very helpful for the staff,” Morreale says.
Another challenge in the battle against C. difficile will be data integration, SAS’s Finerty says. Most hospitals rely on a health information system that keeps information entered by healthcare workers separate from other information such as diagnostic tests. All of that data must be stored in one place to unlock the best predictive ability.
“It doesn’t matter what the scenario is, the data is in many different places in a hospital setting,” he says. “C. difficile is in this case a very deadly scenario.”
A host of proprietary systems and a lack of integration tools are to blame for data fragmentation in hospitals, Morreale says.
“Hospitals are data-rich and information-poor, and that is definitely due to the lack of integration of data,” he says. But over his 15 years of experience, SAS has been his go-to tool for pooling together data from seemingly disparate sources.
The software is able to look at all the factors available to remove the guesswork of what might contribute to an outbreak, Finerty explains. Everything from janitorial services to the demographics of the patient population to the treatments given to administrative functions could be a factor.
“You can’t manage what you can’t measure,” he adds.
Still, the use of SAS to help track or measure data about infection in a surgical setting is only in its infancy stages at Joseph Brant, staff makes clear. It is one option being worked on in the background among many being considered.
“We’re exploring opportunities,” says Julie DePaul, the director of health information and admitting services for the hospital. “It would only be used as a tool to augment what they’re doing now and to help validate what the staff has already identified.”
