A new study suggests that for the past few decades, pressure ulcers in ICU patients may have been either over- or undertreated due to a widely-used predictive tool that may not be sensitive enough for critically ill patients.
Reviewing the electronic health records of 7,790 ICU patients, researchers at Ohio State University Wexner Medical Center found that the Braden scale, a thirty-year old tool that nurses use to assign patients an ulcer risk factor score, wasn’t accurate when it came to evaluating ICU patients.
“The scale told us that every single patient in the ICU was at high risk for a pressure ulcer. But we knew that not every single patient went on to get an ulcer,” said Brenda Vermillion, DNP, RN, a 28-year ICU veteran and clinical nurse specialist who participated in the study. “Going by the score means that most ICU patients would either be under - or over treated for ulcer prevention – and neither is optimal.”
While bedsores – deep tissue damage that develops when patients are unable to move - are largely preventable, an estimated 2.5 million are treated annually in acute care settings, and as many as 39 percent of all ICU patients will get them. In the United States, the annual expenditure for treating pressure ulcers is near $11 billion. Pressure ulcer prevention is a top priority in US hospitals because they impact quality of life, prolong hospital stays and can cause death.
The study validates what Vermillion, like many nurses in ICUs across the country, had suspected for years -- that the Braden scale ‘can’t tell the whole story’ for critically ill patients. The study also represents the culmination of years of cross-discipline clinical and biomedical informatics research that began with a simple linguistics problem.
In 2008, after a review of ICU patients’ electronic medical records showed the Braden scale wasn’t helping, Vermillion and a team of ICU nurses decided to find alternative tools. Vermillion, who was considered one of the unit’s clinical experts, started searching nursing and medical research literature for articles that might sync up with the data she was seeing in the patient charts. There wasn’t much information. When she finally found one that looked promising –it was written in Korean.
Around the same time, just across the street at Ohio State's College of Nursing, Sookyung Hyun, a nurse with a specialty in biomedical informatics, had just been awarded a pilot grant from the Ohio State University Center for Clinical and Translational Science to research the validity of the Braden scale in ICU patients. Back home in Korea, Hyun had worked on a paper which had suggested that the Braden wasn’t effective in an ICU population, and she was hoping to tap into the medical center’s huge EMR database to test her theories.
It was the same paper that Vermillion needed translated.
Both nurses ultimately turned to Ohio State Wexner Medical Center’s Chief Quality and Patient Safety Officer, Susan Moffatt-Bruce, MD, PhD, who was at the time working with units across the entire hospital on a major initiative to reduce the incidence of pressure ulcers. Moffatt-Bruce helped make the introductions, and Hyun’s research took off with the support of a multi-disciplinary team that decoded cryptic EMR acronyms, meshed massive amounts of patient data, and fundamentally, understood how ICU nurses used EMRs and the Braden scale.
Their study, recently published in the American Journal of Critical Care, found that a score of 16 on the Braden Scale, which is the current high risk indicator for ICU patients, would have better predictive validity and accuracy if it was moved closer to a score of 13. However, even with shifting the risk score lower, the researchers hypothesize that the scale still doesn’t sufficiently reflect the characteristics of ICU patients who may have a range of comorbid conditions and medications that make them more at risk for ulcers.
The team’s data pool also reflected the medical center’s ongoing pressure ulcer initiatives, which were able to reduce the incidence of ulcers across the system by 40 percent during the four-year study.
“This research is such a good example of how research can positively contribute to improving clinical care,” noted Moffatt-Bruce, also an associate professor of surgery with Ohio State’s College of Medicine. “While the study and our improvement efforts were running in parallel, it made us all more sensitive to the fact that ICU patients need more scrutiny, and that there are probably other clinical signs we need to find that could help us prevent ulcers in those patients.”
Vermillion says she and her team are doing just that.
“Before, we had experience telling us that the Braden didn’t fully meet our needs. Now we have the data confirming that. Armed with this information and by continuing to use the EMR, we’re beginning to identify factors unique to ICU patients, such as ventilator status, that can help us better predict which patients are really at risk.”
Vermillion acknowledges that the Braden is still very useful in other clinical settings and in nursing homes, and that ICU nurses will continue to use the scale as a starting point, “and from there, use their observational skills and experience” to reduce pressure ulcer risk in ICU patients.
The research team worked closely with the CCTS Biomedical Informatics core to help formulate the research questions, make decisions about which data sets to request and identify available data. They believe this the first large-scale retrospective study to use EMR to evaluate Braden scale efficacy in an ICU setting, and that it demonstrates a relatively quick and low-cost way to test multiple factors while identifying other criteria which could be validated through smaller prospective studies.
The researchers acknowledge that their study has several limitations, including an overall lower rate of pressure ulcers than similar studies, which could have been the result of coding errors in the EMR, or inconsistencies among nurses using the Branden scale. The researchers also did not examine if the admission Braden scores made a difference in the level of care patients received.
Source: Ohio State University Center for Clinical and Translational Science