Researchers sought to find synonymous terms in patient records generated by physicians and nurses. Instead, they came across a glaring gap.
A new algorithmic analysis of electronic health records (EHRs) for the same patients spotlights the relatively little overlap between care provided by physicians and nurses, lending insights into how current quality measures might be missing key information.
Researchers at the University of Illinois, Chicago (UIC), developed an algorithm to analyze the EHRs of 58 patients who were diagnosed with heart failure and treated at a single medical center during an 8-year time span. The algorithm compared each patient’s physician discharge summary with the same patient’s nursing plan of care. The care plans were constructed based on data found in the HANDS nursing documentation software. The analysis pulled out physician-specific terminology and nurse-specific terminology to gauge the extent to which the two types of medical professionals used synonymous terms.
The researchers found synonyms in the physician and nurse documentation in only 26% of cases, indicating that physicians and nurses were focusing on different aspects of patient care. The nursing plans contained an average of 18 medical terms, and physician discharge summaries contained 27 terms. On average, only 4 terms between the 2 professions were related to the same medical concept.
Study co-author Andrew D. Boyd, MD, an assistant professor of biomedical and health information sciences at UIC, told Healthcare Analytics News™ his research isn’t the first to focus on inter-professional communication, but it’s the first to use the rigor of computer analysis.
The previous studies, he said, “are either qualitative, where a researcher is observing the communication, or based on questionnaires about the professionals’ thoughts and opinions about inter-professsional collaboration. People were aware of differences, but the scope of differences in terminology was not well understood. This is the first time a quantitative measure has been used to evaluate the difference in terminology.”
The research suggests that care provided by nurses runs “parallel” to care provided by physicians but with relatively little overlap.
At the start of the study, Boyd said, researchers hoped to ensure patients had the most important information about their hospitalizations. The thinking was that if both the physician and the nurse included a certain medical concept in their documentation, the concept would be important enough for the patient to know. In reality, however, it appears that few concepts are mentioned in both providers’ documentation.
Boyd said this is likely a symptom of the dynamic nature of healthcare today. When nurses and physicians work with the same patients on a regular basis, they tend to develop a common vocabulary.
“However, as both physicians and nurses rotate on and off shifts, when working with new individuals, the chance to build that common nomenclature does not exist,” he said.
He hopes this study raises awareness of that problem.
“So when working with a new nurse or a new physician, being aware of the differences in nomenclature could allow new teams to spend more time to better understand the concerns and treatment of the different health professions,” he said.
Another implication of the study is that it suggests hospital quality metrics might be based on insufficient data, since those metrics are generally based on physician records, and those records apparently leave out most of the care provided by nurses.
The study is titled, “Physician nurse care: A new use of UMLS to measure professional contribution: Are we talking about the same patient a new graph matching algorithm?” It was published last month in the International Journal of Medical Informatics.Get the best insights in healthcare analytics direct to your inbox.