Integrating patient-nurse communication when identifying risks is important for good home healthcare, according to a new study.
The report was published Oct. 17 in Journal of American Medical Informatics Association (JAMIA). The authors said the models for identifying risk can improve care for people with high risks.
About 6 million people in the United States rely on home healthcare, and many are older adults with chronic conditions. About 1 in 3 people who receive care at home will eventually need an emergency department visit or be hospitalized. Clinicians use risk identification models, which are based on electronic health records. Sometimes the models aren’t fully effective or accurate.
That’s why the research team led by Maryam Zolnoori, PhD, a researcher at Columbia University School of Nursing, recorded audio from 126 interactions between patients and nurses. Eight of the 27 patients eventually went to the emergency room for care.
The team updated the risk-assessment model to include three essential components: structured data from the Outcome and Assessment Information Set (OASIS), clinical notes and details from verbal discussions.
The researchers used natural language processing methods to evaluate patient-nurse interactions. This improved risk models by 26%. The analysis also showed that high-risk patients showed more cues indicating their risk. For example, these patients expressed sadness or anxiety, and there were periods of silence during the conversations.
“This development highlights the need for an evolved clinical workflow that routinely records patient-nurse verbal communication within the medical record, potentially improving patient care, reducing hospitalization rates, and enhancing health care providers’ ability to identify and address risks promptly during hospitalizations and emergency department visits,” Zolnoori said.
