Home health providers should be incorporating audio-recorded patient-nurse communication into their patient risk identification models, according to a recent study published in the Journal of American Medical Informatics Association.
The study’s research team — working alongside VNS Health nurses — recorded 126 patient nurse interactions with 47 patients.
Overall, home health care serves more than 6 million patients in the U.S., and a good chunk of those patients end up back in the hospital.
And the majority of patient risk identification models, which are mostly based on electronic health record data, don’t incorporate audio-recorded patient-nurse verbal communication.
In their current form, these models are only seeing modest success when it comes to detecting patients at risk for ED visits or hospitalization.
Still, the research team initially received pushback at the idea of recording patient and nurse encounters.
“One of our challenges when we started this study was convincing nursing to allow us to audio record conversations between nurses and patients,” Maryam Zolnoori, PhD at Columbia University School of Nursing, and leader of the research team, told Home Health Care News. “They told us they document everything in clinical notes, so it’s not necessary to record a conversation between them and the patient. They also had a concern about the privacy of the conversation with the patient.”
The research team redeveloped the risk model to include OASIS data, clinical notes and verbal communication features. They found that incorporating verbal communication data improved risk models by 26%.
Additionally, researchers found that patients that were more at risk typically showed signs of sadness, anxiety and had long periods of silence amid conversations, according to the study.
Zolnoori believes that these findings make the case for beefing up clinical workflows in order to include patient and nurse communication in medical records. She also pointed out the need to train nurses in order to emphasize the importance of audio recording data.
“If [home health providers] embed this into their clinical workflow, it has a substantial impact on the patient’s care,” she said. “We can use this data to not only develop effective risk identification models, it can be used for detecting patients with Alzheimer’s, or it can be used for detecting patients at risk of emergency department visits. It also helps just to reduce documentation time for nurses.”