Patient callers with medical symptoms require medical professionals who can respond within specific callback windows, based on the severity of those symptoms. While it depends on the practice, nonclinical personnel like Answering Service Operators (ASOs) are often the first points of contact for these calls. However, because ASOs aren’t certified medical professionals, it’s easier for them to overlook or underreport symptoms that are actually urgent, leaving patients with delayed care and poorer health outcomes. That’s where TriageLogic’s new software, MedMessage Assist, comes into play. It’s designed to provide operators with intelligent medical scripting to ask the right questions and accurately relay patient messages to nurses and doctors. In fact, TriageLogic recently completed a statistical analysis that compares the performance of ASOs before and after using this artificial intelligence-based tool. The following is what we learned, and why MedMessage Assist is a game changer for patient caller safety.
Initial Investigation for ASO Accuracy
Our initial investigation was used to create a baseline for a medical answering service. It analyzed 186 “Call 911” disposition outcomes from July and August of 2019, and determined that 28.07% of the complaints sent by that service were insufficient to alert a provider of possible high urgency outcomes.
As a result of these insufficient complaints, training sessions were conducted between a clinical education manager and the answering service’s leadership, emphasizing the need for greater clarity for medical message intake.
Example dispositions that were determined to be insufficient included caller symptoms like:
- Difficulty urinating.
- Concerns about a hernia.
- High blood pressure.
- A painful sore throat and all-day fever.
Post-Training Results
Data recorded after this training included 1,101 “Call 911” disposition outcomes from January to December of 2020. Our analysis revealed a marked improvement: only 9.25% of the complaints sent by the answering service were deemed insufficient for possible high urgency outcomes.
Results After One Year
While these initial results were promising, continued analysis of the data from January to March of 2021 indicated that operator training was losing its effectiveness. We determined that 39.50% of the complaints sent by the answering service (including 339 “Call 911” disposition outcomes) were insufficient to alert the charge nurse — even worse than our initial investigation!
As a result, indications were that a “training only” solution would eventually lead to regression, likely due to employee turnover, staff shortages, and a lack of training reinforcement.
Enter MedMessage Assist
Once we saw those results, we knew that we had to come up with a better solution. That’s why we created MedMessage Assist. This patent-pending, standalone module can be used by any nonclinical support staff at a provider’s office or hospital system. It relies on artificial intelligence to review data that an ASO types while the ASO is listening to a patient caller. MedMessage Assist’s Complaint Analyzer then prompts the ASO to ask any additional, necessary questions based on the keywords that it identifies.
Results From MedMessage Assist’s Complaint Analyzer
Once we had the module in place, the next step was to test the accuracy of our tool. In order to do so, we took data from November 2011 that consisted of 200 “Call 911” and “Go to ED” disposition outcomes. Of those, we found that 79 of them contained insufficient complaints (40.50%). However, after we applied MedMessage Assist’s Complaint Analyzer to this test data, only 10 insufficient complaints remained (5.0%). This easily beat the best results we saw from leadership training.
About MedMessage Assist
As noted above, training regression is a big concern for the long-term accuracy of medical intake. From our analysis, we’ve found that MedMessage Assist avoids this by acting as its own training tool, prompting the ASO on the next steps she or he should take while engaging with a patient caller. It standardizes patient interaction so that all symptoms are entered accurately, the severity of those symptoms are documented, and appropriate information is shared with medical providers who can respond.
The end results: better medical messages, timely patient care, and improved health outcomes.
MedMessage Assist also meets all standards for HIPAA compliance and cybersecurity best practices, whether it’s used as a standalone module or integrated with a provider’s internal software.
Improve Your Patient Caller Safety
If you’d like to try this module out with your team, we’d love to hear from you! Contact us today to discuss a program, as well as address any questions or concerns you may have with other areas of nurse triage support.
About TriageLogic
TriageLogic is a URAC-accredited, physician-led provider of top-quality nurse telehealth technology, remote patient monitoring, and medical call center solutions. Founded in 2007, the TriageLogic Group now serves more than 12,000 physicians and covers over 25 million lives nationwide.
