Emergency departments in the U.S. often have too many patients and not enough space or staff. This causes longer wait times and puts stress on the resources available. When patients arrive, triage nurses check them to see how serious their condition is. They then decide how fast and what kind of care each patient needs. The main system they use, called the Emergency Severity Index (ESI), ranks patients from one (most urgent) to five (least urgent).
Even though this system is common, nurse assessments can be based on personal judgment. Two nurses might give different scores to the same patient. This can cause inconsistency and delays in care. This is a bigger problem in busy city hospitals or during times with many patients, like during flu season or after big accidents.
To help fix these issues, many hospitals in the U.S. have started trying AI tools to help or improve the triage process.
Recently, AI tools have been created that look at patients’ digital health records and real-time vital signs. These tools can guess how risky a patient’s condition is and suggest a triage level. One such tool is called TriageGO. It was first made by researchers at Johns Hopkins and the company Stocastic. After being tested in hospitals in Maryland, Florida, Connecticut, and other states, the company Beckman Coulter, which works in clinical diagnostics, bought it.
These AI tools use algorithms to review data like heart rate, blood pressure, medical history, and doctors’ notes. They then give risk results that match clinical best practices. This helps lower the differences seen when only humans do triage assessments.
A study with 758 patients found that GPT-4 models, similar to ChatGPT, worked as well or a little better than nurse triage. The AI had 66.5% accuracy overall, and nurses had 65.2%. More importantly, for very serious cases (ESI levels 1 and 2), AI tools did much better. GPT-4 correctly found about 85% of these serious cases, but humans often missed many of them.
This is important because if a patient is under-triaged, their urgent treatment might be delayed, which can cause serious health problems.
When AI tools give more accurate triage results, it helps move patients through the emergency department faster. AI helps doctors and nurses spot low-risk patients more clearly. These patients can be sent to quicker care routes. This allows earlier discharge or outpatient care, which frees up beds in the ED for sicker patients.
Scott Levin, who helped develop TriageGO, says that finding more low-risk patients “creates improved patient flow” and shortens the time patients stay in the ED. This is good for staff and patients. Staff feel less pressure when the ED is crowded, and patients get care faster, suited to their needs.
AI can also help hospitals plan how to use staff and equipment better, especially during busy times or after mass casualty events. The AI can predict who needs urgent help and who can safely wait. This helps hospitals send staff where they are needed most, avoiding burnout and running out of resources.
These AI tools are already in use at places like The Johns Hopkins Hospital, Johns Hopkins Bayview Medical Center, and Howard County General Hospital. There are also plans to use them in hospitals in Missouri and other states.
Even with benefits, adding AI to triage has challenges. One big concern is the quality of the data AI needs. For AI to make good predictions, it must get complete and correct digital health records. If patient histories are missing or wrong, or if their vital signs change quickly, AI may not work well.
Another issue is called algorithmic bias. This means the AI might unknowingly prefer some patient groups over others because of the data it learned from. Fixing this will take ongoing work to make sure care is fair for everyone.
Doctors and nurses must also trust the AI. Many worry that relying too much on AI could reduce their control or judgment in patient care. For AI to be accepted, its recommendations need to be clear, and healthcare workers must know how to use AI ideas alongside their own.
Privacy and data safety are also important. Hospitals must follow rules like HIPAA and keep strong security to protect patient information.
AI not only helps make triage more accurate but also fits into plans to automate emergency department work. This means making processes faster and easier for staff and patients alike.
Automated triage tools quickly look at patient data, saving time for doctors and nurses. These tools can be linked with phone systems made by companies like Simbo AI, which handle AI-based phone answering and automation. By combining triage tools with phone systems, some patient data and risk checks can happen before patients even get to the ED.
For instance, if a patient calls with chest pain, an AI phone system can ask structured questions and collect symptom details. If linked with telemedicine devices or patient portals, it might record vital signs too. This pre-arrival check helps the emergency team get ready — maybe by preparing beds or calling specialists — so care starts right when the patient arrives.
AI can also handle admin tasks, such as patient registration, insurance checks, and scheduling. This cuts down paperwork for staff, so they can spend more time with patients. Simbo AI’s technology helps make this communication smooth and improves teamwork between front office and clinical workers.
When AI triage tools connect with other hospital systems, they give real-time updates on waiting lists, bed availability, and staff needs. This helps managers adjust schedules, move non-urgent patients to clinics outside the ED, or open extra capacity when many patients come in.
These automations help patient flow run better, improve staff use, and reduce waiting and delays, which makes patients more satisfied.
For hospital leaders and practice owners in the U.S., AI triage can help with running the department smoothly. Less triage mistakes and faster patient prioritization improve how well the hospital performs, which is important for reviews and ratings. This might also affect finances, as better efficiency leads to more appropriate payments under value-based care models.
IT managers are important for setting up AI triage tools. These tools need to work well with existing electronic health record (EHR) systems without causing problems in how staff work. Because patient data is very sensitive, it is critical to keep everything HIPAA compliant and secure.
When choosing an AI triage provider, hospital leaders should think about whether the provider can adjust solutions to fit their hospital’s workflow. They should also check if the data is accurate and if the company offers solid support over time. Tools like TriageGO and those in recent studies show that good AI use needs teamwork among doctors, IT staff, and providers.
The U.S. emergency care system serves all kinds of people and deals with many patients. AI can help hospitals in cities and rural areas by giving fair and steady triage results. This can reduce differences in care and help places with not enough staff.
Artificial intelligence is playing a bigger role in triage at emergency departments in the U.S. AI tools like TriageGO and GPT-4-based systems give clear, data-based patient risk scores that support the work done by triage nurses and emergency doctors.
By making triage more accurate and faster, these tools help patients move through the ED better, help hospitals use limited resources smarter, and lead to improved health results.
For hospital leaders and IT teams, AI triage combined with workflow automation — like Simbo AI’s phone systems — are important steps to modernize emergency care. AI fits well with EHRs and communication tools to help from when patients first call, through to assigning beds.
As AI tools keep improving, emergency departments across the U.S. will find them more useful to handle growing patient numbers while still giving good care.
The AI tool is designed to assist emergency department nurses in triaging incoming patients by predicting their risk of acute outcomes and recommending a triage level of care based on the collected data.
The tool integrates with patients’ digital health records, allowing nurses to input patient information and vital signs, which the AI uses to quickly assess risk and suggest triage levels, enhancing accuracy and efficiency.
The AI tool helps nurses confidently identify low-risk patients, enabling those individuals to receive care more efficiently, ultimately improving patient flow through emergency departments.
The AI tool is used in the emergency departments at The Johns Hopkins Hospital, Johns Hopkins Bayview Medical Center, Howard County General Hospital, and other hospitals in Florida, Connecticut, and Missouri.
The AI tool is called TriageGO, developed by the company Stocastic, which was co-founded by Scott Levin and Eric Hamrock.
The triage level, which ranges from one (the sickest) to five (the least sick), determines the path of care for patients, influencing the urgency and type of treatment they receive.
By efficiently identifying low-risk patients, the AI tool helps streamline care pathways, allowing quicker discharge for those patients and thus optimizing overall patient flow in the emergency department.
Scott Levin, an associate professor of emergency medicine, and Eric Hamrock, a health care administrator, are notable figures in the development of TriageGO and its parent company, Stocastic.
TriageGO and its parent company Stocastic were acquired by Beckman Coulter, a company specializing in clinical diagnostics.
The tool is set to launch in several hospitals in Missouri, expanding its utilization to improve triage and patient care in more emergency departments.
Simbo is using AI agents to automate all phone related workflows. Connect now to know more.
Schedule AI Agents Demo Now© Since 2019, Simbo AI.
