General practitioners in the U.S. face more mental strain because of several reasons. These include difficult patient cases, lots of paperwork, and urgent care needs often outside normal office hours. The extra administrative tasks take up much of the doctors’ time and reduce time for patient care. Studies from North America, Europe, and Asia show these problems happen worldwide. But U.S. doctors feel it more due to a fragmented healthcare system and complex insurance rules.
After-hours care is especially tough. GPs must make quick decisions with fewer resources and incomplete information. Emergencies, many patients, and limited staff can increase chances of mistakes and delays in treatment. High mental demands can also cause doctor burnout, which is a big worry in American medicine as it affects patient outcomes and how smoothly clinics run.
Given these problems, healthcare managers and IT leaders look for ways to help doctors handle information and set priorities without lowering the quality of their decisions.
One new way to test AI tools that help doctors is through simulated patient visits. This means giving AI systems many fake but realistic clinical cases like those GPs see in real life. These cases check how well the AI helps with sorting patients, diagnosing, and making decisions in a safe but useful setting.
One example is NAOMI (Neural Assistant for Optimized Medical Interactions). It is an AI decision support agent built with the GPT-4 model. NAOMI assists mainly in after-hours and places with limited resources by helping triage and diagnose patients.
The NAOMI team, which includes Timothy (Shoon Chan) Hor and Lee Fong, tested it with 80 simulated patient cases. These cases varied from simple routine visits to complex emergencies. This tested how well NAOMI could reason and adapt clinically.
This testing approach has several benefits:
By using simulated cases, healthcare leaders in the U.S. can learn the strengths and limits of AI tools before real use.
The NAOMI project found three main design ideas that help reduce mental strain and improve decisions for GPs:
When adding AI in U.S. practices, managers should choose systems that follow these principles to make care safer and more efficient.
AI decision support tools like NAOMI can improve care quality in different ways:
Testing NAOMI with 80 simulated cases showed it improved diagnosis and triage. Doctors also said it fit well into clinical workflows. This means similar AI tools could help U.S. clinics use doctor time better, get better patient results, and reduce delays caused by workflow problems.
Besides helping with clinical decisions, AI is also used more in front-office tasks. This area is key to making clinics run better. For example, Simbo AI uses AI to automate phone calls and answering services. This can improve patient service and office work in healthcare.
AI phone systems can handle appointment booking, patient questions, prescription refills, and urgent calls without needing staff to answer all the time. This automation can:
For healthcare managers thinking about digital updates, using AI front-office tools along with clinical AI support can improve both care and office work. These technologies help clinics meet growing demands using the resources they have.
Healthcare administrators, practice owners, and IT leaders in the U.S. should think about several things when adding AI:
General practitioners in the U.S. deal with more mental pressure from after-hours work, complex patients, and paperwork. AI decision support tools tested with simulated patient visits, like NAOMI, show they can help improve diagnosis, triage, and clinical decisions. These tools work best when they use full patient data, show clear reasoning, and adjust priorities as needed to ease doctors’ mental load and improve care.
At the same time, AI that automates front-office tasks such as phone management, offered by companies like Simbo AI, helps clinics run better and improves patient contact.
Together, these AI technologies give health managers ways to support doctors, improve care quality, reduce burnout, and make clinic work smoother with the resources they have. Using simulations is a good way to check AI tools fit the complex real world of general practice before using them widely.
By carefully adding AI in both clinical and office work, U.S. healthcare providers can better handle growing demands and keep good standards of patient care in the future.
GPs face increasing cognitive demands, particularly after-hours and in resource-constrained settings, due to urgent decision-making, administrative burdens, and complex patient cases.
NAOMI is an AI-based clinical decision support agent using GPT-4 designed to assist GPs with triage, diagnosis, and decision-making to reduce cognitive overload.
A design science approach was applied, involving 80 simulated patient consultations and clinician feedback to test NAOMI’s effectiveness in clinical support.
They are Comprehensive Data Collection and Analysis, Clinical Reasoning Transparency, and Adaptive Triage and Risk Assessment to support decision-making and workflow integration.
It allows the AI to gather and process complete patient data, enhancing diagnostic precision and supporting informed clinical decisions by providing relevant insights.
Transparency builds trust by explaining AI decision processes clearly, enabling clinicians to understand, verify, and confidently integrate AI recommendations into their workflow.
It dynamically prioritizes patient care based on evolving clinical information, optimizing long-term resource allocation and focusing attention where most needed.
Effectiveness was assessed through 80 simulated patient consultations representing diverse real-world cases, alongside feedback from practicing clinicians.
GPs worldwide face cognitive overload due to administrative tasks, patient complexity, urgent care demands, and resource limitations, which AI can help mitigate.
AI can improve GP efficiency, decision-making quality, equity in healthcare delivery, and address systemic workforce challenges by optimizing clinical workflows.