OB-GYN training lasts several years. Junior residents in their first and second year have little clinical experience. They need to learn how to diagnose and make decisions quickly. Time pressure during work makes this harder. Sometimes this lowers their accuracy and adds stress. As residents get more experience, they do better. Still, the learning process is not simple or smooth. Studies show that residents in the middle of their training can have drops in accuracy because cases become harder or because they have too many tasks.
Healthcare leaders in the U.S. ask two main questions:
The PERFORM Study compared AI large language models (LLMs) with OB-GYN residents in making clinical decisions. It tested 8 AI models and 24 residents using 60 clinical cases in English and Italian. The study focused on accuracy with and without time limits.
Main findings were:
This research, published by groups like the Mayo Clinic, shows AI has a growing role in medical education and practice, especially in teaching OB-GYN residents.
The PERFORM Study’s results suggest leaders in medical practices and healthcare IT in the U.S. should rethink how AI can help with resident training and patient care.
Training programs often struggle to balance teaching goals with patient care. AI tools that assist with real-time diagnosis can lower the mental workload for junior residents. This lets them focus more on hard cases and improves patient results. Administrators might try pilot programs that use AI decision tools for residents, especially early in training.
The data shows AI can make diagnoses more consistent. This could mean fewer missed or late diagnoses. In OB-GYN, quick decisions are very important for both mothers and babies. Clinic owners who are in charge of patient safety might look into AI as a way to improve quality.
The U.S. serves many patients who speak different languages. AI’s strong performance across languages helps doctors work with these patients better. Practices with varied patient groups could use AI tools to make care more equal and support doctors during communication challenges.
AI can also make administrative and clinical work in OB-GYN clinics more efficient. This benefits people who run healthcare operations by saving time and reducing mistakes.
Some companies, like Simbo AI, use AI to handle front office phone calls. OB-GYN clinics get many calls about appointments, health questions, prescriptions, and urgent matters. Staff may not always keep up with calls, leading to delays or missed responses, which affects patients.
AI phone systems understand patient requests and can answer or send calls to the right person. This lowers work for office staff and helps patients get answers faster. It can also stop patients from going to the emergency room when it isn’t needed.
Clinical staff spend a lot of time doing paperwork, entering data, and sending reminders. AI systems can send calls or texts for reminders, schedule appointments, and update records automatically. This gives doctors and residents more time for patient care and learning.
AI tools can help sort patient questions, check symptoms, and recommend initial actions based on rules. When combined with AI diagnosis support like in the PERFORM Study, junior residents get help both with making diagnoses and managing patient flow.
Healthcare must follow many rules about documents and privacy. AI helps make sure paperwork is complete and correct. This lowers errors that can cause problems with compliance. This is useful in training programs where learning documentation is part of education.
Healthcare IT managers have an important job in choosing, setting up, and keeping AI systems working. OB-GYN clinics must choose AI tools that work well with current systems, keep patient data safe, and train users well.
IT teams need to make sure AI connects smoothly to electronic health records and other systems without risking patient privacy. HIPAA rules in the U.S. require strong data protection.
Introducing AI needs good training for residents and staff. IT managers work with teachers and clinical leaders to help users understand AI’s strengths and limits. This helps avoid resistance and makes AI use smoother.
IT managers should watch how well AI works. They can track if diagnoses improve, how fast responses are, and patient satisfaction. Changes can be made to settings or processes as needed.
The PERFORM Study results come at a time when U.S. healthcare faces pressure to offer good, affordable, and patient-centered care. Using AI in OB-GYN training and work might help with challenges such as:
However, some risks exist. For example, senior doctors might rely too much on AI and become less sharp, which some data hints at. It’s best to match AI help to each resident’s skill level.
The PERFORM Study shows AI can be useful in training junior OB-GYN residents in the U.S. Medical leaders and IT managers can use AI not only to help with diagnoses but also to automate tasks that reduce mental load and improve clinic work.
As OB-GYN clinics and training programs think about using AI, they must understand how AI works, its limits, and how best to include it. This will help give better care to patients and better learning for residents in this important medical field.
The primary objective of the PERFORM Study was to systematically evaluate the performance of artificial intelligence (AI) large language models (LLMs) compared to obstetrics-gynecology residents in clinical decision-making, focusing on diagnostic accuracy and error patterns.
The study evaluated 8 AI LLMs and 24 obstetrics-gynecology residents across their first five years of training.
The primary outcome measure was diagnostic accuracy, while secondary endpoints included performance under time constraints and language impact.
AI LLMs reported an overall diagnostic accuracy of 73.75%, while residents achieved 65.35%, highlighting a significant difference (P<.001).
Residents exhibited a marked decline in diagnostic accuracy under time pressure, dropping from 73.2% to 56.5% adjusted accuracy.
Error pattern analysis indicated a moderate correlation between AI and human reasoning with a coefficient of r=0.666, suggesting similarities in decision-making processes.
AI LLMs provided the most significant enhancement in diagnostic accuracy for early-career residents, with an improvement of +29.7% (P<.001).
The AI systems demonstrated high cross-linguistic accuracy (88.33%) with minimal language impact, indicating robustness across different languages.
The residents assessed in the study were from all five years of training in obstetrics and gynecology.
The findings suggest that AI-enhanced decision-making may improve diagnostic consistency and reduce cognitive load, particularly benefitting junior residents in time-sensitive clinical settings.