Stanford Medicine researchers have created an AI tool that analyzes doctors’ notes in electronic medical records to improve care. This tool reviews thousands of records to detect trends, saving time and providing insights that can help doctors.
Instead of manually going through medical charts, this AI tool uses advanced language models to find patterns in written notes. For example, it can track if doctors discuss medication side effects or identify patients who might benefit from specific treatments.
In a study published in Pediatrics, the tool was used to check if children with ADHD received proper follow-up care after starting new medications. It analyzed over 15,000 notes to see if doctors asked about side effects like appetite loss. The AI tool was trained with human-reviewed data and correctly classified about 90% of the notes.
The findings revealed differences in how pediatricians at various practices handled follow-ups. For instance, some asked more about side effects during phone calls, while others didn’t. It also found that follow-up questions were less common for non-stimulant ADHD medications compared to stimulants.
Although the tool identified patterns, it couldn’t explain them. Doctors provided context, noting they had more experience with stimulant medications.
This AI tool offers a faster way to analyze medical records and improve ADHD care. However, it may miss some details, such as unrecorded conversations or notes from specialty care. Despite these limitations, it shows promise for enhancing medical follow-up processes.
Source: medicalxpress
