Precision Education Finally Gets Real Data Sources and Learner Control

Precision education sounded more operational this week: use existing clinical and learning data to show physicians what they need next, then let them act on it. The evidence is narrow—a single AMA education-leadership podcast, not a broad independent clinician conversation—but the implications for CME platforms are clear.

The LMS can’t be the whole learning record

In a Faculty Factory conversation, AMA Chief Academic Officer Sanjay Desai described precision education as using data and technology to personalize education, make it more efficient, and transfer more agency to the learner. The important detail was the data stack: EMR data, LMS data, payer data, and natural language processing from clinical notes could be pulled together to identify competency signals and inform education for faculty and learners (Faculty Factory).

For CME providers, that changes the personalization problem. Many platforms can recommend the next course based on specialty, prior clicks, assessment answers, or stated interests. Desai’s framing starts outside the activity: what patients a clinician has seen, what conditions or procedures they have encountered, where documentation or workflow problems appear, and which competencies need attention.

That is also where learner agency becomes concrete. A dashboard that shows clinicians their own competency-relevant signals is different from an institution assigning another module. It gives the learner a way to see why a pathway is being suggested, choose what to work on, and use a coach or faculty member to interpret the data. We saw a related pattern in last week’s brief on learner co-design; this week’s addition is the operational substrate: machine-readable data, learner-facing dashboards, and education tied to clinical relevance.

The provider implication is straightforward: stop treating personalization as content sequencing alone. Ask whether your learning infrastructure can ingest external signals, show the learner what is known about their needs, let the learner act on those signals, and connect the pathway back to competency rather than mere participation.

What CME Providers Should Do Now

• Map which learner signals you can use today, which require partner access, and which should remain out of scope for privacy or trust reasons.
• Rework one activity into a simple learner-facing pathway that shows why each step is recommended, even if the first version uses limited data.
• Define outcomes that test whether the experience reduced redundant learning, improved patient-relevant confidence or competence, and supported a next action in workflow.

Where this leaves CME teams

Precision education will not be won by adding an AI layer to the same activity catalog. It requires data-sharing relationships, competency taxonomies, learner-visible product surfaces, and outcomes models that can survive outside the LMS. Providers do not need full EMR integration tomorrow. But they should know whether their systems are being built for a future where education starts with what the clinician has actually seen, done, and struggled with—not simply with the next available activity.