Simulation can preserve affirming care practice as exposure shrinks

There is a particular kind of training gap that no amount of curriculum reform fixes on its own: the one where learners understand that the work matters and then never get to do it. Sixty percent of surveyed residency program directors reported no rotation with direct clinical exposure to transgender patients (Baylor University Medical Center Proceedings, 2023). A resident in one of those programs can sit through every lecture on the topic and still finish training without having had the conversation once.

A team presenting at IPSS Rome designed an AI patient scenario to give pediatric residents structured communication practice when clinical exposure is limited. The work is about training access. It is not about replacing patients, standardized patients, or faculty.

The gap is measurable, and it is not a knowledge gap

A multicenter cross-sectional study of OB-GYN and family medicine residents found that 62.4% had received no health education tailored to transgender patients. In the same study, 68.3% reported inadequate knowledge of the clinical implications of gender affirming hormone therapy, and 74.1% reported inadequate knowledge of gender affirming surgeries (Healthcare, 2025).

An internal medicine study puts the mismatch most sharply: 97% of residents considered transgender medical issues relevant to their practice, while 45% reported any prior education on the topic (Transgender Health, 2017).

These samples do not describe every residency program. What they establish is a consistent distance between the clinical responsibility residents expect to carry and the training available to prepare them for it. And the part that lectures address least well is the spoken part.

What a lecture cannot show

A lecture can introduce terminology, clinical principles, and the evidence behind them. It cannot show whether a learner can hold a conversation when a patient corrects an assumption they just made out loud, when a parent asks a hostile question, or when the room becomes emotionally charged and the learner has thirty seconds to decide what to say.

Those are performance skills, and performance skills need repetitions. In a program with no relevant rotation, the only repetitions available are the ones a program builds itself.

What a case built for a charged opening looks like

The affirming care scenario is not one of the encounters we publish, so the design pattern is easier to show through one that is. Different clinical situation, same shape.

In a sickle cell adherence encounter, the parent opens before the fellow has said anything: "Before we start, let me just say it. Nobody in these places really listens to Black patients, and that medicine you keep pushing makes my son feel sick. So tell me why I should keep giving it to him."

The case is built so that the hardest moment arrives in the first ten seconds and cannot be avoided. The learner has to respond to the mistrust, and be accurate about a drug whose evidence is strong in some places and thin in others, without retreating from a clear recommendation. Nine minutes later the report tells them which of those they managed. In that encounter, the fellow named the emotion well and scored 4 out of 5 on it, then spent the rest of the conversation asking about content and logistics rather than the feeling, and scored 3 out of 5 on exploring.

That is the design a program needs when clinical exposure is scarce. Not a patient who behaves pleasantly, but a case that puts the difficult turn first, and a report that separates what the learner did well from the one thing to try differently next time.

What the IPSS Rome work does and does not establish

Gillian Brennan, MB BCh BAO, Vinod Havalad, MD, and colleagues developed a pediatric AI patient scenario for structured communication practice, giving residents a defined situation, a named communication framework, and feedback tied to the transcript.

The presentation describes a simulation design for a documented training gap. It does not establish an effect on patient care, and it will not until somebody runs a study designed to test that question. The practice report does not replace faculty judgment, and structured practice with AI patients extends a standardized patient program rather than substituting for it.

Those are real limits and worth stating in the same breath as the design. The claim here is narrow: residents in programs with no relevant rotation currently get zero repetitions, and a case like this one gets them above zero.

Talk with us about research or curriculum design.