Augmented Reality Training for Pelvic Exams and Pap Smears in Medical Students

BACKGROUND

Contemporary training in gynecology requires a careful balance between foundational scientific knowledge, clinical practice, and the technical skills required in outpatient and inpatient settings. The knowledge and abilities that medical students must acquire during professional training correspond to what is termed clinical competencies. Students have traditionally been assessed on these competencies at the clinical workplace through direct supervision; however, the constraints of clinical practice and healthcare demand often prevent optimal conditions for ensuring that every student attains the minimum required competencies. Clinical simulation has therefore gained relevance across training centers, placing students in a context that replicates some aspect of clinical reality, although simulated settings require physical space and equipment (mannequins) of limited availability and high cost.

In routine gynecological training, competencies are acquired during the execution of clinical procedures. Simulation allows students to acquire basic manual skills and dexterity before facing clinical practice. Simulation is especially valuable in sensitive scenarios such as the gynecological examination, which involves an intrusion into the patient's privacy and makes it difficult to obtain consent for trainees to perform the examination. Published review evidence indicates that simulation of the pelvic examination (bimanual or speculum examination) provides a significant benefit over traditional teaching without simulation in terms of competence, confidence, and communication. Graduating medical students are expected to demonstrate proficiency in speculum examination and Papanicolaou (Pap) sampling; accordingly, simulation of this procedure is part of the current undergraduate curriculum at the participating institution.

Immersive technologies are reshaping health professions education. Virtual reality (VR) and augmented reality (AR) increasingly support clinical specialties, patient care, and health sciences education. VR/AR-based simulation is one of the recognized categories of simulation: interactive computer programs that reproduce situations in physiology, pharmacology, or clinical decision-making, with the principal aim of training and assessing knowledge and decision-making. Because AR and VR allow several students to train simultaneously, programs already exist for team-based training. VR and AR have been used for competency acquisition, emergency-response training, soft-skills training, and psychomotor training. Systematic review evidence on AR for teaching in the health sciences suggests that it is more effective for supporting skill development than knowledge acquisition; most AR systems concentrate on anatomy and surgery, although other practice areas are increasingly being explored.

Few studies have implemented AR/VR systems in obstetrics and gynecology. One study compared AR with traditional mannequin-based simulation for teaching intrapartum fetal monitoring analysis and found similar efficacy between the two modalities. A second study evaluated 360-degree video review before assisting in cesarean delivery and found no measurable benefit for students. Regarding the gynecological examination specifically, only one prior study has examined immersive technology-based training, and it measured exclusively student satisfaction rather than competency attainment; more than 90% of students considered the training useful for recalling task details, and the authors noted that further evaluation was needed to determine its effect on clinical competency. To the investigators' knowledge, no prior study has evaluated AR training for speculum examination and Pap smear sampling using a competency-based outcome.

To address this gap, the investigators designed and piloted a self-directed AR training system for speculum examination and Pap smear sampling, with an evaluation instrument developed and validated through a Delphi panel of gynecology experts. In a preliminary pre-post pilot study among six senior medical interns, AR training was associated with a significant improvement in Direct Observation of Procedural Skills (DOPS) scores (from 35.5 to 38.7; p = 0.017), a short mean training duration (16.9 minutes), and high learner satisfaction, but without a comparison group. Building on these preliminary findings, the present randomized controlled trial, with a blinded evaluator and a conventional-training control group, was conducted.

OBJECTIVES

The primary objective was to evaluate the effect of AR-based pre-simulation training on the acquisition of speculum examination and Pap sampling competencies among undergraduate medical students, compared with conventional text- and video-based preparation. Secondary objectives were to determine the number of simulation repetitions required to reach the minimum competency threshold and to assess student satisfaction with the AR platform.

STUDY DESIGN

Prospective, randomized, controlled trial with a blinded evaluator among undergraduate medical students at the Pontificia Universidad Catolica de Chile, Santiago, Chile. The study was approved by the Health Sciences Scientific Ethics Committee of the Pontificia Universidad Catolica de Chile (approval ID 241002001). The protocol was presented, in part, at the 2025 Latin American Congress on Resident Education.

AUGMENTED REALITY TRAINING SYSTEM

A speculum examination and Pap smear procedure was filmed in a real clinical consultation room, with prior institutional authorization, using a hybrid simulation model that combined a pelvic mannequin with a trained actress portraying the patient. Filming followed a single continuous take, without cuts or edits, using two synchronized cameras (a tripod-mounted 360-degree camera and a first-person head-mounted camera worn by the operator) to preserve procedural flow and realism. The footage was edited (Adobe Premiere Pro, GoPro Player) and developed into an interactive, self-directed AR module (Unity Editor) that allows students to alternate between first-person and external views, explore procedural steps at their own pace, and receive automatic feedback. The module incorporates an error-detection task, contextual infographics on equipment and procedural steps, and a formative quiz, and is hosted on a cloud-based platform (3DVista) accessible from any internet-connected device.

PARTICIPANTS AND RANDOMIZATION

Participants were undergraduate medical students enrolled in the Obstetrics and Gynecology course at the Pontificia Universidad Catolica de Chile during the 2025 academic year. All students entering the course were invited to participate; the eligibility criterion was willingness to provide informed consent, and students who declined were excluded. Consenting students were randomized to the study group (AR-based training) or the control group (conventional training) using a computer-generated randomization sequence with permuted blocks; allocation was concealed from the responsible investigator and from the teaching staff acting as evaluators. Participation, and the decision to participate or not, had no bearing on course grades or clinical teaching.

INTERVENTIONS

Control group: conventional training consisting of a written guide and an instructional video covering speculum examination and Pap smear sampling technique. After reviewing the material, students completed a hands-on simulation session on a high-fidelity pelvic simulator (Noelle) under blinded observation.

Study group: the same written guide and instructional video as the control group, plus self-directed training on the AR platform, which allows students to control their own pace and navigation and provides perspective switching, an error-detection task, contextual infographics, and a formative quiz. After completing the AR training, students underwent the same hands-on simulation session as the control group. The AR platform automatically logged each student's total time of use as an exploratory measure of engagement duration.

OUTCOME MEASURES

Primary outcome: achievement of the minimum competency threshold during a single standardized simulation of speculum examination and Pap smear sampling, assessed using a 20-item DOPS checklist previously developed and validated by the investigators. Minimum competency was defined a priori as achieving 75% or more of the maximum DOPS score. The observer was blinded to group allocation; students were identified only by a numeric code.

Secondary outcomes: mean DOPS score; achievement of a lower performance threshold (60% or more of items); number of simulation repetitions required to reach the minimum competency threshold; and student satisfaction with training, measured using a previously validated 7-item, 5-point Likert-scale perception survey adapted from a published instrument previously used among undergraduate students. Students in the AR group additionally answered supplementary items evaluating the AR platform's specific contribution to their preparation.

SAMPLE SIZE

Based on prior data suggesting 75% DOPS compliance on the first attempt in the control group, and an anticipated 10% absolute improvement (to 85%) in the AR group, a sample of 120 participants (60 per group) was calculated assuming 80% power, a one-sided alpha of 0.05, and a test of proportions. All undergraduate medical students enrolled in the 2025 cohort were invited to participate in order to achieve the target accrual.

STATISTICAL ANALYSIS

Descriptive and inferential analyses were performed using Python (v3.12) with the pandas and SciPy libraries. Continuous variables were compared between groups using the Mann-Whitney U test given non-normal distributions; categorical variables were compared using Fisher's exact test given low expected cell counts. A two-sided p < 0.05 was considered statistically significant.

ETHICAL CONSIDERATIONS

All participants provided written informed consent before enrollment. Participation was voluntary, and students were explicitly informed that declining to participate would have no effect on their grades, teaching, or study opportunities. Each student, whether or not enrolled in the study, was assigned a numeric identifier, and the blinded evaluator completed the DOPS checklist without knowledge of group allocation. Data from students who declined participation were excluded from the analysis dataset. The study posed no additional risk beyond routine simulation training. Students in the control group, as well as those who declined to participate, retained access to the AR platform after study completion. Data were de-identified, stored securely, accessible only to the research team, and will be retained for 10 years before destruction, in accordance with the approved protocol.