Cataract remains the most common cause of blindness globally, and glaucoma is the third after uncorrected refractive error. Surgical management remains a priority, yet surgical training of ophthalmologists continues in the outdated apprentice model. Simulation-based surgical education is yet to be tested to the level of a randomised-controlled trial in ophthalmology. We designed two separate and independent multi-centre multi-country investigator-masked randomised controlled educational-intervention parallel group efficacy trials. Post-graduate doctors in ophthalmology training programmes at collaborating institutions in five East and Southern African countries were assessed for eligibility for inclusion (not having performed the procedure as primary surgeon) into either the OLIMPICS (ophthalmic learning and improvement initiative in cataract surgery) or GLASS (glaucoma simulated surgery) trials. Fifty-one surgical trainees were recruited into the GLASS trial, and 50 into the OLIMPICS trial. Surgical competency was assessed by video recordings, which were double marked by independent experts who were masked to group assignment and timing of the assessment. The intervention was an intense simulation-based cataract or glaucoma surgical training course over 5 days. Primary outcome measure was surgical competency at three-months assessed with validated simulated surgical competency assessment rubrics, the Sim-OSSCARs (ophthalmic simulation surgical competency assessment rubric), for both trials. The trials were registered in March 2017 on the Pan-African Clinical Trial Registry (PACTR201803002159198) and are currently closed to recruitment. Baseline characteristics of age, sex, year of training, baseline knowledge and competency scores were balanced between both arms, for both trials. In total 1,361 surgical videos from across different time-points were independently graded by two separate graders in both trials. In the OLIMPICS trial, 50 participants were recruited between November 2017 and May 2018 and 49 included in the final intention-to-treat analysis with one dropout from the control group. Intervention group participants increased mean simulated surgical competence scores from a baseline of 10.8 of 40 points (27.0%) to 33.7 (84.2%) at 3-months after the training intervention, an increase of 212%. Control group participants’ mean baseline scores were 12.8 (31.9%) and 3-month scores 17.9 (44.7%). We found strong evidence (linear regression p<0.0001) that those in the intervention arm were estimated to have higher scores at three months than those in the control arm, after adjusting for baseline score. Among individuals with the same baseline score, those who received the training were estimated to have scores 16.6 points higher (95%CI 14.5 to 18.8) at three months, compared to those who had not received the training. Intervention participants performed a mean of 22 cataract surgeries as primary surgeon in the one year following the training intervention, compared to 9 by control participants (Poisson regression p<0.0001). Surgical complications were reported for the one year period, and posterior capsule rupture (PCR) rates were 7.4% for the intervention group compared to 26.2% for controls (p<0.0001). Confidence rating scores were assessed using a ten-point Likert scale anchored at 1=’not confident at all’, and 10=’very confident’. Confidence as cataract surgeons increased from 2.2 (of 10) to 6.3 at three-months in the intervention group, compared to 3.4 at baseline to 4.2 for the control group. Among individuals with the same baseline confidence score, those receiving the training were estimated to have scores 2.7 points higher (95%CI 1.6 to 3.7) (p<0.001). In the GLASS trial, 53 trainee ophthalmologists were assessed for eligibility, and 51 were enrolled and randomised. Forty-nine participants were included in the final intention-to-treat analysis: 23 intervention and 26 control, following two drop outs from the intervention group. Baseline surgical competency scores for intervention were a mean of 9.1/40 (22.6%) [median 7.3, IQR 5.4-12.1]; and for control: 8.7/40 (21.8%) [median 8.2, IQR 6.3-12.0] participants. Mean Sim-OSSCAR scores at three-months were 30.4 (76.1%) [median 30.3 IQR 27.8-33.5] and 9.8 (24.4%) [median 9.2 IQR 7.5-11.7] for intervention and control groups respectively. We found strong evidence (linear regression p<0.0001) that those in the intervention arm were estimated to have higher scores at three months than those in the control arm, after adjusting for baseline score as a fixed effect. Among individuals with the same baseline score, those who received the training were estimated to have scores 20.5 points (of 40) higher (95%CI 18.4 to 22.6) at three months, compared to those who had not received the training (linear regression p<0.0001). Baseline mean self-reported confidence in glaucoma surgical skills was 3.0/10 for intervention and 3.2 for control participants. This increased to mean 6.4 and 3.7 at three months respectively (p=0.002). Trainee participants in the intervention group performed a mean of 3.1 live surgical trabeculectomies as primary surgeon over one year following training (median 2, range 0-15, IQR 0-4). Over the same period (and before their simulation training) the control group performed a mean of 0.15 (only one of the 26 control participants performed any glaucoma surgery, compared to 14 of the 23 intervention participants). These are the first multi-centre ophthalmic simulation surgery educational-intervention randomised controlled trials ever conducted. Intense simulation training affords a rapid and sustained increase in surgical competence, confidence as a surgeon, and impacts the number of live surgeries performed. Simulation education in cataract surgery affords a striking benefit in terms of patient safety.
