ABSTRACT
Abstract: Objective: To assess the assumption of 'equity' of Mexico's resident-selection assessment tool, the Examen Nacional para Aspirantes a Residencias Médicas (ENARM). Materials and methods: Official ENARM-2016 and -2017 databases were analyzed. Differences in the absolute number of correct answers (multivariable linear regression) and the number of applicants reaching their specialty minimum score (SMS) per test day (odds ratio [OR]) were calculated. Applicants affected by test-day inequity were estimated. Results: There were 36 114 applicants in 2016, and 38 380 in 2017. In 2016, day-2 applicants had significantly higher scores and more reached the SMS than on days 1-3-4 (OR 1.55), and 5 (OR 3.8); 3 565 non-passing applicants were affected by inequity (equivalent to 44.64% of those selected). In 2017, day-1 and -2 applicants had significantly higher scores and more reached the SMS than on days 3-4 (OR 1.85), and 5 (OR 4.04); 3,155 non-passing applicants were affected by inequity (37.2% of those selected). Conclusion: Analysis of official ENARM databases does not support the official attribution of equity, suggesting the test should be redesigned.
Resumen_ Objetivo: Evaluar el atributo de "equidad" asignado al Examen Nacional para Aspirantes a Residencias Médicas (ENARM). Material y métodos: Se analizaron las bases de datos oficiales del ENARM 2016 y 2017. Se compararon las diferencias inter-día de respuestas correctas (regresión linear multivariable) y de sustentantes que alcanzaron el puntaje mínimo de su especialidad (PME) (razón de momios [RM]). Se estimó a los afectados por la inequidad. Resultados: Hubo 36 114 sustentantes en 2016 y 38 380 en 2017. Los días 2 (ENARM-2016) y 1-2 (ENARM-2017) registraron puntajes significativamente más altos, y más sustentantes alcanzaron el PME que en los días 1-3-4 (RM 1.55) y 5 (RM 3.8) en 2016, y los días 3-4 (RM 1.85) y 5 (RM 4.04) en 2017. Se estimó que cuatro de cada diez sustentantes que aprobaron el ENARM no lo hubieran hecho si el examen fuera equitativo. Conclusión: Los resultados sugieren que el atributo de equidad del ENARM está en duda.