Evaluation of statistical illiteracy in Latin American clinicians and the piloting evaluation of a short course across multiple timepoints

Abstract

Background: All clinicians require statistical interpretation skills to keep up to date with evidence-based recommendations in their field. However, statistical illiteracy among clinicians is a highly prevalent problem with far-reaching consequences. The few available that report statistical literacy improvements after educational interventions do not measure for how long these benefits last. To estimate statistical illiteracy among Latin-American clinicians across multiple levels of training and to evaluate a 10-h course at multiple timepoints. Methods: Using an online questionnaire, we evaluated; self-perceived statistical proficiency, scientific literature reading habits and statistical literacy (using an adaptation of the Quick Risk Test). Separately, we evaluated statistical proficiency after a 10-h statistics course in a group of Internal Medicine residents at a tertiary center in Mexico City across multiple time points between November 2020 and February 2021. Results: Data from 392 clinicians from 9 Latin American countries were analyzed. Most clinicians (85%) failed our adaptation of the Quick Risk Test (mean score = 2.6/10, IQR:1.4). The 10-h course significantly improved the scores of the Internal Medicine Residents (n = 16) from 3.8/10, IQR:1.8 to 8.3/10, IQR:1.4 (p < 0.01). However, scores dropped after one and 2 months to 7.7/10, IQR:1.6 and 6.1 / 10, IQR:2.2, respectively. Conclusion: Statistical Illiteracy is highly prevalent among Latin American clinicians. Short-term educational interventions are effective but, their benefits quickly fade away. Medical boards and medical schools need to periodically teach and evaluate statistical proficiency to ameliorate these issues.