Misalignment of Biostatistics Content Between Licensing Exam Study Aids and Contemporary Medical Research

Abstract

Background and objectives: Medical trainees express difficulty with interpreting statistics in clinical literature. To elucidate educational gaps, we compared statistical methodologies in biomedical literature with biostatistical content in licensing exam study materials.

Methods: In this bibliographic content analysis, we compiled a stratified random sample of articles involving original data analysis published during 2023 in 72 issues of three major medical journals. We recorded all discrete statistical methods and concepts detailed in the methods section of the articles and in three commercial licensing exam study resources. We created a unified list of discrete methods or concepts to define overarching domains and mapped each method to a domain to determine that domain's presence in each resource or article.

Results: In a sample of 273 journal articles and three study resources, we identified 1,057 unique key words mapped onto 20 domains. Statistical error, significance, power analysis, and group comparisons of categorical data were high-frequency domains among the articles. Overall, 63% of articles included methods from domains not covered in any study resource.

Conclusions: Medical licensing exam preparation does not reflect the breadth of contemporary statistics in biomedical research. Future interventions should expand medical students' understanding of research protocols and complex data manipulation.