Medical researchers evaluate their methodological skills

BACKGROUND AND OBJECTIVE: Clinical epidemiology and statistics skills of clinical researchers are often limited. We assessed methodological skills of medical researchers and identified factors associated with higher skill levels. METHODS: In a cross-sectional mail survey at two Swiss teaching hospitals, participants (N=409) rated their ability to perform 26 research-related activities, such as identifying the research question, selecting a study design, computing the required sample size, performing data analysis, and reporting results. RESULTS: The proportion of respondents who were able to perform a specific activity was 33.2% on average, ranging from 1.5% for "numerical statistics (bootstrap, simulation, cross-validation,...)" to 76.0% for "oral presentation of results." The overall skill level (expressed as a percentage of the 26 activities) was associated with principal investigator experience (+8.7%), greater percentage of time devoted to research (+12.4% for near full-time versus no time commitment), years of research experience (+17.6% for 15-40 years versus 0 years), past number of clinical research projects (+18.0% for 15-230 projects versus 0-1 projects), and hours of formal methodological training (+32.6% for 200-1200 hours versus 0-9 hours). CONCLUSION: Self-reported methodological skills were generally modest. The most important covariates of skill levels were current time commitment to research, past experience, and formal training