Evaluation of statistical methods applied in observational studies in neurology

Introduction/Objective: Statistical methods play a key role in the planning, analysis, and interpretation of results in medical research. Due to the complexity of neurological diseases and the heterogeneity of study populations, the proper application of statistical methods in neurology is particularly important for drawing reliable clinical conclusions. This study aims to evaluate the application of statistical methods in observational studies in neurology, with special emphasis on the frequency of the use of different statistical techniques, the adequacy of their application, and the interpretation of results. Materials and methods: A literature search of the MEDLINE/PubMed database was conducted for all publications available up to December 1, 2024, using the term "neurology" with the application of the filter "observational study." After applying the inclusion and exclusion criteria, 95 observational studies published between 2013 and 2024 were randomly selected for analysis. Two researchers independently extracted the data using a previously designed data collection form. Results: A retrospective design was identified in 53.7% of studies, while 46.3% were prospective. Ethical approval was reported in 78.9% of studies. Assessment of the normality of distribution was presented in 29.5% of studies. The most commonly used statistical tests were the chi-square test (46.3%), t-test (36.8%), Fisher's exact test (36.8%), and the Mann-Whitney test (33.7%). Regression analysis was applied in 33.7% of studies. Sample size planning was clearly reported in only 10.5% of studies. The application of the t-test was statistically significantly more frequent in studies that reported a previous assessment of the normality of distribution (p < 0.001). Conclusion: The results indicate that statistical methods in observational neurological studies are generally applied adequately and in accordance with data characteristics. However, greater attention should be devoted to testing statistical assumptions and proper sample size planning in order to further improve the quality and reliability of future research.