Statistical Issues in Malaria Research in Southeast Asia: Systematic Analysis

Abstract

This study systematically analyzed the contribution of statistical misuses in the three main malaria research categories published during 2010-2015 from Southeast Asian countries. Related articles were downloaded from the PubMed and Scopus databases using the predefined keywords. The retrieved articles were checked for compliance with the eligibility criteria and stored in EndNote version X7. Finally, 524 articles were included in the analysis. Most of the research involved was conducted in Thailand (263 articles, 50.19%). Intervention and health care research (n =252, 48.09%) was the primary research category. It was also applied to descriptive and inferential statistics (63.49%). The non-parametric test was the most applied statistical analysis approach (28.63%). Only obvious errors were reported as improper use in this article and analyzed based on the available information. Improper use/misuse of statistics was highest in the intervention and health care research category (47.10%). Results of the systematic analysis revealed a high frequency of inappropriate uses/misuse of statistics. In most articles, there was no rationale for selecting the sample sizes. Sampling issues could be considered the main problem of all the malaria research categories (43.15%). The reliability of the articles and the possibility of incorrect interpretations and conclusions should be noted. To decrease the number of errors stemming from inappropriate uses/misuse of statistical analysis, a clear study protocol and statistical analysis plan should be prepared before a study is conducted, and should include the consultation of statisticians.