Statistical Errors in Medical Journals (A Critical Appraisal)

Abstract

Background:  Importance of Statistics and its appro-priate appliance in every field of life are inevitable. Research conducted in all sciences, particularly in medicine stipulates the usage of statistics to make the results unswerving and authentic. Any research, how so ever expensive and hefty, has no worth and fails to accomplish the tasks for which it is conducted, if sta-tistics involved in it contains error. It has been attested now, that many articles published in medical journals restrain errors in them. Particularly in Pakistan, the use of statistics is not given the suitable worth, resulting many statistical flaws. Thus it has become the need of hour to investigate these errors, highlight and conse-quently make an effort to eradicate them from medical research. This study has been designed to scrutinize such errors in local medical journals of Pakistan.

Study Design:  Cross - sectional study design was used.

Methodology:  80 research articles published in in-dexed and recognized local journals of Pakistan were reviewed. The selection of those research articles was random and those were easily and freely available on-line. Case series and case reports were excluded.

Results:  Descriptive statistics was only used in 32 (28.75%), inferential statistics along with the descript-tive statistics was used in 33 (41.25%) articles. Design of a study was not mentioned in 42 (52.5%) articles, in 74 (92.5%) articles there was no criteria for sample size estimation and power calculation. Wrong statis-tical analysis was done in 23 (28.75%) articles. Out of 80 articles evaluated, 57 (71.2%) articles used infe-rential statistics but did not mention whether the assu-mptions for statistical test used, met or not.

Conclusion:  In order to endorse medical research with authentication and reliability, the role of biosta-tistician needs to be divulged. Thus biostatistician sho-uld be taken in editorial boards or articles should be referred by the biostatisticians or statisticians for app-roval in order to keep the standards of research upgra-ded.

Keywords:  Statistical error, biomedical research and Statistical methods.