Study design, research data and conclusions

Research output: Contribution to journalArticle

Abstract

Not every research paper is worth the time taken to read it. To be of value to the reader, the research question needs to be not only in the reader's field of interest but should also have been addressed using rigorous methodology. Further, absolute care should have been taken in collection of data or outcome assessment to ensure reliability (reproducibility) and validity (accuracy). Unless these precautions are taken, the conclusions of the study may be more speculative than factual, thus misleading the reader with dire consequences at times. Although a randomised controlled trial may be considered to be the" yard-stick” for analytical research in medicine, it is at times not feasible or ethical to use such a design depending on the research question being asked. Even case reports and case series have an important part to play in health research; particularly in the generation of hypothesis. While ecologic studies and surveys may be very useful to the health care planners, conclusions from such studies should not be viewed as conclusive where evidence for causal association or effectiveness is sought. The latter remains the preserve of cohort/case control studies and randomised controlled trials respectively. Since the reader is often faced with publications loaded with methodological flaws despite peer review, the importance of critical appraisal skills must always be emphasised.

Original languageUndefined/Unknown
JournalPaediatrics and Child Health, East Africa
Publication statusPublished - 1 Jan 1998

Cite this