Article Text

PDF
Statistical approaches in published ophthalmic clinical science papers: a comparison to statistical practice two decades ago
  1. Harrison G Zhang1,
  2. Gui-shuang Ying2
  1. 1Conestoga High School, Berwyn, Pennsylvania, USA
  2. 2Department of Ophthalmology, Center for Preventive Ophthalmology and Biostatistics, The University of Pennsylvania, Philadelphia, Pennsylvania, USA
  1. Correspondence to Dr Gui-shuang Ying, Center for Preventive Ophthalmology and Biostatistics, Department of Ophthalmology, The University of Pennsylvania, Philadelphia, PA 19104, USA; gsying{at}pennmedicine.upenn.edu

Abstract

The aim of this study is to evaluate the current practice of statistical analysis of eye data in clinical science papers published in British Journal of Ophthalmology (BJO) and to determine whether the practice of statistical analysis has improved in the past two decades. All clinical science papers (n=125) published in BJO in January–June 2017 were reviewed for their statistical analysis approaches for analysing primary ocular measure. We compared our findings to the results from a previous paper that reviewed BJO papers in 1995. Of 112 papers eligible for analysis, half of the studies analysed the data at an individual level because of the nature of observation, 16 (14%) studies analysed data from one eye only, 36 (32%) studies analysed data from both eyes at ocular level, one study (1%) analysed the overall summary of ocular finding per individual and three (3%) studies used the paired comparison. Among studies with data available from both eyes, 50 (89%) of 56 papers in 2017 did not analyse data from both eyes or ignored the intereye correlation, as compared with in 60 (90%) of 67 papers in 1995 (P=0.96). Among studies that analysed data from both eyes at an ocular level, 33 (92%) of 36 studies completely ignored the intereye correlation in 2017, as compared with in 16 (89%) of 18 studies in 1995 (P=0.40). A majority of studies did not analyse the data properly when data from both eyes were available. The practice of statistical analysis did not improve in the past two decades. Collaborative efforts should be made in the vision research community to improve the practice of statistical analysis for ocular data.

  • epidemiology

Statistics from Altmetric.com

Footnotes

  • Contributors HGZ and GsY contributed to study design, collecting the data and analysed the data. HGZ interpreted the results and drafted the manuscript. GsY critically revised the manuscript.

  • Funding This work is partially supported by the Foundation for Fighting Blindness, the Mackall Trust Funds and P30EY01583-26.

  • Competing interests None declared.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.