Statistics in the ophthalmic literature and the alignment with knowledge, attitudes and practices towards statistics.

Abstract

Clinical relevance: The Entry-Level Competency Standards for Optometry 2022 explicitly state that optometrists must be able to 'find, appraise, analyse, interpret and integrate the best available research to inform clinical reasoning and professional decision-making'.

Background: To enumerate the statistical methods used in the ophthalmic literature and to measure the Knowledge, Attitude, and Practice (KAP) towards statistics in Australian optometrists. A secondary aim was to characterise statistical accessibility.

Methods: The most frequently used statistical methods in articles published in four highly ranked ophthalmic journals over a 12-month period were identified. Two reviewers independently audited articles and assigned analyses to 37 pre-defined categories. A cross-sectional 35-item purposely designed KAP survey was distributed to a random sample of registered Australian practitioners. Descriptive statistics, chi-square tests and odds ratios characterised the statistics used in the ophthalmic literature. KAP results were analysed using Cronbach's alpha and regression (linear, ordinal).

Results: The most frequently used category was descriptive statistics, followed by t-tests, contingency tables, nonparametric tests, other, and ANOVA and 77% of articles could be understood with knowledge of the top 10 statistical methods. Ninety-two of 706 (response rate 13%) optometrists (mean age 44 ± 12 years, 54% female) returned the survey. Optometrists demonstrated moderate knowledge (44% ± 20%) and attitude (47% ± 15%) and a poor practice (31% ± 31%) towards statistics. Females had a lower attitude score, practice, and total scores than males (all p ≤ 0.02). Optometrists had best knowledge of T-test yet 75% had a little or no confidence in 'the knowledge and skills required to interpret all inferential statistics'.

Conclusion: Optometrists must understand descriptive statistics and some basic inferential statistics to critically appraise the literature. The KAP can pinpoint gaps and indicate where and how education should be targeted. The validated KAP is available for use in other settings.