Clinical decision making in nutrition and dietetics requires an ability to understand research evidence, and use these findings to inform practice decisions [1]. Understanding the magnitude (size) of treatment effect (the difference in outcomes between intervention and control groups), is essential to ensure practitioners, health professionals, and researchers can interpret findings and make informed decisions to guide clinical practice. The aim of this study was to explore the correct understanding and perceived confidence and usefulness of five different data presentation approaches to express the magnitude of effect among Accredited Practising Dietitians.
A voluntary online survey was conducted via Qualtrics in Australian Accredited Practising Dietitians. Participants were provided with summary estimates of hypothetical dietary interventions vs. usual diet for reducing myocardial infarction risk and were asked questions on their understanding of five data presentation approaches (two relative i.e., relative risk [RR], odds ratio [OR]; three absolute i.e., risk difference [RD], number needed to treat [NNT], risk in treatment vs. control groups) and perceived confidence and usefulness of the approaches for understanding the magnitude of effect. Participants were randomly assigned to either a trivial or large treatment effect. Descriptive analyses were conducted using Microsoft Excel. Correct understanding of the magnitude of effect from different approaches was determined by calculating the number of correct answers, and perceived confidence and usefulness were reported as means and standard deviations.
Twenty-six respondents were included in the analyses. Respondents primarily worked in clinical nutrition (31%), research (23%), or private practice (15%), with 18 (69%) having previously completed epidemiology or biostatistics courses. Nine (35%) had completed an evidence based practice course or subject as part of their training, while six (23%) had participated in evidence based practice focused continuing professional development. Less than 20% of participants understood the magnitude of effect correctly for RR (12%), OR (14%), and RD (15%), with a higher proportion of participants selecting the correct answer for risk in treatment vs. control groups (38%) and NNT (29%). Participants reported the highest confidence in interpreting effect sizes when using RR (4.2+1.2), but reported highest perceived usefulness for OR (4.7+1.2).
Although many participants had completed training in epidemiology, biostatistics, or evidence-based practice, understanding of the magnitude of effect was low, particularly for RR, OR and RD. Participants understood risk in treatment vs. control groups, an absolute estimate, the most but reported the highest confidence in using RR and highest perceived usefulness for OR. This may reflect the common use of relative estimates in health science literature. While this study is limited by its small sample size, these findings highlight a need for further investigation of understanding of magnitude of effect in Accredited Practising Dietitians, with further training on interpreting absolute effects potentially needed.