Concerns regarding interpretation and reporting in Kolivas et al. mHealth low carbohydrate dietary intervention in T2D

Kolivas et al. recently reported the outcomes of a three-month mHealth low carbohydrate dietary intervention for people with type 2 diabetes (T2D) 1 .Given the limited evidence for digital dietary interventions for T2D, particularly in the Australian context, this is a commendable contribution to an important field.However, several aspects of data interpretation and reporting require clarification or correction.At three months, this single-arm pre-post intervention demonstrated significant reductions in haemoglobin A1c (HbA1c) (-1.0%), fasting plasma glucose (FPG) (-1.3 mmol/L) and systolic blood pressure (-6 mmHg), with no significant changes in diastolic blood pressure, body weight, body mass index (BMI) or waist circumference.In contextualising the findings, the authors cite the T2Diet Study 2 , the only randomised controlled trial of a digital low carbohydrate dietary program for T2D conducted in Australia.They accurately reported the between-group differences of the 16-week intervention-HbA1c -0.6%, body weight -3.3 kg, BMI -1.1 kg/m, and reduced medication requirements-but then misrepresent these results as within-group changes, stating that "while the average reduction in HbA1c in the T2Diet study (-0.6%) was less than the improvement recorded in the current trial (-1.0%)" (p. 6).This comparison understates the T2Diet findings and exaggerates the effect of the current intervention.In fact, the T2Diet within-group changes were HbA1c -0.94%, weight -4.36 kg, and BMI -1.48 kg/m-all statistically significant.Distinguishing between-group differences from within-group changes is essential for fair comparison and accurate scientific and clinical interpretation 3,4 .Being a single-arm trial reporting within-group changes, the manuscript should clearly reflect this distinction to prevent misrepresentation of relative intervention effects.In several parts of the manuscript, the authors present statistically nonsignificant findings as significant, creating potential for misinterpretation.For example, HbA1c changes between predefined carbohydrate intake groups (50 g versus >50 g/day) were assessed.The results on adherence (p.4) reported a non-significant HbA1c reduction of -0.5% (95% CI: -1.0 to 0.0; p = 0.073; Supplementary Table S3) 1 , suggesting no association between the level of carbohydrate intake and HbA1c reduction in the cohort.However, the report then contradicts this finding by stating that "participants who reduced their HbA1csignificantly reduced their dietary carbohydrates when grouped by intake of 50 g a day or less(p = 0.036)"with no additional supporting data.Likewise, multiple statements in the abstract (p.1), discussion (p.5), limitations and conclusion (p.8), described this outcome as a significant finding, implying that participants who