Randomised controlled non-inferiority trial of primary care-based facilitated access to an alcohol reduction website

Brief interventions (BIs) delivered in primary care have been shown to be effective in reducing risky drinking, but implementation is limited. Facilitated access to a digital application offers a novel alternative to face-to-face intervention, but its relative effectiveness is unknown.

Primary care-based, non-inferiority, randomised controlled trial comparing general practitioner (GP) facilitated access to an interactive alcohol reduction website (FA) with face-to-face BI for risky drinking. Patients screening positive on the short Alcohol Use Disorders Identification Test (AUDIT-C) were invited to participate in the trial. Assessment at baseline, 3 months and 12 months was carried out using AUDIT and EQ-5D-5L questionnaires.

58 participating GPs approached 9080 patients of whom 4529 (49.9%) logged on, 3841 (84.8%) undertook screening, 822 (21.4%) screened positive and 763 (19·9%) were recruited. 347 (45.5%) were allocated to FA and 416 (54.5%) to BI. At 3 months, subjects in FA group with an AUDIT score of ≥8 reduced from 95 (27.5%) to 85 (26.8%) while those in BI group increased from 123 (20.6%) to 141 (37%). Differences between groups were principally due to responses to AUDIT question 10. Analysis of primary outcome indicated non-inferiority of FA compared with BI, and prespecified subgroup analysis indicated benefits for older patients and those with higher levels of computer literacy and lower baseline severity. Additional analyses undertaken to take account of bias in response to AUDIT question 10 failed to support non-inferiority within the prespecified 10% boundary.

Prespecified protocol-driven analyses of the trial indicate that FA is non-inferior to BI; however, identified bias in the outcome measure and further supportive analyses question the robustness of this finding. It is therefore not possible to draw firm conclusions from this trial, and further research is needed to determine whether the findings can be replicated using more robust outcome measures.

NCT01638338; Results.