Sebastian Potthoff ^1,2* Håvar Brendryen ^1,3 Haris Bosnic ^1,4 Anne Lill Mjølhus Njå ¹ Tracy Finch ⁵ Torgeir Lid ^1,6,7

• ¹ Centre for Alcohol and Drug Research, Stavanger University Hospital, Stavanger, Norway
• ² Northumbria University, Newcastle upon Tyne, United Kingdom
• ³ Department of Psychology, Faculty of Social Sciences, University of Oslo, Oslo, Oslo, Norway
• ⁴ Norwegian Reading Centre, Faculty of Arts and Education, University of Stavanger, Stavanger, Norway
• ⁵ Department of Nursing, Midwifery and Health, Northumbria University, Newcastle upon Tyne, United Kingdom
• ⁶ Faculty of Health Sciences, University of Stavanger, Stavanger, Norway
• ⁷ Research Unit for General Practice in Bergen, Norwegian Research Institute (NORCE), Bergen, Hordaland, Norway

Excessive alcohol consumption is a leading global risk factor for ill-health and premature death. Digital alcohol interventions can be effective at reducing alcohol consumption, but their widespread adoption is lagging behind. This study aimed to identify factors promoting or inhibiting the implementation of a digital alcohol intervention in Norwegian primary care, by using Normalization Process Theory (NPT). A mixed methods feasibility study combining quantitative and qualitative methods. A digital alcohol intervention called 'Endre' was implemented across four GP practices in Stavanger and Oslo. Usage of the intervention was logged on the digital platform. General practitioners (GPs) reported their perceived uptake of the intervention via a web-based survey. The Normalization MeAsure Development (NoMAD) survey was used to measure support staff's perceived normalization of the intervention. Qualitative data were analyzed using the NPT framework, with quantitative data analyzed descriptively and using χ2 and Wilcoxon signed-rank test for differences in current and future normalization.Results: Thirty-seven GPs worked in the clinics and could recruit patients for the digital intervention. Thirty-six patients registered for the intervention. Nine patients dropped out early and 25 completed the intervention as intended. Low normalization scores at follow-up (n = 27) indicated that Endre did not become fully embedded in and across practices. Nonetheless, staff felt somewhat confident about their use of Endre and thought it may become a more integral part of their work in the future. Findings from six semi-structured group interviews suggested that limited implementation success may have been due to a lack of tailored implementation support, staff's lack of involvement, their diminished trust in Endre, and a lack of feedback on intervention usage. The outbreak of the Covid-19 pandemic further limited opportunities for GPs to use Endre.This study investigated the real-world challenges of implementing a digital alcohol intervention in routine clinical practice. Future research should involve support staff in both the development and implementation of digital solutions to maximize compatibility with professional workflows and needs. Integration of digital solutions may further be improved by including features such as dashboards that enable clinicians to access and monitor patient progress and self-reported outcomes.