Effects of service-wide support on regularity of alcohol screening of clients in Australian Aboriginal and Torres Strait Islander Community Controlled Health Services: a cluster randomised trial

Background: We have previously shown that service-wide support can increase the odds of alcohol screening in any 2-month period in a cluster randomized trial of service-wide support to Aboriginal and Torres Strait Islander Community Controlled Health Services (ACCHS). Here we report an exploratory analysis on whether the resulting pattern of screening was appropriate. Aim: we assess whether that increase in screening was associated with: (i) increased first-time screening, (ii) increased annual screening, (iii) whether frequently screened clients fell into one of four risk categories as defined by national guidelines. Methods: Setting and participants: 22 ACCHS; randomized to receive the support model in the treatment (‘early-support’) arm over 24-months or to the waitlist control arm. Intervention: eight-component support, including training, sharing of experience, audit-and-feedback and resource support. Analysis: records of clients with visits before and after start of implementation were included. Multilevel logistic modelling was used to compare (i) the odds of previously unscreened clients receiving an AUDIT-C screen, (ii) odds of clients being screened with AUDIT-C at least once annually. We describe the characteristics of a sub-cohort of clients who received four or more screens annually, including if they were in a high-risk category. Results: Of the original trial sample, 43,054 met inclusion criteria, accounting for 81.7% of the screening events in the overall trial. The support did not significantly increase the odds of first-time screening (OR = 1.33, 95% CI 0.81–2.18, p = 0.25) or of annual screening (OR = 0.99, 95% CI 0.42–2.37, p = 0.98). Screening more than once annually occurred in 6240 clients. Of the 841 clients with four or more screens annually, over 50% did not fall into a high-risk category. Females were overrepresented. More males than females fell into high-risk categories. Conclusion: The significant increase in odds of screening observed in the main trial did not translate to significant improvement in first-time or annual screening following implementation of support. This appeared to be due to some clients being screened more frequently than annually, while more than half remained unscreened. Further strategies to improve alcohol screening should focus on appropriate screening regularity as well as overall rates, to ensure clinically useful information about alcohol consumption. Trial Registration ACTRN12618001892202, retrospectively registered 16 November 2018 https://anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12618001892202.