Online learning improves substance use care in Kenya: randomized control trial results and implications

Abstract

Background

Alcohol use is the 5th most important risk factor driving the global burden of diseases. WHO identifies a lack of health worker training as one of the main barriers to providing cost-effective brief interventions for alcohol use disorder. This study assesses the impact of online training, using the NextGenU.org model, on the delivery of the WHO Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) and its linked brief intervention (BI). Methods

A randomized control trial (RCT) was performed in two Kenyan counties (ClinicalTrial.gov ID:NCT02388243, ethics from UBC and KEMRI) aiming to recruit 570 patients (sufficient for 80% power to detect a similar difference in alcohol consumption found in past trials, with 30% lost to follow-up). The primary outcome was decreases in alcohol consumption in the last 7 days at baseline, one, three, and six months' follow-up comparing the two trial arms (intention-to-treat analysis with multiple imputation for missing data). Adults presenting to the eight participating facilities were invited to take a lifestyle questionnaire including the ASSIST, and to receive verbal feedback plus written advice from a community health worker (CHW). Those consuming alcohol at moderate or high risk were offered to enroll in the RCT. After obtaining written consent, they were randomized to no further intervention (NFI), or to receive a BI delivered by a nurse. Findings

Of the 696 participants, 91% were male, the average age was 38yo, with the majority having completed only primary school. The average alcohol consumption at baseline was 400 gr in the BI group and 413gr in the NFI group, at one month it decreased respectively by 183gr, and 217 gr; T-test showed a statistically significant decrease in both groups overtime, while the difference between the groups was not significant. Interpretation

CHW trained online to deliver feedback with the ASSIST can help those with moderate to high risk level of alcohol consumption to reduce their consumption as much as those who received a full brief intervention, and both groups decreased their consumption more than observed in the Cochrane Alcohol BI meta-analysis. Funding

Grand Challenges Canada, Annenberg Physician Training Program in Addiction Medicine.