Effectiveness of an internet-delivered cognitive-behavioural aid to improve mental wellbeing: a randomised controlled trial

Abstract

BackgroundInterventions to promote mental wellbeing can bring benefits both to the individual, improving mood and psychological functioning, and to society, in terms of economic prosperity and social cohesion. An approach using the principles of cognitive behavioural therapy to encourage healthier patterns of thinking and behaviour could offer an individual-level intervention to promote positive mental health. The internet allows large-scale and low-cost delivery of health-promoting interventions. We assessed the effectiveness of a self-directed internet-delivered cognitive behavioural skills training aid in improving mental wellbeing in the general population. MethodsWe undertook a randomised trial with a waiting-list control. Using adverts on a national health portal (NHS Choices) and through its mailing list, we recruited 3070 participants aged 18 years or older, resident in England, willing to give their email address, and with access to a fully automated web-based intervention. Recruitment took place during 2 weeks in September, 2010, with final follow-up in December, 2010. Participants were randomly assigned to intervention or control (1:1) with automated software that was integrated into the online trial administration system (block size of 2). The intervention was the self-directed internet-based cognitive behavioural skills programme MoodGYM. Participants in the intervention group received email reminders every week to log into the trial portal to access the intervention. The comparator was a control group on a waiting list who received access to the intervention after the trial was completed. Participants were not masked to assignment. During the trial, the control participants did not receive any specific intervention or email reminders. The primary outcome was mental wellbeing measured with the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS). Secondary outcomes were CES-D depression scores, GAD-7 anxiety scores, EQ5D quality of life scores, physical activity, and health service use. All outcomes were measured at baseline, 6 weeks, and 12 weeks. Findings1529 (50%) of 3070 participants completed final follow-up at 12 weeks (1123 of 1536 in the control group and 406 of 1534 in the intervention group). No relation between baseline measures and withdrawal was noted. The analysis of WEMWBS scores with a linear mixed model for repeated measures showed no difference between intervention and control group at baseline (difference −0·124 points, 95% CI −0·814 to 0·566), and significant improvements (p<0·0001 for both) for the intervention group at 6 weeks (2·542 points, 1·693 to 3·390) and at 12 weeks (2·876 points, 1·933 to 3·819). The model showed a highly significant (p<0·0001) intervention by time interaction effect. There were also significant improvements in self-rated scores of depression and anxiety. In view of the attrition rate, a sensitivity analysis with imputed missing values was undertaken, which also showed a significant positive effect (intervention by time interaction effect p<0·001) of the intervention. InterpretationParticipants assigned to the internet-based cognitive behavioural therapy aid MoodGYM had an average increase of roughly 3 points on the WEMWBS mental wellbeing scale compared with no increase for participants on a waiting list. This difference on the WEMWBS scale is about a third of a SD (SD in general population, 8·3). In a low-cost automated public health intervention designed to shift the whole population distribution of mental wellbeing, a small difference per individual could bring a major benefit in population terms. As to limitations, we undertook an unmasked waiting-list controlled study with no placebo; furthermore (as with other web-based interventions), there was high attrition, particularly in the intervention group. This high attrition is perhaps of less importance in a health promotion intervention in a non-clinical population. FundingNHS Choices.