Commentary on Hyland et al.: Digital treatment for alcohol use disorders - challenges and opportunities.

Abstract

Hyland and colleagues found no significant effects of an internet-delivered cognitive behavioral therapy intervention (iCBT) as addendum to treatment as usual in individuals with moderate alcohol use disorders. While findings can be attributed partially to methodological issues, future studies should focus upon patient preferences and include potential side effects of internet-based interventions. In the field of problematic alcohol consumption, digital interventions have shown promise in non-clinical populations [1]; however, little is known about their efficacy in clinical populations fulfilling criteria for alcohol use disorders (AUDs). Internet-delivered treatment has rapidly evolved as an option for various mental disorders. First meta-analyses show positive effects when compared to non-active controls in disorders such as depression and agoraphobia [2-4], but also for addictive disorders, especially behavioural addictions [5] and tobacco smoking [6]. Given that disorders due to alcohol and/or drugs are highly stigmatized, internet-delivered interventions might be a relevant tool for narrowing the treatment gap in the field. In their excellently conducted randomized controlled trial (RCT), Hyland and colleagues [7] found only weak evidence for an internet-delivered cognitive behavioral therapy (iCBT) intervention as addendum to treatment as usual (TAU) in individuals with AUDs compared to TAU. However, it has to be taken into account that, as comparator in this study, TAU included far more evidence-based components than would usually be expected in most treatment systems world-wide. The high rate of participants willing to take (and receiving) medication is particularly impressive: data from the United States show that only 16.3% of all speciality treatment programmes for alcohol use disorders offered any single medication for AUD treatment [8]. Furthermore, as all participants in the Hyland et al. study were self-selected, the initial motivation to change already might have produced strong effects, especially if we consider the high rates of unassisted recovery that are common in alcohol use disorders [9]. Another encouraging finding is that overall treatment initiation and completion rates are relatively high compared to digital self-help interventions in the field of affective and anxiety disorders [10], where efficacy of these interventions has already been demonstrated. There is no evidence that guided or unguided internet-delivered interventions for AUDs are associated with lower rates of treatment uptake compared to face-to-face delivery. A prior study from the research group of Hyland et al. on a therapist-guided iCBT intervention compared to therapist-delivered CBT showed no differences in uptake numbers between the two conditions, and both interventions did not differ on outcomes [11]. Another RCT on iCBT as a stand-alone intervention for individuals with AUDs [12] even showed that retention rates were better in the iCBT condition compared to therapist-delivered CBT, and that both conditions were superior compared to TAU (which included lower doses of treatment compared to the Hyland et al. study). Given that internet-delivered treatment requires a high level of patient engagement, future studies should take into account the interaction between type of treatment and patient preferences. Internet-delivered interventions might be a good option for some, but certainly not for all, patients suffering from AUDs. A systematic review on effects of matching patient preferences to treatment in the field of substance use disorders suggests that treatment effects might be improved, although findings were inconsistent [13]. Patient preferences have become increasingly important for the development of clinical interventions, and they could play an important role for improving treatment uptake and outcomes by affected individuals [14]. Finally, only recently have studies on the efficacy of treatment assessed potential side effects. A potential risk of digital interventions is that these interventions might incorrectly be viewed as sufficient by some individuals who otherwise would be willing to engage in other evidence-based treatments. The significant difference between the conditions in the Hyland et al. study regarding receipt of medication shows the need for further rigorous, well-conducted studies. For a more detailed understanding on processes of change in newly developed interventions, mixed-methods studies that include the experiences of patients are warranted. GB conceptualized and wrote the manuscript. None.