The commentaries highlight a number of directions for future research and were unanimous that the role of alcohol in recovery from drug use is under-researched . We concur with commentators that the issue is complex, and we look forward to a more focused approach in future research on this topic. A number of specific issues were raised, to which we respond below.
Witkiewitz  raises an important point regarding the casual role that alcohol may play in relapse to drug use. That is, although our review found some evidence that alcohol use was associated with relapse, we agree with Witkiewitz that this could be due to alcohol use co-occurring with other risk factors underlying both alcohol and drug use, and that research evidence is insufficient to distinguish precise causal relationships. We encourage researchers to collect new data accordingly or to re-visit old data that might be useful to shed some light on this. In particular, collection of moment-to-moment data or detailed time-line follow-back data might address this question specifically. On a related point, Ross  notes that psychiatric comorbidity may also influence relapse. This is an important question to address, especially as it may influence the nature and presence of ‘risk factors’. For example, an individual with comorbid trauma issues may drink alcohol to cope with increasing their chances of relapse to illicit drug use . Clearly, the potential causal role of alcohol in relapse is complex, and multiple factors are likely to co-contribute.
Witkiewitz  questions the usefulness of conceptualizing drugs of addiction as ‘primary’ and ‘secondary’. If alcohol does cause relapse to drug use (at least for some people) and if the reality of drug use is polydrug use, then such a framework may encourage treatment providers to focus exclusively on the treatment of just one substance, ignoring these complexities. Indeed, we have seen examples of this in the treatment literature where treatments have focused on amphetamine or cocaine users exclusively. However, we know that, for most, polydrug use is the norm. Conversely, as noted by Gossop , people often use different substances for different reasons and hence we argue that these unique factors related to a ‘primary’ drug of choice still require consideration in treatment. Consistent with Witkiewitz's  view, we have argued that a person-centred analysis of patterns of drug use will help to identify the complexities underlying concurrent alcohol and drug use.
Finally, Roozen  raises some important treatment implications highlighting the importance of finding alternative rewarding activities consistent with our previous work utilizing a personality framework to understand and treat addiction [7-10]. Our work in this area calls for the need to address the role of impulsivity, which is consistent with the suggestion by Ross  that mindfulness techniques may be well suited to the treatment of illicit drug use. Indeed, transdiagnostic approaches that address ‘core’ issues underlying addiction are a promising way forward in drug treatment.
The responses to our paper highlight some of the challenges for future research regarding the role of alcohol in relapse to drug use. We propose that specific analysis of longitudinal data and a treatment approach that takes into account the key personality factors which contribute to drug use will address some of the current limitations in the literature.