In our application, we stated that we would continue without modifications to a full trial if the following success criteria were met:
- at least 50% of eligible participants can be recruited within a 4-week period
- no more than 5% of all recruited participants crossover from one arm to the other
- at least 50% of all recruited participants complete follow-up assessments.
If we did not meet these success criteria, we planned to assess the reasons for this and consider how these could be improved in a main trial. We stated that, if less than 20% of participants complete follow-up assessments, we would consider the main trial unfeasible and would not proceed with a future larger trial application.
Success criterion 1 was met; we recruited 52% of all eligible participants. Although we extended the recruitment period slightly to 6 weeks at some centres, recruitment was completed within the originally specified 5-month period and this target was met despite some researchers working part-time. More participants could have been recruited at some centres (mainly centre 1) if we had had the resources. Success criterion 2 was not met; 67% of UC participants reported any use of an EC during the study period. Unfortunately, we did not ask about extent of use, so whether or not it was regular, sustained use remains unclear. Although some EC use was reported, none of the UC participants received the EC intervention as delivered (i.e. a free EC starter kit offered at homeless centres). Contamination across arms can be a problem for many RCTs; however, supported use of an EC is increasingly being incorporated into the ‘usual care’ offered by SSSs and should not, therefore, invalidate our findings. The pragmatic question that will be explored in a main trial (and for which this feasibility study sought to provide preliminary evidence) will be whether or not the provision of a free EC starter kit offered at a location already being accessed by homeless smokers can increase smoking cessation rates over UC. Success criterion 3 was fully met; we exceeded 50% retention at each and every follow-up point. However, we note that retention did drop below 50% in the UC arm because of poor retention at the Edinburgh site. Given that some of our success criteria were not fully met, we will include an inbuilt pilot study with clear stop–go criteria in our main cRCT.