Rejoinder relating to correspondence by Dr Eleanor Woodward, Professor Robyn Richmond on Brose LS, Simonavicius E, McNeill A (2018). Maintaining abstinence from smoking after a period of enforced abstinence

We thank Dr Woodward and Professor Richmond for their commentary on our literature review and meta-analysis of interventions to maintain abstinence from smoking after a period of enforced abstinence (Brose et al., 2018). We endorse their suggestion to introduce a measure of compliance in settings with smoke-free policies and would like to comment further on how the reduction of cigarette consumption in smokers who undergo a period of enforced abstinence can be utilised in future research and practice.

Woodward and Richmond rightly highlight poor compliance with smoke-free policies across some mental health and prison settings. We had restricted our review to settings with comprehensive smoke-free policies, no access to smoking areas and at least minimal cessation support offered to smokers. Nevertheless, a few studies (Joseph, 1993; Gariti et al., 2002) reported that continued smoking was possible and indeed was happening; other studies in the review did not report on this, and levels of compliance were unclear. However, we do not think that non-compliance with smoke-free policies led to an underestimate of the effectiveness of interventions assessed in our meta-analysis. Of the 10 studies in the review, seven recruited participants for the control and intervention groups from the same wards or units so that the groups likely had similar non-compliance levels, therefore, between-group differences at follow-up should have not been confounded by non-compliance and were likely due to the effects of the study interventions only.

Despite lack of focus and reporting on non-compliance in the included studies, it is an important issue undermining goals of smoke-free policy. As Woodward and Richmond note, the modest provision of cessation support can increase the non-compliance but investing in staff training and providing better support could alleviate this. In England, the National Institute for Health and Care Excellence (NICE) guidelines highlight that comprehensive smoke-free policies across National Health Service (NHS) secondary care settings have to be accompanied by extensive training for staff and evidence-based cessation and abstinence support for both service users and staff members who smoke (2013). Current smoke-free inpatient services and future studies assessing smoking cessation support in these settings should consider the extent of compliance; a standardised verified measure for compliance as suggested by Woodward and Richmond would be useful.

Woodward and Richmond also comment on the relationship between smoking reduction post-discharge, compensatory smoking and the impact of the reduction on participants’ health. Our primary outcome was biochemically verified continuous abstinence while the reduction in cigarettes smoked per day was one of eight secondary outcomes. We agree that compensatory smoking is an important confounder in studies where smokers switch to lower yield tobacco products (Scherer and Lee, 2014) or when their smoking is restricted and reduced (Hughes and Carpenter, 2005); these conditions, however, did not apply to participants of our reviewed studies who were already discharged from a smoke-free environment. We believe that reductions in cigarette consumption can be an important step to smoking cessation (Hughes and Carpenter, 2005) and for some can result in substantial health gains. Woodward and Richmond's suggested biochemical verification of smoke exposure would be an improvement, but, in terms of health impacts, longer follow-ups and biochemical verification of primary outcome measures (i.e. continuous abstinence) should be a priority.

We thank Woodward and Richmond for their interest in our work. Whilst improved smoking cessation support in in-patient settings is essential, smoke-free units and wards seldom support smokers beyond discharge. Our review suggests that extending medication or behavioural support post discharge can help to leverage any gains achieved during a period of enforced abstinence and reduce overall inequalities in harm and mortality among one of the most disadvantaged groups of smokers.