Discussion

These pilot-trial findings strongly suggest that proactive telephone smoking cessation counselling using a motivation enhancing EPPM-based protocol is feasible to integrate into lung cancer screening care delivery, even by a provider with whom the patient has not had a prior relationship. The intervention was well received by participants and was associated with promising increases in utilisation of behavioural cessation counselling. Considering that 40% of Veterans invited to participate in the assessment-only control group agreed to do so, the 28% acceptance rate for the intervention activities suggests that a majority of Veteran smokers who are amenable to research participation were willing to accept a proactively delivered smoking cessation intervention. As a feasibility study, randomisation was not necessary and the scope of the study was not powered to detect differences in gold standard cessation outcomes (e.g. biochemically confirmed abstinence rates). Pragmatic trials, in which patients are randomised between two or more forms of usual care activities, are important in this context to inform delivery systems about real-world effectiveness of alternative approaches. These findings are important because they provide necessary justification for a pragmatic trial to evaluate the effectiveness of proactive telephone counselling routinely integrated into care.

There have been few studies of interventions to improve quit rates in the lung cancer screening context, primarily resulting in null findings (Clark et al., Reference Clark, Cox, Jett, Patten, Schroeder and Nirelli2004; Ferketich et al., Reference Ferketich, Otterson, King, Hall, Browning and Wewers2012; Marshall et al., Reference Marshall, Courtney, Passmore, McCaul, Yang and Bowman2016; van der Aalst, de Koning, van den Bergh, Willemsen, & van Klaveren, Reference Van der Aalst, de Koning, van den Bergh, Willemsen and van Klaveren2012). This pilot trial is innovative in its focus on connecting cessation counselling with screening results. The Centers for Medicare and Medicaid Services requires that cessation is addressed during the initial screening decision making process; however, our study and the emerging literature suggests that continuing to address cessation after the results are reported may also be critical to effective support (Fucito et al., Reference Fucito, Czabafy, Hendricks, Kotsen, Richardson and Toll2016; Piñeiro et al., Reference Piñeiro, Simmons, Palmer, Correa and Brandon2016; Taylor et al., Reference Taylor, Hagerman, Luta, Bellini, Stanton and Abrams2017). This approach has broad dissemination potential through the existing infrastructure of telephone tobacco quitlines, which are available in all 50 states and have been shown to be effective and cost-effective at increasing cessation (Joyce et al., Reference Joyce, Niaura, Maglione, Mongoven, Larson-Rotter and Coan2008). Prior work suggests that the delivery of screening results is a critical time for cessation support, as many current smokers will potentially misinterpret their screening findings as suggesting that their smoking has not harmed them (Zeliadt et al., Reference Zeliadt, Heffner, Sayre, Klein, Simons and Williams2015). Communicating screening results is challenging and must focus on several priorities in addition to smoking cessation, including clearly reporting the findings of the test and encouraging patients to return for repeat annual screening or short-term follow-up if necessary. The use of dedicated telephone counsellors, who have expertise in MI, may help ensure that cessation counselling is repeatedly and consistently delivered allowing primary care clinic staff to focus on clinical activities including the logistics of coordinating future visits (Ferketich et al., Reference Ferketich, Khan and Wewers2006; Jamal et al., Reference Jamal, Dube, Malarcher, Shaw and Engstrom2012; Park et al., Reference Park, Gareen, Japuntich, Lennes, Hyland and DeMello2015; Solberg et al., Reference Solberg, Maciosek, Edwards, Khanchandani and Goodman2006).

Our study findings are consistent with the emerging literature of smoking cessation in the context of lung cancer screening. Notably, our finding that the effect of the intervention was stronger among patients who had a screening result requiring short-term follow-up is similar to studies suggesting positive findings may be a stronger teachable movement (Slatore et al., Reference Slatore, Baumann, Pappas and Humphrey2014). Although the effect size was smaller among patients with normal screening results, the intervention was still positive, suggesting that counselling should be provided to all patients. Additionally, our findings are similar to the recently reported telephone counselling intervention offered by academic screening programs, which included a six session counselling protocol offered by the screening program (Taylor et al., Reference Taylor, Hagerman, Luta, Bellini, Stanton and Abrams2017). This study observed self-reported abstinence rates among patients receiving counselling of 22%, and biochemically confirmed abstinence rates of 17%, which is similar to our self-reported rate of 19%. These findings begin to address gaps in the literature noted by Piñeiro et al. (Reference Piñeiro, Simmons, Palmer, Correa and Brandon2016) and suggest that interventions tailored to this population and setting may provide additional benefit. However, questions remain about the timing, frequency, and intensity of these interventions, although evidence exists elsewhere that extended interventions may be no more effective than brief interventions (Moyer et al., Reference Moyer, Finney, Swearingen and Vergun2002).

There are several potential study limitations. The small size of the trial limits interpretation of the four-fold positive effect. For example, in discussing the unexpected higher rates of reported use of cessation medications in the control group, we discovered that one clinic initiated a policy of providing NRT lozenges to screening participants during the period when control subjects were being assessed. Additionally, we enrolled patients based on being offered screening by their provider and their provider's ascertainment of their smoking history. Notable details such as pack-year history, age of initiation of smoking, or current smoking intensity were infrequently recorded in the EHR when screening was offered. It is possible that smoking intensity may have been different between the intervention and control groups when screening was offered. Such confounding factors in non-randomised pilot trials can only be addressed in a fully-powered, randomised pragmatic trial. However, increased utilisation of NRT in the control group resulting from a change in usual care should lead to an attenuation of the observed effects of the intervention. Thus, our preliminary effect size may be an underestimation of the true effect. Another limitation is that our telephone counsellor was not able to directly provide medications or cessation assistance. Our calls were intentionally brief, intending to evoke smokers’ motivation and connect them to the VA Quitline rather than replicate the cessation support offered by the Quitline or the participant's medical centre. It is possible that arranging medications during the calls or more closely integrating with the Quitline cessation counselling protocol would further increase abstinence rates. Neither participants nor the study team were blind to participation in the intervention or evaluation activities, which may have inadvertently introduced social desirability bias in participants’ survey responses or experimenter bias by the interviewers during evaluation survey. To address this, we used separate telephone interviewers for the evaluation survey, and a uniform script was used across both groups although the last questions about satisfaction with the intervention unblinded the interviewer at the end of the script. Although we attempted to mimic routine care as closely as possible, this was a research study and participants were approached by asking them to volunteer in a research activity. Thus, our participation rate may not be reflective of offering proactive support as part of clinical care. The centre where the study was conducted required outgoing phone calls to be identified on caller ID as ‘Unknown Caller’ which may have limited participation compared to known labels such as ‘VA Medical Center.’ Although the study was intended to determine the feasibility of providing proactive outreach by a third party telephone counsellor who is not part of the patient's care team, our findings may not generalise to other settings if the telephone counsellor is not perceived as being integrated into the healthcare system. We do highlight that demographics of Veterans who participated in the intervention are very similar to Veterans who are eligible for lung cancer screening nationally (e.g. nationally mean age of screening eligible Veterans is 64.8 years and 3.5% are female) (Kinsinger et al., Reference Kinsinger, Anderson, Kim, Larson, Chan and King2017).

Overall, these findings suggest that integrating risk messages with brief motivational interventions when counselling current smokers participating in lung cancer screening, both before screening and after they have their screening results, can be a valuable component of lung cancer screening delivery. This study demonstrates that this counselling can potentially be delivered by a telephone counsellor with whom the patient has not previously interacted during an in-person clinic visit.