Systematic Review of Smoking Cessation Interventions for Smokers Diagnosed with Cancer

The detrimental impact of smoking on health and wellbeing are irrefutable. Additionally, smoking is associated with the development of cancer, a reduction treatment outcomes and poorer health outcomes. Nevertheless, a significant number of people continue to smoke following a cancer diagnosis. Little is understood of the smoking cessation services provided to smokers with cancer or their engagement with them. This systematic review aimed to identify existing smoking cessation interventions for this cohort diagnosed with breast, head and neck, lung and cervical cancers (linked to risk). Systematic searches of Pubmed, Embase, Psych Info and CINAHL from 1 January 2015 to 15 December 2020 were conducted. Included studies examined the characteristics of smoking cessation interventions and impact on referrals and quit attempts. The impact on healthcare professionals was included if reported. Included studies were restricted to adults with a cancer diagnosis and published in English. No restriction was placed on study designs, and narrative data synthesis was conducted due to heterogeneity. A review protocol was registered on PROSPERO CRD 42020214204, and reporting adheres to PRISMA reporting guidelines. Data were screened, extracted in duplicate and an assessment of the quality of evidence undertaken using Mixed Methods Assessment Tool. 23 studies met the inclusion criteria, representing USA, Canada, England, Lebanon, Australia and including randomized controlled trials (9), observational studies (10), quality improvement (3), and one qualitative study. Hospital and cancer clinics [including a dental clinic] were the settings for all studies. 43% (10/23) of studies reported interventions for smokers diagnosed with head and neck cancer, 13% (3/23) for smokers diagnosed with lung cancer, one study provides evidence for breast cancer, and the remaining nine studies (39%) report on multiple cancers including the ones specified in this review. Methodological quality was variable. There were limited data to identify one optimal intervention for this cohort. Key elements included the timing and frequency of quit conversations, use of electronic records, pharmacotherapy including extended use of varenicline, increased counselling sessions and a service embedded in oncology departments. More studies are required to ensure tailored smoking cessation pathways are co-developed for smokers with a diagnosis of cancer to support this population.


Introduction
Leaving no one behind is the focus of the Sustainable Development Goals for improving the planet and the health and well-being of the population [1]. Tobacco control measures are a fundamental component in attaining SDGs. At the same time, global prevalence rates of smoking are decreasing, supported by effective tobacco control mechanisms established by the World Health Organization (WHO) Framework Convention on Tobacco Control (FCTC) [2] and subsequent MPOWER measures [3]. The FCTC uphold human rights [4] and children's rights [5], yet despite the downward trend, 8.7 million deaths criteria, 23 studies are included in this systematic review (Figure 1). Multiple papers ar reported for two studies, and data from Abdelmutti [24] are reported in Guiliani [25 Concurrent and additional reporting for outcomes in Crawford [26] are in four paper [27][28][29][30].
To provide a comprehensive review of the evidence, we report studies that includ data for the specified cancers [head and neck, lung, breast, and cervical cancer] in additio to reporting additional cancers noted in studies. Ten studies report interventions fo smokers diagnosed with head and neck cancers [34][35][36][37]39,45,47,[49][50][51], three report inte ventions for thoracic/lung cancers [40,43,45] and one study provides evidence for brea

Characteristics of Interventions
Studies included participants ranging from 11 to 2652 (Table 2a-c). The interventions described in the evidence vary in content and duration of follow-up (1 month up to 24 months). Most patients are recruited to studies and programmes in hospitals as inpatients or when attending cancer centres or clinics; Conlon [47] enrolled patients attending a dental clinic for oncology patients located within a specialist cancer centre.
Recruitment and enrollment of patients included several options, attending for a first consultation, commencing treatment while awaiting surgery or participants who had a cancer diagnosis for up to 5 years and may have completed a treatment cycle. Seven studies identified recruitment for newly diagnosed patients [24,41,42,[46][47][48]51] and three studies recruited patients to the smoking cessation programme at the pre-surgery phase following diagnosis [40,43,50].
Several developed smoking cessation interventions linked to national smoking cessation programmes and provided participants with information for websites, quit lines, text messaging services and use of quit smart apps [24,31,34,38,45,48,51]. Ten studies reported the availability of free or low-cost varenicline and pharmacotherapies to support quitting [24,26,33,34,36,42,43,45,47,48]. Two studies [33,36] report the availability of nicotine replacement therapy [NRT] for family members. The duration of pharmacotherapies provided free varied per study with Crawford et al. [26] providing up to 12 weeks and further extension of this up to 24 weeks [30]. Other incentives described were monetary and provision of an ipad [37,38,45] (Table 2a-c).
Development of electronic referral systems as part of a smoking cessation programme was reported [24,25,43] specifically identifying 'opt out' systems in seven studies, whereby the smoking status was routinely sought and recorded for all patients attending clinics [24,36,39,41,44,45,47]. Nolan [41] noted including an 'asking and recording smoking cessation status at all touch points' by all healthcare professionals interacting with patients as a key development for their programme. All who were identified as smokers received advice on quitting and were provided with an opportunity for referral to a smoking cessation programme. Where this system was described, it included the development of an electronic health record or embedding of smoking cessation services [24,36,42,44], and the system provided data on referrals, outcomes and follow up. A minority of studies recorded patients' smoking status at each clinic/hospital visit-and those that did, the data were captured from the development of the intervention programme as part of a research study or following commencement of a quality improvement initiative [24,36,39,41,43,44].  Intervention Participants were offered 4 weekly telephone counseling sessions, 4 biweekly telephone sessions delivered over 2 months, and 3 telephone booster sessions monthly. The number/length of sessions based on pilot work with patients from the Massachusetts thoracic oncology clinic. Participants offered a choice of 12 weeks of FDA-approved smoking cessation medication at no cost; they were not required to use any medication. The medication selected by participants was prescribed in the EHR and dispensed (in person or mailed). Participants received an initial 4-week supply of cessation medication (varenicline, bupropion sustained release, single or combination nicotine replacement therapy/patch and/or lozenges) with the option of receiving up to 2 additional 4-week supplies. Based on Self-regulation model and Health Belief model.
Participants offered 4 weekly telephone counseling sessions plus education and cessation medication advice. 1st session, tobacco counselors used a decision aid, designed for patients with cancer, to make medication recommendations.
Follow up 3 months and 6 months.  MD Anderson Tobacco Treatment Programme (TTP) is a comprehensive and personalised intervention including counseling, pharmacotherapy, and management of mental health conditions. TTP participants begin with an in-person consultation followed by 6 to 8 treatment appointments completed within 8 to 12 weeks. Abstinence data prospectively collected at 3-, 6-, 9-, and 12-month intervals beginning at enrollment. Abstinence is defined as 7-day point prevalence of cigarette abstinence at 9-month follow-up. Tobacco treatment specialist (TTS). Opt-in programme. Scheduled face to face counselling arranged (free) if agreement. TTS counselling was 1 hr face to face using motivational interviewing and discuss treatment options. Patients chose type of pharmacotherapy: nicotine replacement (patch, gum, lozenges), bupropion, varenicline, or none. They were also advised to set a quit date. Follow-up with the TTS occurred per patient request and included face-to-face appointments, telephone conversations, or both. Carbon monoxide (CO) levels taken at preoperative visits and on the day of surgery. If CO level >11 ppm surgery could be cancelled or postponed. All patients were given unsupervised home exercise program and asked to complete at least 30 min of moderate exercise daily.

None
Follow up day of surgery, 2 weeks, 6, 12, 24 months. Completion of a survey or participated in a pilot smoking cessation service.
Those that declined to participate in the pilot study invited, to explore the 'difficult to reach individuals and provide representative understanding of views and experiences of the different groups of patients encountered in a head and neck follow up clinic. A Specialist nurse invited participants who were current/former smokers, had completed treatment for HNC and were in the follow-up stage of treatment. The intention was to recruit successful quitters and unsuccessful participants from the pilot study to explore the difference in their experiences and reasons they believe made them successful or not.  None as new process implemented in clinic. Pre-implementation data A minority of studies report smoking cessation services embedded within oncology services providing a bespoke service for patients attending clinics. The practitioners were oncology healthcare professionals (HCP) who completed additional smoking cessation training. McDonnell et al. [40] described support for patients via oncology nurses trained in smoking cessation, support for patients and family members from therapists within oncology services [36,40].

Outcomes of Interventions-Quit Attempts
The outcomes of interventions varied with limitations associated with small sample sizes. Table 3a,b report outcomes, referrals and quit rates. Increased referrals linked to electronic referral systems were reported in four studies [24,25,36,41]. Several studies reported the impact on quit rates [24,27,32,33,38,42,43,45] with Charlot et al. [32] identifying a quit rate at 8 weeks of 50.1% and 44.3% at 3 months. Conlon et al. [47] reported quit rates of 14.8% at 3 to 6 months and 12 months. Almost 50% of current smokers reporting using no smoking cessation supports to help their quit attempt, another 23.8% reported using nicotine replacement therapy [47].
Rettig et al. [45] noted higher quit rates in the intervention group OR 4.83, 95% CI 1. 31-17.76, and for those who were married. Lower odds of quitting were reported in patients with a history of depression, having a co-addiction, and experiencing mucositis during treatment. Higher pain scores and mucositis were associated with increased smoking rates. Rettig suggests a value in integrating smoking supports into cancer treatment.
Cinciripini et al. [33] reported improved abstinence in the intervention group compared to usual care at 9 months RR 1.31 95% CI 1.11-1.56, p = 0.001, while rates of quitting did not differ between patients with cancer and those who did not have cancer, the results present a high rate of abstinence. Crawford et al. [26] used varenicline pharmacotherapy to support quitting but did not report a difference in quit rates at 12 weeks. Subsequently, Schnoll et al. [30] reported point prevalence smoking rates that were not significantly different at 24 and 52 weeks (Table 3a). Price et al. [29] notes that there was no increase in depressed mood scores, and abstinence was associated with improved cognitive function. Overall evidence suggests [26,27,29,30] consistent evidence of the use of varenicline therapy for 24 weeks.
Simmons et al. [46] recruited patients who had relapsed following a previous quit attempt and evidence from the study identifies the positive impact on quit attempts for married or having a partner OR 2.23 (95% CI 1.01 to 4.90). In relation to increased counselling supports, Park [42] identified that increased counselling sessions n = 8 IQR 4 to 11 was associated with use of quit pharmacotherapies observed in treatment group V usual care (77.0% V 59.1%), OR 2.31 95% CI, 1.32-4.04; p = 0.003 (Table 3a,b).     (4) experience of smoking cessation support services.
Themes described guilty habit of smoking, perceived barriers to quit, teachable moment of a diagnosis and social motivation to both smoke and quit. 'Guilty habit' represented the knowledge that smoking was 'wrong' and socially stigmatised, difficulty in adjusting to a life without tobacco and cigarettes. Feelings of guilt and self-blame when smoking after treatment for cancer. Smoking relapse was common. It is unlikely that a simple information giving exercise would impact quitting. Sustained support and encouragement needed. Barriers to quit refusal did not indicate refusal, suggest asking again, wanted ownership of ability to quit. Cost was as barrier, stress was a barrier, cancer treatments cited as a barrier-slowness and challenges eating often relieved by a cigarette. Social motivation-Boredom noted for those with less social supports. However, social support was complex and could encourage smoking. Sometimes exacerbated by the social isolation and unemployment that is often a feature of living with HNC. Teachable moment identified by participants who quit-and in many who hadn't but wish to do so. Shock of diagnosis a motivator as pain in mouth from cancer identified. Understanding the link between smoking head and neck cancers can be paramount. Cessation provisions and effects: For those who had used smoking cessation supports, an abrupt cessation of the smoking cessation supports gave a feeling of being 'left to their own willpower to quit; some patients experienced unwanted side-effects of pharmacotherapy.  Self-report Feasibility Quit rates Low recruitment-50 patients screened. N = 16 reported one quit attempt. n = 5 quit attempt in past year. All reported one quit attempt. n = 5 quit attempt in past year. No attempt to stop a patient smoking, rated smoking as very important to them. Stopping smoking important but low confidence at outset. Face to face meetings associated with adherence. Two additional booster meetings with financial incentive to boost adherence. Challenges included time as recruitment over the summer. Private space to meet. Due to changes in university, lack of referrals due to team members considering it was too late to refer to smoking cessation services-fidelity was retained to reduce team bias through a brief orientation to all new members of the team. Recruitment 44% lower than anticipated despite team support, financial incentives, assurance of privacy,-limitation requirement of family member to participate. Exit interviews identified pre-op timing as ideal, telephone interactions difficult due to competing life issues.
No attempt to stop a patient smoking, rated smoking as very important to them. Family members are less confident than partners in own ability to quit.

Nolan et al., 2019 [41]
Patients and employees N = 45 patients self-reported smoking-data collected, and N = 10 interviews with patients. Not reported Quit rates/attempts Feasibility n = 45 patients. n = 15 smoking at intake call reported quit smoking by time of breast clinic appointment. n = 30 smokers at time of visit 23 (76%) referred to Nicotine Dependency Clinic (NDC). Significant increase to pre-intervention (29% p < 0.0001). Of those referred, 17 (74%) attended NDC consult-an increase from pre-intervention (41% p = 0.026). Qualitative data from n =10 patients. 5 referred to NDC. 4 of 5 planned to attend. System-level interventions-patients recalled stop smoking conversations related to breast reconstruction. Intervention factors identified by patients-gratitude that smoking not discussed front and centre as their primary concern was diagnosis. No patient surprised that smoking was discussed or offered NDC consult. Many could not describe the NDC consult.  [39] Employees N = 117 patient chart reviews. N = 15 care providers surveyed.

Self-reported in charts
Referral quit services and knowledge 13% of staff were aware of tobacco control services. 28% documented. Pre-intervention N = 54 charts identified 6 to 13 smokers/month. Median 8. Pre-intervention lack of knowledge of services and unawareness of smoking cessation services in the community. N = 2 of 15 staff were aware of any resource in the community. Post-intervention 100% of providers could name one cessation resource and 88% felt the intervention prompted discussions.
Nolan et al., 2019 [41] Employees N = 12 qualitative interviews with providers Not reported Factors associated with service plan Provider interviews N = 12 identified systems-level factors-important that smoking stopped before reconstructive surgery (4 to 6 weeks stopped). This surgery used to facilitate discussions about smoking. 2nd theme was perception of team roles-an assumption that previous providers had already discussed smoking status with a patient. Personal decision factors: factors identified as barriers to discussion quitting included: patients who were seen as light smokers (a few a day), advanced cancer, perception of emotional instability, presence of comorbid conditions, geographical distance. Facilitators included: patients' interest in reconstructive surgery, no script for this, but all providers discussed own experiences with family, friends, experiences from other patients who had identified tobacco cessation counselling positive. Intervention level factors from providers included: a lack of knowledge among providers about what happens at NDC consults (same as patients).

Challenges Experienced during Quit Attempts
Ghosh et al. [37] suggests that smoking was valued more by participants than any incentives to quit, noting the particular challenges for those with head and neck cancer and their difficulty travelling a distance to attend a specific smoking cessation clinic. Difficulties impacting successful quit attempts included winter weather, scheduling appointments for late time of the day and the effect of traffic or no parking at hospitals when attending appointments. While electronic or opt out referrals systems generated increased referrals, Guiliani et al. [25] reports low interest in referrals to subsequent smoking cessation programmes. They suggest this may be due to low motivation, lack of confidence to quit, stigma with smoking and having cancer, and fear that care could be impacted if refused referral. Other barriers identified services only available in English and a lack of longer term follow up.
Abdelrahim et al. [49] supports Guiliani et al. [25] as they identified challenges experienced by smokers with cancer trying to quit. Four themes emerged including guilt, barriers to quitting, the teachable moment, and identifying social motivation to quit and smoke. Evidence from this study suggestions that relapse is common, patients who refuse to quit did not indicate refusing overall-they wanted to be asked again. The obstacle was dealing with life and lack of control with cancer diagnosis, they wanted to control if and when they would quit smoking. Sustained support and encouragement are necessary as feelings of guilt and self-blame were reported. Barriers to quitting included the cost of NRT and pharmacotherapies in addition to cancer treatments. Participants with a diagnosis of head and neck cancer identified that side effects of treatment were relieved by smokingprovided relief. They suggest a teachable moment exists at diagnosis and this is supported by Conlon et al [47].
Social support was complex in both being a factor to encourage quitting, but reduced support could result in isolation and increased smoking [46]. Higher smoking rates were reported among those without social support and higher episodes of fatigue, depression, pain, and lower confidence to quit. Abdelrahim et al. [49] noted social support could increase smoking if a partner or friend is a smoker (Table 3a). Table 3b highlights outcomes for healthcare professionals from two studies [39,41] in this review. Ma et al. highlights discrepancies in practice with a lack of knowledge, lack of recording of the smoking status of patients, lack of awareness of smoking cessation services, competing life events, and lack of indoor smoke restrictions in own homes. The sample in the study is small. Despite this limitation, an improvement in staff knowledge is noted. Consistent challenges are reported by Nolan et al. [41], indicating personal barriers and assumptions made by practitioners in limiting their engagement with smoking cessation conversations for patients with cancer; and their perception of limited use of the pre-surgical period for commencing smoking cessation discussions. Table 4 presents a summary of evidence indicating critical components and processes for each of the 23 studies in this review, combining the complex data from Tables 2a-c and 3a,b We identified 13 studies providing successful outcomes, five of the studies reported evidence of increased referral rates but had limited impact on quit rates. Comprehensive development of interventions is critical given the complexity of care required.     SC local = smoking cessation local programme; SC National = smoking cessation national programme; Multilang = available in multiple languages; Opt out = programme developed and all patients included. Anyone who did not want to participate had to opt out; Staff training = availability of training on smoking cessation for staff for new intervention; SC by expert oncology HCPs =smoking cessation provided by oncology healthcare professionals in oncology dept; NRT = Nicotine replacement therapy; Family option = Families specifically included in intervention; Success = was there a positive outcome from intervention; Y = yes, N = no.

Discussion
The evidence in this review foregrounds the range of processes and services offered by smokers with cancer ( Table 4). The variation of methods and success of sustained pharmacotherapies and counselling [26,42] and limited success of many other interventions supporting quitting. What is known is that approximately 69% of smokers express a desire to quit [52]; however, limited engagement with smoking cessation services exists for smokers diagnosed with cancer. Despite what is known, few interventions are routinely implemented [53].
This review presents inconsistent outcomes from a comprehensive array of study designs [including pilot and feasibility studies], interventions and quality improvement programmes developed for smokers diagnosed with cancer. Interventions adopted in cancer centres range from the provision of information sheets-to the development of smartphone applications and the link to national smoking cessation programmes. This review comprehensively reports the characteristics of interventions and their success or other positive change.
Several systems issues are identified as critical in supporting smoking cessation programmes, including developing and using electronic patient records and recording patients' smoking status at each clinical encounter (Tables 2a-c and 4). The evidence in this review highlights the variety of approaches and timing of conversations from referral to pre-surgery to during therapy and post-therapy [41,42,46,47]. Abdelrahim et al. noted the importance of the duration of smoking cessation services and conversations provided and a need for information about the effect of continued smoking on recovery from cancer [25,40]. Conlon et al [47] suggests information about quitting should be consistently offered at all stages of the cancer journey. Identifying the timeline for introducing conversations on quitting was highlighted, with the importance of quitting smoking as a pre-surgical intervention. Table 4 presents evidence that suggests integration of smoking cessation services in cancer care as usual practice, developing electronic referral systems, increased consultation and follow-up reviews, and pharmacotherapies [free or low cost] are essential components of a smoking cessation intervention.
Cinciripini et al. [33] supports inventions in the oncology setting to enable sustained abstinence for patients with cancer and survivors; Rettig et al. [45] suggests the integration of services in oncology is part of the National Cancer Institute Cancer Moonshot initiative [11]. Croyle et al. [54] considers this a high-risk population group, and programmes for cessation should be developed.
In this review, evidence identified significant barriers for healthcare professionals supporting smokers with cancer due to a lack of confidence and perceptions [39]. The findings in this review are consistent with Feuer's [55] review of 29 studies that examined smoking relapse for a population who are cancer survivors. They suggest that smoking cessation after a cancer diagnosis is understudied, and that interpretation of interventions is challenging due to heterogeneity (p. 102237).
Research is required to determine optimal smoking cessation support after treatment ends, the duration of interventions, and cancer-specific tailoring [53,56]. Acknowledging the challenges of patients adjusting to chemotherapy regimens and diagnosis suggests a need to support sustained quit attempts and support patients who relapse to enable confidence-building and reduce stigma. It is crucial to consider the longevity of support that may be needed beyond the usual eight or 12-week programme. Santi et al. [55] confirm the challenges for those quitting and the potential for additional support beyond six months, as noted in the updated National Institute for Health and Care Excellence (NICE) guidelines [57]. Evidence indicates the importance of married/partnered patients for social support when quitting [53]. It is vital to ensure that smoking cessation programmes are inclusive and developed for diverse populations. Consistent evidence on developing population quit programmes is known [58]; what is evident is a lack of adaptability to smokers who have cancer and have different types of cancer, which may result in a variety of side effects and experiences from both cancer and cancer treatments.
The location of quit services is critical. There is evidence in this review supporting embedding smoking cessation services within oncology services (Table 4), and Ghosh et al. [37] presents the challenges for those who have to travel to attend clinics, including time and financial costs. Alternative methods reported in this review include engaging with national smoking cessation services and providing choices to patients on the mode of delivery, including online, email or by phone. The impact of COVID-19 provided opportunities for developing virtual clinics, and this may be a useful addition.
Warren et al. [59], almost ten years ago, identified the use of an automated referral system for those with a diagnosis of cancer into smoking cessation services. Conlon et al.'s [17] systematic review of healthcare professionals also confirms the importance of electronic referral systems, the responsibility of all health care professionals working in oncology to document smoking status, and the need for dedicated referral to a specialist smoking cessation advisor. Evidence in this review confirms the benefit of electronic referral systems and integrated health records.

Quality Review
Quality was assessed in twenty studies, as three studies were quality improvement programmes. Factors associated with lower quality of evidence include the reliance on selfreporting of smoking history, recall bias, and convenience sampling (Supplementary File S3). However, confirmation of quitting was reported in several studies via biochemical verification and carbon monoxide monitoring [26,33,38,42,43,45,46,51]. Several studies included randomisation and a follow-up period of 24 months with statistical adjustment for independent factors, including gender, age, marital status, and comorbidities (Table 3a,b). We did not remove any study from this review, and we acknowledge the breadth of evidence includes studies with a lower quality design, with short-term follow-up of interventions; however, we wished to be comprehensive and acknowledge the tailored responses adopted to support smokers who have cancer from feasibility or pilot studies.

Limitations in the Review Process
A key strength of this review is that it addresses a knowledge gap and has collated evidence from a broad methodological base to report the interventions adopted to support smoking cessation in smokers with cancer. Due to the heterogeneity of included studies, meta-analysis was not performed, while the descriptive nature of studies prevents identifying a causative relationship between measures and outcomes. The duration of interventions varied in the evidence presented, with five studies reporting interventions of less than six months duration [32,38,39,50,51]. We did not exclude studies based on the duration of intervention to present the extent of evidence for this population group. We acknowledge that while a summary of characteristics and outcomes is presented, insufficient evidence is available to statistically evaluate and summarise the relationship between individual measures, and further studies are required to elucidate this. Despite this, the systematic approach to this review has identified the scope of interventions implemented to date for a critical population of smokers diagnosed with cancer. This adds to the limited body of evidence published [17,20,56]. We acknowledge reporting bias from self-reported data and the financial incentives reported in studies. Publication bias was minimised with follow-up contacts with authors for early reporting and by including observational study designs. In addition, we have presented overlapping papers and identified singular primary studies for completeness. We report studies published in English, searching four databases and a timeline of six years, as this was one component of an ongoing research study. However, the methods used to complete this review were complete and adhered to Cochrane review standards [60].
We acknowledge that patient and public involvement (PPI) was not featured in any study included in this systematic review. We do report that PPI was a component of this systematic review. Two experts by experience from the study steering committee were invited to review the preliminary evidence and assist in developing our understanding and reporting as co-authors. Secondly, a summary of the review evidence was presented as part of a PPI Patients Voice in Cancer workshop (April 2022) [a component of the broader research study].

Conclusions
This novel systematic review summarises the evidence base to date, identifying specific factors for consideration in the development of a smoking cessation intervention targeted at smokers who have cancer. As reported by participants, the nuances and sensitivities of the population, the impact of treatment, and the stigma associated provide a strong voice for their input in the subsequent co-development and implementation of interventions to support tobacco control and smoking cessation in cancer care services. A one size fits all approach to development is unhelpful, and further research should embed public and patient involvement at the outset.