System change interventions for smoking cessation in hospitalised patients ThomasDennis 2017 Background: Many smoking cessation interventions have proven efficacy in clinical trials, however they have not been adequately translated into routine clinical practice. Previous reports suggest that smoking cessation support in Australian hospitals is sub-optimal. This warrants cost-effective, pragmatic, and sustainable smoking cessation interventions tailored to the hospital environment. This PhD aimed to systematically review the evidence for system change interventions for smoking cessation, develop and evaluate a pharmacist-led smoking cessation intervention for inpatients, and develop and validate an instrument to assess challenges associated with quitting smoking. Methods: A systematic review was undertaken with the Cochrane Tobacco Addiction Group to identify system change interventions for smoking cessation and to evaluate their effectiveness on both cessation and practice change outcomes. Studies were identified from databases: MEDLINE, EMBASE, CINHAL, PsycINFO, CENTRAL and Cochrane Tobacco Addiction Group Register, and also from grey literature and by hand searching of bibliographies of relevant publications. Each study was evaluated for risk of bias according to the Cochrane handbook and categorised as high, low or unclear risk of bias. The primary endpoint was abstinence from smoking at the longest follow-up. Secondary endpoints included assessment and documentation of smoking status, provision of quitting advice or counselling, prescribing cessation medications, and referral and enrolment in Quitline services. A narrative synthesis was used to describe the study findings due to the presence of significant heterogeneity between studies. A randomised controlled trial (RCT) was designed and conducted in three tertiary public hospitals in Australia to evaluate the effectiveness of a pharmacist-led system change smoking cessation program (GIVE UP FOR GOOD) compared with usual care. In-patients 18 years of age or older, who were self-reported current (daily or occasional) smokers at the time of hospital admission and available for 12 months of follow-up were eligible. Potential participants were identified from the hospital records by a trained research assistant (RA) or referred by the ward staff. A baseline survey assessed the quitting experiences and preferences for a future quit attempt among participants. The multicomponent intervention comprised behavioural counselling (provided during hospital stay and on discharge, with a one month post-discharge telephone follow-up) alone or in combination with the patient preferred pharmacotherapy (free supply during hospital stay and up to one month post-discharge) which included a range of nicotine replacement therapy (NRT) products (patch, gum, inhaler, lozenge and micro tab) and non-NRT prescription medications (bupropion and varenicline). All the intervention participants were offered the counselling sessions and encouraged to use pharmacotherapy. Hospital pharmacists had the authority to initiate NRT products or recommend non-NRT medications. Although the support from the hospitals was restricted to one month due to the pragmatic nature of the intervention (could easily be implemented in routine practice using minimum resources), the participants were referred to Quitline services and primary health care professionals (general practitioners and community pharmacists) for additional and ongoing support beyond one month. Smoking status was assessed at baseline, discharge, one, six and 12 months by a masked RA. Two primary endpoints were assessed in the intention-to-treat population using logistic regression analyses: carbon monoxide (CO) validated one month sustained abstinence at the six month follow-up and verified six months sustained abstinence at the 12 month follow-up. The participants who met the abstinence criteria at six (one month sustained abstinence) and 12 months (six months sustained abstinence) follow-ups were requested to perform the CO verification of smoking status. CO levels were measured by a trained RA using a hand-held device during a hospital or home visit. A participant with CO ≤ 6 parts per million (ppm) was considered abstinent. The item pool for Challenges to Stopping Smoking scale (CSS-21) was generated from a literature search, existing scales, expert opinion, and interviews with smokers and ex-smokers. The questionnaire was administered to participants of GIVE UP FOR GOOD trial at final follow-up. Exploratory factor analysis was performed to identify sub-scales in the questionnaire. Internal consistency, validity and stability of the sub-scales were evaluated. Key findings: The systematic review included seven cluster RCTs. The overall quality of evidence was low. Although the evidence for the primary endpoint – smoking cessation – was uncertain, significant improvements were observed in outcomes such as documentation of smoking status (1 study), Quitline referral (2 studies) and Quitline enrolment (2 studies). The provision of smoking cessation counselling as a result of system change interventions improved in the majority of studies (3 of the 4 studies evaluated). Other system level outcomes such as assessing smoking status and advising smokers to quit also showed some potential. The evidence for prescribing NRT was equivocal. The RCT included 600 smokers (43% participation rate) from three participating hospitals. Participants had a mean (± SD) age of 51±14 years and 64% were male. The common reasons for hospital admission self-reported by participants were disorders of the circulatory system (135, 22.5%), musculoskeletal system and connective tissue (97, 16.2%), respiratory system (75, 12.5%), digestive system (67, 11.2%) and nervous system (65, 10.8%). Majority were daily smokers and had smoked on an average of 18.8±10.8 cigarettes per day. On a scale of 1(low) to 10 (high), current motivation to quit smoking was high (median 9; interquartile range [IQR] 6.5, 10), but confidence was modest (median 5; IQR 3, 8). Of the 386 (64.3%) participants who had attempted quitting in the previous 12 months, 270 (69.9%) had used at least one method to assist their quit attempt. More than half (222, 57.5%) used NRT and almost a quarter (94, 24.4%) used varenicline during their previous quit attempts. Over 80% (n=311) reported experiencing withdrawal symptoms in their previous quit attempts. Among the users of NRT or prescription medications (249), more than half (141, 56.6%) experienced side effects. Most participants (351, 58.5%) believed medications (NRT 322, 53.7%; varenicline 186, 31.0%) would assist them to quit in the future. The GIVE UP FOR GOOD intervention was well accepted among inpatient smokers; 98% of intervention participants received at least one session of the intervention and 79% received all three sessions. Retention rates at six and 12 months were 74% and 72%, respectively. A higher proportion of intervention participants compared to control participants reported ‘satisfaction’ (88.7% vs. 72.1%; p<0.001) with the services received during their hospital stay. However, the use of pharmacotherapy was not adequate; only a minority received any support from Quitline or community health professionals after discharge. The primary endpoints, verified abstinence rates at six (11.6% vs. 12.6%; OR 0.91, 95%CI 0.55 to 1.50) and 12 months (11.6% vs. 11.2%; OR 1.04, 95%CI 0.63 to 1.73) were similar in both groups. The CSS-21 scale was validated in a sub-sample of 182 participants (mean age 55±12.8 years, 70.3% current smokers) from the RCT. Factor analysis of the 21-item instrument resulted in a 2-factor solution representing items measuring intrinsic (9 items) and extrinsic challenges (12 items). All fit indices were acceptable for the 2-factor model (root mean square error of approximation 0.062, comparative fit index 0.948 and Tucker-Lewis index 0.935). Cronbach’s alpha coefficients for the intrinsic and extrinsic sub-scales were 0.86 and 0.82, respectively. The instrument has content and construct validity, was stable in various analyses and has a sound and meaningful factorial structure. Conclusions: The systematic review identified gaps in the literature and the need for well-powered randomised trials of systems level approach to improve the provision of smoking cessation care and cessation outcomes. The RCT demonstrated that a pharmacist-led smoking cessation intervention during hospital stay was feasible but did not achieve long-term abstinence. System change interventions of high intensity with more active post-discharge follow-up, active involvement of primary health professionals and a free supply of a full course of pharmacotherapy might produce more favourable effects. The CSS-21 should be a useful tool to guide smoking cessation support for smokers. The findings of this research will inform smoking cessation interventions for hospitalised smokers, and guide further research in this area.