Background: An admission to hospital provides an opportunity to help people stop smoking. Individuals may be more open to help at a time of perceived vulnerability, and may find it easier to quit in an environment where smoking is restricted or prohibited. Initiating smoking cessation services during hospitalisation may help more people to make and sustain a quit attempt.
Objectives: To determine the effectiveness of interventions for smoking cessation that are initiated for hospitalised patients.
Search strategy: We searched the Cochrane Tobacco Addiction Group register which includes papers identified from CENTRAL, MEDLINE, EMBASE and PSYCINFO in January 2007, and CINAHL in August 2006 for studies of interventions for smoking cessation in hospitalised patients, using terms including (hospital and patient*) or hospitali* or inpatient* or admission* or admitted.
Selection criteria: Randomized and quasi-randomized trials of behavioural, pharmacological or multicomponent interventions to help patients stop smoking, conducted with hospitalised patients who were current smokers or recent quitters (defined as having quit more than one month before hospital admission). The intervention had to start in the hospital but could continue after hospital discharge. We excluded studies of patients admitted for psychiatric disorders or substance abuse, studies that did not report abstinence rates and studies with follow up of less than six months.
Data collection and analysis: Two authors extracted data independently for each paper, with disagreements resolved by consensus.
Main results: Thirty-three trials met the inclusion criteria. Intensive counselling interventions that began during the hospital stay and continued with supportive contacts for at least one month after discharge increased smoking cessation rates after discharge (Odds Ratio (OR) 1.65, 95% confidence interval (CI) 1.44 to 1.90; 17 trials). No statistically significant benefit was found for less intensive counselling interventions. The one study that tested a single brief (<=15 minutes) in-hospital intervention did not find it to be effective (OR 1.16, 95% CI 0.80 to 1.67). Counselling of longer duration during the hospital stay was not associated with a higher quit rate (OR 1.08, 95% CI 0.89 to 1.29, eight trials). Even counselling that began in the hospital but had less than one month of supportive contact after discharge did not show significant benefit (OR 1.09, 95% CI 0.91 to 1.31, six trials). Adding nicotine replacement therapy (NRT) did not produce a statistically significant increase in cessation over what was achieved by intensive counselling alone (OR 1.47, 95% CI 0.92 to 2.35, five studies). The one study that tested the effect of adding bupropion to intensive counselling had a similar nonsignificant effect (OR 1.56, 95% CI 0.79 to 3.06). A similar pattern of results was observed in smokers admitted to hospital because of cardiovascular disease (CVD). In this subgroup, intensive intervention with follow-up support increased the odds of smoking cessation (OR 1.81, 95% CI 1.54 to 2.15, 11 trials), but less intensive interventions did not. One trial of intensive intervention including counselling and pharmacotherapy for smokers admitted with CVD assessed clinical and health care utilization endpoints, and found significant reductions in all-cause mortality and hospital readmission rates over a two-year follow-up period.
Authors' conclusions: High intensity behavioural interventions that begin during a hospital stay and include at least one month of supportive contact after discharge promote smoking cessation among hospitalised patients. These interventions are effective regardless of the patient's admitting diagnosis. lnterventions of lower intensity or shorter duration have not been shown to be effective in this setting. There is insufficient direct evidence to conclude that adding NRT or bupropion to intensive counselling increases cessation rates over what is achieved by counselling alone, but the evidence of benefit for NRT has strengthened in this update and the point estimates are compatible with research in other settings showing that NRT and bupropion are effective.