|The effectiveness of workplace smoking cessation programmes: a meta-analysis of recent studies
|Smedslund G, Fisher K J, Boles S M, Lichtenstein E
This review assessed the effectiveness of recent worksite smoking cessation interventions. The authors concluded that workplace smoking cessation interventions increased short-term quit rates, but appeared to lose effectiveness beyond 12 months. As the quality of the included studies was not adequately reported, it is difficult to comment on the strength of the evidence underpinning the authors' conclusions.
To assess the effectiveness of recent worksite smoking cessation interventions and to compare findings with a meta-analysis published in 1990 (see Other Publications of Related Interest no.1).
ABI/INFORM, BRS Colleague, CHID, Dissertation Abstracts International, ERIC, MEDLINE, an occupational health and safety database, PsycINFO, Smoking and Health database, the Social Sciences Citation Index, and Sociological Abstracts were searched for studies published in English between January 1989 and December 2000; the search terms were reported. Reviews were examined and reference lists in retrieved reports were checked.
Study designs of evaluations included in the review
Controlled studies with at least 6 months' follow-up were eligible for inclusion.
Specific interventions included in the review
Studies of worksite smoking cessation were eligible for inclusion. The included studies used different types of interventions including self-help written materials, cessation groups, incentives, steering committee, non-smoking policy, physician advice and pharmacological treatment. Some worksites in the review had a no smoking policy and others allowed smoking in separate areas, but none had a total ban; other studies did not describe the strictness of the ban. In some studies smoking cessation was part of a larger programme to improve health.
Participants included in the review
Studies of workers were eligible for inclusion. Most of the studies were set in workplaces with high smoking prevalence and white, heavy smokers.
Outcomes assessed in the review
Studies that reported quit rates (QRs) were eligible for inclusion. Two large studies that estimated QRs indirectly from prevalence changes were excluded. The review assessed QRs (7-day point prevalence, 30-day point prevalence, or continuous abstinence) at 6 months, 12 months and more than 12 months. Some studies used some type of biochemical validation of smoking abstinence, while others did not.
How were decisions on the relevance of primary studies made?
Two reviewers independently selected studies for inclusion. Any disagreements were resolved through discussion and with the help of a third author. Inter-rater agreement was 72% for titles and abstracts, and 100% for full publications.
Assessment of study quality
The authors did not explicitly state that they assessed validity. However, they did assess study design, attrition and losses to follow-up. Two reviewers independently extracted validity data. Any disagreements were resolved through discussion and with the help of a third author.
Two reviewers independently extracted the data. Any disagreements were resolved through discussion and with the help of a third author. For each study, the number of participants and the QRs at 6 months, 12 months and more than 12 months were extracted for each treatment group. Only one comparison was extracted from each study for each time period. In studies with more than one treatment, the review compared the simplest treatment (or a no treatment control) with the next simplest treatment.
Methods of synthesis
How were the studies combined?
Pooled odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using a random-effects model. Weighting was done using the standard error of their natural logarithms. The possibility of publication bias was assessed using funnel plots.
How were differences between studies investigated?
Differences between the studies were discussed with respect to type of worksite, presence of smoking policies and intervention. Statistical heterogeneity was assessed using the chi-squared statistic. Results from randomised controlled trials (RCTs) and non-randomised studies were considered separately. There were insufficient data to undertake a subgroup analysis on the influence of potential moderating variables.
Results of the review
Nineteen controlled trials (n=9,578) were included. The number of randomised trials was unclear.
Several methodological inadequacies were noted in the included studies. Six of the 19 studies reported attrition during the intervention, while seven reported losses at follow-up. Potentially important moderating variables were inconsistently reported.
The QR at 6 months ranged from 6.1 to 30.8% with the interventions and from 1.05 to 19.15% with the control.
Workplace smoking cessation significantly increased QRs at 6 months (OR 2.03, 95% CI: 1.42, 2.90) and 12 months (OR 1.56, 95% CI: 1.17, 2.07) compared with control. There was no statistically significant difference between interventions beyond 12 months (OR 1.33, 95% CI: 0.95, 1.87). No statistically significant heterogeneity was detected at 6 or 12 months (P=0.11 and P=0.13, respectively). Statistically significant heterogeneity was detected beyond 12 months (P=0.0004). Funnel plots showed strong evidence for publication bias at the first two follow-ups, but not at the third.
The treatment effect at 6 months was greater and precision was less in non-randomised studies (2 non-randomised trials; OR 4.65, 95% CI: 1.92, 11.28) in comparison with RCTs (6 RCTs; OR 1.74, 95% CI: 1.26, 2.40). No statistically significant heterogeneity was detected (P=0.79 and P=0.23, respectively).
The results were similar at 12 months. There was no statistically significant difference between the interventions beyond 12 months for RCTs or non-randomised studies. Statistically significant heterogeneity was detected for non-randomised studies (P=0.0003), but not for RCTs (P=0.10).
The effectiveness of smoking cessation at 6 months was greater than that shown in the previous meta-analysis in the 1980s.
Workplace smoking cessation interventions increased QRs in the short term but appeared to lose effectiveness beyond 12 months.
The review question was clear in terms of the study design, intervention, participants and outcomes. Many relevant sources were searched, attempts were made to minimise publication bias, and the potential for publication bias was assessed. No attempts were made to minimise language bias and the authors acknowledged this limitation. Two reviewers independently selected studies and extracted the data, thus reducing the potential for bias and errors. Some validity criteria were assessed and some information on the included studies was tabulated. However, there were limitations to the validity assessment and details of the individual studies. For example, there was no information on attrition rates for individual studies, participation rates in interventions, and the baseline comparability of the treatment groups.
In the principal meta-analysis, data from RCTs and non-randomised studies were pooled. However, the authors also analysed RCTs and non-randomised studies separately and statistical heterogeneity was assessed. Attempts were made to explore potential sources of heterogeneity among studies, but these were hampered by insufficient data. A comparison of results from studies using biochemical validation to assess outcomes with studies not using biochemical validation might have been useful in assessing the strength of the evidence. As the quality of the included studies was not adequately reported, it is difficult to comment on the strength of the evidence underpinning the authors' conclusions.
Implications of the review for practice and research
Practice: The authors did not state any implications for practice.
Research: The authors stated that it would be helpful if future researchers adhered to the CONSORT checklist when reporting studies (see Other Publications of Related Interest nos.2-3) and reported attrition and retention rates for the participants.
Smedslund G, Fisher K J, Boles S M, Lichtenstein E. The effectiveness of workplace smoking cessation programmes: a meta-analysis of recent studies. Tobacco Control 2004; 13(2): 197-204
Other publications of related interest
1. Fisher KJ, Glasgow RE, Terborg JR. Work site smoking cessation: a meta-analysis of long-term quit rates from controlled studies. J Occup Med 1990;32:429-39. 2.Altman DG. Better reporting of randomised controlled trials: the CONSORT statement. BMJ 1996;313:570-1. 3. Begg C, Cho M, Eastwood S, Horton R, Moher D, Olkin I, et al. Improving the quality of reporting of randomized controlled trials. The CONSORT statement. JAMA 1996;276:637-9.
Subject indexing assigned by NLM
Adult; Controlled Clinical Trials as Topic; Female; Follow-Up Studies; Humans; Male; Observer Variation; Occupational Health Services /standards; Odds Ratio; Program Evaluation; Randomized Controlled Trials as Topic; Smoking Cessation /methods; Treatment Outcome; Workplace /statistics & numerical data
Database entry date
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.