|
Cost-effectiveness of the Danish smoking cessation interventions: subgroup analysis based on the Danish Smoking Cessation Database |
Olsen K R, Bilde L, Juhl H H, Kjaer N T, Mosbech H, Evald T, Rasmussen M, Hiladakis H |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. Health technology The study examined a group of smoking cessation strategies, such as group or individual courses and nicotine replacement therapy (NRT). Specifically, the interventions were offered by different organisations and included sessions in which an instructor and the smoker met face-to-face to discuss the clinical and motivational/psychological aspects of smoking and smoking cessation. The types of interventions provided were either group courses (7 to 10 participants attending five or six sessions of 2 hours each), individual courses (five or six sessions for a total of 2.5 hours), or quick interventions (1 to 6 participants with one instructor at one or two sessions for a total of 2.5 hours). NRT was also given to course participants.
Type of intervention Primary and secondary prevention.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised the general population of smokers.
Setting The settings were hospitals, pharmacies and the community. The economic study was carried out in Denmark.
Dates to which data relate The effectiveness data and most resource use data were derived from a database covering the period 1995 to 2001 and from studies published between 1997 and 2005. The price year was 2001.
Source of effectiveness data The effectiveness evidence was derived from a synthesis of published studies and authors' opinions.
Modelling A probabilistic Markov model was used to assess the costs and benefits of smoking cessation programmes in comparison with no intervention. The sample size observed in the DNSCD was extended (from 1,818 smokers to 10,000 cases) by a random sampling with replacement. The model followed all individuals until death and the cycle length was one year. Half-cycle correction was also performed. The three main health states considered were smoker, ex-smoker and death. In the intervention strategy, individuals who were not abstinent after the cessation intervention, or who relapsed later, were allowed to have another cessation attempt and also to become abstinent without any intervention (natural quit rate). With the no cessation intervention, the probability of becoming abstinent was equal to the natural quit rate. The structure of the decision model was reported.
Outcomes assessed in the review The outcomes estimated from the literature were:
descriptive statistics of smokers included in the database;
reported abstinence rates in the whole sample as well as in sub-groups of individuals;
the mortality rates for smokers (by age, gender and type of smoker);
the probability of enrolling again after failing to quit;
the lifetime risk of relapse;
the natural cessation rate; and
life expectancy.
Descriptive statistics derived from the database were used as explanatory variables in a regression analysis that evaluated the factors affecting the abstinence rate.
Study designs and other criteria for inclusion in the review The primary studies do not appear to have been identified through a systematic review of the literature, but the main objective of the analysis was to use "real-life" data. Thus, the DNSCD provided data on smoking cessation interventions that were implemented in Denmark from 1995 to 2001 through pharmacies, hospitals, municipalities, and other public or private bodies working with a preventive aim. An overall sample of 3,628 smokers was followed for one year, but data for only 1,877 individuals were available at the end of the follow-up period. Thus, two analyses were carried out, one considering the abstinence rate after one year in the whole group of patients (intent to treat rate) and the other considering the abstinence rate only in patients who completed the follow-up (reported abstinence rate). Details on the other sources of data were not provided.
Sources searched to identify primary studies Criteria used to ensure the validity of primary studies The authors stated that the Danish database represents a valid source of data reflecting real clinical patterns, owing to the naturalistic design.
Methods used to judge relevance and validity, and for extracting data Number of primary studies included Eleven primary studies provided the clinical data.
Methods of combining primary studies Investigation of differences between primary studies Results of the review The mean age of smokers in the database was 48.6 years and the proportion of female was 63%. Most individuals (76%) received group courses as the main intervention. Other descriptive statistics were also reported.
The abstinence rate at 12 months in the whole sample (intention to treat) was 16%. The abstinence rate at 12 months for those for whom data were available at follow-up (reported abstinence rate) was 31%. For sub-groups, the abstinence rate was 32% for men, 30% for women, 29% for heavy smokers, 33% for light smokers, 25% for those younger than 35 years, 31% for age 35 to 54 years, 32% for age 55 years and older, 38% in the hospital setting and 25% in the pharmacy setting.
The regression analysis showed that interventions carried out at hospitals were more effective than interventions carried out at pharmacies, although these results should be interpreted with caution given the limited availability of complete data. Another significant covariate was smoking dependency. Participants smoking more than 20 cigarettes per day had a significantly lower probability of abstinence.
The mortality rates were reported stratified by age, gender and tobacco consumption (ex-smoker, light smoker and heavy smoker).
The probability of enrolling again after failing to quit was 1%.
The lifetime risk of relapse was 10%.
The natural quit rate was 7% per annum.
Life expectancy was 75 years. The expected lifetime as a smoker was 60 years given a starting smoking age of 15 years.
Methods used to derive estimates of effectiveness The authors made a key assumption about the probability of abstinence.
Estimates of effectiveness and key assumptions The probability of abstinence in the second and third cessation attempt was assumed to be the same as the mean abstinence probability of the sub-group to which the individual belonged.
Measure of benefits used in the economic analysis The summary benefit measure used was the number of life-years (LYs) gained with the smoking cessation programmes in comparison with no intervention. The LYs were estimated using the modelling approach. An annual discount rate of 3.5% was used.
Direct costs The perspective of the analysis was restricted to the direct costs associated with NRT and instructor personnel time. The costs of NRT were borne by the smoking cessation units providing the interventions and to some extent by the smokers (both individuals' self-reported expenditures and cost of products delivered to each course participant were considered). The unit costs and quantities of resources used were mainly derived from the Danish database. The costs were estimated from the Danish Medicine Agency, the Danish Association of Bachelors of Pharmacy, and the Association of Danish Pharmacists. Details of cost calculations for time spent by pharmacists and pharmaconomists were reported. Discounting was relevant as the long-term costs were evaluated and an annual rate of 3.5% was used. The price year was 2001.
Statistical analysis of costs A linear regression analysis was run on intervention costs. The covariates used were gender, age, type of cessation intervention (individual, group, quick, other), setting (hospital, pharmacy, other), the Fagerstrom score (a measure of smoking dependency) and cigarette consumption (number of cigarettes per day). The costs were log-transformed because of the skewed nature of cost estimates.
Indirect Costs The indirect costs were not considered.
Currency Danish kroner (DKK). These were converted to euros (EUR) at the average exchange rate in 2001 (EUR 1.00 = DKK 7.45).
Sensitivity analysis A probabilistic sensitivity analysis was carried out to assess the uncertainty surrounding model parameters. Confidence intervals (CIs) and cost-effectiveness acceptability curves were generated by second-order Monte Carlo simulations. Beta-distributions were assigned to annual probabilities of abstinence. A univariate sensitivity analysis was also performed to assess the impact of individual parameters, such as clinical estimates and discount rates, on the results of the model. Most of the alternative estimates were derived from the literature.
Estimated benefits used in the economic analysis In comparison with no intervention, the smoking cessation strategies led to an increase in LYs of:
0.35 in the whole sample,
0.40 for men,
0.34 for women,
0.41 for light smokers,
0.34 for heavy smokers,
0.35 in a pharmacy setting,
0.41 in a hospital setting,
0.05 in the age group 25 to 34 years,
0.24 in the age group 35 to 54 years, and
0.67 in individuals aged 55 years and older.
Cost results The regression analysis showed that covariates gave only a moderate explanation of cost variation. Cessation courses other than individual and group courses, such as quick courses, were slightly cheaper. A similar conclusion was achieved for courses carried out at county services. Group courses were slightly more expensive than individual courses.
The extra cost of smoking cessation strategies over no intervention was:
EUR 450 in the whole sample,
EUR 423 for men,
EUR 450 for women,
EUR 442 for light smokers,
EUR 438 for heavy smokers,
EUR 464 in a pharmacy setting,
EUR 426 in a hospital setting,
EUR 415 in the age group 25 to 34 years,
EUR 446 in the age group 35 to 54 years, and
EUR 443 in individuals aged 55 years and older.
Synthesis of costs and benefits Incremental cost-effectiveness ratios (ICERs; i.e. the incremental cost per LY gained) were calculated in order to combine the costs and benefits of the alternative strategies.
The ICER with the smoking cessation strategies over no intervention was:
EUR 1,358 (95% CI: 1,320 to 1,396) in the whole sample,
EUR 1,090 (95% CI: 1,065 to 1,116) for men,
EUR 1,361 (95% CI: 1,326 to 1,395) for women,
EUR 1,114 (95% CI: 1,090 to 1,137) for light smokers,
EUR 1,362 (95% CI: 1,325 to 1,400) for heavy smokers,
EUR 1,361 (95% CI: 1,326 to 1,396) in a pharmacy setting,
EUR 1,058 (95% CI: 1,036 to 1,081) in a hospital setting,
EUR 9,651 in the age group 25 to 34 years,
EUR 1,984 (95% CI: 1,907 to 2,060) in the age group 35 to 54 years, and
EUR 673 (95% CI: 664 to 681) in individuals aged 55 years and older.
The probabilistic sensitivity analysis showed that the probability of being cost-effective at different thresholds was higher for men, light smokers and participants at hospitals compared with women, heavy smokers and participants at pharmacies. When cost-effectiveness acceptability curves were generated, the analysis suggested that the probability of the intervention being cost-effective approached 1 for the entire population, as well as for the sub-groups related to gender, setting and smoker status, at a willingness-to-pay of EUR 2,300. The worst scenario was observed for participants younger than 35 years, where the probability of cost-effectiveness reached 80% at a willingness-to-pay of EUR 30,000.
The results of the univariate sensitivity analysis indicated that the natural quit rate had a substantial impact on the ICER. Further, the use of the intention to treat abstinence rate partially affected the ICERs, which were slightly higher than those obtained using the reported abstinence rate. This was true especially in the hospital setting, where the most relevant loss to follow-up was observed. However, the ICERs for the intervention always remained lower than the standard threshold for willingness-to-pay.
Authors' conclusions The smoking cessation strategies were cost-effective in Denmark but, with the exception of the youngest group of smokers (less cost-effective in those aged less than 35 years), there were only moderate differences between sub-groups. In general, smoking cessation strategies were more cost-effective when offered to men, older persons and light smokers than when offered to women, young persons and heavy smokers.
CRD COMMENTARY - Selection of comparators The use of smoking cessation programmes was appropriately compared with no intervention, which might represent the usual treatment pattern in several settings. A series of smoking cessation interventions was considered globally without making a distinction between them in terms of their effectiveness or costs. You should decide whether they are valid comparators in your own setting.
Validity of estimate of measure of effectiveness The clinical data came from published studies that might have been identified selectively rather than through a systematic review of the literature, as the methods and conduct of such a review were not reported. In effect, the authors focused on Danish data, especially with respect to the characteristics of the general population of smokers that were derived from a national database and treatment effect (abstinence rate). The authors provided extensive information on this source of data, but few details of the other studies were reported. The use of "real-life" data represents a strength of the analysis, as the authors pointed out. Regression analysis was appropriate for estimating variables that could have an impact on treatment effect. A further strength of the study was the sub-group analysis according to age, gender, tobacco consumption and setting. Further, the authors used two definitions of treatment effect according to an intention to treat analysis and patients available at follow-up. Some key estimates were varied in the probabilistic and deterministic sensitivity analyses.
Validity of estimate of measure of benefit LYs are a valid benefit measure that can be compared with the benefits of other health care interventions. Mortality rates were stratified by age, gender and tobacco consumption. Discounting was carried out, in accordance with guidelines for the economic evaluation of health technologies. The use of alternative discount rates was also investigated.
Validity of estimate of costs The cost analysis was restricted to the direct medical costs. Only two main items were considered in the analysis, but details of the unit costs and quantities of resources used were not reported. This may limit the possibility of replicating the analysis in other settings. The authors reported the sources of costs for all items. Treatment patterns were derived from a large database that reflected resource consumption in Denmark. The cost estimates were specific to the study setting and were not varied in the sensitivity analysis. However, a statistical analysis was carried out to assess the impact of the baseline characteristics of smokers on total costs. The price year was reported, which will facilitate reflation exercises in other time periods. The authors noted that a key assumption in the cost analysis was that the "do nothing" strategy took place at zero cost, which might not be correct if one considers that individual smokers often make self-initiated and self-directed quit-smoking attempts.
Other issues The authors found their results to be consistent with those from other studies. Some economic evaluations of smoking cessation strategies were described and were compared with the current study. The authors pointed out that some of the inconsistencies in other studies could be explained by the use of "real-life" data rather than estimates derived from clinical trials, which might not reflect actual patient management. However, the authors noted that several of the clinical estimates were uncertain, and these were varied in the sensitivity analysis. In addition, some limitations of the sampling technique used in the probabilistic model could have affected the validity of the results of the analysis. In terms of the generalisability of the study results to other settings, the authors stated that a key issue is the local organisational set-up of the intervention, such as the combination of counselling and NRT, as well as the choice of instructor personnel and the intensity of the NRT package.
Implications of the study The study results did not support any form of sub-group differentiation in a health prevention policy since smoking cessation programmes were cost-effective for all analysed sub-groups.
Source of funding Supported by the Danish Ministry of the Interior and Health, the Danish Institute for Health Services Research and the Clinical Unit of Health Promotion at Bispebjerg Hospital.
Bibliographic details Olsen K R, Bilde L, Juhl H H, Kjaer N T, Mosbech H, Evald T, Rasmussen M, Hiladakis H. Cost-effectiveness of the Danish smoking cessation interventions: subgroup analysis based on the Danish Smoking Cessation Database. European Journal of Health Economics 2006; 7: 225-264 Other publications of related interest Because readers are likely to encounter and assess individual publications, NHS EED abstracts reflect the original publication as it is written, as a stand-alone paper. Where NHS EED abstractors are able to identify positively that a publication is significantly linked to or informed by other publications, these will be referenced in the text of the abstract and their bibliographic details recorded here for information.
Godfrey C, Parrott S, Coleman T, Pound E. The cost-effectiveness of the English smoking treatment services: evidence from practice. Addiction 2005;100 Suppl 2:70-83.
Cornuz J, Pinget C, Gilbert A, Paccaud F. Cost-effectiveness analysis of the first-line therapies for nicotine dependence. Eur J Clin Pharmacol 2003;59:201-6.
Crealey GE, McElnay JC, Maguire TA, O Neill C. Costs and effects associated with a community pharmacy-based smoking-cessation programme. Pharmacoeconomics 1998;14:323-33.
Stapleton JA, Lowin A, Russell MA. Prescription of transdermal nicotine patches for smoking cessation in general practice: evaluation of cost-effectiveness. Lancet 1999;354:210-5.
Indexing Status Subject indexing assigned by NLM MeSH Adult; Costs and Cost Analysis; Deductibles and Coinsurance /economics /statistics & Economic Competition; Female; Germany; Health Care Reform /organization & Humans; Male; National Health Programs /economics /organization & Universal Coverage /economics /organization & administration; administration; administration; numerical data AccessionNumber 22007008010 Date bibliographic record published 30/04/2007 Date abstract record published 30/04/2007 |
|
|
|