Health Education Research Advance Access originally published online on August 8, 2005
Health Education Research 2006 21(1):116-132; doi:10.1093/her/cyh048
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The European Smoking prevention Framework Approach (ESFA): effects after 24 and 30 months
1 Department of Health Promotion and Health Education, University of Maastricht, Maastricht, The Netherlands, 2 Health Promotion Service, Institute of Medicine and Health, Agència de Salut Pública de Barcelona, Barcelona, Spain, 3 Portuguese Council for Smoking Prevention, Lisbon, Portugal, 4 Department of Epidemiology and Health Sciences, University of Manchester, Manchester, UK, 5 Department of Cancer Prevention and Documentation, Danish Cancer Society, Copenhagen, Denmark, 6 STIVORO for a Smoke-free Future, The Hague, The Netherlands, 7 National Public Health Institute, Helsinki, Finland and 8 Department of Statistics, University of Maastricht, Maastricht, The Netherlands
9 Correspondence to: H. de Vries; E-mail: hein.devries{at}gvo.unimaas.nl
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
The European Smoking Prevention Framework Approach (ESFA) study in six countries tested the effects of a comprehensive smoking prevention approach after 24 (T3; N = 10 751) and 30 months (T4; N = 9282). The programme targeted four levels, i.e. adolescents in schools, school policies, parents and the community. In Portugal, 12.4% of the T1 non-smokers in the control group had started smoking at T4 compared to 7.9% of the experimental group. Smoking onset in the experimental group was thus 36% lower. In Finland, 32.4% of the T1 non-smokers started smoking compared to 27.6% of the experimental group, implying a 15% lower onset in the experimental group. In Spain, 33.0% of the T1 non-smokers in the control group had started smoking, compared to 29.1% of the experimental group, implying a 12% lower onset. In The Netherlands, the ESFA programme was effective for non-native adolescents with 11.4% new weekly smokers compared to 19.9% in the control group. An opposite effect was found in native Dutch adolescents with 19.0% new weekly smokers in the comparison group compared to 24.0% new smokers in the experimental group. Future programmes should use more standardized ways to assess process evaluations and should assess which elements are responsible for behavioral effects.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Smoking prevention programmes using the social influence (SI) approach can be effective. However, effects have been shown to decay over time (US Department of Health and Human Services, 1994
; Peterson et al., 2000). Sustaining the effectiveness of smoking prevention studies is therefore the most important challenge for current smoking prevention research. However, it is unclear which SI elements result in the (in)effectiveness of such programmes. Smoking prevention projects differ widely in their application of SI elements (De Vries et al., 2003c). For example, programmes differ not only in the number of skills-training sessions provided, but also in the attention afforded to covert and overt skills-training techniques (US Department of Health and Human Services, 1994; Dijkstra et al., 1999). Peer-led teaching methods appear promising; however, results are as yet inconclusive (Mellanby et al., 2000). Booster sessions may prolong effectiveness (Dijkstra et al., 1999), although not in all cases, and discussions about dosage (Glynn, 1989) are as yet not evidence-based (Peterson et al., 2000). Comprehensive community-based approaches have been shown to increase the long-term effects of smoking prevention programmes (Vartiainen et al., 1986; Perry et al., 1992).
The European Smoking Prevention Framework Approach (ESFA) used an integral preventive approach guided by best-practice principles (Glynn, 1989). It targeted adolescents in and out of school as well as their parents and the schools themselves (De Vries et al., 2003c). Short-term effects, 1 year after the pre-test, were found in Finland (smoking onset was 4.7% lower) and Spain (smoking onset was 3.1% lower). However, counter-productive trends were observed in Denmark and the UK (De Vries et al., 2003b). This paper describes the results of ESFA at 24 and 30 months after the pre-test.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Sample and design
The ESFA was initiated in February 1997 as a community intervention trial, the community-based equivalent of a randomized control trial (Bracht, 1999
). Regions within each participating country (Finland, Denmark, The Netherlands, the UK, Spain and Portugal) were approached to participate in the study and in the spring of 1998 were informed that they would be randomly assigned to the experimental or control condition. The National Project Managers (NPMs) then randomly assigned regions willing to participate in the project. Experimental regions received the ESFA project, while control regions received the usual care (De Vries et al., 2003c). In Finland, 27 schools from in and around Helsinki were randomly assigned to the experimental (N = 13) and control (N = 14) conditions. In Denmark, two regions, each consisting of 30 participating schools, were randomly assigned to the experimental or control condition. In the UK, two health authority regions, consisting of 22 and 21 schools, respectively, were randomly assigned to the experimental and control condition. In Portugal, two regions were randomly assigned to the experimental (N = 14) and control (N = 11) conditions. A quasi-experimental design was used in The Netherlands and Spain, as randomization was not possible. In The Netherlands, many schools use a national smoking prevention programme. These schools were included in the control condition (N = 17) while the remaining schools (N = 16) were assigned to the experimental condition. In Spain, all 16 experimental schools were exposed to smoking prevention programmes in the past. The 31 control schools were situated in other districts and were not exposed to these programmes.
Respondents
In the fall of 1998 (T1), 1999 (T2) and 2000 (T3) as well as the early summer of 2001 (T4), questionnaires were distributed to participating schools. Students were invited to participate in the study. They read an introductory letter that indicated that their responses would be treated confidentially and that they could refuse participation at any time. Schools were notified when the questionnaires would arrive and should be filled out; students were not informed of the time when questionnaires would be distributed, thus avoiding biased dropout. Students completed the questionnaire in the classroom, placed their questionnaire in an envelope and sealed it themselves. The teacher placed all of the questionnaires in a larger envelope and sealed it in front of the class.
Questionnaire
The questionnaire was based on a review of the literature, 15 years of work on adolescent smoking behavior and revised according to pilots studies conducted by the NPMs in each country (De Vries and Kok, 1986; De Vries et al., 1988; De Vries, 1995; Dijkstra et al., 1999; Kremers et al., 2001).
Two primary outcome measures were used, i.e. ever-smoking and weekly smoking. These items were assessed by a combination of five questions that were cross-validated (De Vries et al., 2003b). Adolescents were categorized as never-smokers (never smoked a puff), non-smoking deciders (experimented with smoking, but had quit experimenting), experimental smokers (experimenting with smoking, but not smoking weekly), weekly smokers (smoking at least once a week), daily smokers (smoking at least once a day) or as quitters (tried smoking at least weekly, but not smoking anymore). The validity of adolescent self-reported smoking is high in concordance with biological indicators when measurement assures anonymity (Dolcini et al., 1996). Hence, we optimized measurement conditions by assuring respondents of the confidentiality of their responses (Murray et al., 1987; Hansen, 1992; Dolcini et al., 1996).
Secondary outcomes in the study were the pros (six items on a seven-point scale; 
= 0.65) and cons of smoking (five items on a seven-point scale; = 0.68), social self-efficacy (three items about refraining from smoking in social situations; = 0.94), situational self-efficacy (three items about refraining from smoking in various situations; = 0.93), stress self-efficacy (three items measuring self-efficacy when stressed; = 0.96), and intention to smoke in the future (measured by one item). Other items included age, gender, pocket money, religious background, ethnicity, alcohol consumption, soft and hard drug use, family status (disrupted or not), parental occupation, social norms of parents, siblings and peers, social pressure to smoke from parents, siblings and peers, social modeling of smoking from parents, siblings and peers, parental reactions towards smoking, school achievement, school policy towards smoking, and if smoking was discussed during the previous year in school and at home. In order to conduct multilevel analyses, information regarding the respondents' class, school, municipality, region and country were recorded (De Vries et al., 2003b).
Intervention
For the first intervention year (1998–1999) a school-based programme was developed that included information on SI processes and training in refusal skills. Due to the fact that peer-led programmes were uncommon in the ESFA countries, programmes were teacher-led. During the second and third years of the study, classroom lessons were continued and the school, parental and out-of-school levels were elaborated. Programme development was guided through meetings with the NPMs, national board meetings and consensus meetings. As a result of these meetings, interventions were developed for four levels: the individual adolescent level, the school level, the parental level and the out-of-school level. Consensus was reached on core objectives and theoretical methods to be used. Due to differences in the availability of existing materials and country-level needs, the translation of objectives into practical didactic strategies differed per country. Furthermore, some countries added country-specific objectives to their programme (see Table I for country objectives). Further information on the programme content can be found elsewhere (De Vries et al., 2003b,c).
|
Table I shows the topics discussed during the project and Table II provides more detailed information on the activities per country. The range of school lessons regarding refusal skills and role-play differed amongst the countries. Teacher training differed greatly across the countries. The most intensive training was conducted in Portugal (48 hours) and Finland (20 hours). All countries developed and disseminated a school policy guide. Assessment of actual implementation of policies proved to be complicated and time-consuming. Pro-active offers of teacher smoking cessation were accepted by 54 teachers in Spain. All countries informed the parents about the project and its goals. Some countries initiated parent meetings (Finland, Denmark, the UK and Spain). However, attendance was very low. Most countries offered information to parents on how to discuss non-smoking with their children as well as smoking cessation for parents. However, only Spain and Portugal pro-actively offered parents cessation courses. In Spain, ‘Quit and Win’ was offered to parents in the third year; in Portugal, pharmacists offered cessation courses for 150 parents. In the first year of the study, all countries conducted access point analyses (Bullock et al., 1996
) to ascertain where youngsters spent their free time. These results were used to plan out-of-school activities. All countries disseminated posters. Non-smoking activities were developed in out-of-school settings (e.g. at youth camps, sport clubs and discos) in Finland and Denmark. Behavioral journalism was chosen as the out-of-school approach in Finland (McAlister, 1995).
|
Analysis
Dropout from the study was assessed using logistic regression. Where significant interactions were found between the predictors and country, separate analyses were run per country. The t-tests assessed the extent of exposure to the lessons using items from the process evaluation. Differences between the experimental and control groups on attitudes, self-efficacy expectations and intention were analyzed using covariance analyses. Baseline smoking behavior (1 = weekly smoking, 0 = non-smoking) and the demographic variables described above were included as covariates. Adolescents who responded to T1 and T3 or T1 and T4 measurements, had answered at least 90% of the questionnaires, and did not have missing values in the outcome variable were included in the behavioral effect analysis (De Vries et al., 2003b
).
Due to the fact that subjects were nested within classes, schools, quarters, municipalities, regions as well as countries, logistic regression using multilevel procedures were used to analyze differences in ever and weekly smoking prevalence rates ( De Vries et al., 1994, 2003b; US Department of Health and Human Services, 1994; Dijkstra et al., 1999). Non-significant predictors and interactions were deleted using a backward deletion procedure ( = 0.05), with the restriction that predictors were not removed from the model if they were involved as interaction terms. Demographic variables and the adolescent's attitudinal, self-efficacy and intention scores at T1 were included as covariates. Covariates were included to correct for potential baseline differences and to increase power. Previous research has shown the (potential) relationship of these covariates with smoking and the treatment condition (De Vries et al., 1994; US Department of Health and Human Services, 1994; Dijkstra et al., 1999; Ausems et al., 2002). A similar procedure was followed for the analyses of the T3 and T4 data.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Response
At T1, the sample consisted of 19 034 non-smokers; at T3, the sample consisted of 11 119 respondents (58.4%). The final sample consisted of 10 751 adolescents (56.5% of 19 034) following the removal of respondents who had 10% or more missing values in their data or missing values in the outcome variables (N = 368). The response rates were lower amongst adolescents who were male, older, drank more glasses of alcohol per week, had mothers with a 5-day job, were from disrupted families, had higher school achievements and were non-native (P < 0.05). At T3, dropout rates were slightly higher in the experimental condition (44.6%) than in the control condition (44.2%), but did not differ significantly. In some countries, response rates were higher (P < 0.05) in the experimental group than in the control group (Spain: 83 versus 69%; UK: 39 versus 30%). Denmark reported a reversed pattern of response rates between the experimental and control groups (41 versus 60%).
At T4, 10 617 (55.8%) of the subjects who had participated at T1 responded to the study. Due to over 10% missing values in their data or missing values in the outcome variables (N = 1335), the final sample at T4 consisted of 9282 students (48.8% of 19 034). The response rates were lower among adolescents who were male, older, non-religious, with a mother working more than 5 days per week, had more pocket money and who had higher school achievements (P < 0.05). Dropout rates were slightly higher in the experimental condition (52.3%) than in the control condition (50.2%), but did not differ significantly. Denmark and Portugal reported a reverse pattern of response rates between the experimental and control groups (45.9 versus 52.6% for Denmark and 41.7 versus 39.1% for Portugal). In Finland, two control group schools decided not to continue to participate in the program due to time constraints, resulting in a dropout of 45.7% in the control condition versus 27.2% in the experimental condition.
Process evaluation
Inspection of progress reports and NPM minutes revealed that at the individual level, NPMs encountered the most difficulty in developing refusal skills activities, as they were not familiar with them. Schools and national organizations also found these activities to be too time consuming. Furthermore, NPMs were weary of the fact that repeating messages might become boring for adolescents, but found it hard to develop evidence-based alternatives to provide new information.
At the school level, monitoring of the school policy proved to be difficult. The negative attitude of teachers in Denmark towards smoking prevention activities hampered involvement of schools in that country. Most countries offered information on smoking cessation for parents, but few pro-actively offered cessation courses to them, the exception being Spain where ‘Quit and Win’ was offered in the third year.
With regard to the parental level, attempts to actively involve parents through parent meetings were met with low attendance. Brochures describing how to discuss smoking with children were popular. However, the level of usage could not be measured. The proactive recruitment of smoking parents into cessation activities was not successfully implemented in the majority of countries.
NPMs experienced great difficulty in developing effective out-of-school and community-based strategies. Hence, various strategies were used, but were often limited in their reach. Finnish adolescents appreciated behavioral journalism. In the UK, ID cards were distributed to non-smoking students giving them discount in shops; however, process evaluation data revealed that this technique was unsuccessful. Finally, all NPMs and contractors reported that the extensive delays in European funding greatly hindered the implementation of the programme. In fact, some NPMs were forced to seek alternative employment. Dutch participation in the project was delayed for a period of 6 months.
Exposure to smoking prevention activities
Table III shows that adolescents from the experimental condition reported exposure to more non-smoking activities than those from the control condition. Significantly more lessons on smoking prevention were implemented in the experimental schools at T3 and T4 than in the control schools. Portuguese (14.7) and Spanish (14.6) adolescents from the experimental group reported exposure to the most number of lessons, whereas adolescents from the UK (9.0), Finland (9.2) and The Netherlands (9.6) reported exposure to the lowest number of lessons. Whereas more experimental than control group adolescents were exposed to non-smoking activities in schools, the groups hardly differed in terms of their exposure to parental and out-of-school activities. The number of activities on smoking prevention at all four levels increased in both the experimental and control schools from T3 to T4.
|
Cognitive effects
At T3, the experimental group was significantly less convinced of the pros of smoking than the control group. This effect was significant in Finland, Portugal and Spain (see Table IV). Only significant effects were found for the cons in Portugal. Effects on self-efficacy were found for the Portuguese adolescents in that the experimental group felt more confident to cope with social pressures to smoke and to refrain from smoking in various situations (situational self-efficacy). The Portuguese experimental group reported more negative intentions to smoke in the next year than the control group. In The Netherlands, significant counter-effects occurred, showing lower situational self-efficacy and a negative intention in the experimental group.
|
At T4, a significant overall difference was found for social self-efficacy in that the experimental group felt more confident to refuse offers of cigarettes from friends than the control group. This effect was as a result of significant effects in Denmark, Portugal, Spain (P < 0.10) and the UK. The experimental groups from the UK and Portugal reported significantly more negative intentions to smoke in the next year than the respective control groups.
Table IV furthermore shows that most significant effects had effect sizes greater than 0.10. In Portugal, effect sizes larger than 0.10 were found for all the cognitive effects, apart from situational self-efficacy after 30 months. Overall, lower effect sizes were found for the pros (at both tests) and for social self-efficacy after 30 months.
Programme effects on ever-smoking and weekly smoking
Table V shows that overall differences between ever-smokers in the experimental group (44.8%) and the control group (44.1%) who were never-smokers at T1 were not significantly different at T3. Significantly fewer ever-smokers were found in the Portuguese experimental group (33.8%) than the control group (41.5%). A borderline counter-effect (P < 0.10) was found in The Netherlands, with slightly more ever-smokers in the experimental condition (41.7) than in the control condition (36.6). A similar trend was observed in Denmark (49.3% in the experimental group versus in the 43.6% control condition). With regard to experimental smoking at T4, 12% less ever-smokers were found in the experimental condition in Portugal (P < 0.01) and 4.4% less ever-smokers in Spain.
|
With regard to weekly smoking at T3, Table V shows that 18.4% of the non-smokers in the experimental group had begun smoking on a weekly basis compared to 18.8% of the non-smokers in the control group. The lowest percentage of weekly smoking was observed in Portugal (7.3% in the experimental group versus 9.1% in the control group). The highest percentages were found in Finland (24.8% in the experimental group versus 30.1% in the control group). The multilevel analysis showed significant differences in weekly smoking between the experimental (24.8%) and control groups in Finland (30.1%). A significant counter-effect was found in The Netherlands in that more adolescents smoked regularly in the experimental condition (19.6%) than in the control condition (14.6%).
With respect to weekly smoking, a significant overall effect was found at T4, showing 21.9% new smokers in the experimental group versus 23.4% new smokers in the control group. Significant differences were found in Portugal with less new weekly smokers in the experimental group (7.9%) than in the control group (12.4%). Borderline effects were found in Finland (P < 0.06) with 27.6% new smokers in the experimental condition versus 32.4% in the control group and in Spain (P < 0.08) with 29.1% new smokers in the experimental condition versus 33.0% new smokers in the control group. Only in The Netherlands was an interaction effect found, indicating differential significant effects for adolescents with a Dutch and non-Dutch origin. The Dutch ESFA programme was effective for non-native adolescents (N = 289) with 11.4% new weekly smokers compared to 19.9% new weekly smokers in the control group. An opposite effect was found in native Dutch adolescents (N = 2012) with 19.0% new weekly smokers in the comparison group compared to 24.0% new smokers in the experimental group.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
The ESFA study adopted a comprehensive approach that targeted smoking at four levels, i.e. adolescents, schools, parents and the out-of-school setting.
Significant effects were found for the pros of non-smoking 24 and 30 months after the pre-test, and for social self-efficacy after 30 months. In Portugal, significant effects on cognitions were found at 24 and 30 months on all indicators apart from situational self-efficacy at 30 months. Whereas significant effects were found in Portugal on the cognitions, and to a lesser extent in Spain and Denmark, relatively small effect sizes and insignificant changes were found in Finland, The Netherlands and the UK, which raised the question as to whether our interventions were powerful enough in changing the cognitions of smoking.
With regard to the behavioral effects, the results showed that at T4, 2.5 years after the initiation of the study, the project resulted in a significant, although marginal, effect in that 23.4% of the T1 non-smokers from the control group had become weekly smokers compared to 21.9% in the experimental group. This figure represents a 6% lower increase in smoking in the experimental group. The strongest effects were observed in Portugal where the findings imply that smoking onset was 36% lower in the experimental group than in the control group. Smaller effects were also observed in Finland and Spain. In Finland, the findings imply a 15% lower onset of smoking in the experimental group and a 12% lower smoking onset in the experimental group. The Dutch ESFA programme was effective for non-native adolescents, but an opposite effect was found in native Dutch adolescents. In conclusion, the significant behavioral effects of the programme found for Finland and Spain 12 months after initiation of the study (De Vries et al., 2003b) continued to exist, but reached borderline significance. The effects of the programme on regular smoking only became apparent after 3 years in Portugal. The latter observation illustrates the importance of continued implementation of actions and their evaluation to be able to detect sleeper effects. What may have caused the effects in the three countries? The skills-training components in the lessons may have contributed substantially to this effect, since they were well elaborated in these countries, but weaker in the UK and The Netherlands. The large number of lessons may also have contributed to the positive effect. Teacher training was also very well elaborated in the three countries. Teachers' commitment during the whole project was crucial in the countries, and required intensive coordination and involvement of different stakeholders, but also incentives for teachers to participate. In Finland, behavioral journalism probably had substantial added value and was appreciated to a great extent. Moreover, the integration of the activities on all four levels, although innovative, was probably most successful in these three countries.
The ESFA project was the first to develop a best-practice smoking prevention approach at the European level that resulted in an efficacy study (Connelly et al., 2003). It was subject to several limitations. First, NPMs had to fulfill many roles (such as public relations, programme development and research), which restricted complete elaboration of all responsibilities. Although programme development was based on core objectives and theoretical methods, many differences emerged amongst the countries (De Vries et al., 2003b). Such differences were also noted in other SI projects (US Department of Health and Human Services, 1994; De Vries et al., 2003b). However, significantly more school activities were developed in the experimental schools. Second, random assignment was not possible in The Netherlands and Spain. The Spanish experimental schools had been already exposed to smoking prevention programmes. It is therefore plausible that the observed effects in our study may be the result of higher teacher motivation and experience. The lack of results in The Netherlands may be due to the fact that the control group participated in an existing drug abuse programme that is known to significantly reduce smoking behavior (Ausems et al., 2004). Third, the NPMs did not immediately recognize peer-led programmes as a viable alternative to teacher-led programmes, which have been shown to enhance programme effects (Telch et al., 1990; Klepp et al., 1993; Black et al., 1998; Erhard, 1999), although the results are still inconclusive (Mellanby et al., 2000). Peer leaders were included in the programme at a later stage in Portugal. Fourth, teacher training varied greatly. Countries that reported the strongest effects had spent much more time on teacher training, although this relationship was not tested experimentally. Intensive training and monitoring of teacher activities, while beyond the feasibility of our project, should be included in future projects (Peterson et al., 2000). Fifth, although overall response rates did not differ significantly between the experimental and control group, they did vary within countries and were sometimes higher in the experimental schools or in the control schools. Although it cannot be excluded that effects, e.g. in Barcelona and Portugal, may have been influenced by somewhat higher response rates in the experimental group, reports from teachers did not suggest that smokers deliberatively did not fill out questionnaires. Moreover, students were not informed in advance about the timing of questionnaire administration, and our process evaluation and the in-depth inspection of the data showed that dropout was mostly attributable to school dropout from the project due to time constraints (Finland, Denmark and the UK) or students failing to graduate to the next grade (Portugal). Although we provided newsletters to schools and NPMs contacted schools regularly, the necessity of smoking prevention as well as analyzing long-term effects may need more attention in future EU projects. An additional method may also be to provide financial incentives for participating schools, although our budget did not allow for this. Sixth, the development of out-of-school activities was not very successful as many access points had to be considered. Process evaluation during the various waves of the study across the six countries revealed that streets, shops, sport clubs, swimming pools and snack bars were the most popular places were adolescents spent their free time. In-depth analysis also revealed differing patterns of access points for smokers and non-smokers (De Vries et al., 2003b). Further research on access points is therefore required prior to the development of effective out-of-school prevention activities. Finally, delays in project funding seriously stalled programme development.
Connelly and colleagues also described strengths and weaknesses in their Commentary on the ESFA study (Connelly et al., 2003). First, Green (pp. 664–665) correctly indicated that the ESFA study illustrates the difficulty of obtaining appropriate control or comparison groups as well as random assignment in community trials. The ‘golden’ randomized controlled trial evaluation standard was not always feasible and appropriate as it did not account for the four levels of intervention, and should have been supplemented at minimum with more focused process evaluation data [see also the comments of Green (p. 664), Lechner (p. 666) and Roberts (p. 672)]. At the outset of the project, we carefully planned process evaluation procedures including the development of logbooks for NPMs, teachers, school-contact persons and other intermediaries. The collection and analysis of such data, however, proved to be time intensive and received varying support amongst the countries. Moreover, capacity limitations at the national level hindered the in-depth analyses of the process evaluation. Hence, future European projects should allocate more funds to conduct in-depth process evaluation and employ more standardized ways to assess process data. Second, the strength as well as a limitation of ESFA is that it was an effectiveness study (Connelly et al., 2003). Our strength was that we were able to realize the development of a first EU project on smoking prevention with common goals, objectives and research methodology. A further strength was that real-life implementation showed the potential of the programme in three countries. Third, dissemination implies reinvention (Rogers, 1983). It is not always clear when adaptation renders new programmes. As Lechner stated (p. 667) in this Commentary (Connelly et al., 2003), in retrospect, many elements and adaptations should have been piloted before implementation. However, time and financial constraints prevented pilot testing. Fourth, our sample size estimations did not take into account the cluster research design (see Rigby's comment on p. 667), as in 1998 all the relevant information to make these calculations was not forthcoming. Fifth, programme reach, acceptability and integrity ranged from country to country (see Connelly's comment on p. 668). While it is always useful to anticipate these issues, the answers are not always that easy to find. While many of these issues were anticipated and extensively discussed during project development, controlling these elements sometimes proved to be extremely complicated. Finally, alternative approaches to the empirical approach are relevant (Connelly et al., 2003). Further research is required to demonstrate their effects.
What is recommendable; should we use best-practice or dismantling designs? We believe that two main streams of research should be promoted. First, we agree with Mittelmark (p. 667) that we need to consolidate 25 years of experience and disseminate the state-of-the-art science for tobacco prevention. The ESFA study attempted to achieve this for the first time in Europe and was partly successful. Research of this nature requires sound effectiveness, process evaluation and cost-effectiveness research as well as timely allocation of sufficient funds. Second, we disagree with Mittelmark's suggestion that dismantling studies are not needed. Although stated somewhat differently, Robert's comments (p. 672) in the same Commentary (Connelly et al., 2003) also point towards the need for dismantling research. Illuminative studies are needed, not for the sake of ‘techno-tweaking’ results, but to improve the quality of interventions and the understanding of working mechanisms. For instance, our Finnish behavioral journalism approach was innovative; however, we could not analyze its impact. Moreover, our study as well as other research showed that peer influences as well as other SI mechanisms may be operating simultaneously [e.g. the influences of parents and selection of friends (De Vries et al., 2003a)]. This requires adaptations of current approaches. In addition, the identification of causal factors of smoking onset does not automatically imply the identification of effective communication strategies. For example, we still need to discern which groups will benefit from self-efficacy-enhancing information and which groups will benefit from other types of information (e.g. attitude-reinforcing information). Prospect theory suggests differential impacts of gain and loss-framed messages (Kahneman and Tversky, 1984). Internet-based approaches are increasing in popularity. An understanding of how to best use this channel is therefore required. Furthermore, we need strong evidence indicating that a school-based approach will benefit from an integral approach, addressing teachers smoking behavior and school policies as well. Whereas it is logical to suppose synergetic effects, experimental evidence is scarce and is also needed to be able to better convince school management.
Finally, was the ESFA worth the effort? The project started collaboration at the EU level that was not reached before. Capacity building was realized with regard to program development and research methodology. This also allowed for additional analyses, such as comparing smoking stages in youngsters (Kremers et al., 2004), and analyzing the gateway hypothesis about smoking and alcohol (Wetzels et al., 2003), the importance of parents in addition to peers (De Vries et al., 2003a), and similarities in parenting rules (Huver et al., 2005) in the six EU countries. However, such efforts need to be continued and it is recommended to invest in a strong EU smoking prevention network that stimulates collaboration between the various EU members. However, this requires stronger pro-active initiatives at the EU level than currently exist in order to be able to develop second-generation studies focusing on best-practice and/or experimental studies, and to be able to translate research findings into practice.
In sum, best-practice approaches use a ‘container approach’, which is somewhat inevitable. Consequently ‘container programmes’ limit the possibility of identifying elements that work best for particular groups of people. Hence, best-practice and innovative dismantling prevention studies are both needed. Collaboration at the European level has the potential to realize these objectives.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
None declared.
|
Acknowledgments |
|---|
The ESFA project was financed by a grant from the European Commission (The Tobacco and Information Fund; contract 96/IT/13-B96 SOC96201157). We thank Martijntje Bakker, Kerstin van der Groot, and Ingrid Leijs who served as the project managers. We thank the contractors Jenaro Astray Mochales, Goof Buijs, Anne Charlton, Manuel Pais Clemente, Manel Nebot, Andrew Povey, Trudy Prins, António Romeiro, Hans Storm and Errki Vartiainen. We also thank Saadhna Panday for correcting our English. Finally, we thank all teachers, health intermediaries, students and all other individuals who participated in the ESFA project, and the reviewers for their constructive comments.
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Ausems, M., Mesters, I., Van Breukelen, G. and De Vries, H. (2004) Effects in-school and tailored out-of-school smoking prevention among Dutch vocational school students. Health Education Research, 19, 51–63.
Ausems, M., Mesters, I., Van Breukelen, G., De Vries, H., Brug, J., Steenhuis, I. and Van Assema, P. (2002) Short-term effects of a randomized computer-based out-of-school smoking prevention trial aimed at elementary schoolchildren. The impact of a computer-tailored nutrition intervention. Preventive Medicine, 34, 581–589.[CrossRef][ISI][Medline]
Black, D.R., Tobler, N.S. and Sciacca, J.P. (1998) Peer helping/involvement: an efficacious way to meet the challenge of reducing alcohol, tobacco, and other drug use among youth? Journal of School Health, 68, 87–93.
Bracht, N. (1999) Health Promotion at the Community Level: New Advances, 2nd edn. Sage, Thousand Oaks, CA.
Bullock, A.D., De Vries, H., Lopez, M.L., Thomas, H. and Charlton, A. (1996) Smoking prevention and young people: using research to identify out-of-school intervention sites in three countries. Educational Review, 48, 143–152.
Connelly, J., Green, J., Lechner, L., Mittelmark, M.B., Rigby, A.S. and Roberts, C. (2003) The European Smoking Prevention Framework Approach (ESFA) project: observations by six commentators. Health Education Research, 18, 664–677.
De Vries, H. (1995) Socio-economic differences in smoking: Dutch adolescents' beliefs and behaviour. Social Science and Medicine, 41, 419–424.
De Vries, H. and Kok, G.J. (1986) From determinants of smoking behaviour to the implications for a prevention programme. Health Education Research, 1, 85–94.
De Vries, H., Backbier, E., Dijkstra, M., Van Breukelen, G., Parcel, G. and Kok, G. (1994) A Dutch social influence smoking prevention approach for vocational school students. Health Education Research, 9, 365–374.
De Vries, H., Dijkstra, M. and Kuhlman, P. (1988) Self-efficacy: the third factor besides attitude and subjective norm as a predictor of behavioural intentions. Health Education Research, 3, 273–282.
De Vries, H., Engels, R., Kremers, S., Wetzels, J. and Mudde, A. (2003a) Parents' and friends' smoking status as predictors of smoking onset: findings from six European countries. Health Education Research, 18, 627–636.
De Vries, H., Mudde, A., Kremers, S., Wetzels, J., Uiters, E., Ariza, C., Vitoria, P.D., Fielder, A., Holm, K., Janssen, K., Lehtuvuori, R. and Candel, M. (2003b) The European Smoking Prevention Framework Approach (ESFA): short-term effects. Health Education Research, 18, 649–663.
De Vries, H., Mudde, A., Leijs, I., Charlton, A., Vartiainen, E., Buijs, G., Clemente, M.P., Storm, H., Gonzalez Navarro, A., Nebot, M., Prins, T. and Kremers, S. (2003c) The European Smoking Prevention Framework Approach (EFSA): an example of integral prevention. Health Education Research, 18, 611–626.
Dijkstra, M., Mesters, I., De Vries, H., Van Breukelen, G. and Parcel, G.S. (1999) Effectiveness of a social influence approach and boosters to smoking prevention. Health Education Research, 14, 791–802.
Dolcini, M.M., Adler, N.E. and Ginsberg, D. (1996) Factors influencing agreement between self-reports and biological measures of smoking among adolescents. Journal of Research on Adolescence, 6, 515–542.
Erhard, R. (1999) Peer-led and adult-led programs—student perceptions. Journal of Drug Education, 29, 295–308.[CrossRef][ISI][Medline]
Glynn, T. (1989) Essential elements of school-based smoking prevention programs. Journal of School Health, 59, 181–188.
Hansen, W.B. (1992) School-based substance abuse prevention: a review of the state of the art in curriculum 1980–1990. Health Education Research, 7, 403–430.
Huver, R.M.E., Engels, R.C.M.E. and De Vries, H. (2005) Are anti-smoking parenting practices related to adolescent smoking cognitions and behavior? Health Education Research, in press (cyh045).
Kahneman, D. and Tversky, A. (1984) Choices, values and frames. American Psychologist, 39, 341–350.[CrossRef]
Klepp, K.I., Tell, G. and Vellar, O.D. (1993) Ten-year follow-up of the Oslo youth study smoking prevention program. Preventive Medicine, 22, 453–462.[CrossRef][ISI][Medline]
Kremers, S.P.J., de Vries, H., Mudde, A.N. and Candel, M. (2004) Motivational stages of adolescent smoking initiation: predictive validity and predictors of transitions. Addictive Behaviors, 29, 781–789.[CrossRef][ISI][Medline]
Kremers, S.P.J., Mudde, A.N. and De Vries, H. (2001) ‘Kicking the initiation’: do adolescent ex-smokers differ from other groups within the initiation continuum? Preventive Medicine, 33, 392–401.[CrossRef][ISI][Medline]
Lipsey, M.W. (1990) Design Sensitivity: Statistical Power for Experimental Research. Sage, Newbury Park, CA.
McAlister, A. (1995) Behavioral journalism: beyond the marketing model for health communication. American Journal of Health Promotion, 9, 417–420.[ISI][Medline]
Mellanby, A.R., Rees, J.B. and Tripp, J.H. (2000) Peer-led and adult-led school health education: a critical review of available comparative research. Health Education Research, 15, 533–545.
Murray, D.M., O'Connell, C.M., Schmid, L.A. and Perry, C.L. (1987) The validity of smoking self-reports by adolescents: a reexamination of the bogus pipeline procedure. Addictive Behaviors, 12, 7–15.[CrossRef][ISI][Medline]
Perry, C.L., Kelder, S.H., Murray, D.M. and Klepp, K.I. (1992) Communitywide smoking prevention: Long-term outcomes of the Minnesota Heart Health Program and the Class of 1989 Study. American Journal of Public Health, 82, 1210–1216.
Peterson, A.V., Kealey, K.A., Mann, S.L., Marek, P.M. and Sarason, I.G. (2000) Hutchinson smoking prevention project: long-term randomized trial in school-based tobacco use prevention. Results on smoking. Journal of the National Cancer Institute, 92, 1979–1991.
Rogers, E.M. (1983) Diffusion of Innovations. Collier Macmillan, New York.
Telch, M.J., Miller, L.M., Killen, J.D., Cooke, S. and Maccoby, N. (1990) Social influences approach to smoking prevention: the effects of videotape delivery with and without same-age peer leader participation. Addictive Behaviors, 15, 21–28.[CrossRef][ISI][Medline]
US Department of Health and Human Services (1994) Preventing Tobacco Use Among Young People: A Report of the Surgeon General. USDHH Services, Public Health Services, Washington, DC.
Vartiainen, E., Puska, P., Koskela, K. and Nissinen, A. (1986) Ten-year results of a community-based anti-smoking program (as part of the North Karelia Project in Finland). Health Education Research, 1, 175–184.
Wetzels, J.J.L., Kremers, S.P.J., Vitoria, P.D. and De Vries, H. (2003) The alcohol–tobacco relationship: a prospective study among adolescents in six European countries. Addiction, 98, 1755–1763.[CrossRef][ISI][Medline]
Received on January 31, 2005; accepted on July 4, 2005
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Ariza, M. Nebot, Z. Tomas, E. Gimenez, S. Valmayor, V. Tarilonte, and H. De Vries Longitudinal effects of the European smoking prevention framework approach (ESFA) project in Spanish adolescents Eur J Public Health, June 18, 2008; (2008) ckn050v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||