Hawthorne G
Centre for Health Program Evaluation, University of Melbourne, Australia.
Addiction. 1996 Aug;91(8):1151-9. doi: 10.1046/j.1360-0443.1996.91811517.x.
Evaluations of drug education programmes focus on the effect on participating students. Although this is appropriate, where programmes are institutionalized the wider population impact should also be considered. This paper examines this argument and provides a practical methodology for examining programme impacts on populations. It illustrates this using the example of the Life Education (LE) programme-a mature early intervention programme operating internationally-on the Victorian (Australia) population of Year 6 schoolchildren. Data from a recent evaluation of Life Education in Victoria, Australia, were re-analysed enabling estimation of adjusted attributable risk percentage (which provide an estimate of programme effect among those who are exposed to it), population prevalence estimates and population attributable risk percentage (which provide estimates of programme impacts on the entire population). In 1992 when the data were collected LE operated in 43% of all Victorian primary schools. The findings showed that within LE-schools 43% of girls' and 35% of boys' recent smoking was attributable to participation in the LE-programme. For drinking within LE-schools, while there was no apparent effect on girls, the programme was associated with 40% of boys' recent drinking. When the data were extrapolated to the state-wide smoking and drinking estimates, these showed that of all smoking among Year 6 schoolchildren, 25% of girls' and 19% of boys' smoking could be attributed to participation in Life Education, as could 22% of all boys' recent drinking. This re-analysis showed there was no preventive effect associated with the programme, at either the school or the population level. The findings suggest that intervention programmes should be thoroughly evaluated prior to widespread implementation, and that such evaluations should include reference to the community-wide impacts of such programmes.
药物教育项目评估主要关注对参与项目学生的影响。虽然这是恰当的，但在项目制度化的情况下，还应考虑对更广泛人群的影响。本文探讨了这一观点，并提供了一种检验项目对人群影响的实用方法。文章以“生命教育”（LE）项目为例进行说明——这是一个在国际上开展的成熟早期干预项目，针对澳大利亚维多利亚州六年级学童群体。对澳大利亚维多利亚州近期“生命教育”评估的数据进行了重新分析，从而能够估算调整后的归因风险百分比（用于估计接触该项目人群中的项目效果）、人群患病率估计值以及人群归因风险百分比（用于估计项目对整个人群的影响）。1992年收集数据时，“生命教育”项目在维多利亚州所有小学中的43%开展。研究结果表明，在参与“生命教育”项目的学校中，女生近期吸烟的43%和男生近期吸烟的35%可归因于参与该项目。对于参与“生命教育”项目学校中的饮酒情况，虽然对女生没有明显影响，但该项目与男生近期饮酒的40%有关。当将数据外推至全州吸烟和饮酒估计值时，结果显示，在六年级学童的所有吸烟情况中，女生吸烟的25%和男生吸烟的19%可归因于参与“生命教育”项目，所有男生近期饮酒的22%也可归因于此。这种重新分析表明，该项目在学校层面和人群层面均未产生预防效果。研究结果表明，干预项目在广泛实施之前应进行全面评估，且此类评估应包括对项目在社区层面影响的参考。
