The role of randomized trials in testing interventions for the prevention of youth suicide

The role of randomized trials in testing interventions for the prevention of youth suicide
复制标题

DOI:
10.1080/09540260701797779
复制
发表时间:
2007-12-01
影响因子:
2.8
通讯作者:
Gibbons, Robert D.
Gibbons, Robert D.
中科院分区:
医学4区
文献类型:
--
作者:
Brown, C. Hendricks;Wyman, Peter A.;Gibbons, Robert D.

文献摘要

被引文献

相似文献

流行病学方面的考虑表明,少数预防策略有可能大幅降低自杀率。几乎所有这些预防方法都涉及基于人群的策略,以发现更多的自杀高风险个体,或减少总体自杀率相对较低的人群中危险因素的流行率。这些方法很少经过严格的试验评估。我们认为,存在严格的随机试验设计,既可行又符合道德,可用于测试基于人群的自杀预防方案和实施策略。我们回顾了现有的自杀预防试验,并引入了两种新的随机试验设计,这些设计可能会达到足够的统计功效。 “动态候补设计”在不同时间段进行随机分配,目前正用于测试 32 所学校的看门人培训计划。它还可以用于检查农村地区的自杀预防计划。这项多试验后续研究建立在大量成功的基于人群的预防干预措施的基础上,旨在减少已知的青少年自杀危险因素,看看这些干预措施是否也会导致自杀完成率降低。
Epidemiological considerations point to a small handful of prevention strategies that have the potential for dramatically reducing suicide rates. Nearly all of those prevention approaches involve population-based strategies to either find an increased number of individuals at high risk for suicide or to reduce the prevalence of risk factors in members of a population that, as a whole, has a relatively low rate of suicide. Few of these approaches have been evaluated in rigorous trials. We argue that there are rigorous randomized trial designs that are both feasible and ethical and can be used to test both programmes and implementation strategies for population-based suicide prevention. We review existing suicide prevention trials and introduce two new randomized trial designs that are likely to achieve sufficient statistical power. The 'dynamic wait-listed design' randomizes across different time periods and is now being used to test a gatekeeper training programme in 32 schools. It could also be used to examine suicide prevention programmes in rural areas. The multi-trial follow-up study builds on the large number of successful population-based preventive interventions aimed at reducing known risk factors for suicide in youths to see whether these also cause a reduction in rates of completed suicide.