Steinert Tilman
Zentrum für Psychiatrie Weissenau, Abt. Psychiatrie I der Universität Ulm.
Psychiatr Prax. 2009 Jul;36(5):238-42. doi: 10.1055/s-0028-1090237. Epub 2009 May 19.
Together with a multi-professional working group, the author was responsible for the development of the German Psychiatric Association's treatment guideline on Aggressive Behaviour. Using his experience during this process, he makes some critical comments on the current methods of developing evidence-based treatment recommendations.
Treatment recommendations derived from established levels of evidence were examined regarding their possible bias.
The following sources of bias were identified: 1. The levels of evidence are related to the quality of studies, not to reported effect sizes; 2. external validity of studies are not taken into account systematically; 3. absence of evidence is not evidence of absence; 4. the ethical framework of many clinically relevant objectives cannot be represented adequately in randomized controlled trials; 5. achieved consensus between professionals, users, and relatives may be highly important and receives inadequately low levels of recommendation due to the definition of evidence.
There is an urgent need for a clear and transparent method to get from the analysis of existing evidence to treatment recommendations.
作者与一个多专业工作小组共同负责制定德国精神科协会关于攻击行为的治疗指南。基于他在此过程中的经验，他对当前制定循证治疗建议的方法提出了一些批判性意见。
对源自既定证据水平的治疗建议进行了可能存在的偏倚检查。
识别出以下偏倚来源：1. 证据水平与研究质量相关，而非与报道的效应量相关；2. 未系统考虑研究的外部效度；3. 缺乏证据并非没有证据；4. 许多临床相关目标的伦理框架在随机对照试验中无法得到充分体现；5. 专业人员、使用者和亲属之间达成的共识可能非常重要，但由于证据的定义，其推荐级别过低。
迫切需要一种清晰透明的方法，以从现有证据分析得出治疗建议。
