Strahl A, Gerlich C, Wolf H-D, Gehrke J, Müller-Garnn A, Vogel H
Abteilung für Med. Psychologie und Psychotherapie, Med. Soziologie und Rehabilitationswissenschaften, Universität Würzburg, Würzburg.
Bereich Sozialmedizin, Deutsche Rentenversicherung Bund, Berlin.
Gesundheitswesen. 2016 Mar;78(3):156-60. doi: 10.1055/s-0034-1390450. Epub 2014 Dec 22.
The sociomedical evaluation by the German Pension Insurance serves the purpose of determining entitlement to disability pensions. A quality assurance concept for the sociomedical evaluation was developed, which is based on a peer Review process. Peer review is an established process of external quality assurance in health care. The review is based on a hierarchically constructed manual that was evaluated in this pilot project.
The database consists of 260 medical reports for disability pension of 12 pension insurance agencies. 771 reviews from 19 peers were included in the evaluation of the inter-rater reliability. Kendall's coefficient of concordance W for more than 2 raters is used as primary measure of inter-rater reliability.
Reliability appeared to be heterogeneous. Kendalls W varies for the particular criteria from 0.09 to 0.88 and reached for primary criterion reproducibility a value of 0.37.
The reliability of the manual seemed acceptable in the context of existing research data and is in line with existing peer review research outcomes. Nevertheless, the concordance is limited and requires optimisation. Starting points for improvement can be seen in a systematic training and regular user meetings of the peers involved.
德国养老保险的社会医学评估旨在确定领取残疾抚恤金的资格。已制定了基于同行评审流程的社会医学评估质量保证概念。同行评审是医疗保健领域既定的外部质量保证流程。该评审基于在此试点项目中进行评估的分层构建手册。
数据库包含12家养老保险机构的260份残疾抚恤金医疗报告。19位同行的771份评审意见纳入了评分者间信度评估。肯德尔和谐系数W用于评估超过2名评分者时的评分者间信度。
信度似乎存在异质性。特定标准下的肯德尔W值在0.09至0.88之间变化,主要标准可重复性的肯德尔W值为0.37。
在现有研究数据背景下,手册的信度似乎可以接受,并且与现有同行评审研究结果一致。然而,一致性有限,需要优化。改进的起点可以是对相关同行进行系统培训并定期召开用户会议。