Kim Son Chae, Young Lori, Berry Brigette
St. David's School of Nursing, Texas State University, Round Rock, Texas, USA.
Extendicare (Canada) Inc., Unit 227, Calgary, Alberta, Canada.
J Adv Nurs. 2017 Jul;73(7):1747-1756. doi: 10.1111/jan.13247. Epub 2017 Jan 31.
The aim of this study was to revise the 10-item Aggressive Behaviour Risk Assessment Tool for predicting aggressive events among residents newly admitted to long-term care homes.
The original tool had acceptable sensitivity and specificity for identifying potentially aggressive patients in acute care medical-surgical units, but its usefulness in long-term care homes is unknown.
A retrospective cohort study design was used.
All residents admitted to 25 long-term care homes in western Canada were assessed for the risk of aggression using the original tool within 24 hours of admission from January 2014 - December 2014 (n = 724). Incident reports of aggressive events occurring within 30 days of admission were collected. Multiple logistic regression and receiver operating characteristics analyses were performed.
Fifty-three residents of 724 exhibited aggressive behaviours. The demographic variable of age less than 85 years was found to be a positive predictor of aggressive events in multivariate logistic regression model and was added to the tool. The revised six-item Aggressive Behaviour Risk Assessment Tool for Long-Term Care consists of one new item, age less than 85 years and five items from the original tool: History of physical aggression, physically aggressive/threatening, anxiety, confusion/cognitive impairment and threatening to leave. The receiver operating characteristics of the revised tool with weighted scoring showed a good discriminant ability with satisfactory sensitivity and specificity at the recommended cut-off score of 4.
The revised six-item tool may be useful in identifying potentially aggressive residents newly admitted to long-term care homes.
本研究旨在修订10项攻击行为风险评估工具，以预测新入住长期护理机构的居民中的攻击事件。
原始工具在识别急性护理内科-外科病房中潜在的攻击性患者方面具有可接受的敏感性和特异性，但其在长期护理机构中的实用性尚不清楚。
采用回顾性队列研究设计。
对2014年1月至2014年12月期间加拿大西部25家长期护理机构收治的所有居民，在入院后24小时内使用原始工具评估攻击风险（n = 724）。收集入院后30天内发生的攻击事件的事件报告。进行多因素逻辑回归和受试者工作特征分析。
724名居民中有53名表现出攻击行为。在多因素逻辑回归模型中，年龄小于85岁的人口统计学变量被发现是攻击事件的正向预测因子，并被添加到工具中。修订后的长期护理机构六项攻击行为风险评估工具包括一项新项目，即年龄小于85岁，以及原始工具中的五项：身体攻击史、身体攻击性/威胁性、焦虑、意识混乱/认知障碍和威胁离开。经加权评分的修订工具的受试者工作特征显示出良好的判别能力，在推荐的临界值4时具有令人满意的敏感性和特异性。
修订后的六项工具可能有助于识别新入住长期护理机构的潜在攻击性居民。
