Background. Recent studies have reported lower statistics of upper limb (UL) weakness (48-57%) compared to widely cited values collected over 2 decades ago (70-80%). Objective. To explore potential factors contributing to the accuracy of prevalence values of UL weakness using a case study from a single regional centre. Methods. All patients admitted to the acute stroke unit with suspected diagnosis of stroke were screened from February 2016 to August 2017. Upper limb weakness was captured (a) prospectively using the Shoulder Abduction and Finger Extension (SAFE) score performed by unit physical therapists within 7 days post-stroke and (b) retrospectively via chart review using the National Institutes of Health Stroke Scale (NIHSS) arm score at admission and 24 hours post-admission. Results. A total of 656 patients were admitted with a first-ever stroke, and 621 (95%) individuals were administered the SAFE score. A total of 40% of individuals had UL weakness using the SAFE score (SAFE ≤8) at a mean time of 1.9 (SD 1.5) days post-stroke. In the same sample, 57% and 49% had UL weakness using the admission and 24-hour post-admission NIHSS arm score, respectively. Conclusions. The accuracy of population-level UL weakness prevalence values can be affected by weakness measure and score cut-off, time post-stroke weakness is captured, sample characteristics and use of single or multiple sites. Researchers using prevalence values for clinical trial planning should consider these attributes when using prevalence data for estimating recruitment rates and resource needs.

背景。最近的研究报告显示，与 20 多年前收集的广泛引用的值 (70-80%) 相比，上肢 (UL) 无力的统计数据 (48-57%) 较低。客观的。使用来自单个区域中心的案例研究来探讨影响 UL 弱点患病率值准确性的潜在因素。方法。 2016 年 2 月至 2017 年 8 月期间，对所有入住急性卒中病房疑似诊断为中风的患者进行了筛查。(a) 使用单位理疗师在 7 天内进行的肩外展和手指伸展 (SAFE) 评分前瞻性地捕捉上肢无力情况(b) 使用入院时和入院后 24 小时的美国国立卫生研究院卒中量表 (NIHSS) 手臂评分通过图表回顾进行回顾。结果。共有 656 名首次中风患者入院，其中 621 名 (95%) 人接受了 SAFE 评分。根据 SAFE 评分（SAFE ≤ 8），在中风后平均 1.9 (SD 1.5) 天时，共有 40% 的个体出现 UL 无力。在同一样本中，根据入院时和入院后 24 小时 NIHSS 手臂评分，分别有 57% 和 49% 的人患有 UL 无力。结论。人群水平 UL 肌无力患病率值的准确性可能受到肌无力测量和评分截止值、捕获中风后肌无力的时间、样本特征以及单个或多个部位的使用的影响。使用流行率值进行临床试验计划的研究人员在使用流行率数据估计招募率和资源需求时应考虑这些属性。