Department of Medicine, McMaster University, Hamilton, Ontario, Canada.
J Stroke Cerebrovasc Dis. 2010 Sep-Oct;19(5):410-6. doi: 10.1016/j.jstrokecerebrovasdis.2009.08.002.
Proxy respondents are an important aspect of stroke medicine and research. We performed a systematic review of studies evaluating the reliability of proxy respondents for stroke patients. Studies were identified by searches of MEDLINE, Google, and the Cochrane Library between January 1969 and June 2008. All were prospective or cross-sectional studies reporting the reliability of proxy respondents for patients with a history of previous stroke or transient ischemic attack. One author abstracted data. For each study, intraclass correlation (ICC) or the k-statistic was categorized as poor (<or=0.40), moderate (0.41-0.60), substantial (0.61-0.80), or excellent (>0.80). Thirteen studies, with a total of 2618 participants, met our inclusion criteria. Most studies recruited patients >3 months after their stroke. Of these studies, 5 (360 participants; 5 scales) evaluated reliability of proxy respondents for activities of daily living (ADL), and 9 (2334 participants; 9 scales) evaluated reliability of proxy respondents for quality of life (QoL). One study evaluated both. In studies, the ICC/k for scales ranged from 0.61 to 0.91 for ADL and from 0.41 to 0.8 for QoL. Most studies reported that proxy respondents overestimated impairments compared with patient self-reports. Stroke severity and objective nature of questions were the most consistent determinants of disagreement between stroke patient and proxy respondent. Our data indicate that beyond the acute stroke period, the reliability of proxy respondents for validated scales of ADL was substantial to excellent, while that of scales for QoL was moderate to substantial.
代理应答者是卒中医学和研究的一个重要方面。我们对评估卒中患者代理应答者可靠性的研究进行了系统综述。通过对 1969 年 1 月至 2008 年 6 月间 MEDLINE、Google 和 Cochrane Library 的检索，确定了研究。所有研究均为前瞻性或横断面研究，报告了有既往卒中或短暂性脑缺血发作史的患者的代理应答者的可靠性。一位作者提取数据。对于每项研究，采用组内相关系数（ICC）或 k 统计量来分类，ICC 或 k 统计量<0.40 为差（<=0.40），0.41-0.60 为中（0.41-0.60），0.61-0.80 为优（0.61-0.80），>0.80 为极好（>0.80）。共有 13 项研究，总计 2618 名参与者，符合我们的纳入标准。大多数研究在卒中后 3 个月以上招募患者。这些研究中，5 项（360 名参与者；5 个量表）评估了代理应答者对日常生活活动（ADL）的可靠性，9 项（2334 名参与者；9 个量表）评估了代理应答者对生活质量（QoL）的可靠性。有一项研究同时评估了这两个方面。在这些研究中，ADL 量表的 ICC/k 值范围为 0.61-0.91，QoL 量表的 ICC/k 值范围为 0.41-0.8。大多数研究报告称，与患者的自我报告相比，代理应答者高估了损伤程度。卒中严重程度和问题的客观性质是卒中患者和代理应答者之间意见不一致的最一致决定因素。我们的数据表明，在急性卒中期过后，ADL 验证量表的代理应答者的可靠性为优到极好，而 QoL 量表的可靠性为中到优。
