ASSESSING THE CLINICAL IMPORTANCE OF SYMPTOMATIC IMPROVEMENTS - AN ILLUSTRATION IN RHEUMATOLOGY
ASSESSING THE CLINICAL IMPORTANCE OF SYMPTOMATIC IMPROVEMENTS - AN ILLUSTRATION IN RHEUMATOLOGY
复制标题
DOI：
10.1001/archinte.153.11.1337
复制
发表时间：
1993-06-14
影响因子：
--
通讯作者：
LORIG, K
中科院分区：
文献类型：
--
作者：
REDELMEIER, DA;LORIG, K
目的：估计残疾症状的差异何时足够大，对个体患者有重要意义。设计：两组横断面分析：衍生集（n=46）和验证集（n=57）。环境：关节炎基金会，北加州分会。参与者：居住在社区的关节炎患者志愿者样本。主要结果测量：我们采用斯坦福健康评估问卷来评估个体的功能状态。然后，参与者彼此进行一对一的对话，并对他们遇到的每个人的残疾程度进行评估，判断他们的残疾程度是“好多了，‘’ ‘’稍微好一点，‘’ ‘’差不多，‘’ ‘’稍微差一点”还是“差得多”。对于每一次谈话，我们都计算了两个参与者的健康评估问卷得分之间的差异，并将差异与两人中每个人的主观比较评分联系起来。结果：健康评估问卷得分差异与主观比较评分显著相关（相关系数为0.41；95%可信区间为0.31 ~ 0.50）。我们估计，健康评估问卷得分需要相差约0.19个单位，平均受访者才会停止将自己评为“大致相同”，并开始将自己评为“稍好一些”（95%置信区间，0.10至0.28）。对第二组患者的分析显示了类似的阈值（平均值，0.23个单位；95%置信区间，0.13至0.23）。在两组中，健康评估问卷得分差异是个体评分的不完美预测因子，残疾程度较轻的参与者的阈值往往低于残疾程度较高的参与者的阈值。结论：一些在功能状态评分上有统计学意义的差异可能非常小，以至于它们代表了微不足道的症状缓解程度。意识到症状评分的最小差异对患者很重要，可以提供一个粗略的指导，帮助临床医生解释医学文献。
Objective: To estimate when a difference in disability symptoms is sufficiently large to be important to individual patients.Design: Cross-sectional analysis of two groups: derivation set (n=46) and validation set (n=57).Setting: The Arthritis Foundation, Northern California Chapters.Participants: Volunteer sample of patients with arthritis who live in the community.Main Outcome Measures: We applied the Stanford Health Assessment Questionnaire to assess the functional status of individuals. Participants then conductd one-on-one conversations with each other and rated whether their disability was ''much better,'' ''somewhat better,'' ''about the same,'' ''somewhat worse,'' or ''much worse' relative to each person they met. For every conversation we calculated the difference between the two participants' health assessment questionnaire scores and linked the difference to the subjective comparison ratings of each individual in the pair.Results: Health assessment questionnaire score differences were significantly correlated with subjective comparison ratings (correlation coefficient, .41; 95% confidence interval, 0.31 to 0.50). We estimated that health assessment questionnaire scores needed to differ by about 0.19 units for average respondents to stop rating themselves as ''about the same'' and start rating themselves as ''somewhat better'' (95% confidence interval, 0.10 to 0.28). Analysis of a second group of patients revealed a similar threshold (mean, 0.23 units; 95% confidence interval, 0.13 to 0.23). In both groups, health assessment questionnaire score differences were imperfect predictors of individual ratings and the threshold for less disabled participants tended to be lower than the threshold for more disabled participants.Conclusions: Some statistically significant differences in functional status scores may be so small that they represent trivial degrees of symptom relief. An awareness of the smallest difference in symptom scores that is important to patients can provide a rough guide to help clinicians interpret the medical literature.