Idvall Ewa, Berg Katarina, Unosson Mitra, Brudin Lars
Department of Medicine and Care, Division of Nursing Science, Linköping University, Sweden.
J Eval Clin Pract. 2005 Oct;11(5):444-51. doi: 10.1111/j.1365-2753.2005.00555.x.
Differences between patient and professional assessments on pain and pain management have been reported, but no further analysis has described the statistical problems of pseudocorrelation concerning the nature of these differences.
The aim of the present study was: (1) to investigate the differences between nurse and patient assessments of postoperative pain management in two hospitals, and (2) to discuss the nature and scope of these differences.
The subjects were 209 inpatients and 63 nurses from a central county hospital and 77 inpatients and 34 nurses from a university hospital. The 'Strategic and Clinical Quality Indicators in Postoperative Pain Management' questionnaire was used, comprising 14 items in four sub-scales (communication, action, trust and environment) and two questions concerning the worst pain experienced during the past 24 hours and general satisfaction.
Except for the trust sub-scale in one hospital, the correlations between patient and nurse ratings concerning all assessments were significant in both hospitals (r = 0.22-0.59). Both groups of patients had significantly higher (better) scores than judged by the nurses on the environment sub-scale and general satisfaction. In contrast, nurses from both hospitals tended to significantly underestimate patients' worst pain during the past 24 hours. Other differences between patient and nurse assessments were either non-significant or inconsistent between hospitals. Using so-called Oldham plots nurses tended to underestimate severe pain more often than mild pain, as judged by the patients, but this association was weak and statistically significant in one hospital only.
Although the effects of pseudocorrelation are minimized by using Oldham plots, they are not cancelled. This issue is discussed, and we conclude that this study does not support the notion that the nurses tend to underestimate severe pain more often than mild pain.
已有报道称患者与专业人员在疼痛及疼痛管理评估方面存在差异，但尚无进一步分析描述这些差异本质上的伪相关性统计问题。
本研究的目的是：（1）调查两家医院护士与患者对术后疼痛管理评估的差异，（2）探讨这些差异的性质和范围。
研究对象为一家中心县医院的209名住院患者和63名护士，以及一家大学医院的77名住院患者和34名护士。使用了“术后疼痛管理中的战略和临床质量指标”问卷，该问卷包括四个子量表（沟通、行动、信任和环境）中的14个项目，以及两个关于过去24小时内经历的最严重疼痛和总体满意度的问题。
除一家医院的信任子量表外，两家医院中患者与护士在所有评估方面的评分相关性均显著（r = 0.22 - 0.59）。两组患者在环境子量表和总体满意度方面的得分均显著高于护士的判断（更好）。相比之下，两家医院的护士往往显著低估患者过去24小时内最严重的疼痛。患者与护士评估之间的其他差异要么不显著，要么在两家医院之间不一致。根据患者的判断，使用所谓的奥尔德姆图时，护士往往比轻度疼痛更频繁地低估重度疼痛，但这种关联较弱，仅在一家医院具有统计学意义。
尽管使用奥尔德姆图可将伪相关性的影响降至最低，但并未消除。对这一问题进行了讨论，我们得出结论，本研究不支持护士往往比轻度疼痛更频繁地低估重度疼痛这一观点。
