利用Rasch測量分析英國2003年病人為中心住院滿意度調查

Rasch Analysis of 2003 England Inpatient Questionnaire of the Picker Institute Europe

近年來，幾乎所有的醫療提供者都需要做病患的滿意度調查。值得一提 的是病患的滿意度調查已由傳統的滿意問項轉為對病人親身體驗流程事項的 答與悶，因而可以更具體的掌控住改善的主題與內容。Picker 歐洲機構即以此 精神由病人的眼光來設計滿意度問項，而非只是李克式量表滿意程度上的表 示。另一李克式量表問卷分析的主要問題，是傳統的順序量表當作等距量表 來做因素分析，而導致假設等距方法上的錯誤。 因此，本研究利用Rasch （1960） 測量的客觀等距的特性，來分析2003 年英國169 家醫院住院患者的滿意度調查。經由Picker 歐洲機構的同意，由網 站上下載問卷結果的資料，將各項拈標的0 至100 言十分，轉換為5 點計分，數值 愈大表示愈滿意。利用Winsteps （Linacre, 1999） 軟體進行等級量尺模式的檢 測，當問z真的outfit 或infit MNSQ 超出0.5 及1. 5 之外者，即予刪列而期以符合 其單一向度及部份獨立性的試題反應理論假設。 結果顯示的個問項經效度分析後剩下29 題符合其單一向度的Rasch 模 式。最困難的問項是醫病關條間的溝通，一般基層醫院的滿意度最高，在教 學醫院內或是在基層醫院間的滿意度是沒有統計檢定上的差異。

In recent years, almost all caregivers are required to conduct surveys of patients satisfaction. The critiques of patient satisfaction surveys have led to a new emphasis on measuring patients experience rather than satisfaction. The Picker Institute Europe has conducted such questionnaires to ask patients to report in detail on their experience of a particular provider at a specific point in time instead of asking patients to rate their care on a Likert scale. The main problem is that Picker as traditional others directly operated factor analysis on ordinal scores to justify its inference. In this study, we apply Rasch analysis (Rasch, 1960) to examine whether all items conform to the model s expectation so that the underlying latent trait can be quantified. We analyzed the data of the 2003 England Inpatient Questionnaire from a total 169 hospitals. With permission, we downloaded a total of 169 hospital datasheets from the 2003 England Inpatient Questionnaire of the Picker Institute Europe. The original indicator questions have been scored using a scale of 0 to 100. The scores represent the extent to which the patient s experience could have been improved. The rating scale model was fitted to these items using the computer software Winsteps (Linacre, 1999). Items with outfit or infit MNSQ not in the range of .05 and 1.5 are usually diagnosed as potential misfits to Rasch model conditions and considered for deletion from the assessed sequence. Successive Rasch analyses were performed until a final set of items satisfied the model fit requirements. The results show that 29 of the 45 items fitted the model s expectation fairly well. The most difficult items for the hospitals to earn inpatients satisfaction were staff-patient communication, whereas general practices were easier to receive patients satisfaction. No statistically significant differences between specialist hospitals and teaching hospitals were found.