【作者】 陈学涛；

【导师】 易东；

【作者基本信息】 第三军医大学 ， 流行病与卫生统计学， 2009， 博士

【摘要】 世界卫生组织(WHO)对医疗服务质量有效性的解释是:“病人得到期望的和满意的效果。”随着医疗机构之间的竞争愈演愈烈,医疗服务的满意研究受到了更多的重视。然而,病人满意研究的现状是:一方面各类病人满意研究得出了异常高的的满意度结果,另一方面却是社会对医院服务的强烈批评。“高满意低忠诚”的现象无法从病人满意研究中得到解释,病人满意并不是病人忠诚的充分条件,病人满意研究面临着前所未有的困惑和越来越多的质疑。病人忠诚研究是病人满意研究的扩展。从国内外研究现状看,病人满意的研究文献众多,但病人忠诚的系统研究却十分缺乏,国内外现有病人忠诚研究存在的问题主要表现为:一是模型设计过于简单或者只以病人满意为核心变量,对病人忠诚的解释效果差。事实上,病人忠诚的形成受多种因素的影响和调节,是一个非常复杂的过程,现有研究主要以病人满意的大小作为病人忠诚的影响因素,而未考虑病人满意不确定性、市场因素等对病人忠诚形成的影响。二是照搬市场研究中顾客忠诚的研究模型和方法,在构建病人忠诚模型时,对医疗服务固有的特殊性考虑不够,研究结论难以令人信服。三是定性研究多,定量研究少,主要对病人忠诚的重要性、病人满意与忠诚的关系等进行定性分析,研究结论缺乏实证研究的验证,我们认为对于病人忠诚而言,实证研究是不可或缺的,实证研究不仅可以验证理论研究结论,而且可更精确地描述概念间的关系强度。本文通过理论和实证研究,系统研究病人忠诚形成的影响因素,构建了适合医疗服务特点的病人忠诚意向模型。首先从病人质量感知、满意评价和忠诚意向的角度深入讨论医疗服务的特点,对于医疗服务特点的考虑,贯穿于病人忠诚意向理论模型的构建和概念的定义与测量研究的始终;在此基础上,结合市场研究中顾客满意和顾客忠诚理论,特别是最新理论研究成果,系统研究病人忠诚意向、病人感知质量、感知价格、病人满意、转换成本、替代选择等要素的内涵和外延,建立病人忠诚意向理论模型;然后通过统计学研究和实证研究对理论模型进行求解和验证,最终建立起适合医疗服务特点的病人忠诚意向模型。本研究完全从病人感知的角度研究病人忠诚,不仅体现在概念选择和模型构建上,而且体现在测量工具的形成上,都是从病人的观点出发研究问题,研究结论具有重要的实践指导意义,有助于医疗机构理解病人忠诚的形成机制,建立病人忠诚的培育机制,提高医疗服务质量,提升医疗市场竞争力。整个研究形成了较为完善的病人忠诚研究模式,为病人忠诚的研究作出了理论和研究方法上的贡献。在统计学方法上,本文深入研究了病人忠诚意向模型的PLS-SEM求解方法,重点研究了PLS-SEM的模型评价方法,基于PLS-SEM的细分算法以及PLS-SEM的调节效应和调节效应的估计准确性与统计功效问题。在模型评价上,推导了PLS-SEM的内部关系、外部关系和总体预测效能的评价指标,用R~2评价内部关系,用共同因子评价外部关系,用冗余评价整体模型的预测效能,还研究了blindfolding的交互验证评价方法。在基于PLS-SEM的细分算法研究上,对CLR算法进行了改进,用有限混合模型建立细分模型,用EM算法对细分模型进行求解,并提出用AIC,CAIC,BIC和信息熵对细分效果进行评价,进而确定细分组数。在PLS-SEM的调节效应研究上,首先研究了调节变量的模型设定问题,即当预测LV和调节LV均满足PLS-SEM模型设定时,交互效应LV满足模型设定的情形;然后针对调节效应的低统计功效和PLS-SEM的估计有偏性的问题,对PLS-SEM的调节效应的估计准确性和统计功效进行了Monte Carlo模拟研究。在估计准确性上,PLS-SEM在估计交互效应大小时,虽然存在低估,但估计偏差小于10%,估计偏差与总体交互效应的大小关系密切;在统计功效上,PLS-SEM估计交互效应的统计功效较低,平均为0.53,交互效应的统计功效既与样本量有关,也与总体交互效应的大小有关,较低的交互效应统计功效提示我们,从模型中删除不显着的调节变量应慎重。本文理论和实证研究获得的主要研究结果有:1.感知质量、感知价格、病人满意和病人忠诚构成了病人忠诚模型的核心关系链病人感知质量直接影响病人满意,并通过病人满意影响病人忠诚,而感知价格对病人满意和病人忠诚都有直接影响。感知质量、感知价格、病人满意和病人忠诚构成了病人忠诚模型的核心关系链。本文用SERVQUAL五维(有形性、可靠性、响应性、保证性、移情性)度量了医疗服务的感知质量,五维感知质量的操作性定义概括为“技术可靠、响应需求、安全放心、关怀礼貌、有形感知”,全面反映了病人对医疗服务的期望与评价准则。实证研究中的整体样本情形,验证了五维质量对病人满意的显着性影响,其中可靠性、保证性、移情性是五维感知质量中最为重要的质量维度。本文用感知价格取代经典模型中的感知价值,避免了概念上的重复,即质量和价值的同义反复,理清了概念间关系。2.信息沟通、就医经验与病人满意不确定性对病人忠诚的影响本文理论和实证研究发现,病人满意的高低和病人满意评价的不确定程度都会对病人忠诚产生重要影响。医疗服务具有复杂性、专业性的特点,病人医疗知识和就医经验的缺乏,导致满意评价的不确定性,而病人满意的不确定性是导致高满意、低忠诚的重要原因之一。本文研究结果显示信息沟通对病人就医经验有显着影响,通过充分、有效的信息沟通,增强病人的就医经验,可降低病人满意的不确定程度,虽然在整体样本情形下,就医经验对病人满意和病人忠诚之间的调节作用并不显着,但在细分样本情形下,约占样本三分之一的“综合型”病人群体在满意向忠诚转化的过程中,就医经验起了显着性调节作用。3.转换障碍(转换成本和替代选择性)对病人忠诚的影响不仅病人满意影响病人忠诚意向,转换成本、替代选择性也是病人忠诚形成的重要的直接因素。转换成本和替代选择性是病人择医转换障碍的两个方面,病人感知的转换成本越高,病人忠诚度越高;病人感知的替代选择性越高,转向竞争医院的可能性越大。实证研究显示转换成本比替代选择性对病人忠诚的作用更为重要。4.关于病人忠诚影响因素的细分特征本文将病人细分为三组,分别命名为“全面质量型”、“价格敏感型”和“综合型”。从病人忠诚的直接影响因素看,病人满意和转换成本是各组病人忠诚形成的共同因素,感知价格只对“价格敏感型”和“综合型”病人忠诚形成有显着影响,而替代选择性、就医经验只对“综合型”病人忠诚形成作用显着;从病人满意的直接影响因素看,可靠性、保证性、移情性感知质量和信息沟通是病人满意形成的共同因素,而响应性质量只对“全面质量型”和“价格敏感型”病人满意形成有显着影响,有形性质量只对“全面质量型”病人满意形成有显着影响,病人感知价格只对“综合型”病人满意形成作用显着。5.病人忠诚及其维度病人忠诚意向体现为重返就诊意愿和口碑传播(推荐)意愿两个维度。在病人忠诚模型中应将二者并为同一概念——“病人忠诚意向”来研究。医疗服务对象具有广泛性的特点,医院存在着大量有潜在需求的顾客。每一个接受服务且满意的病人,不仅自身具有较高的忠诚意向,而且会成为医院最有效和廉价的口碑宣传者,因此,口碑传播意愿在医疗市场开拓中的地位和作用是不可或缺的。更多还原

【Abstract】 World Health Organization explained the validity of medical service quality: " patients get the satisfied effect as expecting. " As medical market competition is becoming more fierce, study on patient satisfaction is regarded more highly. However, on one hand, all kinds of satisfaction study have reached extraordinarily high degree of satisfaction result, on the other hand, the intense criticism is given to the hospital. In fact, patient satisfaction is not the sufficient condition of patient loyalty. How to explain the phenomenon of ’high satisfaction and low loyalty’? The study on patient satisfaction is confronted with unknown puzzle.The study on patient loyalty is the expansion of patient satisfaction, as the ’ultimate goal’, loyalty is more important than satisfaction. Researchers and managers acknowledge that small changes in loyalty and retention can yield disproportionately large changes in profitability. A lot of studies on customer satisfaction and loyalty in service industries have been published, but relatively few in healthcare organizations. Problems existing in the previous paper are as follow:①The models are too simple to explain patient loyalty.②The conclusions are too difficult to make people convinced as lacking for considering particularity of healthcare organizations.③By few quantitative study and much qualitative analysis , the conclusion is short of the verification.We construct patient conative loyalty model by theoretical and empirical study. Firstly, this paper analyzes service attributes of medical treatment about patient perceived quality, patient satisfaction and patient loyalty. Secondly, systematic studying is devoted to the connotation and extension of patient conative loyalty, perceived quality, perceived price, patient satisfaction, switching costs, substitute selectivity et al. Then we build patient conative loyalty theoretical model. Lastly, we verify theoretical model by statistical and empirical study. In this research, not only conception selection and measurement but also model construction is based on the patient’s view. The conclusions of this research have important practical meanings which can help medical service providers catching on interaction mechanism of patient loyalty, improving medical service quality and enhancing market competitive ability. The whole research mode for patient loyalty is formed which devoted to the theory and methods about loyalty research.We study profoundly the partial least squares approach to structural equation modeling including model validation, segmentation algorithm based PLS-SEM and moderating effect in PLS-SEM:①The communality index measures the quality of the measurement model for each block, and the R~2 index measure the inner model. Then the redundancy index measures the quality of the whole model for each endogenous block, taking into account the measurement model. It is also useful to slightly modify the communality index and express the redundancy index in terms of the LVs prediction so as to set the basis for validating the values by means of the cross-validation procedure named blindfolding.②A modified finite mixture is the approach that integrates customer segmentation into partial least squares based analyses of path models which is realized by the EM algorithm. When applying EM algorithm, the number of segments is unknown and the identification of an appropriate number of K classes is not straightforward. We propose using AIC, CAIC, BIC and entropy statistic for determining the number of classes in a mixture model.③We discuss the model specification on predictor, moderator and dependent variables. In succession, based of poor statistical power of moderating effect and biased estimation of PLS-SEM, Monte Carlo simulation is adopted to investigate the power and point estimate accuracy.Through the above research work, the main conclusions are as follows:1. The core relationship chain of the patient loyalty modelPatient perceived quality has a direct and positive effect on patient satisfaction, then quality has an indirect effect on patient loyalty via satisfaction. Both patient satisfaction and patient loyalty are affected directly by patient perceived price. The core relationship chain is constructed by patient perceived quality ,perceived price, patient satisfaction and patient loyalty .SERVQUAL is adopted to measure patient perceived quality in our study. The operational definition of SERVQUAL is technology reliability, responsiveness willingness, security and credibility, politeness and consideration, tangibles. We find that ’reliability’, ’assurance’ and ’empathy’ affect patient satisfaction more strongly. In the current models the relationship from quality to value maybe tautological as well as causal because quality is related to value by definition, so we use ’perceived price’ instead of "perceived cost’ in the patient loyalty model.2. The effect of information communication, medical experience and uncertainty of satisfaction evaluation on patient loyaltyThrough this research work, we find that both patient satisfaction scores and uncertainty of satisfaction evaluation affect patient loyalty profoundly. If a patient is not very knowledgeable about the medical service at stake, uncertainty of satisfaction evaluation may happen. uncertainty of satisfaction evaluation could cause the phenomenon of ’high satisfaction and low loyalty’. Medical experience is not significant for the full set of data, but in segmentation study, medical experience has a significant moderating effect when satisfaction translates into loyalty.3. The effect of switching barriers(switching costs and substitute selectivity) on patient loyaltyNot only patient satisfaction but also patient switching barriers, including switching costs and substitute selectivity, have direct effect on patient loyalty. There is a positive relationship between switching costs and patient loyalty, while a negative relationship exists between substitute selectivity and patient loyalty. Our study shows that the strength of the relationship between switching costs and patient loyalty is stronger than that between substitute selectivity and patient loyalty.4. The segmentation on the antecedent variables of patient loyaltyWe divide patients into three groups which named ’quality oriented group’ , ’price oriented group’ and ’compositive group’ respectively. Through the segmentation research, we find that satisfaction and switching costs are common direct factors of patient loyalty for all kinds of patients, but substitute selectivity and medical experience are significant antecedents only for ’compositive group’, ’responsiveness’ ,’assurance’, ’empathy’ and information communication are common factors of patient satisfaction for all patient groups, but the effect of ’responsiveness’ is only significant for ’quality oriented group’ and ’price oriented group’, while the effect of ’tangibles’ is only significant for ’quality oriented group’. 5. The dimension of patient loyaltyPatient loyalty is composed of revisiting intention and word-of-mouth communication. A satisfied patient not only has a great probability of revisiting behaviors, but also more likely become a free word-of-mouth advertising agents that informally bring networks of friends, relatives and other potential patients.更多还原