【作者】 孟亚军；

【导师】 李宁秀；

【作者基本信息】 四川大学 ， 社会医学与卫生事业管理， 2007， 博士

【摘要】 随着高效抗病毒疗法等的应用，HIV(Human immunodeficiency virus)感染者的生存时间大为延长。社会歧视的存在，加上对治疗的绝望、职业影响、经济困难、体像(体貌)影响、性生活影响、负罪感、悔恨感、对感染责任的认定、抗病毒药物的副作用等等诸多因素，都会严重影响HIV感染者的生命质量，甚至导致感染者产生对外界的敌意心理，采取反社会的行为。国外至今已经有很多量表用于测定HIV感染者的生命质量，但受制于其产生、验证的社会文化背景，国外量表很难真正经过翻译、文化调适后就适用于中国大陆地区的HIV感染者。中国的传统文化习惯和经济文化背景不仅与上述量表的研制地西方发达国家不同，也与其它发展中国家颇有不同，因此有必要发展一个适合中国HIV感染者特点的特异性量表，并且量表包括有反映感染者敌意心理趋势的维度。一、中国HIV感染者生命质量测定量表的研制1量表的形成1．1问题库的产生问题库的内容主要来自三个方面，一是对国内外相关量表及文献的参考。二是对HIV感染者本人的访谈。三是对相关专家以及基层防治人员的访谈，包括有防治经验的社会医学专家、流行病学专家、基层防治人员。1．2条目池及初始量表的形成问题库的内容，经过整理和调整，选择合适的表述方式及5点等距离尺度，形成条目池，所有条目尽量选用同一答案用词，调查所涉及到的是感染者最近四周的情况。形成的初始量表具有64个条目，其中有3个条目用于反映药物的疗效和副作用(正在服用抗病毒药物的感染者填写)，不纳入下一步的条目精简过程，直接进入最终量表。1．3初始量表的应用及最终量表的内容形成在四川、湖北、贵州及江西等地，共调查了443名HIV感染者。调查对象为非住院病人，年满18周岁以上，能够清晰的表达自己。初中以上文化程度的，由感染者自填，其它的由工作人员进行访谈填写。问卷填写前要获得感染者的知情同意，并给予感染者一定的补偿。感染者年龄范围为18～67岁，平均年龄35.9岁(标准差为8.53岁)，男性占65.9％。调查对象共来自12个民族，汉族占91.1％。平均查出血阳时间为25.4个月(标准差为22.9月)。其中的334名感染者还同时填写了普适性生命质量量表SF-36(Short form-36)。1．3．1条目精简过程反映性欲和性生活的3个条目拒答率都超过了10％，在访谈中也发现调查对象多回避此类问题，因此将其删除。所有条目的天花板、地板效应都没有超过70％。对剩下的条目进行因子分析，按特征根大于1的标准，形成了12个公因子，有12个条目在任意一个公因子上的负荷都小于0.50，删除这12个条目。根据公因子划分维度，维度11的Cronbach’sα值仅为0.46，删除此维度的2个条目。维度7及维度8中各有一个条目删掉后α值增加较多，因此删除这2个条目。维度4中有一个条目删掉后α值有所增加，但专家认为此条目比较重要，予以保留。对所有条目进行Spearmean相关分析，在维度5中，有2个条目之间的相关系数大于0.80，删掉其中1个条目。1．3．2最终量表的内容剩下的41个条目根据因子分析的结果形成10个维度：精神状况和精力，对健康和责任的担忧，家庭社会支持，敌意心理趋势，身体状况和活力，食欲和疼痛，经济状况的担忧，医生支持度，歧视感，生活满意度，加上评价药物疗效和副作用的3个条目，最终形成的针对中国HIV感染者的特异性生命质量量表(Quality of life for the Chinese HIV-infected，QOL-CHIV)共有44个条目。各条目得分范围分别为1～5分，得分较高表示相对好的生命质量。各维度得分为维度内条目得分之和，然后转化为0～100分的尺度。2 QOL-CHIV量表的心理测量学特点2．1各维度得分情况及天花板、地板效应QOL-CHIV量表各个维度都没有发现明显的天花板和地板效应。精神状况和精力，对健康和责任的担忧，家庭社会支持，敌意心理趋势，身体状况和活力，食欲和疼痛，经济状况的担忧，医生支持度，歧视感，生活满意度10个维度得分平均值±标准差分别为：42.96±22.29，34.62±19.81，52.41±27.69，70.52±21.63，40.48±21.56，61.59±20.66，28.75±23.36，65.61±27.89，50.96±27.49，46.75±25.78。反映药物疗效和副作用的三个条目为“服用药物使自己很难过正常的生活、觉得药物有效、服药的实际情况比想像的要难受些”，得分平均值±标准差分别为：42.93±31.76，73.36±24.11，52.35±26.58。2．2 QOL-CHIV量表的重测信度、内部一致性信度10个维度的2周重测信度系数(Intraclass correlation coefficient，ICC)分别为：0.75、0.58、0.76、0.79、0.80、0.67、0.63、0.61、0.58、0.63(N=40)。前9个非单条目维度的Cronbach’sα分别为0.91、0.84、0.89、0.78、0.70、0.73、0.80、0.90、0.76。按照全量表计算的Cronbach’sα为0.90，重测信度系数为0.80。2．3量表的效度进行多特质/多条目相关分析，各个维度的成维率都为100％，条目与本维度的相关系数在0.57～0.96之间，对量表进行因子分析，特征根大于1的9个维度共解释了总变异的64.74％。量表内容来自于国外HIV感染者专用量表，以及对相关专家及感染者的访谈内容，专家评议认为具有较好的表面效度和内容效度。以SF-36作为效标，两量表所计算的总分(各维度按相同的权重)的Spearman相关系数为0.70。在两量表中反映相近内容的维度，如疼痛、疲劳等，其Spearman相关系数都在0.6以上。二QOL-CHIV量表对感染者生命质量的评价1量表的总体得分情况分析有3个维度的平均分超过60分：敌意心理趋势维度得分平均分和中位数都在70以上，但在3个条目中41人次所得分为最低分；感染者医生支持度得分相对较高，表明感染者对防治人员的认同；食欲和疼痛维度得分也超过60。得分最低的几个维度所反映的内容和访谈内容相当一致，包括经济状况的担忧、对健康和责任的担忧等。反映药物的副作用及对感染者生活影响的两个条目得分平均值和中位数都不到60分。说明抗病毒药物疗效的条目平均分为73.4分，中位数也有75分，这3个条目得分分布与定性访谈的内容一致。2维度得分的单因素和多因素分析有4个维度的得分在不同性别感染者之间有显著差别，都是女性得分高于男性。≤30岁、30～40岁、>40岁组的感染者中，5个维度得分有显著区别，都是≤30岁组生命质量较差，>40岁组生命质量较好。将感染者按照婚姻状况分成3组：未婚组、在婚组、离异或丧偶组，在7个维度中感染者的生命质量有差别，全部都是在婚者的生命质量要好。感染者按工作状况分为三组，无工作或务农组、短期或偶尔工作组、长期稳定工作组，有7个维度的生命质量不一致，两两比较，往往都是长期稳定工作组生命质量最好，无工作或务农组生命质量最差。以首次检测血阳的时间长短，将感染者分为两组：≤1年和>1年。在敌意心理趋势、身体状况和活力中，感染时间短者生命质量稍好，在歧视感维度中，感染时间长者得分要高。不同感染方式的感染者在7个维度得分有明显区别，其中生命比较好的都是受血或售血途径的感染者。家里还有其它感染者的话，感染者的生命质量在4个维度中得分较高。正在服用抗病毒药物的感染者，有5个维度得分高于未服用药物者。按感染者所选的不同医疗负担方式分组，在6个维度中各组得分有差异，都是完全免费组的生命质量较好。按调查对象所来自的四个省份四川、湖北、贵州和江西分组，各个维度得分都不一致，贵州的生命质量相对较差，湖北和江西的相对较好。本次研究中，大部分感染者教育水平较低，具体到不同教育水平两两比较时，各维度没有发现明显的区别。多元线性回归分析的结果表明，在QOL-CHIV量表的10个维度中，被纳入次数最多的是医疗负担方式和工作状况。感染时间、所属地域、性别也是比较重要的影响因素，对几个维度的生命质量也有作用。其它的因素如年龄、是否正在服用药物、教育水平、婚姻状况、家中有其它感染者、感染方式等，也对1个或2个维度的生命质量有影响。三SF-36量表对感染者生命质量的评价在4省的344名感染者中应用了SF-36量表，这些感染者年龄在18～67岁之间，平均34.6岁，标准差为8.3岁。共来自12个民族，大多数为汉族(302人，88.6％)。SF-36在躯体角色、情感角色维度中地板效应较高。各个维度的Cronbach’sα系数都大于0.7。2周重测信度方面，在躯体功能维度低于0.6，另有2个维度的重测系数在0.6～0.7之间，其它维度都在0.8以上。各维度的成维率均为100％。用因子分析法来分析结构效度，提取出了7个主成分，解释了总变异的65.3％，条目按照较大的因子负荷划入相应维度的话，与原SF-36量表划分有一定差异。与具有相似经济文化背景的四川省普通农村居民比较，本次调查中HIV感染者的所有8个维度得分明显较低。单因素和多因素分析的结果与QOL-CHIV量表的结果类似。从两个量表单因素和多因素分析的结果来看，QOL-CHIV量表由于特异性强，更能够区分出不同特征的HIV感染者。四总结和结论QOL-CHIV量表是一个可信、可靠的HIV感染者的特异性量表，而且易于填写，容易获得好的依从性。首次在量表中添加了一个重要的维度：敌意心理趋势。QOL-CHIV量表能单独应用于感染者，也可与其它普适量表联合使用。QOL-CHIV量表内容易于理解，对于接受过初中教育的感染者能采取自填、面对面访谈和电话访谈等填写方式。本次调查对象多居住在乡镇，受教育程度相对较低，而且本次研究没有采取随机抽样的方法，在今后的不断推广使用中，QOL-CHIV量表将得到进一步验证、调整，使其更加完善。SF-36在中国HIV感染人群中的信度和效度是可以接受的。作为应用最为广泛的生命质量测定工具之一，SF-36的使用有利于不同文化背景或是同一文化背景内不同特征人群之间的比较。更多还原

【Abstract】 BackgroundsEffective treatments such as highly active antiretroviral therapy(HAART) have prolonged the life span and reduced opportunistic infections for many HIV-infected individuals. As a result, the quality of life(QOL) and social well-being of patients with HIV/AIDS(Acquired immune deficiency syndrome) has become a major concern of both patients and health services providers.The people living with HIV/AIDS(PLWHA) have suffered from serious stress and depression. These factors have contributed to the deterioration of QOL of the patients and impact their treatment and rehabilitation, sometimes even led to hostile mentality to outside and anti-social behaviors.While there are currently many measures for assessing QOL of the PLWHA, the majority of these measures have been developed in developed countries. As a result, it is unclear as to how applicable the translations of such measures might be for the majority of HIV sufferers that live in China with vastly different cultures and incomparable health resources. Therefore, it is necessary to develop a QOL mesaure which is specific for Chinese PLWHA and this measure has the dimension assessing the hostile mentality trend. The development of a quality of life measure specific for Chinese people living with HIV/AIDS1 The formation of the measure1.1 The formation of question databaseThe contents of question database mainly came from the following three sources: First source was the relevant measures and literatures from China and other countries, second source was the contents of in-depth interviews with the PLWHA, and the third was the contents of in-depth interviews with some experts and health workers.1.2 The formation of the item pool and primary measureAfter adjustment, the contents in question database were transformed to item pool. The items asking patients about their feeling in the recent four weeks. A five-point scale was designed to score the items. Nearly all items used the same set of Likert-style frequency descriptors("all of the time","a lot of time","some of the time","a little of the time","none of the time"). The highest possible score of each item was 5 and the lowest was 1. The primary measure including 64 items. Three items out of them were used to reflect the effect and side-effect of the anti-virus medicine(only used for the patients who were receiving HAART) and were saved to the final measure directly.1.3 The application of the primary measure and the formation of the final measureThis study was undertaken in Sichuan, Hubei, Guizhou and Jiangxi provinces. 443 eligible PLWHAs participated the study. The participatants were older than 18 years old and intellectual intact. They were all outpatients. Informed consents were obtained before the survey. For those who had received the education of junior high school, the questionnaires were self-administered. Otherwise, the questionnaires were administered through face-to-face interviews. Each participant was given some money to compensate for their cost of participation. The mean age of the participants was 35.9 years(ranging from 18 to 67 years, standard deviation is 8.53 year), 65.9% of the patiens were male. These participants belonged to 12 nationalities, 91.1% of them came from the Han nationality. Mean length from the first time of being tested as HIV positive was 25.4 months. The generic QOL measure SF-36 was also be applied in 344 patients out of the 443 participants.1.3.1 The dimension-identification and item-removal processesFirstly, the response rate of all items were calculated. Response rates of three items which reflected the sexuality and sex life was lower than 90.0%. In the course of interviewing with patients, we also found they avoided questions which involving sex.These three items were removed. Distributions of all items’ response were assessed, no item had≥70% of the sample choosing the same extreme response option.Next, all remaining items were placed into principal components analysis(varimax rotation). 12 items that loaded <0.50 on any factor were removed. For each newly identified factor, an internal consistency reliability coefficient(Cronbach’sα) was computed. Cronbach’sαvalue of factor 11 was only 0.46, two items in this factor were removed. Coefficients then were recomputed with items removed serially. Two items of facotr 7 and factor 8 were removed, because when the two items were removed, the Cronbach’sαincreased. The same conditions happened on one item of factor 4, but after discussing the significance of this item with experts, it was saved. Spearman correlation analysis was done in all items. Two items within factor 5 with a correlation coefficient value >0.80 were identified, one item of them was removed.1.3.2 Contents of the final measureAfter the process of item reduction, 41 items were remained. Factor analysis was done to structure the dimension. There were nine factors formed(eigen value≥1.0), explained 64.74% of the total variance. Another item which reflecting the satisfection of their life loading <0.40 on every factor formed a single dimension. In additional of 3 items reflecting the effect and side effect of anti-virus medicine, there were ten dimensions and one facet being formed. The ten dimensions and one facet included Mentality and energy, ME; Health and responsibility worries, HRW; Support of family and society, SFS; Hostility mentality trend, HMT; Physical function and vitality, PFV; Appetite and pain, AP; Financial worries, FW; Opinion on the health worker, OHW; Perception of stigma, PS; Life satisfaction, LS; Effect and side-effect of anti-virus medicine, ESM. So the meaure Quality of life for the Chinese HIV-infected(QOL-CHIV)including 44 items. Score of each item ranged from 1 ot 5, higher score reflecting a better QOL. The dimension score was the sum of its items’ scores and could be changed into 0～100 scale.2 Psychometric properties of the QOL-CHIV2.1 Score distribution of all dimensionsNo dimension displayed obvious floor and ceiling effect. The patients got the following scores of the ten dimensions: 42.96±22.29, 34.62±19.81, 52.41±27.69, 70.52±21.63, 40.48±21.56, 61.59±20.66, 28.75±23.36, 65.61±27.89, 50.96±27.49, 46.75±25.78(mean+standard deviation), respectively. Scores of the three items which reflecting the effect and side effect of anti-virus medicine were displayed:42.93±31.76, 73.36±24.11, 52.35±26.58, respectively.2.2 Internal consistency reliability and test-retest reliability40 participants were sampled to repeat the questionnaire survey two weeks after the first round of survey. The test-retest reliability was assessed with intra-class correlation coefficients(ICC), of the ten dimensions the ICC ranging from 0.58 to 0.80. The first 9 dimensions weren’t one-item dimension, their Cronbach’sαcoefficients ranged from 0.70 to 0.91. For the whole measure, the ICC reached 0.80 and the Cronbach’ sαcoefficient reached 0.90.2.3 Validity of this measureThe result of factor analysis showed that there were nine factors formed(eigen value≥1.0), explained 64.74% of the total variance. Multitrait/Multiitem correlation analysis gave the result that the scaling success rates of all dimension were 100%. Pearson correlation coefficients between items with their dimension ranged from 0.57 to 0.96(not single-item dimension).Contents of this measure came from foreign measure specific for PLWHA and interviews with PLWHA, experts and health worker, the content validity should be acceptable. Taking the SF-36 as criterion, spearman correlation coefficient between the sum score of the two measures was 0.70. Dimensions in the two measures which reflecting the similar content had spearman correlation coefficients exceeding 0.6.Assessment of QOL of the PLWHA with QOL-CHIV1 Overall analysis of the score of QOL-CHIVThere were three dimensions had the mean score exceeding 60. Althrough mean and median of the score of hostility mentality trend dimension exceeding 70, there were still 41 times of patients getting lowest score on three items of this dimension. Score of the opinion on the health worker was relative high, indicating the approval to the health worker from the PLWHA. Particpants in this study were out-patients, HIV may not aggrieve their body severely, which could explain the result that mean score of the dimension appetite and pain exceeding 60. Mean scores of the patients were relatively low in the following dimensions: health and responsibility worries, financial worries, which was agree with the contents of interviews with the PLWHA. There were three items used to reflect the effect and side effect of the anti-virus medicine. two of the three items reflecting the side-effect and the impact on patients’ life got relatively low scores, mean and median of the scores didn’t reach 60. Another item which reflecting the effect got higher score, mean score was 73.4 and median was 75. The distribution of these socres were also agreed with the contents of inerviews.2 Single-factor analysis and multivariate analysisFirstly single-factor analysis was done. The scores of four dimensions showed significant difference between the different sex, female had the higher score than male. The participatants were divided into<30 years old, 30-40 years old, and 40-year-old group. There were noticeable difference in five dimensions between the groups,≤30 age group showed poorest quality of life and 40-year-old group showed the best. Infected patients could be divided into three groups according to their marital status: unmarried, in marriage, divorced or widowed. There were significant difference in seven dimensions and the in marriage group got the hignest score. Patients were divided into three groups according to the status of work: unemployed or farming, short-term or occasional work, long-term and stable work. QOL was inconsistent in seven dimensions, usually were the first group showed the worst QOL and the third showed the best. According to the first serum-positive time the patients were divided into two groups:≤one year and >one year. In the dimension of hostility mentality trend and physical function and vitality,≤one year group showed better QOL. But in the dimension of perception of stigma, >one year group showed better QOL. Patients with different infecting ways scored differently in seven dimensions, receiving blood group and saling blood group showed better QOL. If there were another PLWHA in family, the patient got higher scores in four dimensions. The patiens who were receiving HAART service showed better QOL in five dimensions. Patients with completely free treatment had better QOL in six dimensions. Patients in different provinces got inconsistent scores, QOL of patients from Guizhou was relatively poor and from Hubei and Jiangxi was relatively good. Majority patients in this study had poor educational backgrounds, so they did not showed different QOL according to different educatoional level.Results of multiple linear regression analysis formed ten regression equations respective to the tern dimensions. There were two factor had the most times to be included in the equation: the way of paying for treatment and working status. Length of being infected, where they came from and gender also be included in several equations. Other factors such as age, receiving HAART service, educational level, marital status, having another infected patient in family and way of being infected also be included in one or two regerssion equation.Assessment the QOL of the patients with SF-36SF-36 was applied in 344 HIV-infected persons in four provinces. The mean age of the participants was 34.6years(ranging from 18 to 67 years, standard deviation was 8.3 years). These patients came from 12 nationalities, most was the Han nationality. Substantial flooring effects were observed for the two role-limited doamins(50.5% for"role limitations due to physical health problems" and 62.2% for"role limitations due to emotional problems"). A satisfactory internal consistency of the SF-36 has been demonstrated in our study, with Cronbach’sαexceeding 0.74. Test-retest correlation coefficients ranged from 0.54 to 0.80("Physical functioning" to"vitality")for the eight dimensions and less than 0.7 for three dimensions. The scaling success rate were 100% for all dimensions. Factor analysis showed that there were 7 factor and explain 65.3% of total variance. The item-factor distribution of the factor analysis was not very coincident with the SF-36.The PLWHA experienced poorer QOL in all of the eight dimensions of SF-36 compared to the general populations. Rusults of single-factor analysis and multivariate analysis made the familiar conclusions of QOL-CHIV. QOL-CHIV showed stronger specificity than SF-36 and could distingusih more patients with different characteristics.ConclusionsQOL-CHIV is a reliable and valid measure specific for Chinese PLWHA. It is easy to administer and acquire well compliance. This measure could capture the special properties of Chinese PLWHA. It is the first time for the important and concerned dimension"hostility mentality trend" being included to QOL measure. QOL-CHIV could be applied in the PLWHA alone, and also could be applied with other generic measure. This measure could be used in the way of self-administration, face to face interview and phone interview. The participants of this study were not randomly selected, QOL-CHIV measure will be further verified and adjusted to make it perfect.Although it may not be able to capture some unique aspects of QOL associated with HIV infection, the SF-36 has its merit in assessing QOL of PLWHA. Through using SF-36 we can compare the QOL of PLWHA with other people which has different characteristics.更多还原