Misspecification tests for binomial and beta-binomial models

The IOS test of Presnell and Boss (J. Am. Stat. Assoc. 2004; 99(465):216-227) is a general-purpose goodness-of-fit test based on the ratio of in-sample and out-of-sample likelihoods. For large samples, the IOS statistic can be approximated by a multiplicative contrast between two estimates of the information matrix, and in this way the IOS test is connected to White's (Econometrica 1982; 50:1-26) information matrix test, or IM test, which is based directly on the difference of two estimates of the information matrix. In this paper, we compare the performance of IOS to that of the IM test and of other goodness-of-fit tests for binomial and beta-binomial models, in both examples and simulations. Our findings suggest that IOS is strongly competitive, not only against the IM test but also against tests designed for specific binomial and beta-binomial models.     

SF-36健康调查量表中文版的研制及其性能测试

目的 研制SF-36健康调查量表中文版并验证量表维度建立及记分假设、信度和效度。方法 采用多阶段混合型等概率抽样法,用SF-36健康调查量表中文版对1000户家庭的居民进行自评量表式调查;参照国际生命质量评价项目的标准程序,进行正式的心理测验学试验。结果 在收回的1985份问卷中,18岁以上的有效问卷1972份,其中应答者1688人(85．6％),1316人回答了所有条目,372人有1个或以上的缺失答案,无应答者中文盲、半文盲占65．5％。等距假设在活力(VT)和精神健康(MH)维度被打破了,按重编码后值计算维度分数;条目集群的分布接近源量表及其他2个中文译本;除了生理功能(PF)、躯体疼痛(BP)、社会功能(SF)维度,其余维度有相似的标准差;除了SF、VT维度,其余6个维度条目维度相关一致;除了SF维度,7个维度集合效度成功率范围为75％～100％,,区分效度成功率范围为87．5％～100％。一致性信度系数除了SF、VT维度,其余6维度变化范围为0．72～0．88,满足群组比较的要求。两周重测信度变化范围为0．66～0．94。因子分析产生了2个主成分,分别代表生理健康和心理健康,解释了56．3％的总方差。结论 为SF-36健康调查量表适用于中国提供了证据,已知群效度试验将为量表效度提供更有意义的证据。     

SF-36问卷应用于老年人群生命质量的研究

目的：利用SF－36问卷调查社区老年人群的生命质量，探讨影响生命质量的因素。方法：面对面访谈调查生命质量及影响因素，利用Logistic逐步回归分析调查资料。结果：社区老年人群的生理健康评分较好，但心理健康较差。年龄、性别、婚姻状况和医疗费用是生理健康的主要影响因素；年龄、性别、经济收入和子女状况是心理健康的主要影响因素。结论：一些人口社会学特征影响老年人的生命质量，因此，家庭和社会应该更多地关注老年人的生活环境和医疗条件，提高老年人的健康水平。     

Assessing the validity of the SF-36 General Health Survey

Our objective was to assess the validity of the SF-36 General Health Survey against the Social Maladjustment Schedule (SMS) and two questionnaire measures, the Social Problem Questionnaire and the Nottingham Health Profile (NHP) in a random subsample of 206 men and women from the Whitehall II study, a longitudinal survey of health and disease amongst 10,308 London-based civil servants. We found that social functioning on the SF-36 correlated significantly with social contacts, total satisfaction and total management scores on the SMS, and social isolation and emotional reactions on the NHP. General mental health on the SF-36 was associated with marriage, social contacts, leisure scores, total satisfaction and total management scores on the SMS, and emotional reactions, energy level and social isolation on the NHP. Conversely, physical functioning and physical role limitations were generally not associated with the SMS but were associated with physical abilities and pain on the NHP. In conclusion, this study offers evidence of the discriminant validity of the general mental health and physical functioning scales of the SF-36. We also found moderate construct and criterion validity for the social functioning scale of the SF-36 and considerable overlap between the general mental health and social functioning scales.     

Sample Sizes for the SF-6D Preference Based Measure of Health from the SF-36: A Comparison of Two Methods

Background: Health Related Quality of Life (HRQoL) measures are becoming more frequently used in clinical trials. Investigators are now asking statisticians for advice on how to plan and analyse studies using HRQoL measures, which includes questions on sample size. Sample size requirements are critically dependent on the aims of the study, the outcome measure and its summary measure, the effect size and the method of calculating the test statistic. The SF-6D is a new single summary preference-based measure of health derived from the SF-36 suitable for use in clinical trials. Objectives: To describe and compare two methods of calculating sample sizes when using the SF-6D in comparative clinical trials and to give pragmatic guidance to researchers on what method to use. Methods: We describe two main methods of sample size estimation. The parametric (t-test) method assumes that the SF-6D data is continuous and Normally distributed and that the effect size is the difference between two means. The non-parametric (Mann-Whitney or MW) method makes no distributional assumptions about the data and the effect size is defined in terms of the probability that an observation drawn at random from population Y would exceed an observation drawn at random from population X. We used bootstrap computer simulation to compare the power of the two methods for detecting a shift in location. Results: Computer simulation suggested that if the distribution of the SF-6D is reasonably symmetric then the t-test appears to be more powerful than the MW test at detecting differences in means. If the distribution of the SF-6D is skewed then the MW test appears to be more powerful at detecting a location shift (difference in means) than the t-test. However the differences in power (between the t and MW tests) are small and decrease as the sample size increases. Conclusions: Computer simulation has suggested that parametric methods work reasonably well. Therefore pragmatically we would recommend that parametric methods be used for sample size calculation and analysis when using the SF-6D.     

The use of bootstrap methods for estimating sample size and analysing health-related quality of life outcomes

Health-related quality of life (HRQoL) measures are increasingly used in trials as primary outcome measures. Investigators are now asking statisticians for advice on how to plan and analyse studies using such outcomes. HRQoL outcomes, like the SF-36, are usual measured on an ordinal scale, although most investigators assume that there exists an underlying continuous latent variable and that the actual measured outcomes (the ordered categories) reflect contiguous intervals along this continuum. The ordinal scaling of HRQoL measures means they tend to generate data that have discrete, bounded and skewed distributions. Thus, standard methods of analysis that assume Normality and constant variance may not be appropriate. For this reason, conventional statistical advice would suggest non-parametric methods be used to analyse HRQoL data. The bootstrap is one such computer intensive non-parametric method for estimating sample sizes and analysing data.We describe three methods of estimating sample sizes for two-group cross-sectional comparisons of HRQoL outcomes. We then compared the power of the three methods for a two-group cross-sectional study design using bootstrap simulation. The results showed that under the location shift alternative hypothesis, conventional methods of sample size estimation performed well, particularly Whitehead's method. Whitehead's method is recommended if the HRQoL outcome has a limited number of discrete values (<7) and/or the expected proportion of cases at either of the bounds is high. If a pilot data set is readily available then bootstrap simulation will provide a more accurate and reliable estimate, than conventional methods.Finally, we used the bootstrap for hypothesis testing and the estimation of standard errors and confidence intervals for parameters, in an example data set. We then compared and contrasted the bootstrap with standard methods of analysing HRQoL outcomes. In the data set studied, with the SF-36 outcome, the use of the bootstrap for estimating sample sizes and analysing HRQoL data produces results similar to conventional statistical methods. These results suggest that bootstrap methods are not more appropriate for analysing HRQoL outcome data than standard methods.     

Psychometric evaluation of a Chinese (Taiwanese) version of the SF-36 health survey amongst middle-aged women from a rural community

To test the psychometric properties of the Chinese (Taiwanese) version of the short form 36 health survey (SF-36), 1439 women, aged 40–54 years and living in Kinmen (a Taiwanese island reflecting a predominantly rural community) were recruited to participate in this survey. The rate of unavailable data points for the 36 tested items remained consistently low, and item-discriminate validity was high (95%) for all subscales. Cronbach's α coefficient remained above the 0.70 threshold criterion for all scales except for social functioning and bodily pain. Principal components analysis supported the two major dimensions of health, physical and mental, in the internal structure of the SF-36 scales, although the dimensions did not match the hypothesized association very well. Poorer health profiles were associated with physical and mental conditions. The mental health subscores in the SF-36 test correlated highly with the associated hospital anxiety and depression score (Spearman rank correlation coefficient = −0.62). In conclusion, the reliability and validity tests performed on the data collected support the cross-cultural application of the Chinese (Taiwanese) version of the SF-36 test. This revised version was published online in June 2006 with corrections to the Cover Date.     

The SF-36 Profile and Health-Related Quality of Life: An Interpretative Analysis

The purpose of this investigation was to explore the relationship between the SF-36 scales and a direct, category-scaling, self-evaluation of health-related quality of life in a sample of healthy persons. The study of the relationship between the two provides a deeper insight into the structure and meaning of the SF-36 profile and explores its interpretability in terms of a comprehensive, subjective evaluation of health. Furthermore, this study leads to a preliminary interpretation of the profile in terms of a utility scale. This revised version was published online in June 2006 with corrections to the Cover Date.     

SF-36量表在浙江省麻风治愈者中的应用

目的探讨健康调查量表SF-36在浙江省麻风治愈者中应用的信度、效度和可行性。方法以问卷访谈的形式调查892名麻风治愈者，用分半信度、内部一致性、构想效度和结构效度等指标对SF-36量表进行评判。结果该量表具有良好的内部一致性，8个维度的Chronbach’a系数均≥0．8。除心理健康（MH）外，各维度的分半信度较好，Pearson相关系数均≤0．7。构想效度比较满意。因子分析产生两个因子，能解释总方差的62．05％，且除精神影响（RE）外其他各维度在相应因子有较满意的因子载荷量（≥0．4）。结论SF-36量表基本适用于麻风治愈者生活质量评价，但部分条目需进行调整。     

SF‐6D POPULATION NORMS

The derivation of population norms using simple generic health‐related quality of life measures to inform policy has been recommended in the literature. This letter illustrates the derivation of population norms for the SF‐6D in the United Kingdom. It uses a sample of 22,166 respondents from the 2010 wave of the study Understanding Society. Understanding Society is a national representative sample of British citizens. The survey of this study contains the SF‐12. It is possible to derive health state utilities from the SF‐12 (and from the SF‐36) using a relatively new instrument, the SF‐6D. The SF‐12 and the SF‐36 belong to the most widely used generic health‐related quality of life measures. Mean SF‐6D utility scores for males and females are 0.81 and 0.79, respectively. Especially the older age categories have lower utility scores. The younger age categories have slightly higher utility scores. From a list of 17 conditions, people with congestive heart failure had the lowest (0.60) and people with diabetes the highest (0.76) SF‐6D scores. This letter encourages the health economics research community to derive SF‐6D population norms to inform policy. Copyright © 2012 John Wiley & Sons, Ltd.     

SF-36量表在疾病生命质量谱构建中应用

目的 建立疾病生命质量谱,反映不同疾病人群的生命质量状态.方法 采用健康调查简表(the MOSitem short from health survey,SF-36)研究疾病人群生命质量的文献,对不同疾病人群生命质量的相对位置进行研究,采用系统聚类方法建立疾病生命质量谱.以1～4级生命质量谱段代表严重疾病-较重疾病-中度疾病-轻度疾病的生命质量.结果 生命质量的4级谱段反映了生命质量的不同水平:1级生命质量谱段的疾病主要造成躯体活动受限,总体生命质量很差;2级生命质量谱段为精神系统疾病,以心理健康受到严重影响为特点;3级生命质量谱段以慢性迁延性疾病为主,生命质量一般;4级生命质量谱段以轻度疾病或早期疾病为主,生命质量受疾病影响开始下降,但程度不重.结论 疾病生命质量谱从轻度疾病到严重疾病的生命质量变化,为不同疾病人群的生命质量评价及疗效评价提供了简便直观的参照体系,同时也可以作为非疾病人群生命质量评价的参考体系,为描述不同群体的生命质量水平提供了新的表达方式.     

Use of both short musculoskeletal function assessment questionnaire and short form-36 among tibial-fracture patients was redundant

ObjectiveTo compare the Short Musculoskeletal Function Assessment Dysfunction Index (SMFA DI) and the Short Form-36 Physical Component Summary (SF-36 PCS) scores among patients undergoing operative management of tibial fractures.Study Design and SettingBetween July 2000 and September 2005, we enrolled 1,319 skeletally mature patients with open or closed fractures of the tibial shaft that were managed with intramedullary nailing. Patients were asked to complete the SMFA Questionnaire and SF-36 at discharge and 3, 6, and 12 months post–surgical fixation.ResultsThe SMFA DI and SF-36 PCS scores were highly correlated at 3, 6, and 12 months post–surgical fixation. The difference in the mean standardized change scores for SMFA DI and SF-36 PCS, from 3 to 12 months post–surgical fixation, was not statistically significant. Both the SMFA DI and SF-36 PCS scores were able to discriminate between healed and nonhealed tibial fractures at 3, 6, and 12 months postsurgery.ConclusionIn patients with tibial-shaft fractures, the SMFA DI offered no significant advantages over the SF-36 PCS score. These results, along with the usefulness of SF-36 for comparing populations, recommend the SF-36 for assessing physical function in studies of patients with tibial fractures.     

Measuring change over time: a comparison of results from a global single item of health status and the multi-dimensional SF-36 health status survey questionnaire in patients presenting with menorrhagia

This paper compares the sensitivity to change of a multi-item, multi-dimensional health status measure with a single global health status question, in the assessment of treatment for menorrhagia. A cohort study of patients recruited by general practitioners, was carried out, with a follow up at eighteen months. Questionnaires were administered postally at baseline and follow up. General practices in Berkshire, Buckinghamshire, Northamptonshire and Oxford-shire supplied three hundred and nine women who reported heavy menstrual bleeding, and received either drug treatment alone or both drug and surgical treatment (endometrial resection or hysterectomy) during the eighteen months between the two admini-strations of the questionnaires. A single global question was given to patients asking them to rate their overall health status as excellent, very good, good, fair or poor. The eight dimensions of the SF-36 health survey questionnaire were also given to patients to complete. The dimensions of the SF-36 indicated only small levels of improvement for patients who received drug treatment. However, on many dimensions of the SF-36, a moderate to large improvement was detected for the surgical group. However, small changes were reported in overall health status, as indicated by the single global question, for both groups. Single item measures of health status may not provide a sufficiently accurate indication of health status to be appropriate for use in longitudinal studies.     

Using the SF-36 and Euroqol on an elderly population

An important methodological issue in measuring health-related quality of life is whether instruments such as SF-36 and EQ can be used on an elderly population. This paper aims to test the completion, reliability and validity of the SF-36 and Euroqol on an elderly female population, and to compare them with the OPCS Disability Survey. Three hundred and eighty women aged 75 and over participated in a randomized controlled trial of the use of clodronate provided the sample. As part of the trial, patients were asked to complete the UK SF-36 and Euroqol, and the OPCS disability survey instrument administered by interview in a hospital clinic at baseline. A random subsample of respondents were retested six months later. The SF-36 achieved poorer levels of completion by dimension (68.1%–88.9%) than the OPCS (99.2%) and Euroqol (84%–93.5%) instruments. There were no major floor effects in the distribution of scores, except for the role dimensions of SF-36. Correlation between test-retest were significant for all instruments, but lower for the role dimensions and social functioning of SF-36, and these dimensions also had 95% Cls for the mean differences in excess of 10 points. There was substantial agreement between the three instruments, and evidence for their construct validity against age and recent use of health services. The sensitivities of the instruments were tested through hypothetical changes in health status. There was some evidence of greater sensitivity to lower levels of morbidity in the SF-36. Where brevity is required and the health changes are expected to be substantial, then EQ may be sufficient. For greater sensitivity SF-36 seems to have an advantage, however lower completion rates and problems with consistency suggest it requires adaptation. One solution would be to use interviewer administration. Another would be to change the SF-36 to make it more suitable for use in elderly people, although this may reduce its usefulness as a generic instrument.     

Estimation and comparison of derived preference scores from the SF-36 in lung transplant patients

The purpose of this study was to estimate and compare preference scores derived from MOS Short Form-36 (SF-36) data for a sample of lung transplant patients using three methodologies: Fryback et al. (Med Decis Making 1997; 17: 1-9), Nichol et al. (Med Decis Making 2001; 21: 105-112) and Brazier et al. (J Health Econ 2002: 21: 271-292). Data were gathered from 99 lung transplant recipients using a mail survey, which included the SF-36 and other health-related quality of life (HRQL) measures. The mean preference score for the sample was 0.643 (range 0.43-0.83), 0.765 (range 0.36-1.0), and 0.697 (range 0.33-1.00) for Fryback, Nichol and Brazier methods, respectively. Correlations between the derived scores and visual analogue ratings of health (0.58-0.68) and pulmonary symptoms (-0.59 to -0.62) were moderate to good and in the expected directions. The mean preferences of patients grouped by levels of dyspnea, depression symptoms, illness burden, and self-rated general health differed significantly with all methods and supported the construct validity of the derived scores as measures of preference. The Nichol and Brazier scores, both derived with standard gamble utilities, were generally higher than Fryback scores, which are not utility-based. Given the popularity of the SF-36, these three methods could be useful where direct elicitation of preferences is not feasible. Researchers must be cognizant of the derivation method used, as absolute preference levels, hence quality adjusted life years (QALYs), will differ by method.     

The Short Form Health Survey (SF-36): Translation and validation study of the Iranian version

This was a large population-based study to develop and validate the Iranian version of the Short Form Health Survey (SF-36) for use in health related quality of life assessment in Iran. A culturally comparable questionnaire was developed and pilot tested. Then, the Iranian version of the SF-36 was administered to a random sample of 4163 healthy individuals aged 15 years and over in Tehran. The mean age of the respondents was 35.1 (SD=16.0) years, 52% were female, mostly married (58%) and the mean years of their formal education was 10.0 (SD=4.5). Reliability was estimated using the internal consistency and validity was assessed using known groups comparison and convergent validity. In addition factor analysis was performed. The internal consistency (to test reliability) showed that all eight SF-36 scales met the minimum reliability standard, the Cronbachs coefficients ranging from 0.77 to 0.90 with the exception of the vitality scale (=0.65). Known groups comparison showed that in all scales the SF-36 discriminated between men and women, and old and the young respondents as anticipated (all p values less than 0.05). Convergent validity (to test scaling assumptions) using each item correlation with its hypothesized scale showed satisfactory results (all correlation above 0.40 ranging from 0.58 to 0.95). Factor analysis identified two principal components that jointly accounted for 65.9% of the variance. In general, the Iranian version of the SF-36 performed well and the findings suggest that it is a reliable and valid measure of health related quality of life among the general population.     

Equivalence of Chinese and US–English Versions of the SF-36 Health Survey

This study evaluated the equivalence of Chinese and US–English versions of the SF-36 Health Survey in a convenience sample of 309 Chinese nationals bilingual in Chinese and English living in a US city. Snowball sampling was used to generate sufficient sample size. Internal consistency, test–retest, and equivalent-forms reliability were estimated. Patients were randomized to one of four groups: (1) English version completed first, followed by Chinese version (same occasion); (2) Chinese version completed first, followed by English version (same occasion); (3) English version completed once and then again 1-week later; (4) Chinese version completed once and then again 1-week later. Internal consistency reliability estimates for the Chinese and US–English versions of the SF-36 scales ranged from 0.60 to 0.88; test–retest reliability estimates (1 week time interval) ranged from 0.67 to 0.90. Reliability estimates for corresponding Chinese and US–English SF-36 scales tended to be similar and not significantly different. Equivalent-forms reliability estimates (product–moment correlations) ranged from 0.81 to 0.98. Mean SF-36 scale scores were comparable for both versions of the instrument. This study provides support for the equivalence of the Chinese and US–English versions of the SF-36.     

Responsiveness of the EuroQol (EQ 5-D) and the SF-36 in elderly patients with displaced femoral neck fractures

Objectives: To evaluate the responsiveness of the EuroQol (EQ-5D) and Short-Form 36 (SF-36) instruments, i.e. their ability to capture clinically important changes, in elderly patients with a displaced femoral neck fracture. The study was part of a prospective randomised study comparing two different surgical procedures, internal fixation (IF) and total hip replacement (THP). Setting: University hospital. Patients: A total of 110 patients, mean age 80 years with an acute displaced femoral neck fracture (Garden III and IV). The inclusion criteria were age 70, absence of severe cognitive dysfunction, independent living status and independent walking capability. Intervention: The patients were randomised to IF or to a THR. Main outcome measurements: Health-related quality of life according to EQ-5D and SF-36. Responsiveness measured by the ability of the EQ-5D and the SF-36 to detect clinically relevant differences in the study population according to an external criterion (EC) for outcome (good or less good clinical outcome). Responsiveness was measured in terms of change scores, standardised effect size (SES) and standardised response mean (SRM). Results: The rated prefracture EQ-5D_index scores and SF-36 scores showed good correspondence with the scores of age-matched Swedish reference populations. The relationship between the EC and EQ-5D_index score and the SF-36 global score showed significant differences in both comparisons (p < 0.001). The responsiveness expressed with the SES and SRM were large for both the EQ-5D (1.37 and 0.90, respectively) and for the SF-36 global score (0.89 and 0.82, respectively). The correlation between the change scores for the SF-36 global score and the EQ-5D was 0.39 (p < 0.001). Conclusion: The results showed high responsiveness for both the EQ-5D and the SF-36, indicating that both instruments are suitable for use as outcome measures in clinical trials in elderly hip fracture patients.