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.     

Sample size and power estimation for studies with health related quality of life outcomes: a comparison of four methods using the SF-36

We describe and compare four different methods for estimating sample size and power, when the primary outcome of the study is a Health Related Quality of Life (HRQoL) measure. These methods are: 1. assuming a Normal distribution and comparing two means; 2. using a non-parametric method; 3. Whitehead's method based on the proportional odds model; 4. the bootstrap. We illustrate the various methods, using data from the SF-36. For simplicity this paper deals with studies designed to compare the effectiveness (or superiority) of a new treatment compared to a standard treatment at a single point in time. The results show that if the HRQoL outcome has a limited number of discrete values (< 7) and/or the expected proportion of cases at the boundaries is high (scoring 0 or 100), then we would recommend using Whitehead's method (Method 3). Alternatively, if the HRQoL outcome has a large number of distinct values and the proportion at the boundaries is low, then we would recommend using Method 1. If a pilot or historical dataset is readily available (to estimate the shape of the distribution) then bootstrap simulation (Method 4) based on this data will provide a more accurate and reliable sample size estimate than conventional methods (Methods 1, 2, or 3). In the absence of a reliable pilot set, bootstrapping is not appropriate and conventional methods of sample size estimation or simulation will need to be used. Fortunately, with the increasing use of HRQoL outcomes in research, historical datasets are becoming more readily available. Strictly speaking, our results and conclusions only apply to the SF-36 outcome measure. Further empirical work is required to see whether these results hold true for other HRQoL outcomes. However, the SF-36 has many features in common with other HRQoL outcomes: multi-dimensional, ordinal or discrete response categories with upper and lower bounds, and skewed distributions, so therefore, we believe these results and conclusions using the SF-36 will be appropriate for other HRQoL measures.     

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.     

Methods for Determining Sample Sizes for Studies Involving Health-Related Quality of Life Measures: A Tutorial

Health Related Quality of Life (HRQoL) measures are becoming more frequently used in clinical trials and health services research, both as primary and secondary endpoints. 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 and objectives of the study, the proposed summary measure and effect size and the method of calculating the test statistic.We present a tutorial on methods of sample size calculation for HRQoL outcomes. We also briefly review the HRQoL literature to see what has been done in practice. The aim of this tutorial is provide pragmatic guidance to researchers on the methods of calculating sample sizes when using HRQoL measures as outcomes.HRQoL measures such as the SF-36, NHP and QLQ-C30 are usually measured on an ordered categorical (ordinal) scale. We argue that it is often incorrect to treat the scales as if they were continuous and normally distributed and that the mean score may not be a good summary measure of HRQoL data. However the ordinal scaling of HRQoL measures leads to problems in determining sample size, and we suggest that the odds ratio (OR) may be a more suitable summary measure for comparing groups (rather than the mean difference) and therefore methods suitable for ordinal data be used for analysis.The frequency distribution of HRQoL scores should be assessed to see if parametric assumptions are satisfied and whether or not the sample mean is a good summary measure of the data. Given the non-normal distribution of the majority HRQoL outcome measures, summary measures such as means and standard deviations are difficult to interpret. Thus standardised differences (effect sizes) and parametric methods may not be a suitable basis for calculation of sample size. Finally we argue, that any sample size calculation (with all its attendant assumptions) leads to better research than no sample size calculation at all.     

Analysis of the short form-36 (SF-36): the beta-binomial distribution approach

Health-related quality of life (HRQoL) is an important indicator of health status and the Short Form-36 (SF-36) is a generic instrument to measure it. Multiple linear regression (MLR) is often used to study the relationship of HRQoL with patients' characteristics, though HRQoL outcomes tend to be not normally distributed, skewed and bounded (e.g. between 0 and 100). A sample of 193 patients with eating disorders has been analysed to assess the performance of the MLR under non-normality conditions. Normal distribution was rejected for seven out of the eight domains. A beta-binomial distribution is suggested to fit the SF-36 scores. The beta-binomial distribution is not rejected for five out of the eight domains. Thus, a beta-binomial regression (BBR) is suggested to analyse the SF-36 scores. Results using MLR and BBR have been compared for real and simulated data. Performance of the BBR is shown to be better than MLR in the HRQoL domains with few ordered categories and very similar to MLR in the more continuous domains. Moreover, the interpretation of the estimates obtained with BBR is clinically more meaningful. A common technique of statistical analysis is preferable for all the HRQoL dimensions. Therefore, the BBR approach is recommended not only to detect significant predictors of HRQoL when SF-36 is used, but also to analyse and interpret the effect of several explanatory variables on HRQoL. Further work is required to test the better performance of BBR against standard methods for other HRQoL outcomes, populations or interventions.     

A cross-sectional study on self-reported physical and mental health-related quality of life in rheumatoid arthritis and the role of illness perception

Background

Psychosocial models including illness perception might explain individual differences in health-related quality of life (HRQoL) and daily functioning in chronically ill patients. The aim of this study was to assess the association of illness perception among rheumatoid arthritis (RA) patients with physical and mental HRQoL, adjusted for demographic variables, clinical variables and social support.

Methods

A cross-sectional study conducted at a Viennese rheumatology outpatient clinic on 120 RA patients. Participants completed questionnaires on demographic and clinical characteristics, HRQoL (SF-36 Questionnaire), illness beliefs (Brief Illness Perception Questionnaire) and social support (Social Support Scale-8). Analyses were performed with multivariate linear regression.

Results

The mean physical was lower (38.38) than the mean mental SF-36 summary score (46.94). In univariate analysis, all domains of illness perception except belief in a chronic disease course were associated with physical and mental HRQoL. In multivariate analyses, illness perception accounted for 51% of variance in physical HRQoL. A stronger belief in the consequences of RA (consequences, β?=???0.33) and a stronger belief in repeated disease recurrence (timeline cyclical, β =???0?.31) were significantly associated with physical HRQoL in the fully adjusted model. Illness perception accounted for 45% of variance in mental HRQoL. Emotional representation (β?=???0?.27) and fatigue (β?=???0?.36) were significantly associated with mental HRQoL in the fully adjusted model.

Conclusion

This study highlights the importance of RA patients’ beliefs about their illness and symptoms in relation to HRQoL. Identification of patients’ perception of RA may be a way to positively influence disease outcomes such as quality of life as illness perception is amenable to intervention.

     

Analysis of cost data in randomized trials: an application of the non-parametric bootstrap

Health economic evaluations are now more commonly being included in pragmatic randomized trials. However a variety of methods are being used for the presentation and analysis of the resulting cost data, and in many cases the approaches taken are inappropriate. In order to inform health care policy decisions, analysis needs to focus on arithmetic mean costs, since these will reflect the total cost of treating all patients with the disease. Thus, despite the often highly skewed distribution of cost data, standard non-parametric methods or use of normalizing transformations are not appropriate. Although standard parametric methods of comparing arithmetic means may be robust to non-normality for some data sets, this is not guaranteed. While the randomization test can be used to overcome assumptions of normality, its use for comparing means is still restricted by the need for similarly shaped distributions in the two groups. In this paper we show how the non-parametric bootstrap provides a more flexible alternative for comparing arithmetic mean costs between randomized groups, avoiding the assumptions which limit other methods. Details of several bootstrap methods for hypothesis tests and confidence intervals are described and applied to cost data from two randomized trials. The preferred bootstrap approaches are the bootstrap-t or variance stabilized bootstrap-t and the bias corrected and accelerated percentile methods. We conclude that such bootstrap techniques can be recommended either as a check on the robustness of standard parametric methods, or to provide the primary statistical analysis when making inferences about arithmetic means for moderately sized samples of highly skewed data such as costs.     

Neonatal treatment philosophy in Dutch and German NICUs: health-related quality of life in adulthood of VP/VLBW infants

Purpose

Although survival after very preterm birth (VP)/very low birth weight (VLBW) has improved, a significant number of VP/VLBW individuals develop physical and cognitive problems during their life course that may affect their health-related quality of life (HRQoL). We compared HRQoL in VP/VLBW cohorts from two countries: The Netherlands (n = 314) versus Germany (n = 260) and examined whether different neonatal treatment and rates of disability affect HRQoL in adulthood.

Method

To analyse whether cohorts differed in adult HRQoL, linear regression analyses were performed for three HRQoL outcomes assessed with the Health Utilities Index 3 (HUI3), the London Handicap Scale (LHS), and the WHO Quality of Life instrument (WHOQOL-BREF). Stepwise hierarchical linear regression was used to test whether neonatal physical health and treatment, social environment, and intelligence (IQ) were related to VP/VLBW adults’ HRQoL and cohort differences.

Results

Dutch VP/VLBW adults reported a significantly higher HRQoL on all three general HRQoL measures than German VP/VLBW adults (HUI3: .86 vs .83, p = .036; LHS: .93 vs. .90, p = .018; WHOQOL-BREF: 82.8 vs. 78.3, p < .001). Main predictor of cohort differences in all three HRQoL measures was adult IQ (p < .001).

Conclusions

Lower HRQoL in German versus Dutch adults was related to more cognitive impairment in German adults. Due to different policies, German VP/VLBW infants received more intensive treatment that may have affected their cognitive development. Our findings stress the importance of examining effects of different neonatal treatment policies for VP/VLBW adults’ life.

     

Intraclass correlation coefficients and bootstrap methods of hierarchical binary outcomes

Intraclass correlation coefficients are designed to assess consistency or conformity between two or more quantitative measurements. When multistage cluster sampling is implemented, no methods are readily available to estimate intraclass correlations of binomial-distributed outcomes within a cluster. Because statistical distribution of the intraclass correlation coefficients could be complicated or unspecified, we propose using a bootstrap method to estimate the standard error and confidence interval within the framework of a multilevel generalized linear model. We compared the results derived from a parametric bootstrap method with those from a non-parametric bootstrap method and found that the non-parametric method is more robust. For non-parametric bootstrap sampling, we showed that the effectiveness of sampling on the highest level is greater than that on lower levels; to illustrate the effectiveness, we analyse survey data in China and do simulation studies.     

Health-related quality of life in coronary heart disease: a systematic review and meta-analysis mapped against the International Classification of Functioning,Disability and Health

Background

Health-related quality of life (HRQoL) is an important outcome in coronary heart disease (CHD). However, variability in HRQoL indicators suggests a need to consider domain coverage. This review applies a globally accepted framework, the International Classification of Functioning, Disability and Health (ICF), to map HRQoL measures that are reliable and valid among people with CHD.

Methods

The Embase, Pubmed and PsycInfo databases were searched, with 10 observational studies comparing HRQOL among 4786 adults with CHD to 50949 controls identified. Study reporting quality was examined (QualSyst). Hedges’ g statistic (with 95% CIs and p values) was used to measure the effect size for the difference between group means (≤?0.2 small, ≤?0.5 medium, ≤?0.80 large difference), and between-study heterogeneity (tau, I² test) examined using a random effects model.

Results

Adults with CHD reported lowered HRQoL (g_w?=???0.418, p?<?0.001). Adjusted mean differences in HRQoL ratings, controlling for socio-demographics, were smaller but remained significant. Large group differences were associated with individual measures of activity and participation (WHOQOL g?=???1.199, p?<?0.001) and self-perceived health (SF 36 g?=???0.616, p?<?0.001).

Conclusions

The ICF provides a framework for evaluating and understanding the impact of CHD on HRQoL. The results demonstrate that HRQoL goes beyond physical symptoms, with activity limitations, social support and participation, and personal perceptions identified as key ICF domains in CHD assessment. Further investigations are needed to unravel the dynamic and inter-relationships between these domains, including longitudinal trends in HRQoL indicators.

     

Correlates of health-related quality of life in young-old and old–old community-dwelling older adults

Purpose

This study aimed (1) to examine the role of potential correlates of HRQoL in a large representative sample of older adults, and (2) to investigate whether the relationships between HRQoL and potential factors differ as a function of HRQoL component (physical vs. mental) and/or age cohort (young-old vs. old–old).

Methods

This cross-sectional study included 802 older adults aged 60–79 years old. HRQoL was assessed using the SF-36 questionnaire. Functional fitness was assessed using the Senior Fitness Test. Physical activity was measured via the Baecke questionnaire. Demographic information, mental and health features were obtained through questionnaires.

Results

A multiple regression analysis showed that BMI (β = ?0.15, p?=?0.001), body strength (β =?0.21, p?<?0.001), aerobic endurance (β =?0.29, p?<?0.001), physical activity (β =?0.11, p?=?0.007), depressive symptoms (β = ?0.19, p?<?0.001), falls (β = ?0.19, p?<?0.001), and living alone (β = ?0.16, p?<?0.001) were all significantly related to HRQoL-SF-36 total score. The positive relation with aerobic endurance was significantly higher for the physical component of HRQoL, while the negative relation with living alone was significantly higher for the mental component. The positive relation of HRQoL with physical activity was significantly higher in old–old compared to young-old adults.

Conclusion

This data suggest that body composition, functional fitness, psycho-social factors, and falls are important correlates of HRQoL in old age. There are HRQoL-component and age-cohort differences regarding these correlates, underlying the need for specific strategies at the community level to promote HRQoL in older adults.

     

Patient-reported measures of health: The Veterans Health Study

The goal of the Veterans Health Study (VHS) was to extend the work of the Medical Outcomes Study (MOS) into the VA, by developing methodology for monitoring patient-based outcomes of care for use in ambulatory outpatient care. The principal objective of the VHS was developing valid and reliable measures to assess general health-related quality of life (HRQoL) and identifying the presence of selected health conditions, their severity, and their impact on HRQoL. In this article, we provide an overview of the historical context, framework, objectives, and applications of the VHS for the purpose of assessing the health outcomes of veteran patients. The VHS is a prospective observational study that has followed 2425 VA patients for up to 2 years. The patients were sampled from users of the Veterans Affairs (VA) ambulatory care system in the Boston area. The health conditions selected were hypertension, diabetes, chronic lung disease, osteoarthritis of the knee, chronic low-back pain, and alcohol-related problems. These conditions were chosen because they are both prevalent in the VA and have measurable impacts on HRQoL. One of the cornerstones of the VHS was the development of the Veterans SF-36, modified from the MOS SF-36 for use in veteran ambulatory populations. Other key accomplishments included the development of patient-based disease-specific measures of health and the establishment of methods and logistics for comprehensive health outcomes research in large health care systems such as the VA, using these patient-based measures. Selected measures developed in the VHS, eg, the Veterans SF-36, have been integrated into the VA outcomes measurement system. The scope of the VHS is unique; it resulted in the development of a broad range of patient-focused process and outcome measures, as well as methodologies for assessing large numbers of patients, that have been widely used in the VA outpatient health care system for monitoring health outcomes across the nation.     

Perceived Financial Satisfaction,Health Related Quality of Life and depressive Symptoms in Early Pregnancy

Objectives To assess the associations of perceived financial satisfaction and health-related quality of life (HRQoL) and depressive symptoms in an unselected pregnant population in early pregnancy. Methods 750 consecutive pregnant women attending the first communal ultrasound examination before gestational week 14 were invited to participate. Questionnaires assessing HRQoL (15D), depressive symptoms (Edinburgh Depression Scale, EPDS), medical, obstetric, and socioeconomic status were handed out. The participants were divided into three groups according to their satisfaction with their financial status, (unsatisfied, somewhat satisfied, and satisfied). Main outcome measures were 15D and EPDS-scores and dimensions of HRQoL. Results 325 (43,3%) questionnaires were returned. The mean 15D-score for HRQoL was 0,926 (SD 0,056). The financially unsatisfied women had lower HRQoL than women in more satisfied groups (0.906, 0.923 and 0.931, p?=?0.012). The result remained significant, even after adjusting for age and education(p?=?0.032). The unsatisfied women had a higher mean body mass index (BMI) (25.4, 24.4 and 23.2 kg/m², p for linearity?=?0.002), were more often smokers, (13 vs. 4 and 3%, p?=?0.029), and had experienced at least one abortion (18, 14 and 7%, p?=?0.017). Dimensions of depression, distress and sleep explained the differences between the groups. 27% of unsatisfied women scored EPDS ≥10 points suggesting increased risk of depression. Conclusions Financial satisfaction in early pregnancy associates with HRQoL and risk of perinatal depressive symptoms. Unsatisfied women more often have risk factors for unfavourable pregnancy outcomes which may influence the later health and wellbeing of the mother and child.     

Depression treatment and health-related quality of life among adults with diabetes and depression

Background

Previous findings regarding depression treatment and its consequences on health-related quality of life (HRQoL) of adults with diabetes were inconsistent and targeted certain groups of population. Therefore, there is a critical need to conduct a population-based study that focuses on a general population with diabetes and depression.

Objective

The primary aim of this study was to examine the physical and mental HRQoL associated with depression treatment during the follow-up year.

Methods

We adopted a longitudinal design using multiple panels (2005–2011) of the Medical Expenditure Panel Survey to create a baseline year and follow-up year. We included adults with diabetes and depression. We categorized the baseline depression treatment into: (1) antidepressant use only; (2) psychotherapy with or without antidepressants; and (3) no treatment. HRQOL was measured using SF-12 version 2 physical component summary (PCS) and SF-12 mental component summary (MCS) scores during both baseline year and follow-up year. Ordinary least squares (OLS) were used to estimate the association between depression treatment and the HRQoL measures. The OLS regression controlled for predisposing, enabling, need, external environment factors, personal health practices, and baseline HRQoL measures.

Results

After controlling for all the independent variables and the baseline PCS, individuals who received psychotherapy with or without antidepressants had higher PCS scores as compared to those without any treatment for depression (beta = 1.28, p < 0.001). Individuals who reported using only antidepressants had lower PCS scores (beta = ?0.54, p < 0.001) as compared to those without depression treatment. On the contrary, individuals who reported receiving psychotherapy with or without antidepressants had lower MCS scores as compared to those without depression treatment (beta = ?1.43, p < 0.001). Those using only antidepressants had higher MCS scores as compared to those without depression treatment (beta = 0.56, p < 0.001).

Conclusion

The associations between depression treatment and the HRQoL varied by the type of depression treatment and the component of the HRQoL measures.

     

Health-related quality of life in multiple sclerosis: role of cognitive appraisals of self,illness and treatment

Purpose

Health-related quality of life (HRQoL) is considered an important measure of treatment and rehabilitation outcomes in multiple sclerosis (MS) patients. In this study, we used multivariate regression analysis to examine the role of cognitive appraisals, adjusted for clinical, socioeconomic and demographic variables, as correlates of HRQoL in MS.

Methods

The cross-sectional study included 257 MS patients, who completed Multiple Sclerosis Impact Scale, Generalized Self-Efficacy Scale, Rosenberg Self-Esteem Scale, Brief Illness Perception Questionnaire, Treatment Beliefs Scale, Actually Received Support Scale (a part of Berlin Social Support Scale) and Socioeconomic Resources Scale. Demographic and clinical characteristics of the participants were collected with a self-report survey. Correlation and regression analyses were conducted to determine associations between the variables.

Results

Five variables, illness identity (β = 0.29, p ≤ 0.001), self-esteem (β = ?0.22, p ≤ 0.001), general self-efficacy (β = ?0.21, p ≤ 0.001), disability subgroup “EDSS” (β = 0.14, p = 0.006) and age (β = 0.12, p = 0.012), were significant correlates of HRQoL in MS. These variables explained 46 % of variance in the dependent variable. Moreover, we identified correlates of physical and psychological dimensions of HRQoL.

Conclusions

Cognitive appraisals, such as general self-efficacy, self-esteem and illness perception, are more salient correlates of HRQoL than social support, socioeconomic resources and clinical characteristics, such as type and duration of MS. Therefore, interventions aimed at cognitive appraisals may also improve HRQoL of MS patients.

     

Functional Measures Developed for Clinical Populations Identified Impairment Among Active Workers with Upper Extremity Disorders

Purpose Few studies have explored measures of function across a range of health outcomes in a general working population. Using four upper extremity (UE) case definitions from the scientific literature, we described the performance of functional measures of work, activities of daily living, and overall health. Methods A sample of 573 workers completed several functional measures: modified recall versions of the QuickDASH, Levine Functional Status Scale (FSS), DASH Work module (DASH-W), and standard SF-8 physical component score. We determined case status based on four UE case definitions: (1) UE symptoms, (2) UE musculoskeletal disorders (MSD), (3) carpal tunnel syndrome (CTS), and (4) work limitations due to UE symptoms. We calculated effect sizes for each case definition to show the magnitude of the differences that were detected between cases and non-cases for each case definition on each functional measure. Sensitivity and specificity analyses showed how well each measure identified functional impairments across the UE case definitions. Results All measures discriminated between cases and non-cases for each case definition with the largest effect sizes for CTS and work limitations, particularly for the modified FSS and DASH-W measures. Specificity was high and sensitivity was low for outcomes of UE symptoms and UE MSD in all measures. Sensitivity was high for CTS and work limitations. Conclusions Functional measures developed specifically for use in clinical, treatment-seeking populations may identify mild levels of impairment in relatively healthy, active working populations, but measures performed better among workers with CTS or those reporting limitations at work.     

Proxy reliability: Health-related quality of life (HRQoL) measures for people with disability

Objectives: Research and surveillance activities sometimes require that proxy respondents provide key exposure or outcome information, especially for studies of people with disability (PWD). In this study, we compared the health-related quality of life (HRQoL) responses of index PWD to proxies. Methods: Subjects were selected from nursing home, other assisted living residences, and from several clinic samples of PWD. Each index identified one or more proxy respondents. Computer-assisted interviews used a random order of measures. Proxy reliability was measured by intraclass correlation (ICC) and κ statistics. HRQoL measures tested included the surveillance questions of the Behavioral Risk Factor Surveillance System (BRFSS), basic and instrumental activities of daily living (ADLs and IADLs), medical outcomes study short-form 36 and 12 (SF-36 and SF-12). Results: A total of 131 index-proxy sets were completed. In general, agreement and reliability of proxy responses to the PWD tended to be best for relatives, with friends lower, and health care proxies lowest. For example, the ICC for the physical functioning scale of the SF-36 was 0.68 for relatives, 0.51 for friends, and 0.40 for healthcare proxies. There was a tendency for proxies to overestimate impairment and underestimate HRQoL. This pattern was reversed for measures of pain, which proxies consistently underestimated. The pattern among instruments, proxy types, and HRQoL domains was complex, and individual measures vary from these general results. Conclusions: We suggest caution when using proxy respondents for HRQoL, especially those measuring more subjective domains. This revised version was published online in June 2006 with corrections to the Cover Date.     

The application of a network approach to Health-Related Quality of Life (HRQoL): introducing a new method for assessing HRQoL in healthy adults and cancer patients

Purpose

Health-Related Quality of Life (HRQoL) research has typically adopted either a formative approach, in which HRQoL is the common effect of its observables, or a reflective approach—defining HRQoL as a latent variable that determines observable characteristics of HRQoL. Both approaches, however, do not take into account the complex organization of these characteristics. The objective of this study was to introduce a new approach for analyzing HRQoL data, namely a network model (NM). An NM, as opposed to traditional research strategies, accounts for interactions among observables and offers a complementary analytic approach.

Methods

We applied the NM to samples of Dutch cancer patients (N = 485) and Dutch healthy adults (N = 1742) who completed the 36-item Short Form Health Survey (SF-36). Networks were constructed for both samples separately and for a combined sample with diagnostic status added as an extra variable. We assessed the network structures and compared the structures of the two separate samples on the item and domain levels. The relative importance of individual items in the network structures was determined using centrality analyses.

Results

We found that the global structure of the SF-36 is dominant in all networks, supporting the validity of questionnaire’s subscales. Furthermore, results suggest that the network structure of both samples was highly similar. Centrality analyses revealed that maintaining a daily routine despite one’s physical health predicts HRQoL levels best.

Conclusions

We concluded that the NM provides a fruitful alternative to classical approaches used in the psychometric analysis of HRQoL data.

     

Norwegian reference values for the Short-Form Health Survey 36: development over time

Purpose

Reference values for patient-reported outcome measures are useful for interpretation of results from clinical trials. The study aims were to collect Norwegian SF-36 reference values and compare with data from 1996 to 2002.

Methods

In 2015, SF-36 was sent by mail to a representative sample of the population (N = 6165). Time trends and associations between background variables and SF-36 scale scores were compared by linear regression models.

Results

The 2015 response rate was 36% (N = 2118) versus 67% (N = 2323) in 1996 and 56% (N = 5241) in 2002. Only 5% of the youngest (18–29 years) and 27% of the oldest (>70 years) responded in 2015. Age and educational level were significantly higher in 2015 relative to 1996/2002 (p < .001). The oldest age group in 2015 reported better scores on five of eight scales (p < 0.01), the exceptions being bodily pain, vitality, and mental health compared to 1996/2002 (NS). Overall, the SF-36 scores were relatively stable across surveys, controlled for background variables. In general, the most pronounced changes in 2015 were better scores on the role limitations emotional scale (7.4 points, p < .001) and lower scores on the bodily pain scale (4.6 points, p < .001) than in the 1996/2002 survey.

Conclusions

The low response rate in 2015 suggests that the results, especially among the youngest, should be interpreted with caution. The high response rate among the oldest indicates good representativity for those >70 years. Despite societal changes in Norway the past two decades, HRQoL has remained relatively stable.