Multilevel latent variable models for global health-related quality of life assessment

A multilevel structural equation model for assessing a drug effect on a patient-reported outcome measure in on-demand medication data

We analyze data from a clinical trial investigating the effect of an on-demand drug for women with low sexual desire. These data consist of a varying number of measurements/events across patients of when the drug was taken, including data on a patient-reported outcome consisting of five items measuring an unobserved construct (latent variable). Traditionally, these data are aggregated prior to analysis by composing one sum score per event and averaging this sum score over all observed events. In this paper, we explain the drawbacks of this aggregating approach. One drawback is that these averages have different standard errors because the variance of the underlying events differs between patients and because the number of events per patient differs. Another drawback is the implicit assumption that all items have equal weight in relation to the latent variable being measured. We propose a multilevel structural equation model, treating the events (level 1) as nested observations within patients (level 2), as alternative analysis method to overcome these drawbacks. The model we apply includes a factor model measuring a latent variable at the level of the event and at the level of the patient. Then, in the same model, the latent variables are regressed on covariates to assess the drug effect. We discuss the inferences obtained about the efficacy of the on-demand drug using our proposed model. We further illustrate how to test for measurement invariance across grouping covariates and levels using the same model.

Predicting functional disability in patients with spondyloarthritis using a CRP-based algorithm: A 3-year prospective study

This prospective study explored the link between values of C-reactive protein (CRP) in patients with SpA (ankylosing spondylitis, psoriatic arthritis, reactive arthritis, or arthritis-related inflammatory bowel disease) and functional disability in order to derive an algorithm that may predict functional disability based on disease activity. Patients diagnosed with Spa were classified into five groups based on the type of therapy and they were followed up for 3 years. Group 1: Symptomatic medication alone; Group 2: Disease-modifying antirheumatic drugs (DMARDs); Group 3: DMARDs and 30 rehabilitation sessions twice a year; Group 4: Group 3 therapy and biologic anti-tumor necrosis factor-alpha (anti-TNF-α) drugs; and Group 5: Group 4 therapy and, in addition, a daily home-adapted kinesiotherapy program. CRP, modified Health Assessment Questionnaire (mHAQ-S), Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), and T-score of the patients were recorded. Correlation and multivariate regression analyses were conducted using demographic data, CRP, and mHAQ-S scores to derive the CRP-mHAQ-S correlation algorithm. Statistical analysis included the chi-square, Mann-Whitney, and multiple regression tests and repeated measures analysis of variance. A total of 144 patients were enrolled, all of whom completed the study. The best predictive model (P<0.001) provided the algorithm mHAQ-S₃₆=17.14+0.12xCRP₀-0.24xCRP₁₂-0.15xCRP₃₆ (CRP₀, CRP₁₂, and CRP₃₆ correspond to CRP levels at baseline, 12, and 36 months, respectively, and mHAQ-S₃₆ to mHAQ-S score at 36 months). This derived algorithm based on objective CRP assessment may have implications in the prediction of functional disability evolution in patients with SpA.

Health-related quality-of-life assessment in dementia: Evidence of cross-cultural validity in Latin America

All health-related quality-of-life (HRQL) measures for dementia have been developed in high-income countries and none were validated for cross-cultural use. Yet, the global majority of people living with dementia reside in low- and middle-income countries. We therefore investigated the measurement invariance of a set of self- and informant-report HRQL measures developed in the United Kingdom when used in Latin America. Self-reported HRQL was obtained using (DEMQOL) at a memory assessment service in the United Kingdom (n = 868) and a population cohort study in Latin America (n = 417). Informant reports were collected using DEMQOL-Proxy at both sites (n = 909 and n = 495). Multiple-group confirmatory bifactor models for ordered categorical item responses were estimated to evaluate measurement invariance. Results support configural, metric, and scalar invariance for the concept of general HRQL in DEMQOL and DEMQOL-Proxy. The dominant impact of general HRQL on item responses was evident across U.K. English and Ibero American Spanish versions of DEMQOL (ωh = 0.87-0.90) and DEMQOL-Proxy (ωh = 0.88-0.89). Ratings of "positive emotion" did not show a major impact on general HRQL appraisal, particularly for Latin American respondents. Item information curves show that self- and informant-reports were highly informative about the presence rather than the absence of HRQL impairment. We found no major difference in conceptual meaning, sensitivity, and relevance of DEMQOL and DEMQOL-Proxy for older adults in the United Kingdom and Latin America. Further replication is needed for consensus over which HRQL measures are appropriate for making cross-national comparisons in global dementia research. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

EM algorithm estimation of a structural equation model for the longitudinal study of the quality of life

Health-related quality of life (HRQoL) data are measured via patient questionnaires, completed by the patients themselves at different time points. We focused on oncology data gathered through the use of European Organization for Research and Treatment of Cancer questionnaires, which decompose HRQoL into several functional dimensions, several symptomatic dimensions, and the global health status (GHS). We aimed to perform a global analysis of HRQoL and reduce the number of analyses required by using a two-step approach. First, a structural equation model (SEM) was used for each time point; in these models, the GHS is explained by two latent variables. Each latent variable is a factor that summarizes, respectively, the functional dimensions and the symptomatic dimensions to the global measurement. This is achieved through the maximization of the likelihood of each SEM using the EM algorithm, which has the advantage of giving an estimation of the subject-specific factors and the influence of additional explanatory variables. Then, to consider the longitudinal aspect, the GHS variable and the two factors were concatenated for each patient visit at which the questionnaire was completed. The GHS and the two factors estimated in the first step can then be explained by additional explanatory variables using a linear mixed model.

Quality-of-life assessment in dementia: the use of DEMQOL and DEMQOL-Proxy total scores

PURPOSE

There is a need to determine whether health-related quality-of-life (HRQL) assessments in dementia capture what is important, to form a coherent basis for guiding research and clinical and policy decisions. This study investigated structural validity of HRQL assessments made using the DEMQOL system, with particular interest in studying domains that might be central to HRQL, and the external validity of these HRQL measurements.

METHODS

HRQL of people with dementia was evaluated by 868 self-reports (DEMQOL) and 909 proxy reports (DEMQOL-Proxy) at a community memory service. Exploratory and confirmatory factor analyses (EFA and CFA) were conducted using bifactor models to investigate domains that might be central to general HRQL. Reliability of the general and specific factors measured by the bifactor models was examined using omega (ω) and omega hierarchical (ω h) coefficients. Multiple-indicators multiple-causes models were used to explore the external validity of these HRQL measurements in terms of their associations with other clinical assessments.

RESULTS

Bifactor models showed adequate goodness of fit, supporting HRQL in dementia as a general construct that underlies a diverse range of health indicators. At the same time, additional factors were necessary to explain residual covariation of items within specific health domains identified from the literature. Based on these models, DEMQOL and DEMQOL-Proxy overall total scores showed excellent reliability (ω h > 0.8). After accounting for common variance due to a general factor, subscale scores were less reliable (ω h < 0.7) for informing on individual differences in specific HRQL domains. Depression was more strongly associated with general HRQL based on DEMQOL than on DEMQOL-Proxy (-0.55 vs -0.22). Cognitive impairment had no reliable association with general HRQL based on DEMQOL or DEMQOL-Proxy.

CONCLUSIONS

The tenability of a bifactor model of HRQL in dementia suggests that it is possible to retain theoretical focus on the assessment of a general phenomenon, while exploring variation in specific HRQL domains for insights on what may lie at the 'heart' of HRQL for people with dementia. These data suggest that DEMQOL and DEMQOL-Proxy total scores are likely to be accurate measures of individual differences in HRQL, but that subscale scores should not be used. No specific domain was solely responsible for general HRQL at dementia diagnosis. Better HRQL was moderately associated with less depressive symptoms, but this was less apparent based on informant reports. HRQL was not associated with severity of cognitive impairment.