Dataset Information

Longitudinal beta regression models for analyzing health-related quality of life scores over time.

ABSTRACT:

Background

Health-related quality of life (HRQL) has become an increasingly important outcome parameter in clinical trials and epidemiological research. HRQL scores are typically bounded at both ends of the scale and often highly skewed. Several regression techniques have been proposed to model such data in cross-sectional studies, however, methods applicable in longitudinal research are less well researched. This study examined the use of beta regression models for analyzing longitudinal HRQL data using two empirical examples with distributional features typically encountered in practice.

Methods

We used SF-6D utility data from a German older age cohort study and stroke-specific HRQL data from a randomized controlled trial. We described the conceptual differences between mixed and marginal beta regression models and compared both models to the commonly used linear mixed model in terms of overall fit and predictive accuracy.

Results

At any measurement time, the beta distribution fitted the SF-6D utility data and stroke-specific HRQL data better than the normal distribution. The mixed beta model showed better likelihood-based fit statistics than the linear mixed model and respected the boundedness of the outcome variable. However, it tended to underestimate the true mean at the upper part of the distribution. Adjusted group means from marginal beta model and linear mixed model were nearly identical but differences could be observed with respect to standard errors.

Conclusions

Understanding the conceptual differences between mixed and marginal beta regression models is important for their proper use in the analysis of longitudinal HRQL data. Beta regression fits the typical distribution of HRQL data better than linear mixed models, however, if focus is on estimating group mean scores rather than making individual predictions, the two methods might not differ substantially.

SUBMITTER: Hunger M

PROVIDER: S-EPMC3528618 | biostudies-literature | 2012 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Longitudinal beta regression models for analyzing health-related quality of life scores over time.

Hunger Matthias M Döring Angela A Holle Rolf R

BMC medical research methodology 20120917

<h4>Background</h4>Health-related quality of life (HRQL) has become an increasingly important outcome parameter in clinical trials and epidemiological research. HRQL scores are typically bounded at both ends of the scale and often highly skewed. Several regression techniques have been proposed to model such data in cross-sectional studies, however, methods applicable in longitudinal research are less well researched. This study examined the use of beta regression models for analyzing longitudina ...[more]

PMID: 22984825

Similar Datasets

Project description:BackgroundEvery two years, long-term care organizations for the elderly are obliged to evaluate and publish the experiences of residents, representatives of psychogeriatric patients, and/or assisted-living clients with regard to quality of care. Our hypotheses are that publication of this quality information leads to improved performance, and that organizations with substandard performance will improve more than those whose performance is relatively good.MethodsThe analyses included organizational units that measured experiences twice between 2007 (t(0)) and 2009 (t(1)). Experiences with quality of care were measured with Consumer Quality Index (CQI) questionnaires. Besides descriptive analyses (i.e. mean, 5(th) and 95(th) percentile, and 90% central range) of the 19 CQI indicators and change scores of these indicators were calculated. Differences across five performance groups (ranging from 'worst' to 'best') were tested using an ANOVA test and effect sizes were measured with omega squared (?(2)).ResultsAt t0 experiences of residents, representatives, and assisted-living clients were positive on all indicators. Nevertheless, most CQI indicators had improved scores (up to 0.37 change score) at t(1). Only three indicators showed a minor decline (up to -0.08 change score). Change scores varied between indicators and questionnaires, e.g. they were more profound for the face-to-face interview questionnaire for residents in nursing homes than for the other two mail questionnaires (0.15 vs. 0.05 and 0.04, respectively), possibly due to more variation between nursing homes on the first measurement, perhaps indicating more potential for improvement. A negative relationship was found between prior performance and change, particularly with respect to the experiences of residents (?(2) = 0.16) and assisted-living clients (?(2) = 0.15). However, the relation between prior performance and improvement could also be demonstrated with respect to the experiences reported by representatives of psychogeriatric patients and by assisted-living clients. For representatives of psychogeriatric patients, the performance groups 1 and 2 ([much] below average) improved significantly more than the other three groups (?(2) = 0.05).ConclusionsBoth hypotheses were confirmed: almost all indicator scores improved over time and long-term care organizations for the elderly with substandard performance improved more than those with a performance which was already relatively good.

Project description:BackgroundQuality of life (QoL) is profoundly impaired in patients with idiopathic pulmonary fibrosis (IPF). However, data is limited regarding the course of QoL. We therefore analysed longitudinal data from the German INSIGHTS-IPF registry.MethodsClinical status and QoL were assessed at enrollment and subsequently at 6- to 12-months intervals. A range of different QoL questionnaires including the St. George's Respiratory Questionnaire (SGRQ) were used.ResultsData from 424 patients were included; 76.9% male; mean age 68.7 ± 9.1 years, mean FVC% predicted 75.9 ± 19.4, mean DLCO% predicted 36.1 ± 15.9. QoL worsened significantly during follow-up with higher total SGRQ scores (increased by 1.47 per year; 95% CI: 1.17 to 1.76; p < 0.001) and higher UCSD-SOBQ scores and lower EQ-5D VAS and WHO-5 scores. An absolute decline in FVC% predicted of > 10% was associated with a significant deterioration in SGRQ (increasing by 9.08 units; 95% CI: 2.48 to 15.67; p = 0.007), while patients with stable or improved FVC had no significantly change in SGRQ. Patients with a > 10% decrease of DLCO % predicted also had a significant increase in SGRQ (+ 7.79 units; 95% CI: 0.85 to 14.73; p = 0.028), while SQRQ was almost stable in patients with stable or improved DLCO. Patients who died had a significant greater increase in SGRQ total scores (mean 11.8 ± 18.6) at their last follow-up visit prior to death compared to survivors (mean 4.2 ± 18.9; HR = 1.03; 95% CI: 1.01 to 1.04; p < 0.001). All QoL scores across the follow-up period were significantly worse in hospitalised patients compared to non-hospitalised patients, with the worst scores reported in those hospitalised for acute exacerbations.ConclusionsQoL assessments in the INSIGHTS-IPF registry demonstrate a close relationship between QoL and clinically meaningful changes in lung function, comorbidities, disease duration and clinical course of IPF, including hospitalisation and mortality.

Dataset Information

Longitudinal beta regression models for analyzing health-related quality of life scores over time.

Background

Methods

Results

Conclusions

Publications

Longitudinal beta regression models for analyzing health-related quality of life scores over time.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets