Spatio-temporal Rasch analysis of quality of life outcomes in the French general population: measurement invariance and group comparisons

Testing gender and longitudinal measurement invariance of the SF-36 in American Indian older adults: The strong heart study

Information about the equality of psychometric properties of the medical outcomes study (MOS) Short Form-36 (SF-36), a health status measure, across gender and across the lifespan for American Indian adults is lacking. We tested measurement invariance (configural, metric, scalar invariance) of the physical and mental components between gender and over time in a sample of 2,709 (1,054 men, 1,654 women) American Indian older adults at three time points, and across a 6-year time frame. Measurement invariance of a 2-factor higher-order model was demonstrated between gender at each time point. Tests of longitudinal invariance indicated longitudinal measurement invariance over time. Multiple-group latent means analysis indicated men had significantly higher physical and mental component latent means compared to women at each time point, and longitudinal latent means analysis found physical and mental component latent means decreased over time. The 2-factor higher-order model SF-36 is valid for American Indian older adults over a 6-year time frame. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

How to Investigate the Effects of Groups on Changes in Longitudinal Patient-Reported Outcomes and Response Shift Using Rasch Models

In order to investigate patients’ experience of healthcare, repeated assessments of patient-reported outcomes (PRO) are increasingly performed in observational studies and clinical trials. Changes in PRO can however be difficult to interpret in longitudinal settings as patients’ perception of the concept being measured may change over time, leading to response shift (longitudinal measurement non-invariance) and possibly to erroneous interpretation of the observed changes in PRO. Several statistical methods for response shift analysis have been proposed, but they usually assume that response shift occurs in the same way in all individuals within the sample regardless of their characteristics. Many studies aim at comparing the longitudinal change of PRO into two groups of patients (treatment arm, different pathologies, …). The group variable could have an effect on PRO change but also on response shift effect and the perception of the questionnaire at baseline. In this paper, we propose to enhance the ROSALI algorithm based on Rasch Measurement Theory for the analysis of longitudinal PRO data to simultaneously investigate the effects of group on item functioning at the first measurement occasion, on response shift and on changes in PRO over time. ROSALI is subsequently applied to a longitudinal dataset on change in emotional functioning in patients with breast cancer or melanoma during the year following diagnosis. The use of ROSALI provides new insights in the analysis of longitudinal PRO data.

Contrasted trends in health-related quality of life across gender, age categories and work status in France, 1995–2016: repeated population-based cross-sectional surveys using the SF-36

Background

Interest in monitoring health-related quality of life (HRQoL) in general populations has increased in the past 20 years, reinforced by population ageing and repeated economic crises. This study aims to identify temporal trends in HRQoL in France between 1995 and 2016 and to assess disparities according to demographic and socioeconomic characteristics.

Methods

Data from repeated population-based cross-sectional surveys conducted in 1995, 2003 and 2016 were used. HRQoL was measured using the Medical Outcomes Study 36-item Short Form (SF-36) questionnaire.

Results

A substantial decrease in score was observed between 1995 and 2016 for both genders in almost all subscales of the SF-36, with the largest decrease being in the mental health dimension for men. However, the age group 18–54 years were the most affected with persistent negative or even worsening trends in HRQoL. The largest decreases were among men aged 45–54 years and women aged 35–44 years in most dimensions, and among the age group 18–24 years in vitality. Conversely, an overall improvement was noted among the age group 65–84 years. People in employment were more affected than the unemployed by the decline in several HRQoL dimensions.

Conclusion

A general decline in HRQoL was found between 1995 and 2016 in the French population, but with wide disparities in trends between age groups. Young and especially middle-aged, employed people exhibited persistent negative and worsening trends. Consistent with evidence from traditional mental health morbidity and mortality indicators, our findings raise questions about the potential influence of macro-socioeconomic factors, especially the 2008 crisis; these observations deserve special attention from health policy-makers.

Item response models for the longitudinal analysis of health-related quality of life in cancer clinical trials

Background

The use of health-related quality of life (HRQoL) as an endpoint in cancer clinical trials is growing rapidly. Hence, research into the statistical approaches used to analyze HRQoL data is of major importance, and could lead to a better understanding of the impact of treatments on the everyday life and care of patients. Amongst the models that are used for the longitudinal analysis of HRQoL, we focused on the mixed models from item response theory, to directly analyze raw data from questionnaires.

Methods

We reviewed the different item response models for ordinal responses, using a recent classification of generalized linear models for categorical data. Based on methodological and practical arguments, we then proposed a conceptual selection of these models for the longitudinal analysis of HRQoL in cancer clinical trials.

Results

To complete comparison studies already present in the literature, we performed a simulation study based on random part of the mixed models, so to compare the linear mixed model classically used to the selected item response models. As expected, the sensitivity of the item response models to detect random effects with lower variance is better than that of the linear mixed model. We then used a cumulative item response model to perform a longitudinal analysis of HRQoL data from a cancer clinical trial.

Conclusions

Adjacent and cumulative item response models seem particularly suitable for HRQoL analysis. In the specific context of cancer clinical trials and the comparison between two groups of HRQoL data over time, the cumulative model seems to be the most suitable, given that it is able to generate a more complete set of results and gives an intuitive illustration of the data.

Electronic supplementary material

The online version of this article (doi:10.1186/s12874-017-0410-9) contains supplementary material, which is available to authorized users.

A simple ratio-based approach for power and sample size determination for 2-group comparison using Rasch models

BACKGROUND

Despite the widespread use of patient-reported Outcomes (PRO) in clinical studies, their design remains a challenge. Justification of study size is hardly provided, especially when a Rasch model is planned for analysing the data in a 2-group comparison study. The classical sample size formula (CLASSIC) for comparing normally distributed endpoints between two groups has shown to be inadequate in this setting (underestimated study sizes). A correction factor (RATIO) has been proposed to reach an adequate sample size from the CLASSIC when a Rasch model is intended to be used for analysis. The objective was to explore the impact of the parameters used for study design on the RATIO and to identify the most relevant to provide a simple method for sample size determination for Rasch modelling.

METHODS

A large combination of parameters used for study design was simulated using a Monte Carlo method: variance of the latent trait, group effect, sample size per group, number of items and items difficulty parameters. A linear regression model explaining the RATIO and including all the former parameters as covariates was fitted.

RESULTS

The most relevant parameters explaining the ratio's variations were the number of items and the variance of the latent trait (R2 = 99.4%).

CONCLUSIONS

Using the classical sample size formula adjusted with the proposed RATIO can provide a straightforward and reliable formula for sample size computation for 2-group comparison of PRO data using Rasch models.