Measurement error, time lag, unmeasured confounding: Considerations for longitudinal estimation of the effect of a mediator in randomised clinical trials

Green and blue spaces and lung function in the Generation XXI cohort: a life-course approach

BACKGROUND

Exposure to natural environments may affect respiratory health. This study examined the association of exposure to green and blue spaces with lung function in children, and assessed the mediation effect of air pollution and physical activity.

METHODS

The study used data from the Generation XXI, a population-based birth cohort from the Porto Metropolitan Area (Portugal). Residential Normalised Difference Vegetation Index (NDVI) at different buffers (100, 250 and 500 m), the accessibility to urban green spaces (UGS) within 400 and 800 m and the minimum distance to the nearest UGS and to the nearest blue spaces were assessed at birth, 4, 7 and 10 years of age. Three life-course measures were calculated: averaged exposure, early-life exposure (birth) and exposure trend over time (change in exposure). Forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1) and forced expiratory flow between 25% and 75% of FVC (FEF25-75%) at 10 years were used as outcomes. To assess associations, linear regression models and path analysis were used.

RESULTS

This study included 3278 children. The adjusted models showed that increasing the NDVI exposure over time within 100 m of the child's residence was associated with higher values of FEV1 (L) and FEF25-75% (L·s-1) (β 0.01, 95% CI 0.0002-0.03 and β 0.02, 95% CI 0.001-0.05, respectively). No significant associations were observed for the remaining measures of exposure, and no mediation effect was found for pollution or physical activity.

CONCLUSION

Increasing exposure to greenness at close proximity from residences was associated with improved lung function. While the mechanism remains unknown, this study brings evidence that city greening may improve children's respiratory health.

Estimation and inference for the mediation effect in a time-varying mediation model

Background

Traditional mediation analysis typically examines the relations among an intervention, a time-invariant mediator, and a time-invariant outcome variable. Although there may be a total effect of the intervention on the outcome, there is a need to understand the process by which the intervention affects the outcome (i.e., the indirect effect through the mediator). This indirect effect is frequently assumed to be time-invariant. With improvements in data collection technology, it is possible to obtain repeated assessments over time resulting in intensive longitudinal data. This calls for an extension of traditional mediation analysis to incorporate time-varying variables as well as time-varying effects.

Methods

We focus on estimation and inference for the time-varying mediation model, which allows mediation effects to vary as a function of time. We propose a two-step approach to estimate the time-varying mediation effect. Moreover, we use a simulation-based approach to derive the corresponding point-wise confidence band for the time-varying mediation effect.

Results

Simulation studies show that the proposed procedures perform well when comparing the confidence band and the true underlying model. We further apply the proposed model and the statistical inference procedure to data collected from a smoking cessation study.

Conclusions

We present a model for estimating time-varying mediation effects that allows both time-varying outcomes and mediators. Simulation-based inference is also proposed and implemented in a user-friendly R package.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-022-01585-x).

Dynamic change meets mechanisms of change: Examining mediators in the latent change score framework

Researchers in behavioral sciences are often interested in longitudinal behavior change outcomes and the mechanisms that influence changes in these outcomes over time. The statistical models that are typically implemented to address these research questions do not allow for investigation of mechanisms of dynamic change over time. However, latent change score models allow for dynamic change (not just linear or exponential change) over time and have flexibility in parameter constraints that other longitudinal models do not have. Developmental researchers also frequently utilize mediation analyses to investigate mechanisms of influence in longitudinal research implemented in path analytic or latent growth curve models. In this article, we provide three examples of how mediation can be tested in the latent change score framework by combining aspects of traditional mediation models with latent change score models of repeated measures outcomes (and mediators and predictors) with more than two timepoints. We also provide the Mplus syntax to complete these analyses and practical considerations of latent change score mediation (LCSM) models.

Discernment of mediator and outcome measurement in the PACE trial

OBJECTIVE

When measuring latent traits, such as those used in psychology and psychiatry, it can be unclear whether the instruments used are measuring different concepts. This issue is particularly important in the context of mediation analysis, since for a sound mediation hypothesis the mediator and outcome should be distinct. We sought to assess the extent of measurement overlap between cognitive and behavioural mediators and physical functioning and fatigue outcomes in a large trial of treatments for chronic fatigue syndrome.

METHODS

A secondary analysis of 640 participants in the "Pacing, graded Activity, and Cognitive behaviour therapy" trial was conducted. Potential measurement overlap was assessed using generalised linear latent variable models where confirmatory factor models quantified the extent to which the addition of cross-loading items significantly improved model fit.

RESULTS

We considered 13 mediators and two outcomes, giving a total of 26 mediator-outcome pairs. Of these, only six showed evidence of cross-loading items, supporting the suggestion that the chosen mediator and outcome constructs were conceptually distinct.

CONCLUSION

This study highlights how established psychometric methods can be applied to assess measurement overlap in mediation analyses. We found minimal overlap between cognitive and behavioural mediators and physical functioning and fatigue outcomes that are often used in studies of chronic fatigue syndrome. The use of such methods in mediational studies where questionnaires are used to quantify latent traits would add to their robustness and transparency.

Mediators of the Disparities in Depression Between Sexual Minority and Heterosexual Individuals: A Systematic Review

Evidence suggests that sexual minorities (e.g., those identifying as lesbian, gay, or bisexual) experience increased rates of depression compared to heterosexual individuals. Minority stress theory suggests that this disparity is due to stigma experienced by sexual minorities. Stigma processes are proposed to contribute to reduced coping/support resources and increased vulnerability processes for mental health problems. This review provided a systematic examination of research assessing the evidence for mediating factors that help explain such disparities. A literature search was conducted using the databases PubMed, PsycINFO, and Web of Science. The review included 40 identified studies that examined mediators of sexual minority status and depressive outcomes using a between-group design (i.e., heterosexual versus sexual minority participants). Studies of adolescents and adult samples were both included. The most common findings were consistent with the suggestion that stressors such as victimization, harassment, abuse, and increased stress, as well as lower social and family support, may contribute to differing depression rates in sexual minority compared to heterosexual individuals. Differences in psychological processes such as self-esteem and rumination may also play a role but have had insufficient research attention so far. However, caution is needed because many papers had important methodological shortcomings such as the use of cross-sectional designs, inferior statistical analyses for mediation, or measures that had not been properly validated. Although firm conclusions cannot be drawn, the current evidence base highlights many factors potentially suitable for further exploration in high-quality longitudinal research or randomized studies intervening with the potential mediators.

Benefit-Finding Intervention Delivered Individually to Alzheimer Family Caregivers: Longer-Term Outcomes of a Randomized Double-Blind Controlled Trial

OBJECTIVES

To examine the longer-term effects of benefit-finding on caregivers' depressive symptoms (primary outcome), and global burden, role overload, psychological well-being, and positive aspects of caregiving (secondary outcomes).

METHOD

Ninety-six Hong Kong Chinese caregivers of relatives with Alzheimer's disease were randomly assigned to receive the benefit-finding intervention (BFT) or one of the two control conditions, namely, simplified psychoeducation (lectures only; SIM-PE) or standard psychoeducation (STD-PE). Caregivers received four biweekly one-to-one interventions of 3 hours each at their own homes. We focused on outcomes measured at 4- and 10-month follow-ups. The trajectories of intervention effects were modeled by BFT × time and BFT × time2 interaction terms.

RESULTS

Mixed-effects regression showed significant BFT × time2 interaction effects on depressive symptoms against both control conditions, suggesting diminishing BFT effects over time. Z tests showed that, compared with controls, BFT participants reported substantial reductions in depressive symptoms at 4-month follow-up (d = -0.85 and -0.75 vs. SIM-PE and STD-PE, respectively). For depressive symptoms measured at 10-month follow-up, BFT was indistinguishable from STD-PE, whereas a moderate effect was observed in comparison with SIM-PE (d = -0.52). Moreover, positive aspects of caregiving, but not other secondary outcomes, continued to show intervention effect up to 10-month follow-up.

DISCUSSION

Benefit-finding is an efficacious intervention for depressive symptoms in Alzheimer caregivers, with strong effects in the medium-term post-intervention and possible moderate effects in the longer-term post-intervention.

Estimation of Controlled Direct Effects in Longitudinal Mediation Analyses with Latent Variables in Randomized Studies

In a randomized study with longitudinal data on a mediator and outcome, estimating the direct effect of treatment on the outcome at a particular time requires adjusting for confounding of the association between the outcome and all preceding instances of the mediator. When the confounders are themselves affected by treatment, standard regression adjustment is prone to severe bias. In contrast, G-estimation requires less stringent assumptions than path analysis using SEM to unbiasedly estimate the direct effect even in linear settings. In this article, we propose a G-estimation method to estimate the controlled direct effect of treatment on the outcome, by adapting existing G-estimation methods for time-varying treatments without mediators. The proposed method can accommodate continuous and noncontinuous mediators, and requires no models for the confounders. Unbiased estimation only requires correctly specifying a mean model for either the mediator or the outcome. The method is further extended to settings where the mediator or outcome, or both, are latent, and generalizes existing methods for single measurement occasions of the mediator and outcome to longitudinal data on the mediator and outcome. The methods are utilized to assess the effects of an intervention on physical activity that is possibly mediated by motivation to exercise in a randomized study.

How and for whom does supportive adjustment to multiple sclerosis cognitive-behavioural therapy work? A mediated moderation analysis

The supportive adjustment for multiple sclerosis (saMS) randomised controlled trial showed cognitive behavioural therapy (CBT) reduced distress at 12-months compared to supportive listening (SL). Larger changes in distress and functional impairment following CBT occurred in participants with clinical distress at baseline. This secondary analysis investigates whether CBT treatment effects occur through pre-defined CBT mechanisms of change in the total cohort and clinically distressed subgroup. 94 participants were randomised to saMS CBT or SL. Primary outcomes were distress and functional impairment (12 months). Mediators included cognitive-behavioural variables at post-treatment (15 weeks). Structural equation mediation and mediated-moderation models adjusting for baseline confounders assessed mediation overall and by distress level. Significant mediation was found but only for those with clinical distress at baseline. Illness acceptance (-0.20, 95% confidence interval -0.01 to -0.46) and reduced embarrassment avoidance behaviours (-0.22, -0.02 to -0.58) mediated CBT's effect on distress. Changes in beliefs about processing emotions (-0.19, -0.001 to -0.46) mediated CBT's effect on functional impairment. saMS CBT had effects on distress and functional impairment via some of the hypothesised mechanisms drawn from a theoretical model of adjustment for MS but only among participants with clinical distress at baseline. Increasing acceptance and emotional expression and decreasing embarrassment avoidance improves MS adjustment.

Beyond intervention into daily life: A systematic review of generalisation following social communication interventions for young children with autism

Researchers have generally considered autistic individuals to have difficulties generalising learned skills across novel contexts. Successful generalisation is necessary for an intervention to have benefits in everyday life beyond the original learning environment. We conducted a systematic review of randomised controlled trials of early social communication interventions for children with autism in order to explore generalisation and its measurement. We identified nine RCTs that provided evidence of initial target learning and measured generalisation, of which eight demonstrated at least some successful generalisation across people, settings, and/or activities. The findings did not support the widely reported generalisation 'difficulties' associated with autism. However, generalisation was not consistent across all skills within studies, and one study found no generalisation despite evidence for initial target learning within the intervention context. In general, there are few methodologically sound social communication intervention studies exploring generalisation in autism and no consensus on how it should be measured. In particular, failure to demonstrate initial learning of target skills within the intervention setting and an absence of formal mediation analyses of the hypothesised mechanisms limit current research. We outline a framework within which measurement of generalisation can be considered for use in future trials. To maximise the effectiveness of interventions, the field needs to gain a better understanding of the nature of generalisation among autistic individuals and what additional strategies may further enhance learning. Autism Res 2020, 13: 506-522. © 2020 The Authors. Autism Research published by International Society for Autism Research published by Wiley Periodicals, Inc. LAY SUMMARY: It is generally considered that autistic individuals experience difficulties applying things they have learned in one context into different settings (e.g. from school to home). This is important to consider for intervention studies. Our review does not support a complete lack of generalisation but instead suggests that after early social communication intervention, autistic children can transfer some skills to new contexts. Overall, there is limited research in this area and further work is needed.

Cognitive change in cognitive-behavioural therapy v. pharmacotherapy for adult depression: a longitudinal mediation analysis

BACKGROUND

Although cognitive-behavioural therapy (CBT) is a well-established treatment for adult depression, its efficacy and efficiency may be enhanced by better understanding its mechanism(s) of action. According to the theoretical model of CBT, symptom improvement occurs via reductions in maladaptive cognition. However, previous research has not established clear evidence for this cognitive mediation model.

METHODS

The present study investigated the cognitive mediation model of CBT in the context of a randomized controlled trial of CBT v. antidepressant medication (ADM) for adult depression. Participants with major depressive disorder were randomized to receive 16 weeks of CBT (n = 54) or ADM (n = 50). Depression symptoms and three candidate cognitive mediators (dysfunctional attitudes, cognitive distortions and negative automatic thoughts) were assessed at week 0 (pre-treatment), week 4, week 8 and week 16 (post-treatment). Longitudinal associations between cognition and depression symptoms, and mediation of treatment outcome, were evaluated in structural equation models.

RESULTS

Both CBT and ADM produced significant reductions in maladaptive cognition and depression symptoms. Cognitive content and depression symptoms were moderately correlated within measurement waves, but cross-lagged associations between the variables and indirect (i.e. mediated) treatment effects were non-significant.

CONCLUSIONS

The results provide support for concurrent relationships between cognitive and symptom change, but not the longitudinal relationships hypothesized by the cognitive mediation model. Results may be indicative of an incongruence between the timing of measurement and the dynamics of cognitive and symptom change.

Long-Term Outcomes of the Benefit-Finding Group Intervention for Alzheimer Family Caregivers: A Cluster-Randomized Double-Blind Controlled Trial

OBJECTIVES

To examine the effects of the group benefit-finding therapeutic intervention (BFT) for Alzheimer family caregivers up to 10-month follow-up.

METHODS

This was a cluster-randomized double-blind controlled trial in social centers and clinics. Participants included 129 caregivers. Inclusion criteria were 1) primary caregiver aged 18 years and older and without cognitive impairment, 2) providing 14 or more care hours per week to a relative with mild-to-moderate Alzheimer disease, and 3) scoring 3 or more on the Hamilton Depression Rating Scale. Exclusion criterion was care-recipient having parkinsonism or other forms of dementia. BFT (using cognitive reappraisal to find positive meanings) was evaluated against two forms of psychoeducation as controls-standard and simplified (lectures only) psychoeducation. All interventions had eight weekly sessions of 2 hours each. Primary outcome was depressive symptoms, whereas secondary outcomes were global burden, role overload, and psychological well-being. Measures were collected at baseline, postintervention, and 4- and 10-month follow-up.

RESULTS

Mixed-effects regression showed that BFT's effect on depressive symptoms conformed to a curvilinear pattern, in which the strong initial effect leveled out after postintervention and was maintained up to 10-month follow-up; this was true when compared against either control group. The effect on global burden was less impressive but moderate effect sizes were found at the two follow-ups. For psychological well-being, there was an increase in the BFT group at 4-month follow-up and a return to baseline afterward. No effect on role overload was found.

CONCLUSION

Benefit-finding reduces depressive symptoms as well as global burden in the long-term and increases psychological well-being in the medium-term.

Systemic low-grade inflammation and subsequent depressive symptoms: Is there a mediating role of physical activity?

OBJECTIVE

Systemic low-grade inflammation has been associated with the onset of depression, but the exact mechanisms underlying this relationship remain elusive. This study examined whether physical activity (PA) explained the association between elevated serum levels of inflammatory markers and subsequent depressive symptoms.

DESIGN

Prospective cohort design.

METHOD

The sample consisted of 3809 non-depressed men and women (aged 50+) recruited from the English Longitudinal Study of Ageing (ELSA). Serum levels of inflammatory markers (C-reactive protein (CRP), fibrinogen) and covariates (age, sex, education, wealth, body mass index, smoking, cholesterol, triglycerides) were measured at baseline (wave 4, 2008/09). Self-reported weekly moderate/vigorous (high) PA versus no weekly moderate/vigorous (low) PA was examined at a four-year follow-up (wave 6, 2012/13), using a single-item question. Depressive symptoms were assessed at baseline, four years (wave 6, 2012/13) and six years post baseline (wave 7, 2014/15), using the 8-item version of the Centre for Epidemiological Studies Depression Scale (CES-D).

RESULTS

Participants with higher baseline concentrations of inflammatory markers were significantly more likely to report low PA levels four years later (CRP: OR: 1.25; 95% CI, 1.05-1.48; fibrinogen: OR: 1.18; 95% CI, 1.05-1.39). Moreover, low PA was associated with higher odds of elevated depressive symptoms at follow-up (OR: 1.59; 95% CI, 1.15-2.19). Mediation analyses revealed that low PA explained a total of 36.71% of the relationship between high CRP and elevated depressive symptoms, and 33.26% between higher levels of fibrinogen and elevated depressive symptoms six years later. No direct association was found between systemic low-grade inflammation and future depressive symptoms.

CONCLUSION

These results suggest that low PA is a significant partial mediator of the relationship between systemic low-grade inflammation and subsequent elevated depressive symptoms in a nationally representative cohort of older adults.

Impact of telephone delivered case-management on the effectiveness of collaborative care for depression and anti-depressant use: A systematic review and meta-regression

Background

The health service delivery framework collaborative care is an effective intervention for depression. However, uncertainties remain about how to optimise its delivery at scale. Structured case management is a core component of collaborative care; its delivery via the telephone may improve access.

Aims

To examine using meta-regression if telephone delivered case management diminishes the clinical effectiveness of collaborative care on depressive symptoms and anti-depressant use relative to face-to-face delivery methods.

Methods

Randomised controlled trials were eligible if they included collaborative care interventions for adults with depression identified using self-report measures or diagnostic interviews and reported depression outcomes. Sociodemographics, intervention characteristics, depressive symptoms, and anti-depressant use were extracted. Random effects univariable and multivariable meta-regression analyses were used to examine the moderating effect of telephone delivered case-management on outcomes.

Results

Ninety-four trials were identified comprising of 103 comparisons across 24, 132 participants with depression outcomes and 67 comparisons from 15,367 participants with anti-depressant use outcomes. Telephone delivered case management did not diminish the effects of collaborative care on depressive symptoms (β = -0.01, 95% CI -0.12 to 0.10; p = 0.86). Telephone delivered case management decreased anti-depressant medication use (relative risk 0.76, 95% CI 0.63 to 0.92; p = 0.005); this effect remained when assessed simultaneously alongside other study-level moderators of collaborative care.

Conclusion

Using remote platforms such as the telephone to deliver case management may be a feasible way to implement collaborative care with no loss of effectiveness on depressive symptoms. However, adherence to anti-depressant medication may decrease when telephone case management is used.

Illustration of Measurement Error Models for Reducing Bias in Nutrition and Obesity Research Using 2-D Body Composition Data

OBJECTIVE

This study aimed to illustrate the use and value of measurement error models for reducing bias when evaluating associations between body fat and having type 2 diabetes (T2D) or being physically active.

METHODS

Logistic regression models were used to evaluate T2D and physical activity among adults aged 19 to 80 years from the Photobody Study (n = 558). Self-reported T2D and physical activity were categorized as "yes" or "no." Body fat measured by two-dimensional photographs was adjusted for bias using dual-energy x-ray absorptiometry scans as a reference. Three approaches were applied: regression calibration (RC), simulation extrapolation (SIMEX), and multiple imputation (MI).

RESULTS

Unadjusted two-dimensional measures of body fat had upward biases of 30% and 233% for physical activity and T2D, respectively. For the physical activity model, RC-adjusted values had a 13% upward bias, whereas MI and SIMEX decreased the bias to 9% and 91%, respectively. For the T2D model, MI reduced the bias to 0%, whereas RC and SIMEX increased the upward bias to > 300%.

CONCLUSIONS

Of three statistical approaches to reducing bias due to measurement errors, MI performed best in comparison to RC and SIMEX. Measurement error methods can improve the reliability of analyses that look for relations between body fat measures and health outcomes.

Tutorial: The practical application of longitudinal structural equation mediation models in clinical trials

The study of mediation of treatment effects, or how treatments work, is important to understanding and improving psychological and behavioral treatments, but applications often focus on mediators and outcomes measured at a single time point. Such cross-sectional analyses do not respect the implied temporal ordering that mediation suggests. Clinical trials of treatments often provide repeated measures of outcomes and, increasingly, of mediators as well. Repeated measurements allow the application of various types of longitudinal structural equation mediation models. These provide flexibility in modeling, including the ability to incorporate some types of measurement error and unmeasured confounding that can strengthen the robustness of findings. The usual approach is to identify the most theoretically plausible model and apply that model. In the absence of clear theory, we put forward the option of fitting a few theoretically plausible models, providing a type of sensitivity analysis for the mediation hypothesis. In this tutorial, we outline how to fit several longitudinal mediation models, including simplex, latent growth and latent change models. This will allow readers to learn about one type of model that is of interest, or about several alternative models, so that they can take this sensitivity approach. We use the Pacing, Graded Activity, and Cognitive Behavioral Therapy: A Randomized Evaluation (PACE) trial of rehabilitative treatments for chronic fatigue syndrome (ISRCTN 54285094) as a motivating example and describe how to fit and interpret various longitudinal mediation models using simulated data similar to those in the PACE trial. The simulated data set and Mplus code and output are provided. (PsycINFO Database Record