Addressing unmeasured confounders in cohort studies: Instrumental variable method for a time-fixed exposure on an outcome trajectory.

Instrumental variable methods, which handle unmeasured confounding by targeting the part of the exposure explained by an exogenous variable not subject to confounding, have gained much interest in observational studies. We consider the very frequent setting of estimating the unconfounded effect of an exposure measured at baseline on the subsequent trajectory of an outcome repeatedly measured over time. We didactically explain how to apply the instrumental variable method in such setting by adapting the two-stage classical methodology with (1) the prediction of the exposure according to the instrumental variable, (2) its inclusion into a mixed model to quantify the exposure association with the subsequent outcome trajectory, and (3) the computation of the estimated total variance. A simulation study illustrates the consequences of unmeasured confounding in classical analyses and the usefulness of the instrumental variable approach. The methodology is then applied to 6224 participants of the 3C cohort to estimate the association of type-2 diabetes with subsequent cognitive trajectory, using 42 genetic polymorphisms as instrumental variables. This contribution shows how to handle endogeneity when interested in repeated outcomes, along with a R implementation. However, it should still be used with caution as it relies on instrumental variable assumptions hardly testable in practice.

Time and change: a typology for presenting research findings in qualitative longitudinal research.

BACKGROUND: Qualitative longitudinal research (QLR) is an emerging methodology used in health research. The method literature states that the change in a phenomenon through time should be the focus of any QLR study, but in empirical studies, the analysis of changes through time is often poorly described, and the emphasis on time/change in the findings varies greatly. This inconsistency might depend on limitations in the existing method literature in terms of describing how QLR studies can present findings. The aim of this study was to develop and describe a typology of alternative approaches for integrating time and/or change in QLR findings. METHODS: In this method study, we used an adapted scoping review design. Articles were identified using EBSCOhost. In total, methods and results sections from 299 QLR articles in the field of health research were analyzed with inspiration from content analysis. RESULTS: We constructed a typology of three types and seven subtypes. The types were based on the underlying structural principles of how time/change was presented: Type A) Findings have a low utilization of longitudinal data, Type B) Findings are structured according to chronological time, and Type C) Findings focus on changes through time. These types differed in 1) the way the main focus was on time, change or neither; 2) the level of interpretation in the findings; and 3) how theoretical understandings of time/change were articulated in the articles. Each type encompassed two or three subtypes that represented distinct approaches to the aim and results presentation of QLR findings. CONCLUSIONS: This method study is the first to describe a coherent and comprehensive typology of alternative approaches for integrating time/change into QLR findings in health research. By providing examples of various subtypes that can be used for results presentations, it can help researchers make informed decisions suitable to their research intent.

Tweedie Compound Poisson Models with Covariate-Dependent Random Effects for Multilevel Semicontinuous Data.

Multilevel semicontinuous data occur frequently in medical, environmental, insurance and financial studies. Such data are often measured with covariates at different levels; however, these data have traditionally been modelled with covariate-independent random effects. Ignoring dependence of cluster-specific random effects and cluster-specific covariates in these traditional approaches may lead to ecological fallacy and result in misleading results. In this paper, we propose Tweedie compound Poisson model with covariate-dependent random effects to analyze multilevel semicontinuous data where covariates at different levels are incorporated at relevant levels. The estimation of our models has been developed based on the orthodox best linear unbiased predictor of random effect. Explicit expressions of random effects predictors facilitate computation and interpretation of our models. Our approach is illustrated through the analysis of the basic symptoms inventory study data where 409 adolescents from 269 families were observed at varying number of times from 1 to 17 times. The performance of the proposed methodology was also examined through the simulation studies.

Qualitative longitudinal research in health research: a method study.

BACKGROUND: Qualitative longitudinal research (QLR) comprises qualitative studies, with repeated data collection, that focus on the temporality (e.g., time and change) of a phenomenon. The use of QLR is increasing in health research since many topics within health involve change (e.g., progressive illness, rehabilitation). A method study can provide an insightful understanding of the use, trends and variations within this approach. The aim of this study was to map how QLR articles within the existing health research literature are designed to capture aspects of time and/or change. METHODS: This method study used an adapted scoping review design. Articles were eligible if they were written in English, published between 2017 and 2019, and reported results from qualitative data collected at different time points/time waves with the same sample or in the same setting. Articles were identified using EBSCOhost. Two independent reviewers performed the screening, selection and charting. RESULTS: A total of 299 articles were included. There was great variation among the articles in the use of methodological traditions, type of data, length of data collection, and components of longitudinal data collection. However, the majority of articles represented large studies and were based on individual interview data. Approximately half of the articles self-identified as QLR studies or as following a QLR design, although slightly less than 20% of them included QLR method literature in their method sections. CONCLUSIONS: QLR is often used in large complex studies. Some articles were thoroughly designed to capture time/change throughout the methodology, aim and data collection, while other articles included few elements of QLR. Longitudinal data collection includes several components, such as what entities are followed across time, the tempo of data collection, and to what extent the data collection is preplanned or adapted across time. Therefore, there are several practices and possibilities researchers should consider before starting a QLR project.

A longitudinal cohort study of gestational diabetes mellitus and perinatal depression.

BACKGROUND: The association between gestational diabetes mellitus (GDM) and perinatal depression (PND) remains controversial. Our study aimed to comprehensively assess this association in a longitudinal cohort study with repeated measurements of depression. METHODS: Our cohort study was nested in a pilot study of an implementation study aiming to screen and manage perinatal depression within the primary health system in China. Women were recruited in the first trimester from May-September 2019 and followed four times up to 1 year postpartum. Data on sociodemographic characteristics and depression were collected using self-developed questionnaires incorporating the Edinburgh Postnatal Depression Scale (EPDS). Oral glucose tolerance test at 24 ~ 28 weeks and fasting plasma glucose (FPG) data were extracted from medical records. Depression throughout the whole period was divided into different trajectories. Associations of GDM with PND at different time periods and PND of different trajectories were determined by logistic regression. The path of association between blood glucose and depression over time was estimated with an autoregressive cross-lagged model. RESULTS: In total, 1043 women were included in this analysis and 313 (30.0%) were diagnosed with GDM. The prevalence of depression in the first, second, and third trimesters and postpartum period were 17.2, 6.9, 6.8 and 9.0%, respectively. GDM was neither significantly associated with PND at any time point nor with any specific trajectory of depression. Except for autoregressive paths, no cross-lagged path of FPG and scores of EPDS was significant. CONCLUSIONS: Our study indicates no association between GDM/blood glucose and PND.

Decreased serial scores of severe organ failure assessments are associated with survival in mechanically ventilated patients; the prospective Maastricht Intensive Care COVID cohort.

BACKGROUND: The majority of patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection are admitted to the Intensive Care Unit (ICU) for mechanical ventilation. The role of multi-organ failure during ICU admission as driver for outcome remains to be investigated yet. DESIGN AND SETTING: Prospective cohort of mechanically ventilated critically ill with SARS-CoV-2 infection. PARTICIPANTS AND METHODS: 94 participants of the MaastrICCht cohort (21% women) had a median length of stay of 16 days (maximum of 77). After division into survivors (n = 59) and non-survivors (n = 35), we analysed 1555 serial SOFA scores using linear mixed-effects models. RESULTS: Survivors improved one SOFA score point more per 5 days (95% CI: 4-8) than non-survivors. Adjustment for age, sex, and chronic lung, renal and liver disease, body-mass index, diabetes mellitus, cardiovascular risk factors, and Acute Physiology and Chronic Health Evaluation II score did not change this result. This association was stronger for women than men (P-interaction = 0.043). CONCLUSIONS: The decrease in SOFA score associated with survival suggests multi-organ failure involvement during mechanical ventilation in patients with SARS-CoV-2. Surviving women appeared to improve faster than surviving men. Serial SOFA scores may unravel an unfavourable trajectory and guide decisions in mechanically ventilated patients with SARS-CoV-2.