A review on longitudinal data analysis with random forest.

In longitudinal studies variables are measured repeatedly over time, leading to clustered and correlated observations. If the goal of the study is to develop prediction models, machine learning approaches such as the powerful random forest (RF) are often promising alternatives to standard statistical methods, especially in the context of high-dimensional data. In this paper, we review extensions of the standard RF method for the purpose of longitudinal data analysis. Extension methods are categorized according to the data structures for which they are designed. We consider both univariate and multivariate response longitudinal data and further categorize the repeated measurements according to whether the time effect is relevant. Even though most extensions are proposed for low-dimensional data, some can be applied to high-dimensional data. Information of available software implementations of the reviewed extensions is also given. We conclude with discussions on the limitations of our review and some future research directions.

The Value of Serial Measurement of Arterial Stiffness in Cardiovascular Disease.

Clinically assessing arterial stiffness is valuable because it aids in predicting future cardiovascular events. There are several methods for measuring arterial stiffness, including pulse wave velocity (PWV), augmentation index, and pulse pressure. Numerous studies have shown that these indicators of arterial stiffness possess prognostic value for various patient groups as well as the general population. In cross-sectional studies, arterial stiffness was also linked to organ damage indices. However, most studies related to arterial stiffness have relied on a single measurement. Taking multiple serial measurements of arterial stiffness offers several advantages. Through repeated assessments, one can confirm the variability of arterial stiffness and observe changes over time, which is beneficial for understanding its pathophysiology. Such repeated measurements are also invaluable in evaluating the efficacy of interventions aimed at improving arterial stiffness. However, caution is needed, as there is no standardized method for measuring arterial stiffness. For instance, with PWV, the values can be influenced by numerous external factors. Therefore, the external conditions during the measurement must be noted. It's essential to recognize the pros and cons of repeated arterial stiffness measurements and integrate them effectively into clinical practice.

Urinary Equol and Equol-Predicting Microbial Genera Are Favorably Associated with Body Fat Measures among Chinese Adults.

BACKGROUND: Many studies have investigated the intake of dietary isoflavones in relation to obesity risk, whereas the association using objective biomarkers of isoflavones, particularly equol (a gut-derived metabolite of daidzein with greater bioavailability than other isoflavones) has been less studied. In addition, the associations between equol and gut microbiota profile at the population level remain to be fully characterized. OBJECTIVES: We aimed to identify equol-predicting microbial species and to investigate the associations of equol-predicting microbial species and urinary excretion of isoflavones including glycitein, genistein, daidzein, and equol with diverse obesity markers in free living-individuals. METHODS: In this 1-y longitudinal study of 754 community-dwelling adults, urinary isoflavones, fecal microbiota, height, weight, and circumferences of waist and hip were measured at baseline and again after 1 y. Liver fat [indicated by the controlled attenuation parameter (CAP)] and other body composition were also measured after 1 y. Linear models and linear mixed-effects models were used to analyze the associations for single measure and repeated measures, respectively. RESULTS: Among 305 participants (median age: 50 y, IQR, 37-59 y) including 138 males and 167 females, higher urinary excretion of equol was associated with lower CAP (ß = -0.013, P < 0.001) and body fat mass (ß= -0.014, P = 0.046). No association was found between any other urinary isoflavones and obesity markers (all P > 0.05). We identified 21 bacterial genera whose relative abundance were positively associated with urinary equol concentrations (all Pfalsediscovery rate < 0.05), and constructed an equol-predicting microbial score to reflect the overall equol-producing potential of host gut microbiota. This score was inversely associated with CAP (ß = -0.040, P = 0.011). CONCLUSIONS: High urinary equol concentrations and equol-predicting microbial species could be favorably associated with liver fat and other obesity markers.

Association of urinary and seminal plasma vanadium concentrations with semen quality: A repeated-measures study of 1135 healthy men.

Although animal studies have shown the reproductive toxicity of vanadium, less is known about its effects on semen quality in humans. Among 1135 healthy men who were screened as potential semen donors, we investigated the relationships of semen quality with urinary and seminal plasma vanadium levels via inductively coupled plasma-mass spectrometry (ICP-MS). Spearman rank correlation tests and linear regression models were used to assess the correlations between average urinary and within-individual pooled seminal plasma vanadium concentrations (n = 1135). We utilized linear mixed-effects models to evaluate the associations of urinary and seminal plasma vanadium levels (n = 1135) with repeated sperm quality parameters (n = 5576). Seminal plasma vanadium concentrations were not significantly correlated with urinary vanadium concentrations (r = 0.03). After adjusting for possible confounders, we observed inverse relationships of within-individual pooled seminal plasma vanadium levels with total count, semen volume, and sperm concentration (all P values for trend < 0.05). Specifically, subjects in the highest (vs. lowest) tertile of seminal plasma vanadium concentrations had - 11.3% (-16.4%, -5.9%), - 11.1% (-19.1%, -2.4%), and - 20.9% (-29.0%, -11.8%) lower sperm volume, concentration, and total count, respectively; moreover, urinary vanadium levels appeared to be negatively associated with sperm motility. These relationships showed monotonically decreasing dose-response patterns in the restricted cubic spline analyses. Our results demonstrated a poor correlation between urinary and seminal plasma levels of vanadium, and elevated vanadium concentrations in urine and seminal plasma may be adversely related to male semen quality.

Fast approximate inference for multivariate longitudinal data.

Collecting information on multiple longitudinal outcomes is increasingly common in many clinical settings. In many cases, it is desirable to model these outcomes jointly. However, in large data sets, with many outcomes, computational burden often prevents the simultaneous modeling of multiple outcomes within a single model. We develop a mean field variational Bayes algorithm, to jointly model multiple Gaussian, Poisson, or binary longitudinal markers within a multivariate generalized linear mixed model. Through simulation studies and clinical applications (in the fields of sight threatening diabetic retinopathy and primary biliary cirrhosis), we demonstrate substantial computational savings of our approximate approach when compared to a standard Markov Chain Monte Carlo, while maintaining good levels of accuracy of model parameters.

Health-related quality of life in patients with primary sclerosing cholangitis: A longitudinal population-based cohort study.

BACKGROUND & AIMS: Data regarding health-related quality of life (HRQoL) in primary sclerosing cholangitis (PSC) are sparse and have only been studied cross-sectionally in a disease which runs a fluctuating and unpredictable course. We aim to describe HRQoL longitudinally by using repeated measurements in a population-based cohort. METHODS: Every 3 months from May 2017 up to August 2020, patients received digital questionnaires at home. These included the EQ-5D, 5-D Itch, patient-based SCCAI and patient-based HBI. The SF-36, measuring HRQoL over eight dimensions as well as a physical component summary (PCS) and mental component summary (MCS) score, was sent annually. Data were compared with Dutch reference data and a matched IBD disease control from the population-based POBASIC cohort. Mixed-effects modelling was performed to identify factors associated with HRQoL. RESULTS: Three hundred twenty-eight patients completed 2576 questionnaires. A significant reduction of small clinical relevance in several mean HRQoL scores was found compared with the Dutch reference population: 46.4 versus 48.0, p = .018 for PCS and 47.5 versus 50.5, p = .004 for MCS scores. HRQoL outcomes were significantly negatively associated with coexisting active IBD (PCS -12.2, p < .001 and MCS -12.0, p < .001), which was not the case in case of quiescent IBD. Decreasing HRQoL scores were also negatively associated with increasing age (PCS -0.1 per 10 years, p = .002), female sex (PCS -2.8, p < .001), diagnosis of AIH overlap (PCS -3.7, p = .059), end-stage liver disease (PCS -3.7, p = .015) and presence of itch (PCS -9.2, p < .001 and MCS -3.1, p = .078). The odds of reporting a clinically relevant reduction in EQ-5D scores showed seasonal variation, being lowest in summer (OR = 0.48 relative to spring, p = .037). In patients with liver transplant, HRQoL outcomes were comparable to the Dutch general population. CONCLUSIONS: PSC patients report impaired HRQoL of small clinical relevance compared with the general population. After liver transplantation, HRQoL scores are at comparable levels to the general population. HRQoL scores are associated with potentially modifiable factors such as itch and IBD activity.

NT-proBNP and changes in cognition and global brain structure: The Rotterdam Study.

OBJECTIVE: To investigate the association between N-terminal pro-B-type natriuretic peptide (NT-proBNP) and changes in cognition and global brain structure. METHODS: In the Rotterdam Study, baseline NT-proBNP was assessed at baseline from 1997 to 2008. Between 1997 and 2016, participants without dementia or stroke at baseline (n = 9566) had repeated cognitive tests (every 3-6 years) for global cognitive function, executive cognitive function, fine manual dexterity, and memory. Magnetic resonance imaging of the brain was performed repeatedly at re-examination visits between 2005 and 2015 for 2607 participants to obtain brain volumes, focal brain lesions, and white matter microstructural integrity as measures of brain structure. RESULTS: Among 9566 participants (mean age 65.1 ± 9.8 years), 5444 (56.9%) were women, and repeated measures of cognition were performed during a median follow-up time of 5.5 (range 1.1-17.9) years, of whom 2607 participants completed at least one brain imaging scan. Higher levels of NT-proBNP were associated with a faster decline of scores in the global cognitive function (p value = 0.003) and the Word-Fluency test (p value = 0.003) but were not related to a steeper deterioration in brain volumes, global fractional anisotropy, and mean diffusivity, as indicators of white matter microstructural integrity, or focal brain lesions. CONCLUSIONS: Higher baseline NT-proBNP levels were associated with a faster decline in cognition; however, no association with global brain structure was found.

Longitudinal profile of circulating endothelial cells in post-acute coronary syndrome patients.

IntroductionPatients who have experienced an acute coronary syndrome (ACS) are at risk of a recurrent event, but their level of risk varies. Because of their close temporal relationship with vascular injury, longitudinal measurements of circulating endothelial cells (CECs) carry potential to improve individual risk assessment.MethodsWe conducted an explorative nested case-control study within our multicenter, prospective, observational biomarker study (BIOMArCS) of 844 ACS patients. Following an index ACS, high-frequency blood sampling was performed during 1-year follow-up. CECs were identified using flow cytometric analyses in 15 cases with recurrent event, and 30 matched controls.ResultsCases and controls had a median (25th-75thpercentile) age of 64.1 (58.1-75.1) years and 80% were men. During the months preceding the endpoint, the mean (95%CI) CEC concentration in cases was persistently higher than in controls (12.8 [8.2-20.0] versus 10.0 [7.0-14.4] cells/ml), although this difference was non-significant (P = 0.339). In controls, the mean cell concentration was significantly (P = 0.030) lower in post 30-day samples compared to samples collected within one day after index ACS: 10.1 (7.5-13.6) versus 17.0 (10.8-26.6) cells/ml. Similar results were observed for CEC subsets co-expressing CD133 and CD309 (VEGFR-2) or CD106 (VCAM-1).ConclusionDespite their close relation to vascular damage, no increase in cell concentrations were found prior to the occurrence of a secondary adverse cardiac event.

Power and sample size analysis for longitudinal mixed models of health in populations exposed to environmental contaminants: a tutorial.

BACKGROUND: When evaluating the impact of environmental exposures on human health, study designs often include a series of repeated measurements. The goal is to determine whether populations have different trajectories of the environmental exposure over time. Power analyses for longitudinal mixed models require multiple inputs, including clinically significant differences, standard deviations, and correlations of measurements. Further, methods for power analyses of longitudinal mixed models are complex and often challenging for the non-statistician. We discuss methods for extracting clinically relevant inputs from literature, and explain how to conduct a power analysis that appropriately accounts for longitudinal repeated measures. Finally, we provide careful recommendations for describing complex power analyses in a concise and clear manner. METHODS: For longitudinal studies of health outcomes from environmental exposures, we show how to [1] conduct a power analysis that aligns with the planned mixed model data analysis, [2] gather the inputs required for the power analysis, and [3] conduct repeated measures power analysis with a highly-cited, validated, free, point-and-click, web-based, open source software platform which was developed specifically for scientists. RESULTS: As an example, we describe the power analysis for a proposed study of repeated measures of per- and polyfluoroalkyl substances (PFAS) in human blood. We show how to align data analysis and power analysis plan to account for within-participant correlation across repeated measures. We illustrate how to perform a literature review to find inputs for the power analysis. We emphasize the need to examine the sensitivity of the power values by considering standard deviations and differences in means that are smaller and larger than the speculated, literature-based values. Finally, we provide an example power calculation and a summary checklist for describing power and sample size analysis. CONCLUSIONS: This paper provides a detailed roadmap for conducting and describing power analyses for longitudinal studies of environmental exposures. It provides a template and checklist for those seeking to write power analyses for grant applications.

Influence of analytic methods, data sources, and repeated measurements on the population attributable fraction of lifestyle risk factors.

Population attributable risk (PAR%) reflects the preventable fraction of disease. However, PAR% estimates of cancer have shown large variation across populations, methods, data sources, and timing of measurements. Three statistical methods to estimate PAR% were identified from a systematic literature review: the Levin's formula, the comparative incidence rate method, and the comparative risk assessment method. We compared the variations in PAR% of postmenopausal breast cancer in the Nurses' Health Study to evaluate the influence by method choice, source of prevalence data, use of single vs repeated exposure measurements, and potential joint effects of obesity, alcohol, physical activity, fruit and vegetable intake. Across models of the three methods, the estimated PAR% using repeated measurements were higher than that using baseline measurement; overall PAR% for the baseline, simple update, and cumulative average models were 13.8%, 21.1%, 18.6% by Levin's formula; 13.7%, 28.0%, 31.2% by comparative risk assessment; and 17.4%, 25.2%, 29.3% by comparative incidence rate method. The estimated PAR% of the combination of multiple risk factors was higher than the product of the individual PAR%: 18.9% when assuming independence and 31.2% when considering the risk factors jointly. The three methods provided similar PAR% based on the same data source, timing of measurements, and target populations. However, sizable increases in the PAR% were observed for repeated measures over a single measure and for calculations based on achieving all recommendations jointly rather than individually.

The healthy beverage index is not associated with insulin resistance, prediabetes and type 2 diabetes risk in the Rotterdam Study.

PURPOSE: Whether beverage quality affects changes in glycaemic traits and type 2 diabetes (T2D) risk is unknown. We examined associations of a previously developed Healthy Beverage Index (HBI) with insulin resistance, and risk of prediabetes and T2D. METHODS: We included 6769 participants (59% female, 62.0 ± 7.8 years) from the Rotterdam Study cohort free of diabetes at baseline. Diet was assessed using food-frequency questionnaires at baseline. The HBI included 10 components (energy from beverages, meeting fluid requirements, water, coffee and tea, low-fat milk, diet drinks, juices, alcohol, full-fat milk, and sugar-sweetened beverages), with a total score ranging from 0 to 100. A higher score represents a healthier beverage pattern. Data on study outcomes were available from 1993 to 2015. Multivariable linear mixed models and Cox proportional-hazards regression models were used to examine associations of the HBI (per 10 points increment) with two measurements of HOMA-IR (a proxy for insulin resistance), and risk of prediabetes and T2D. RESULTS: During follow-up, we documented 1139 prediabetes and 784 T2D cases. Mean ± SD of the HBI was 66.8 ± 14.4. Higher HBI score was not associated with HOMA-IR (ß: 0.003; 95% CI - 0.007, 0.014), or with risk of prediabetes (HR: 1.01; 95% CI 0.97, 1.06), or T2D (HR: 1.01; 95% CI 0.96, 1.07). CONCLUSION: Our findings suggest no major role for overall beverage intake quality assessed with the HBI in insulin resistance, prediabetes and T2D incidence. The HBI may not be an adequate tool to assess beverage intake quality in our population.

Overall lifestyle changes in adulthood are associated with cancer incidence in the Norwegian Women and Cancer Study (NOWAC) - a prospective cohort study.

BACKGROUND: Cancer is a leading cause of premature death worldwide and incidence is expected to rise in the coming decades. Many cohort studies, measuring lifestyle factors at one time-point, have observed that overall healthy lifestyles were inversely related to cancer incidence. However, there is little knowledge on the impact of lifestyle modification within adulthood. METHODS: Using the Norwegian Women and Cancer study, two repeated self-reported assessments of lifestyle behaviours were used to calculate healthy lifestyle index scores at each time-point (N = 66 233). The associations between change in healthy lifestyle index score and lifestyle-related cancer incidence, including alcohol-, tobacco-, obesity-, and reproductive-related, and site-specific breast and colorectal cancer incidence were estimated using Cox proportional hazard regression models. To assess nonlinearity in the dose-response relationships, restricted cubic spline models were used. RESULTS: Independent of baseline lifestyle, positive lifestyle changes were inversely related to the incidence of overall lifestyle-related cancers, as well as alcohol-related, tobacco-related, obesity-related, and reproductive-related cancers, but not breast and colorectal site-specific cancers. An association between lifestyle worsening and cancer incidence compared to stable lifestyle was observed. CONCLUSIONS: This study provides evidence that overall lifestyle changes among cancer-free women between the ages of 41 and 76 impact the incidence of many cancer types. Regardless of baseline lifestyle, there was a negative dose-response relationship between magnitude of positive lifestyle change and the incidence of overall lifestyle-related cancers. We observed that underlying this trend was an especially clear association between lifestyle worsening and increased risk compared to stable lifestyle. For adult women, maintaining a stable healthy lifestyle and lifestyle improvement are important for preventing the occurrence of many cancer types.

Exposing the confounding in experimental designs to understand and evaluate them, and formulating linear mixed models for analyzing the data from a designed experiment.

Comparative experiments involve the allocation of treatments to units, ideally by randomization. This necessarily confounds treatment information with unit information, which we distinguish from the other forms of information blending, in particular aliasing and marginality. We outline a factor-allocation paradigm for describing experimental designs with the aim of (i) exhibiting the confounding in a design, using analysis-of-variance-like tables, so as to understand and evaluate the design and (ii) formulating a linear mixed model based on the factor allocation that the design involves. The approach exhibits the dispersal of treatments information between units sources, allows designers a choice in the strategy that they adopt for including block-treatment interactions, clarifies differences between experiments, accommodates systematic allocation of factors, and provides a consolidated analysis of nonorthogonal designs. It provides insights into the process of designing experiments and issues that commonly arise with designs. The paradigm has pedagogical advantages and is implemented using the R package dae.

Blood pressure patterns of hypertensive disorders of pregnancy in first and second trimester and contributing factors: a retrospective study.

We aimed to investigate the blood pressure (BP) patterns of hypertensive disorders of pregnancy (HDP) in the first and second trimesters and its contributing factors, which may help us understand its pathogenesis and identify this group of diseases in a timely manner. SPSS 21.0 was used to describe the BP patterns of 688 HDP as well as 2050 normotensive pregnancies respectively before 28 gestational weeks, and the repeated measurements and two-way ANOVA was used to decide the significant difference of blood pressure in the same period. The results revealed blood pressure in HDP underwent a mid-pregnancy drop as normal while the drop was unremarkable in advanced-age or obesity pregnancies. Besides, we found blood pressure was significantly higher in patients during first and second trimesters, not just after 20 weeks. In conclusion, our study indicated a significant elevation of blood pressure had appeared before 20 weeks in HDP pregnancies, we should pay more attention to monitoring blood pressure before 20 weeks, especially for advanced age and obese women.IMPACT STATEMENTWhat is already known on this subject? Gestational hypertension, preeclampsia as well as eclampsia were supposed to have the similar pathogenesis and their time of onset was strictly defined after 20 gestational weeks, while the reason for the time point was not clear. On the other hand, higher blood pressure in the first trimester was associated with increasing risk of developing HDP, while the blood pressure(BP) pattern of normal as well as HDP pregnancy was still controversial, especially for the existence of mid-trimester drop.What do the results of this study add? Firstly, we found blood pressure in HDP underwent a mid-pregnancy drop as normal while the BP drop was unremarkable in advanced-age or obesity pregnancies. Secondly, we noticed the blood pressure in HDP was significantly higher than the normal before 20 weeks, which had not been proved before.What are the implications of these findings for clinical practice and/or further research? On one hand, both the abnormal elevation of BP and the development of the placenta happened in the first trimester suggested toxic substances caused by the defective placenta played a vital role in the onset and aggravation of HDP, which guides us to pay more attention to monitor blood pressure before 20 weeks, especially for advanced age and obesity pregnancies. On the other hand, our results about BP patterns in HDP help us identify this group of diseases in time which can contribute to a better outcome.

A functional mixed model for scalar on function regression with application to a functional MRI study.

Motivated by a functional magnetic resonance imaging (fMRI) study, we propose a new functional mixed model for scalar on function regression. The model extends the standard scalar on function regression for repeated outcomes by incorporating subject-specific random functional effects. Using functional principal component analysis, the new model can be reformulated as a mixed effects model and thus easily fit. A test is also proposed to assess the existence of the subject-specific random functional effects. We evaluate the performance of the model and test via a simulation study, as well as on data from the motivating fMRI study of thermal pain. The data application indicates significant subject-specific effects of the human brain hemodynamics related to pain and provides insights on how the effects might differ across subjects.

Test-retest reliability of the Epworth Sleepiness Scale in clinical trial settings.

The present analysis examined the test-retest reliability of the Epworth Sleepiness Scale in participants with excessive daytime sleepiness associated with narcolepsy or obstructive sleep apnea in three clinical trials. Intraclass correlation coefficient estimates for Epworth Sleepiness Scale scores from two solriamfetol 12-week placebo-controlled trials (one narcolepsy, one obstructive sleep apnea) and one long-term open-label extension trial (narcolepsy or obstructive sleep apnea) were calculated using postbaseline time-point pairs for the overall population in each trial, by treatment, and by primary obstructive sleep apnea therapy adherence. In the 12-week narcolepsy trial, intraclass correlation coefficients (95% confidence intervals) were 0.83 (0.79, 0.87) for weeks 4 and 8 (n = 199), 0.87 (0.83, 0.90) for weeks 8 and 12 (n = 196), and 0.81 (0.76, 0.85) for weeks 4 and 12 (n = 196). In the 12-week obstructive sleep apnea trial, intraclass correlation coefficients (95% confidence intervals) were 0.74 (0.69, 0.78) (n = 416), 0.80 (0.76, 0.83) (n = 405), and 0.74 (0.69, 0.78) (n = 405), respectively. In the open-label extension trial, intraclass correlation coefficients (95% confidence intervals) were 0.82 (0.79, 0.85) for weeks 14 and 26/27 (n = 495), 0.85 (0.82, 0.87) for weeks 26/27 and 39/40 (n = 463), and 0.78 (0.74, 0.81) for weeks 14 and 39/40 (n = 463). Placebo/solriamfetol treatment or adherence to primary obstructive sleep apnea therapy did not affect reliability. In conclusion, across three large clinical trials of participants with narcolepsy or obstructive sleep apnea, Epworth Sleepiness Scale scores demonstrated a robust acceptable level of test-retest reliability in evaluating treatment response over time.

Persistently elevated levels of sST2 after acute coronary syndrome are associated with recurrent cardiac events.

PURPOSE: Higher soluble ST2 (sST2) levels at admission are associated with adverse outcome in acute coronary syndrome (ACS) patients. We studied the dynamics of sST2 over time in post-ACS patients prior to a recurrent ACS or cardiac death. METHODS: We used the BIOMArCS case cohort, consisting of 187 patients who underwent serial blood sampling during one-year follow-up post-ACS. sST2 was batch-wise quantified after completion of follow-up in a median of 8 (IQR: 5-11) samples per patient. Joint modelling was used to investigate the association between longitudinally measured sST2 and the endpoint, adjusted for gender, GRACE risk score and history of cardiovascular diseases. RESULTS: Median age was 64 years and 79% were men. The 36 endpoint patients had systematically higher sST2 levels than those that remained endpoint free (mean value 29.6 ng/ml versus 33.7 ng/ml, p-value 0.052). The adjusted hazard ratio for the endpoint per standard deviation increase of sST2 was 1.64 (95% confidence interval: 1.09-2.34; p = 0.019) at any time point. We could not identify a steady or sudden increase of sST2 in the run-up to the combined endpoint. CONCLUSION: Asymptomatic post-ACS patients with persistently higher sST2 levels are at higher risk of recurrent ACS or cardiac death during one-year follow-up.

Longitudinal changes in concentrations of persistent organic pollutants (1986-2016) and their associations with type 2 diabetes mellitus.

BACKGROUND: Positive associations have been reported between persistent organic pollutants (POPs) and type 2 diabetes mellitus (T2DM); however, causality has not been established. Over the last decades, environmental exposure to legacy POPs has decreased, complicating epidemiological studies. In addition, physiological risk factors for T2DM may also influence POP concentrations, contributing to a complex network of factors that could impact associations with T2DM. Longitudinal studies on this topic are lacking, and few have assessed prospective and cross-sectional associations between repeated POP measurements and T2DM in the same individuals, which may shed light on causality. OBJECTIVES: To compare longitudinal trends in concentrations of polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs) in T2DM cases and controls, and to examine prospective and cross-sectional associations between PCBs, OCPs and T2DM at different time-points before and after T2DM diagnosis in cases. METHODS: We conducted a longitudinal, nested case-control study (1986-2016) of 116 T2DM cases and 139 controls from the Tromsø Study. All participants had three blood samples collected before T2DM diagnosis in cases, and up to two samples thereafter. We used linear mixed-effect models to assess temporal changes of POPs within and between T2DM cases and controls, and logistic regression models to investigate the associations between different POPs and T2DM at different time-points. RESULTS: PCBs, trans-nonachlor, cis-nonachlor, oxychlordane, cis-heptachlor epoxide, p,p'-DDE, and p,p'-DDT declined more slowly in cases than controls, whereas ß-HCH and HCB declined similarly in both groups. Most POPs showed positive associations between both pre- and post-diagnostic concentrations and T2DM, though effect estimates were imprecise. These associations were most consistent for cis-heptachlor epoxide. DISCUSSION: The observed positive associations between certain POPs and T2DM may be because of higher POP concentrations within prospective T2DM cases, due to slower temporal declines as compared to controls.

Exploratory assessment of treatment-dependent random-effects distribution using gradient functions.

In analyzing repeated measurements from randomized controlled trials with mixed-effects models, it is important to carefully examine the conventional normality assumption regarding the random-effects distribution and its dependence on treatment allocation in order to avoid biased estimation and correctly interpret the estimated random-effects distribution. In this article, we propose the use of a gradient function method in modeling with the different random-effects distributions depending on the treatment allocation. This method can be effective for considering in advance whether a proper fit requires a model that allows dependence of the random-effects distribution on covariates, or for finding the subpopulations in the random effects.

Short-term risk assessments in an acute psychiatric inpatient setting: A re-examination of the Brøset Violence Checklist using repeated measurements - Differentiating violence characteristics and gender.

This study investigated the association between short-term risk assessment measured by the Brøset Violence Checklist (BVC) and imminent violence using repeated measurements and differentiating violence characteristics and gender. All patients admitted to an acute psychiatric ward during one year (N = 528) were included. Logistic regression and generalized linear mixed model (GLMM) analyses were conducted. Results confirmed BVC's suitability for both male and female inpatients throughout their hospitalization also when differentiating threats and physical violence, and adjusting for diagnostic subpopulations and circadian variability. Results point to modified interpretations of the BVC sum scores. Future research should adjust for repeated measurements.