Estimating Equation Approach Research Articles

A partially linear single index model for longitudinal data with terminal event

In longitudinal studies, we may be interested in some non-terminal event, such as health related quality of life. However, before the outcome of interest can be observed, a terminal event, such as death, can occur. If this situation occurs, we call this problem as truncation by death. In addition, observed times for subjects may be irregular, and the effect of covariates on the outcome of interest may be complicated. In this paper, we propose a new semiparametric partially linear single index model in the presence of truncation by death. A local linear approximation is used for the nonparametric link function, and estimating equation approaches are developed. The obtained estimator for the link function is proved to be uniformly consistent with convergence rate under root $n$. A data-splitting technique is developed to eliminate the side effect of this low convergence rate to establish the asymptotic normality of the parametric part.The finite sample behavior is examined by simulation studies, and an application to Alzheimers disease data is illustrated.

Fluorodeoxyglucose Detected Changes in Brain Metabolism After Chemotherapy in Pediatric Non-Hodgkin Lymphoma

BackgroundPotential neurocognitive dysfunction after chemotherapy is a worrisome long-term outcome. Our objective was to evaluate the effect on brain metabolism in pediatric patients with non–central nervous system cancer treated with chemotherapy by analyzing brain data from serial whole-body fluorodeoxyglucose positron emission tomography/computed-tomography (FDG-PET/CT) scans taken before and sequentially after therapy. MethodsFourteen pediatric patients diagnosed with lymphoma and treated with systemic and prophylactic intrathecal chemotherapy were included. All patients had baseline pretreatment whole-body FDG-PET/CT and at least one post-therapy study preformed as part of standard surveillance. Brain positron emission tomography data were quantitatively analyzed for normalized fluorodeoxyglucose uptake in various brain regions. A generalized estimating equation approach was used to evaluate temporal changes after chemotherapy. ResultsMedian time of follow-up surveillance positron emission tomography–computed-tomography was 456 days after chemotherapy course. Various brain regions demonstrated significant changes in fluorodeoxyglucose uptake as a function of time passed since chemotherapy. Increased fluorodeoxyglucose uptake was noted in the parietal and cingulate cortexes. Decreased fluorodeoxyglucose uptake was demonstrated in deep gray matter nuclei and in the brainstem. ConclusionsOur study provides novel insights into long-standing and progressive changes in regional glucose metabolism after chemotherapy in pediatric cancer population, lasting long after the end of therapy and reaching clinical remission. Expanding the utility of regular surveillance fluorodeoxyglucose positron emission tomography to a detailed quantitative assessment of regional brain metabolism after chemotherapy can provide valuable information on individual chemotherapy-related neuromodulation and facilitate the development of strategies to minimize neurocognitive side effects.

Healthcare utilization and cost trajectories post-stroke: role of caregiver and stroke factors

BackgroundIt is essential to study post-stroke healthcare utilization trajectories from a stroke patient caregiver dyadic perspective to improve healthcare delivery, practices and eventually improve long-term outcomes for stroke patients. However, literature addressing this area is currently limited. Addressing this gap, our study described the trajectory of healthcare service utilization by stroke patients and associated costs over 1-year post-stroke and examined the association with caregiver identity and clinical stroke factors.MethodsPatient and caregiver variables were obtained from a prospective cohort, while healthcare data was obtained from the national claims database. Generalized estimating equation approach was used to get the population average estimates of healthcare utilization and cost trend across 4 quarters post-stroke.ResultsFive hundred ninety-two stroke patient and caregiver dyads were available for current analysis. The highest utilization occurred in the first quarter post-stroke across all service types and decreased with time. The incidence rate ratio (IRR) of hospitalization decreased by 51, 40, 11 and 1% for patients having spouse, sibling, child and others as caregivers respectively when compared with not having a caregiver (p = 0.017). Disability level modified the specialist outpatient clinic usage trajectory with increasing difference between mildly and severely disabled sub-groups across quarters. Stroke type and severity modified the primary care cost trajectory with expected cost estimates differing across second to fourth quarters for moderately-severe ischemic (IRR: 1.67, 1.74, 1.64; p = 0.003), moderately-severe non-ischemic (IRR: 1.61, 3.15, 2.44; p = 0.001) and severe non-ischemic (IRR: 2.18, 4.92, 4.77; p = 0.032) subgroups respectively, compared to first quarter.ConclusionHighlighting the quarterly variations, we reported distinct utilization trajectories across subgroups based on clinical characteristics. Caregiver availability reducing hospitalization supports revisiting caregiver’s role as potential hidden workforce, incentivizing their efforts by designing socially inclusive bundled payment models for post-acute stroke care and adopting family-centered clinical care practices.

Association of Transcatheter Mitral Valve Repair With Quality of Life Outcomes at 30 Days and 1 Year

Improvements in symptoms, functional capacity, and quality of life are among the key goals of edge-to-edge transcatheter mitral valve repair (TMVR) for mitral regurgitation. To examine health status outcomes among patients undergoing TMVR in clinical practice and the factors associated with improvement. This cohort study used the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry, which contains data on patients with severe mitral regurgitation treated with TMVR from 2013 through 2017 in 217 US hospitals. Change in disease-specific health status (Kansas City Cardiomyopathy Questionnaire-Overall Summary score [KCCQ-OS]; range 0-100 points, with higher scores indicating better health status) at 30 days and 1 year after TMVR. We also examined factors associated with health status at 30 days after TMVR, by means of multivariable linear regression using a generalized estimating equations approach to account for clustering of patients within sites. The KCCQ data were available in 81.2% at baseline, 69.3% of survivors at 30 days, and 47.4% of survivors at 1 year. Among 4226 patients who underwent TMVR, survived 30 days, and completed the KCCQ at baseline and follow-up, the KCCQ-OS increased from 41.9 before TMVR to 66.7 at 30 days (mean change 24.8 [95% CI, 24.0-25.6] points; P < .001), representing a large clinical improvement. The KCCQ scores remained stable from 30 days to 1 year after TMVR, with no further significant increase or decline. On multivariable analysis, atrial fibrillation (-2.2 [95% CI, -3.7 to -0.6] points; P = .01), permanent pacemaker (-2.1 [95% CI, -3.7 to -0.4] points; P = .01), severe lung disease (-3.9 [95% CI, -6.2 to -1.5] points; P = .001), home oxygen (-2.7 [95% CI, -4.9 to -0.4] points; P = .02), and lower KCCQ scores at baseline (3.9 points for each 10-point increase [95% CI, 3.6-4.2]; P < .001) were independently associated with lower 30-day KCCQ-OS scores. In-hospital renal failure was uncommon but was also associated with significant reductions in 30-day KCCQ-OS scores (-7.3 [95% CI -13.3 to -1.2] points). In estimates calculated with inverse probability weighting, after 1 year after TMVR, 54.2% (95% CI 52.2%-56.1%) of patients were alive and well; 23.0% had died, 21.9% had persistently poor health status (KCCQ-OS <60 points), 5.5% had a health status decline from baseline, and 4.6% had both poor health status and health status decline. In a national cohort of US patients undergoing edge-to-edge TMVR in clinical practice, health status was impaired prior to the procedure, improved within 30 days, and remained stable through 1 year among surviving patients with available data. While long-term mortality remains high, most surviving patients demonstrate improvements in symptoms, functional status, and quality of life, with only modest differences by patient-level factors.

Intraplaque Neovascularization Attenuated Statin Benefit on Atherosclerotic Plaque in CAD Patients: A Follow-Up Study with Combined Imaging Modalities

Background: Plaque progression increases the risk of a cardiovascular event. This study aims to determine whether intraplaque neovascularization (NV) attenuates the response of plaque progression to statin therapy. Methods: Baseline and 12-month follow-up IVUS was used in combination with baseline OCT to assess 164 non-culprit plaques in 118 patients with CAD. A generalized estimating equation approach with exchangeable correlation structure was used to correct for the dependency of repeated measurements. Findings: Patients were divided into two groups according to NV (52 patients with 62 NV plaques, 66 patients with 102 non-NV plaques). Nonculprit plaques in the NV group exhibited a more frequent occurrence of TCFA (P = 0.004), macrophage (P = 0.005), cholesterol crystal (P = 0.012), calcification (P = 0.030), thinner fibrous cap thickness (FCT) [(86.8 ± 55.1) vs. (127.4 ± 70.1) μm, P = 0.015], larger lipid arc [(219.5 ± 66.9) vs. (179.8 ± 61.4), P = 0.002] compared to non-NV group. A large change in percent atheroma volume (PAV), plaque plus media crosssectional area (P&M CSA), plaque volume, and plaque burden was observed from baseline to follow-up in the NV group. Changes in P&M CSA, plaque volume, and plaque burden showed significant differences in fibroatheroma with NV. Intraplaque NV could predict a high risk of plaque progression despite statin therapy [OR 6.454 (95% CI 2.412-17.267), P < 0.001]. Interpretation: NV might attenuate the benefits of statin therapy in plaque progression. This study may provide a new basis for anti-angiogenic strategies to prevent atherosclerotic plaque progression. Funding Statement: This study was supported by grants from the National Natural Science Foundation of China (to B.Y., Grant No. 81330033; to J.W.T., Grant No. 81571749 and to X.X.L., Grant No. 81703502), China Postdoctoral Science Foundation (2018M630376) and Heilongjiang Postdoctoral Science Foundation (LBH-Z17120). Declaration of Interest: The authors have no financial conflicts of interest. Ethics Approval Statement: The study protocol was approved by the ethics committee of the Second Affiliated Hospital of Harbin Medical University (Harbin, China). All patients provided written informed consent.

Demographic senescence in the aquatic plant Lemna gibba L. (Araceae)

Senescence is progressive, age-related bodily deterioration, accompanied at the population level by declines in average survival and fecundity (i.e., ‘demographic senescence’). Demographic senescence of plants has been investigated in only a few species, including small, floating macrophytes in the genus Lemna (family Araceae, subfamily Lemnoideae – the ‘duckweeds’). Unlike most plant species, Lemna ramets exhibit determinate growth, potentially rendering them more likely to experience demographic senescence. Here, our objective was to investigate senescence in a Lemna species not previously studied in this context, L. gibba L., toward the long-term goal of conducting cross-species comparative analyses. In a longitudinal lab study, we investigated a cohort of 334 individual L. gibba fronds, whose survival and reproduction we followed daily from birth (defined by the date a focal frond detached from its parent) to death (defined by the date a focal frond’s last daughter detached). We fit survival data to exponential, Weibull, Gompertz, and logistic models, the first of which represents ‘no senescence’. The logistic model was found to have the greatest support (AICC weight >0.99), indicating strong age-related declines in survival. We fit reproduction data using a generalized estimating equation approach, which showed a significant age-related decline in the predicted probability of daily reproduction – from 0.61 at age 3 days to 0.23 at age 52 days (i.e., after excluding the first two days of reproduction data to account for the initial, pre-reproductive phase of the L. gibba lifecycle). These age-related declines provide strong evidence that L. gibba does exhibit demographic senescence, consistent with evidence from congeneric species.

Are Complications After the Bernese Periacetabular Osteotomy Associated With Subsequent Outcomes Scores?

The Bernese periacetabular osteotomy (PAO) continues to be a commonly performed nonarthroplasty option to treat acetabular dysplasia, but only a few short-term studies have evaluated complications rigorously after PAO. (1) What complications are observed at 10-year mean followup of the Bernese PAO in patients with symptomatic acetabular dysplasia? (2) What factors are associated with these complications? (3) Do these complications affect clinical outcome scores? We reviewed 238 hips in 206 patients treated with PAO from July 1994 to August 2008. Only PAOs performed for symptomatic acetabular dysplasia and those that had at a minimum 4-year followup were included. Patients who went on to THA before 4 years were included in the study. Patients with hip pain who presented with a clinical presentation of symptomatic acetabular dysplasia, radiographic evidence of femoral head uncovering, and a lateral center-edge angle < 25° were considered for PAO and no other juxtaacetabular osteotomy was offered other than PAO. Sixty-two hips had diagnoses other than acetabular dysplasia and 22 were lost to followup. The remaining 154 hips (129 patients) were evaluated by chart review at a mean of 10 years (range, 1.7-20.5 years) using the UCLA Activity Score, modified Harris hip score (mHHS), WOMAC, and radiographic analysis. The mean age at PAO was 26 years (range, 10-60 years) and consisted of 113 female patients (132 hips [86%]) and 16 male patients (22 hips [14%]). Complications were graded using the validated Clavien-Dindo system. Complications were assessed for each hip and the highest complication grade was assigned to the hip if multiple complications occurred. We divided complication grades into three groups for analysis: no complications, Grade 1 complications, and complications that deviated from the standard postoperative course (Grades 2, 3, and 4). There were no Grade 5 complications. Variables with significant (p < 0.05) univariable associations with complications were considered for inclusion in a multivariable model. Outcome variables (mHHS and WOMAC) at the most recent followup visit were analyzed using a generalized estimating equation approach. Analysis of variance was used to compare UCLA at the most recent followup among the complication classes. Major complications defined as Clavien-Dindo Grade 3/4 occurred in 14 hips (9%). After controlling for potential confounding variables, we found that increasing body mass index (BMI) (odds ratio [OR], 1.16; 95% confidence interval, 1.05-1.25; p = 0.004) was associated with increased risk of complication. In contrast, greater surgeon experience was associated with a decreased risk (OR, 0.3; p = 0.002). Complications were associated with postoperative pain and activity, WOMAC (mean ± SD: 0 complications = 1.5 ± 15.1, 1 complication = 4.3 ± 4.1, 2-3 complications = 3.8 ± 4.6; p = 0.020) and UCLA scores (mean ± SD: 0 complications = 7.8 ± 2, 1 complication = 6.7 ± 2.1, 2-3 complications = 6.5 ± 2; p = 0.003). Most hips undergoing PAO have few complications. The most common major surgical complication is nonunion. Increasing BMI was a predictor of having a complication, and surgeon experience decreased complication risk. Having a complication adversely affected long-term pain and activity. To minimize complications and maximize outcomes, a patient's BMI should be assessed preoperatively and those with excessive BMI should be counseled on the increased risk of complications. In an experienced surgeon's hands, PAO has few complications at mean 10-year followup and a low risk of permanent disability. Level III, therapeutic study.

The Cox–Aalen model for recurrent‐event data with a dependent terminal event

Recurrent events with a dependent terminal event arise frequently in a wide variety of fields. In this paper, we propose a new joint model to analyze these data and model the dependence between recurrent and terminal events through shared gamma frailty. Specifically, a Cox–Aalen rate frailty model is specified for the recurrent event, and an additive hazards frailty model is specified for the terminal event. An estimating equation approach is developed for the parameters in the joint model, and the asymptotic properties of the proposed estimators are established. Simulation studies demonstrate that the proposed estimators perform well with finite samples. An application to a medical cost study of chronic heart failure patients is illustrated.

Hospitalization for mental health related ambulatory care sensitive conditions: what are the trends for First Nations in British Columbia?

BackgroundIndigenous peoples globally experience a disproportionate burden of mental illness due to forced policies and practices of colonization and cultural disruption. The objective of this study was to provide a baseline profile of hospitalization rates for mental health-related Ambulatory Care Sensitive Conditions among First-Nations living both on and off reserve in British Columbia, Canada, and explore the relationship between local access to health services and mental health-related hospitalization rates.MethodsA population-based time trend analysis of mental health-related Ambulatory Care Sensitive Conditions hospitalizations was conducted using de-identified administrative health data. The study population included all residents eligible under the universal British Columbia Medical Services Plan and living on and off First Nations reserves between 1994/95 and 2009/10. The definition of mental health-related Ambulatory Care Sensitive Conditions included mood disorders and schizophrenia, and three different change measures were used to operationalize avoidable hospitalizations: 1) rates of episodes of hospital care, 2) rates of length of stay, and 3) readmission rates. Data were analyzed using generalized estimating equations approach, controlling for age, sex, and socio-economic status, to account for change over time.ResultsOur findings show that First Nations living on reserve have higher hospitalization rates for mental disorders compared to other British Columbia residents up until 2008. Those living off reserve had significantly higher hospitalization rates throughout the study period. On-reserve communities served by nursing stations had the lowest rates of hospitalization whereas communities with limited local services had the highest rates. Compared to other British Columbia residents, all First Nations have a shorter length of stay and lower readmission rates.ConclusionsThis study suggests that despite reduced rates of hospitalization for mental-health related Ambulatory Care Sensitive Conditions over time for First Nations, gaps in mental health care still exist. We argue greater investments in primary mental health care are needed to support First Nations health. However, these efforts should place equal importance on prevention and the social determinants of health.

Factors Related to the Acute Cellular Rejection During the First Year After Heart Transplant

IntroductionThe acute cellular rejection is recognized as a factor related to the long-term viability of the heart graft. We intend to establish which factors are associated with the acute cellular rejection during the first year post heart transplant using a longitudinal model with repeated measures. MethodsA retrospective cohort study was performed with all the patients who underwent heart transplant between 2005–2018 at the Hospital Universitario San Ignacio in Bogota, Colombia. In order to determine the factors associated with the development of acute cellular rejection, a generalized estimating equation approach was used, with an interchangeable correlation structure. The lowest value of quasi-likelihood information criterion and P < .05 was considered significant. ResultsFifty-five patients (49.3 ± 11.1 years old) were included. The mortality during the first month was 16.3% and the accumulated mortality during the first year was 23.6%. The incidence of the acute cellular rejection was higher during the third month after the transplant (79.9%); most of them were acute cellular rejection grade 1. The factors associated with the development of the rejection were the cyclosporine levels out of the therapeutic range in several periods of evaluation (P < .03) and the age of the receptor (P = .049). ConclusionsUsing advanced modeling methodologies of longitudinal data we identified that the factors associated with acute cellular rejection during the first year after the transplant are related to the therapeutic levels of the calcineurin inhibitor (cyclosporin) during the first 6 months of follow-up and the age of the receptor.

General retrospective mega‐analysis framework for rare variant association tests

Here, we describe a retrospective mega-analysis framework for gene- or region-based multimarker rare variant association tests. Our proposed mega-analysis association tests allow investigators to combine longitudinal and cross-sectional family- and/or population-based studies. This framework can be applied to a continuous, categorical, or survival trait. In addition to autosomal variants, the tests can be applied to conduct mega-analyses on X-chromosome variants. Tests were built on study-specific region- or gene-level quasiscore statistics and, therefore, do not require estimates of effects of individual rare variants. We used the generalized estimating equation approach to account for complex multiple correlation structures between family members, repeated measurements, and genetic markers. While accounting for multilevel correlations and heterogeneity across studies, the test statistics were computationally efficient and feasible for large-scale sequencing studies. The retrospective aspect of association tests helps alleviate bias due to phenotype-related sampling and type I errors due to misspecification of phenotypic distribution. We evaluated our developed mega-analysis methods through comprehensive simulations with varying sample sizes, covariates, population stratification structures, and study designs across multiple studies. To illustrate application of the proposed framework, we conducted a mega-association analysis combining a longitudinal family study and a cross-sectional case-control study from Genetic Analysis Workshop 19.

A Joint Modeling Approach for Longitudinal Data with Informative Observation Times and a Terminal Event

In this article, we propose a new joint modeling approach for the analysis of longitudinal data with informative observation times and a dependent terminal event. We specify a semiparametric mixed effects model for the longitudinal process, a proportional rate frailty model for the observation process, and a proportional hazards frailty model for the terminal event. The association among the three related processes is modeled via two latent variables. Estimating equation approaches are developed for parameter estimation, and the asymptotic properties of the proposed estimators are established. The finite sample performance of the proposed estimators is examined through simulation studies, and an application to a medical cost study of chronic heart failure patients is illustrated.

Poor birth outcomes among Aboriginal Western Australians and smoking, alcohol and substance misuse, and assault

IntroductionOn average, Aboriginal neonates in Western Australia (WA) weigh 200g less than non-Aboriginal infants and are 2-3 times more likely to be preterm, stillborn, or die neonatally. They are also more likely to be exposed in utero to maternal behaviour risks like smoking, due to factors such as intergenerational trauma.&#x0D; Objectives and ApproachWe aimed to estimate the proportion of small for gestational age (SGA) births, preterm births, and perinatal deaths of Western Australian Aboriginal infants from 1998-2010 attributable to maternal smoking, alcohol misuse, drug misuse, and assault against the mother. We used linked birth, hospital, mental health, and death records of all Aboriginal singletons and their parents. Using logistic regression with a generalized estimating equation approach, associations between birth outcomes and the four risk factors of interest were estimated after adjusting for maternal age, height and health. Using coefficients from these models, we estimated adjusted population attributable fractions (PAFs).&#x0D; ResultsOf 28,119 births, 16% of infants were SGA, 13% were preterm and 2% died perinatally. 51% of infants were exposed to maternal smoking, alcohol misuse, drug misuse, and/or assault, and 37% [95% CI: 35%, 40%] of SGA births, 16% [95% CI: 14%, 19%] of preterm births and 20% [95% CI: 12%, 28%] of perinatal deaths were attributable to these factors, predominantly smoking. The PAFs for alcohol misuse (for example, for SGA, 3% [95% CI: 2%, 3%]) are likely to be underestimates as it is difficult to identify alcohol misuse using administrative data.&#x0D; Conclusion/ImplicationsWhile smoking rates have dropped considerably, reduction measures have been less successful among Aboriginal women than non-Aboriginal women. Significant improvements in perinatal health are possible with identification and support of effective risk reduction approaches for Aboriginal women, as well as their communities and families.

Using linked data and family studies to understand multigenerational causes of low birthweight among Australian Aboriginal infants

IntroductionLow birthweight is common among Aboriginal infants. Birthweight is correlated across generations due to the transmission of genetic and environmental factors. Another cause may be fetal programming, where a fetus’ response to a hostile environment leads to poor adult health and, in turn, a poor uterine environment for her offspring.&#x0D; Objectives and ApproachIdentifying a causal relationship between maternal birthweight and offspring birthweight is difficult. However, we can gain insights by approaching the question from several different angles, including family studies involving both parents and cousins who share maternal grandparents. Family studies using linked data are possible with data from Western Australia (WA), the only Australian state with a database of family relationships.&#x0D; We used linked birth, hospital, mental health, and family relationship records of 12,865 Aboriginal singletons born 1998 to 2011 in WA whose mother also linked to a WA birth record from 1980 onwards, and their parents’ records.&#x0D; Results17% of births were small for gestational age. Using a linear regression model with a generalised estimating equation approach for offspring birthweight z-score (BWZ), the coefficient for maternal BWZ was 0.17 (95% CI: 0.14, 0.20), compared to 0.13 (95% CI: 0.10, 0.16) for paternal BWZ. The difference in coefficients (0.03 [95% CI: -0.01, 0.08]) provides only limited support for the fetal programming hypothesis. Other associations with offspring BWZ were much larger, including maternal smoking (-0.39 [95% CI: -0.45, -0.34]). After restricting the sample to cousins with shared maternal grandparents (fixed-effects model), the mother-offspring association was fully attenuated (-0.01 [95% CI: -0.07, 0.05]), suggesting transmission of maternal genetic and environmental factors alone can explain the association, though the 95% confidence interval was wide.&#x0D; Conclusion/ImplicationsIf fetal programming is an important cause of low birthweight, Aboriginal people would be disproportionately affected, following generations of low birthweight and chronic disease. However, the family studies indicate fetal programming has a limited role compared to other risk factors in the current pregnancy.

Assessment of fenofibrate-methylation interactions on triglycerides using longitudinal family data

BackgroundTriglyceride (TG) concentrations decrease in response to fenofibrate treatment, and also are associated with DNA methylation. But how interactions between fenofibrate response and DNA methylation affect TGs remains unclear.MethodsIn the present study, we identified and compared differential methylation sites associated with TG concentrations in individuals before and after fenofibrate treatment. We then estimated interactions between fenofibrate treatment and methylation to identify differential methylation effects associated with fenofibrate treatment on TG concentrations using the entire longitudinal family sample. To account for within-family and within-individual corrections, the generalized estimating equations approach was used to estimate main and interaction effects between methylation sites and fenofibrate treatment, adjusting for potential confounders. Analyses were also performed with and without adjusting for high-density lipoprotein (HDL) concentrations.ResultsPrior to fenofibrate treatment, 23 cytosine-phosphate-guanine (CpG) sites were significantly associated with TG concentrations, while only 13 CpG sites were identified posttreatment, adjusting for HDL. Without adjusting for HDL, pretreatment, 20 CpG sites were significantly associated with TG concentrations, while only 12 CpG sites were identified posttreatment. Among these sites, only one differential site (cg19003390 in the CPT1A gene) overlapped from pre- and posttreatment measurements regardless of HDL adjustment. Furthermore, 11 methylation sites showed substantial interaction effects (p < 1.43 × 10−7with Bonferroni correction) with or without HDL adjustment when using the whole longitudinal data.ConclusionsOur analyses suggest that DNA methylation likely modified the effect of fenofibrate on TG concentrations. Differential fenofibrate-associated methylation sites on TGs differed with and without adjusting for HDL concentrations, suggesting that these HDLs and TGs might share some common epigenetic processes.

Simulated versus traditional occupational therapy placements: A randomised controlled trial.

Professional practise placements in occupational therapy education are critical to ensuring graduate competence. Australian occupational therapy accreditation standards allow up to 200 of a mandated 1000 placement hours to include simulation-based learning. There is, however, minimal evidence about the effectiveness of simulation-based placements compared to traditional placements in occupational therapy. We evaluated whether occupational therapy students completing a 40hour (one week block) Simulated Clinical Placement (SCP) attained non-inferior learning outcomes to students attending a 40 hour Traditional Clinical Placement (TCP). A pragmatic, non-inferiority, assessor-blinded, multicentre, randomised controlled trial involving students from six Australian universities was conducted. Statistical power analysis estimated a required sample of 425. Concealed random allocation was undertaken with a 1:1 ratio within each university. Students were assigned to SCP or TCP in one of three settings: vocational rehabilitation, mental health or physical rehabilitation. SCP materials were developed, manualised and staff training provided. TCPs were in equivalent practice areas. Outcomes were assessed using a standardised examination, unit grades, the Student Practice Evaluation Form-Revised and student confidence survey. A generalised estimating equationapproach was used to assess non-inferiority of the SCP to the TCP. Of 570 randomised students (84% female), 275 attended the SCP and 265 the TCP (n=540, 94.7% retention). There were no significant differences between the TCP and SCP on (i) examination results (marginal mean difference 1.85, 95% CI: 0.46-3.24; P=0.087); (ii) unit score (mean (SD) SCP: 71.9 (8.8), TCP: 70.34 (9.1); P=0.066); or (iii) placement fail rate, assessed using the Student Practice Evaluation Form-Revised (100% passed both groups). Students can achieve equivalent learning outcomes in a 40 hour simulated placement to those achieved in a 40 hour traditional placement. These findings provide assurance to students, educators and professional accreditation bodies that simulation can be embedded in occupational therapy education with good effect.

Exploring Different Strategies of Assessing the Economic Impact of Multiple Diabetes-Associated Complications and Their Interactions: A Large Claims-Based Study in Germany.

In the context of an aging population with increasing diabetes prevalence, people are living longer with diabetes, which leads to increased multimorbidity and economic burden. The primary aim was to explore different strategies that address the economic impact of multiple type 2 diabetes-related complications and their interactions. We used a generalized estimating equations approach based on nationwide statutory health insurance data from 316,220 patients with type 2 diabetes (baseline year 2012, 3 years of follow-up). We estimated annual total costs (in 2015euros) for type 2 diabetes-related complications and, in addition, explored different strategies to assess diabetes-related multimorbidity: number of prevalent complications, co-occurrence of micro- and macrovascular complications, disease-disease interactions of prevalent complications, and interactions between prevalent/incident complications. The increased number of complications was significantly associated with higher total costs. Further assessment of interactions showed that macrovascular complications (e.g., chronic heart failure) and high-cost complications (e.g., end-stage renal disease, amputation) led to significant positive effects ofinteractions on costs, whereas early microvascular complications (e.g., retinopathy) caused negative interactions. The chronology of the onset of these complications turned out to have an additional impact on the interactions and their effect on total costs. Health economic diabetes models and evaluations of interventions in patients with diabetes-related complications should pay more attention to the economic effect of specific disease interactions. Politically, our findings support the development of more integrated diabetes care programs that take better account of multimorbidity. Further observational studies are needed to elucidate the shared pathogenic mechanisms of diabetes complications.

Quebec physicians' perspectives on medical aid in dying for incompetent patients with dementia.

To elicit Quebec physicians' attitudes towards extending medical aid in dying (MAiD) to incompetent patients and to compare the attitudes of family physicians to those of other medical specialists. We conducted a postal survey among physicians caring for patients with dementia. We used hypothetical vignettes to elicit their attitudes towards MAiD and continuous deep sedation (CDS) to relieve suffering at end of life. Two patients were depicted in the vignettes: one with cancer eligible for MAiD and one with dementia. The generalized estimating equation approach was used to investigate factors associated with attitudes, including the stage of the illness (advanced vs terminal dementia) and the presence or absence of a prior written request. A total of 136 physicians out of 653 returned the questionnaire. Physicians favoured CDS over MAiD for relieving suffering in the cancer vignette (93% vs 79%; p = 0.002). In advanced dementia, 45% of physicians supported giving the patient access to MAiD with a written request and 14% without such request. At the terminal stage of dementia, these proportions increased to 71% and 43%, respectively (p < 0.001), reaching 79% and 52% among family physicians. Support for CDS in terminal dementia was lower than in end-stage cancer (68% vs 93%; p < 0.001) and equal to MAiD with a written request (68% vs 71%; p = 0.623). Many Quebec physicians support extending MAiD to incompetent patients with dementia to relieve suffering at the terminal stage. This finding will inform current deliberations as to whether MAiD should be extended to these patients.

An evaluation of school-based e-cigarette control policies' impact on the use of vaping products.

INTRODUCTIONElectronic cigarette (e-cigarette) use among youth is common, and so efforts to regulate its use and availability are continually being made. The school environment represents an important domain for advancing health policy among youth populations. This study examines the impact of school-based e-cigarette control policies on student e-cigarette use in the context of a natural experiment.METHODSUsing three years of longitudinal student and school level data (2013/2014 to 2015/2016), from a sample of 69 secondary schools in Ontario, Canada, a generalized estimating equation approach examined the impact of school-based e-cigarette control policy changes on the prevalence of youth e-cigarette use. The main outcome of interest was current e-cigarette use, while covariates included age, gender, ethnicity, and amount of spending money in dollars per week the student has. Tests of proportion (t-tests) were used to examine whether there were any significant differences in the changes for each intervention school relative to the sample of schools that report no changes in school-level e-cigarette control policies.RESULTSEstimates from the generalized estimating equation approach suggest that students had lower odds of using e-cigarettes in schools where an e-cigarette control policy was implemented. That is, the e-cigarette control policy decreased the adjusted odds of being an e-cigarette user (OR=0.68; 95% CI: 0.48–0.97). Examining school-specific impact, at four of six schools that had an e-cigarette control policy, the ban on the use of e-cigarettes may have lowered the prevalence of e-cigarette use.CONCLUSIONSThis is the first study to use longitudinal data to study school-level e-cigarette use and the impact of e-cigarette control policy. These results provide new evidence that school-level policies banning the use of e-cigarettes on school property may be effective in reducing e-cigarette use (or preventing it) in their current form, as seen in this natural experiment.

Covariate-Adjusted Regression for Distorted Longitudinal Data With Informative Observation Times

ABSTRACTIn many longitudinal studies, repeated response and predictors are not directly observed, but can be treated as distorted by unknown functions of a common confounding covariate. Moreover, longitudinal data involve an observation process which may be informative with a longitudinal response process in practice. To deal with such complex data, we propose a class of flexible semiparametric covariate-adjusted joint models. The new models not only allow for the longitudinal response to be correlated with observation times through latent variables and completely unspecified link functions, but they also characterize distorted longitudinal response and predictors by unknown multiplicative factors depending on time and a confounding covariate. For estimation of regression parameters in the proposed models, we develop a novel covariate-adjusted estimating equation approach which does not rely on forms of link functions and distributions of frailties. The asymptotic properties of resulting parameter estimators are established and examined by simulation studies. A longitudinal data example containing calcium absorption and intake measurements is provided for illustration. Supplementary materials for this article are available online.