Dimitris Rizopoulos | Erasmus University Rotterdam (original) (raw)

Books by Dimitris Rizopoulos

The primary goal of this monograph is to introduce this joint modeling framework. In particular, we will focus on the type of research questions joint models attempt to answer and the circumstances under which these models are appropriate to answer those questions, we will explain which are the key assumptions behind them, and how they can be optimally utilized to extract relevant information from the data. To facilitate exposition of all the theoretical material covered in this book, all illustrations put forward are available within the R software environment for statistical computing and graphics, using package JM.

Talks by Dimitris Rizopoulos

◃ Longitudinal Part: Reconstruct M i (t) = {m i (s), 0 ≤ s < t} using y i (t) and a mixed effects... more ◃ Longitudinal Part: Reconstruct M i (t) = {m i (s), 0 ≤ s < t} using y i (t) and a mixed effects model (we focus on continuous markers)

We define a standard joint model

Teaching Documents by Dimitris Rizopoulos

Papers by Dimitris Rizopoulos

Journal of Human Hypertension, Jan 13, 2011

It has been demonstrated that aortic stiffness is an independent predictor of cardiovascular dise... more It has been demonstrated that aortic stiffness is an independent predictor of cardiovascular disease. We investigated whether this measure is of use in cardiovascular risk stratification in clinical practice for elderly subjects (mean age 71.5 years). Within the framework of the Rotterdam Study, we stratified subjects free of coronary heart disease (CHD) at baseline into categories of low (< 10%), intermediate (10–20%) and high (> 20%) 10-year risk of CHD based on Framingham risk factors. Within each risk category, we ...

Until recently, latent variable models such as the factor analysis model for metric responses, th... more Until recently, latent variable models such as the factor analysis model for metric responses, the two-parameter logistic model for binary responses, the multinomial model for nominal responses considered only main effects of latent variables without allowing for interaction or ...

European heart journal. Acute cardiovascular care, Jan 19, 2016

Baseline values of N-terminal pro B-type natriuretic peptide (NT-proBNP) and carbohydrate antigen... more Baseline values of N-terminal pro B-type natriuretic peptide (NT-proBNP) and carbohydrate antigen 125 (CA125) predict all-cause mortality in acute heart failure (AHF). However, there is limited information about the added prognostic benefit of using longitudinal values, and how this predictive ability is modified when modelling together. The aim of this study was to determine the mutually-adjusted association between the longitudinal trajectories of NT-proBNP and CA125 with all-cause mortality after an episode of AHF. We included 946 consecutive patients discharged for AHF. NT-proBNP and CA125 were measured at each physician-patient encounter (median (interquartile range (IQR)):3 (2-4)). The effect on mortality (time-dependent modelling) was assessed using joint modelling (JM) and multi-state Markov. The mean age was 71±11 years and 51% exhibited left ventricular systolic dysfunction. At a median follow-up of 2.64 years (IQR=1.20-5.36), 498 patients died (52.6%). The observed trajec...

Prostaglandins Leukotrienes and Essential Fatty Acids, 2014

Prenatal polyunsaturated fatty acid (PUFA) concentrations may be involved in the prenatal program... more Prenatal polyunsaturated fatty acid (PUFA) concentrations may be involved in the prenatal programming of adiposity. In this study we therefore explored the association between maternal PUFA concentrations, measured up to four times during pregnancy, and offspring adiposity at age 7 in 234 mother-child pairs of the Maastricht Essential Fatty Acid Birth cohort. Only dihomo-gamma-linolenic acid (DGLA, an n-6 fatty acid) concentration was associated with adiposity: per standard deviation increase in relative DGLA concentration, BMI increased by 0.44kg/m(2) (CI95: 0.16, 0.72), sum of skinfolds increased by 3.41mm (CI95: 1.88, 4.95), waist circumference increased by 1.09cm (CI95: 0.40, 1.78), and plasma leptin concentration increased by 0.66µg/l (CI95: 0.20, 1.11). In conclusion, maternal DGLA throughout gestation was associated with increased BMI and some additional measures of adiposity at age 7. This suggests that maternal DGLA might play a role in or reflect the prenatal programming of adiposity. Copyright © 2014 Elsevier Ltd. All rights reserved.

Journal of the American Statistical Association, Dec 22, 2014

The joint modeling of longitudinal and time-to-event data is an active area of statistics researc... more The joint modeling of longitudinal and time-to-event data is an active area of statistics research that has received a lot of attention in the recent years. More recently, a new and attractive application of this type of models has been to obtain individualized predictions of survival probabilities and/or of future longitudinal responses. The advantageous feature of these predictions is that they are dynamically updated as extra longitudinal responses are collected for the subjects of interest, providing real time risk assessment using all recorded information. The aim of this paper is two-fold. First, to highlight the importance of modeling the association structure between the longitudinal and event time responses that can greatly influence the derived predictions, and second, to illustrate how we can improve the accuracy of the derived predictions by suitably combining joint models with different association structures. The second goal is achieved using Bayesian model averaging, which, in this setting, has the very intriguing feature that the model weights are not fixed but they are rather subject-and time-dependent, implying that at different follow-up times predictions for the same subject may be based on different models.

Circulation Journal, Jan 30, 2014

JAMA, Jan 10, 2015

Supervised exercise is recommended as a first-line treatment for intermittent claudication. Combi... more Supervised exercise is recommended as a first-line treatment for intermittent claudication. Combination therapy of endovascular revascularization plus supervised exercise may be more promising but few data comparing the 2 therapies are available. To assess the effectiveness of endovascular revascularization plus supervised exercise for intermittent claudication compared with supervised exercise only. Randomized clinical trial of 212 patients allocated to either endovascular revascularization plus supervised exercise or supervised exercise only. Data were collected between May 17, 2010, and February 16, 2013, in the Netherlands at 10 sites. Patients were followed up for 12 months and the data were analyzed according to the intention-to-treat principle. A combination of endovascular revascularization (selective stenting) plus supervised exercise (n = 106) or supervised exercise only (n = 106). The primary end point was the difference in maximum treadmill walking distance at 12 months ...

We define a standard joint model

Journal of Human Hypertension, Jan 13, 2011

European heart journal. Acute cardiovascular care, Jan 19, 2016

Prostaglandins Leukotrienes and Essential Fatty Acids, 2014

Journal of the American Statistical Association, Dec 22, 2014

Circulation Journal, Jan 30, 2014

JAMA, Jan 10, 2015

Leukemia, 2015

Deletions in IKZF1 are found in ~15% of children with B-cell precursor acute lymphoblastic leukem... more Deletions in IKZF1 are found in ~15% of children with B-cell precursor acute lymphoblastic leukemia (BCP-ALL). There is strong evidence for the poor prognosis of IKZF1 deletions affecting exons 4-7 and exons 1-8, but evidence for the remaining 33% of cases harboring other variants of IKZF1 deletions is lacking. In an international multicenter study we analyzed the prognostic value of these rare variants in a case-control design. Each IKZF1-deleted case was matched to three IKZF1 wild-type controls based on cytogenetic subtype, treatment protocol, risk stratification arm, white blood cell count and age. Hazard ratios for the prognostic impact of rare IKZF1 deletions on event-free survival were calculated by matched pair Cox regression. Matched pair analysis for all 134 cases with rare IKZF1 deletions together revealed a poor prognosis (P&amp;lt;0.001) that was evident in each risk stratification arm. Rare variant types with the most unfavorable event-free survival were DEL 2-7 (P=0.03), DEL 2-8 (P=0.002) and DEL-Other (P&amp;lt;0.001). The prognosis of each type of rare variant was equal or worse compared with the well-known major DEL 4-7 and DEL 1-8 IKZF1 deletion variants. We therefore conclude that all variants of rare IKZF1 deletions are associated with an unfavorable prognosis in pediatric BCP-ALL.Leukemia advance online publication, 11 August 2015; doi:10.1038/leu.2015.199.

Journal of Statistical Software, 2006

The R package ltm has been developed for the analysis of multivariate dichotomous and polytomous ... more The R package ltm has been developed for the analysis of multivariate dichotomous and polytomous data using latent variable models, under the Item Response Theory approach. For dichotomous data the Rasch, the Two-Parameter Logistic, and Birnbaum's Three-Parameter models have been implemented, whereas for polytomous data Semejima's Graded Response model is available. Parameter estimates are obtained under marginal maximum likelihood using the Gauss-Hermite quadrature rule. The capabilities and features of the package are illustrated using two real data examples.

Pattern mixture models are frequently used to analyze longitudinal data where missingness is indu... more Pattern mixture models are frequently used to analyze longitudinal data where missingness is induced by dropout. When interest lies in the marginal model, inferences are mainly based on assumptions, in which case the two most common model fitting techniques are, the imposition of identifying restrictions and the extrapolation method. Here we consider the extrapolation method under a Bayesian model formulation and propose a suitable prior distribution for the model parameters, which provides a flexible tool for sensitivity analysis. In particular, a parameter that controls the difference between average trends among neighbouring patterns is introduced and sensitivity is assessed by comparing results obtained from analyses with different values for this parameter.

$Research paper thumbnail of Supplementary Material for \Non-Ignorable Models for Intermittently Missing Categorical Longitudinal Responses$

According to Section 3 of the paper, obtaining a uniqueĜ requires fixing its support points µ c .... more According to Section 3 of the paper, obtaining a uniqueĜ requires fixing its support points µ c . SinceĜ has a finite number of support points, a finite grid with points µ 1 , . . . , µ C is specified and thus the choice of an appropriate range and choice of the number of points is required. Regarding the range, a grid is chosen for the scaled random-effect b * i = S −1 b b i instead of b i to avoid specification of an unnecessarily huge grid. The matrix S b represents an approximate estimate of the Choleski decomposition of the random-effects covariance matrix and is chosen equal to the Choleski decompo-

Latent variable models represent a useful tool for the analysis of complex data when the construc... more Latent variable models represent a useful tool for the analysis of complex data when the constructs of interest are not observable. A problem related to these models is that the integrals involved in the likelihood function cannot be solved analytically. We propose a computational approach, referred to as Dimension Reduction Method (DRM), that consists of a dimension reduction of the multidimensional integral that makes the computation feasible in situations in which the quadrature based methods are not applicable. We discuss the advantages of DRM compared with other existing approximation procedures in terms of both computational feasibility of the method and asymptotic properties of the resulting estimators.

Statistics in medicine, Jan 30, 2015

Statistical methods in medical research, Jan 9, 2015

Nowadays there is an increased medical interest in personalized medicine and tailoring decision m... more Nowadays there is an increased medical interest in personalized medicine and tailoring decision making to the needs of individual patients. Within this context our developments are motivated from a Dutch study at the Cardio-Thoracic Surgery Department of the Erasmus Medical Center, consisting of patients who received a human tissue valve in aortic position and who were thereafter monitored echocardiographically. Our aim is to utilize the available follow-up measurements of the current patients to produce dynamically updated predictions of both survival and freedom from re-intervention for future patients. In this paper, we propose to jointly model multiple longitudinal measurements combined with competing risk survival outcomes and derive the dynamically updated cumulative incidence functions. Moreover, we investigate whether different features of the longitudinal processes would change significantly the prediction for the events of interest by considering different types of associa...

BMC Cardiovascular Disorders, 2015

Background: Physicians utilize different types of information to predict patient prognosis. For e... more Background: Physicians utilize different types of information to predict patient prognosis. For example: confronted with a new patient suffering from severe aortic stenosis (AS), the cardiologist considers not only the severity of the AS but also patient characteristics, medical history, and markers such as BNP. Intuitively, doctors adjust their prediction of prognosis over time, with the change in clinical status, aortic valve area and BNP at each outpatient clinic visit. With the help of novel statistical approaches to model outcomes, it is now possible to construct dynamic event prediction models, employing longitudinal data such as AVA and BNP, and mimicking the dynamic adjustment of prognosis as employed intuitively by cardiologists. We illustrate dynamic prediction of patient survival and freedom from intervention, using baseline patient characteristics and longitudinal BNP data that are becoming available over time, from a cohort of patients with severe aortic stenosis. Methods: A 3-step approach was employed: (1) construction of a mixed-effects model to describe temporal BNP progression, (2) jointly modeling the mixed-effects model with time-to-event data (death and freedom from intervention), and (3) using the joint model to build subject-specific prediction risk models. The dataset used for this purpose includes 191 patients with severe aortic stenosis who were followed over a 3-year time period. Results: In the mixed-effects model BNP was significantly influenced by time, baseline patient age, gender, LV fractional ejection fraction and creatinine. Additionally, the joint model showed that an increasing BNP trend over time was found to be a significant predictor of death. Conclusions: By jointly modeling longitudinal data with time-to-event outcomes it is possible to construct individualized dynamic event prediction models that renew over time with accumulating evidence. It provides a potentially valuable evidence-based tool for everyday use in medical practice.

Plastic and reconstructive surgery, Jan 20, 2015

To examine the relationship between head growth, obstructive sleep apnea (OSA) and intracranial h... more To examine the relationship between head growth, obstructive sleep apnea (OSA) and intracranial hypertension in patients with syndromic or complex craniosynostosis, and to evaluate our standardized treatment protocol for the management of intracranial hypertension in these patients. Prospective observational cohort study of patients with syndromic craniosynostosis at a national referral center, treated according to a standardized protocol. occipital-frontal head circumference (OFC), with growth arrest defined as downward deflection in OFC trajectory ≥0.5 SD fall from baseline over 2 years, or lack of change in OFC growth curve; sleep studies, with results dichotomized into no/mild versus moderate/severe OSA; fundoscopy to indicate papilledema, supplemented by optical coherence tomography and/or intracranial pressure monitoring to identify intracranial hypertension. We included 62 patients, of whom 21 (33.9%) had intracranial hypertension, 39 (62.9%) OSA, and 20 (32.3%) OFC growth ar...

Transfusion, Jan 10, 2015

A too short recovery time after blood donation results in a gradual depletion of iron stores and ... more A too short recovery time after blood donation results in a gradual depletion of iron stores and a subsequent decline in hemoglobin (Hb) levels over time. This decline in Hb levels may depend on individual, unobserved characteristics of the donor. We used a data set of 5388 Dutch blood donors from the Donor InSight study. The statistical analysis is based on a Bayesian growth mixture model, which assumes that each donor belongs to one of several groups. Each group implies a different Hb trajectory, and donors with similar longitudinal trajectories belong to the same group. Analyses were performed for male and female donors separately. For both sexes the model identified four groups of donors. Stable Hb trajectories were found among 14% of male donors and 15% of female donors; declining Hb trajectories were observed in the remaining groups of donors. The percentage of donor deferrals differed strongly between groups. The model can be used to predict to which group a donor belongs, an...

Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology, Jan 26, 2015

Nicardipine permits rapid blood pressure control in severe pre-eclamptic women with a hypertensiv... more Nicardipine permits rapid blood pressure control in severe pre-eclamptic women with a hypertensive crisis. Our objective was to investigate its maternal and fetal haemodynamic effects. Ten severe pre-eclamptic pregnant women who required intravenous nicardipine for severe hypertension were included in this prospective observational trial. Maternal macrocirculation was assessed by transthoracic echocardiography. Maternal microcirculatory perfusion was examined sublingually with the Sidestream Dark Field imaging technique. Fetal haemodynamics were assessed by Doppler examination of the uteroplacental and fetal circulation. Cardiac output, total vascular resistance, mitral E/A ratio, uterine artery pulsatility index, cerebroplacental ratio and capillary heterogeneity index were considered primary outcomes. Paired measurements, obtained before the start of nicardipine infusion and after stabilisation of blood pressure were compared. Nicardipine significantly reduced blood pressure (26 m...

Circulation, Jan 25, 2014

Prospective data on long-term survival and clinical outcome beyond 30 years after surgical correc... more Prospective data on long-term survival and clinical outcome beyond 30 years after surgical correction of tetralogy of Fallot are nonexistent. This longitudinal cohort study consists of the 144 patients with tetralogy of Fallot who underwent surgical repair at <15 years of age between 1968 and 1980 in our center. They are investigated every 10 years. Cumulative survival (data available for 136 patients) was 72% after 40 years. Late mortality was due to heart failure and ventricular fibrillation. Seventy-two of 80 eligible survivors (90%) participated in the third in-hospital investigation, consisting of ECG, Holter, echocardiography, cardiopulmonary exercise testing, N-terminal pro-brain natriuretic peptide measurement, cardiac magnetic resonance (including dobutamine stress testing), and the Short Form-36 questionnaire. Median follow-up was 36 years (range, 31-43 years). Cumulative event-free survival was 25% after 40 years. Subjective health status was comparable to that in the ...

Minerva anestesiologica, 2014

Dynamic arterial lactate indices predict mortality more accurately than static arterial lactate m... more Dynamic arterial lactate indices predict mortality more accurately than static arterial lactate measurements in children with septic shock or congenital cardiac defects. The current study evaluates whether this also applies to children with primary respiratory disease requiring extracorporeal membrane oxygenation (ECMO). Static arterial lactate levels (LACabs) were prospectively collected before and during ECMO support for this single center, observational cohort study. Also, time-weighted arterial lactate (LACtw) and lactate change over time (LACdelta) were calculated as dynamic indices for, respectively, the duration and the trend over time of lactate derangement. Intensive care mortality was the primary endpoint. Analyses were performed for neonatal and pediatric patients separately. Fifty-six neonatal and 39 pediatric patients were included. Eighteen (32%) neonatal and 12 (31%) pediatric patients died. The evolution of LACabs and LACdelta differed between the pediatric survivors...