{"id":10045,"date":"2026-04-08T15:00:28","date_gmt":"2026-04-08T15:00:28","guid":{"rendered":"https:\/\/www.proefschriftmaken.nl\/portfolio\/rana-dandis\/"},"modified":"2026-04-23T07:45:27","modified_gmt":"2026-04-23T07:45:27","slug":"rana-dandis","status":"publish","type":"us_portfolio","link":"https:\/\/www.proefschriftmaken.nl\/en\/portfolio\/rana-dandis\/","title":{"rendered":"Rana Dandis"},"content":{"rendered":"","protected":false},"excerpt":{"rendered":"","protected":false},"author":8,"featured_media":12910,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_acf_changed":false,"footnotes":""},"us_portfolio_category":[45],"class_list":["post-10045","us_portfolio","type-us_portfolio","status-publish","has-post-thumbnail","hentry","us_portfolio_category-new-template"],"acf":{"naam_van_het_proefschift":"","samenvatting":"Er is geen Nederlandse samenvatting beschikbaar. De Engelse samenvatting vind je hier<\/a>.","summary":"This thesis aims to provide an overview of the currently utilized approaches to incorporate repeated measurements for prediction. More specifically it provides understanding, comparison, and applications of several available methods for (dynamic) prediction of a prospective outcome based on repeated measurements, namely the: two-stage method, joint modeling approach, longitudinal discriminant analysis and the latent class analysis. These methods can be used to address several types of public health problems. First, they can help to understand how changes in a biomarker or other longitudinal variable are related to changes in status of a subject. Second, they can be used to predict the outcome of a subject based on the trajectory of the longitudinal outcome, thus providing information that can be used in a personalized medicine setting. This allows investigators to identify potentially harmful patterns and intervene at an earlier stage.\n\nThe thesis is composed of six chapters, each of them dealing with different research objectives. Chapter 1 is introductory and reviews the available approaches for prediction based on repeatedly measured predictors and discusses our motivation to study them further.\n\nChapter 2 is a tutorial that provides an illustrative overview of the two-stage method and the joint modelling approach to obtain dynamic predictions using both the maximum likelihood and Bayesian frameworks. The chapter is divided into two parts. Part 1 illustrates the four modeling approaches using a real example data set and evaluate their predictive performance. Part 2 investigates the situations where the predictive performance of these models is different using several simulation scenarios. The findings of this study showed that the performance of these approaches varies with the different subject-specific variability of the biomarker and suggested that the joint model is preferred over the two-stage method when the within- and the between-patients variability of the biomarker is high.\n\nChapter 3 looks at the differences in the predictive performance between the joint model (JM-Bin) and the longitudinal discriminant analysis (LoDA). Even though the two approaches use the same information for prediction, both methods handle them differently. In JM-Bin, predictions typically are based on the subject-specific deviations from the overall mean trajectory of the longitudinal biomarker, while LoDA models the distribution of longitudinal biomarker per group separately. The influence of such difference on the predictive performance of the two approaches is expected and is explored using different simulation scenarios. The results showed better predictive performance of LoDA compared to JM-Bin when the within- and\/or between-subject variability of the biomarker is different between the outcome groups. The chapter ends by stressing the importance of exploring the variability of the longitudinal biomarker before choosing which model to apply.\n\nThe next two chapters of the thesis focus on the latent class mixed modeling (LCMM) and its application for prediction. The latent class analysis approaches, including the latent class mixed modeling (LCMM), deal with hidden heterogeneity by assuming that a population under study may consist of \u201clatent\u201d subpopulations or classes with distinct patterns of longitudinal trajectories of biomarkers that can also have different effects on the outcome in each subpopulation. Chapter 4 presents an interesting application of latent class analysis to predict response to early high-intensity physiotherapy training after Total Knee Arthroplasty based on longitudinal trajectories of walking speed during the training. In this study we used the latent class mixed model (LCMM) to classify patients into groups according to their walking speed trajectories during the program. Then the association between the outcomes and the identified groups was assessed using multivariable regression. Chapter 5 describes another application using data from the same cohort as used in chapter 4. This time the Latent class analysis was used to distinguish classes of patients based on recovery trajectories over 6 weeks. Multivariable regression analyses were used to identify associations between the classes and one-year outcomes. Based on the findings in chapter 4 and chapter 5, we recommend using the LCMM followed by a suitable regression model when the aim is to first identify distinct groups of subjects in a heterogenous population based on their longitudinal profiles and then use the groups for prediction.\n\nDiscussion and conclusion are drawn in Chapter 6. The main aim of the thesis has been reached by providing guidance to help the researchers with the choice between different methods for risk prediction with longitudinal predictors. When the appropriate method is used, prediction models can make optimal utilization of the available information.","auteur":"Rana Dandis","auteur_slug":"rana-dandis","publicatiedatum":"8 april 2026","taal":"EN","url_flipbook":"https:\/\/ebook.proefschriftmaken.nl\/ebook\/ranadandis?iframe=true","url_download_pdf":"https:\/\/ebook.proefschriftmaken.nl\/download\/3e350fcd-bf20-4dae-85b6-1a7bed9c7fe8\/optimized","url_epub":"","ordernummer":"FTP-202604081456","isbn":"","doi_nummer":"","naam_universiteit":"Overig","afbeeldingen":12910,"naam_student:":"","binnenwerk":"","universiteit":"Overig","cover":"","afwerking":"","cover_afwerking":"","design":""},"_links":{"self":[{"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/us_portfolio\/10045","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/us_portfolio"}],"about":[{"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/types\/us_portfolio"}],"author":[{"embeddable":true,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/comments?post=10045"}],"version-history":[{"count":1,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/us_portfolio\/10045\/revisions"}],"predecessor-version":[{"id":10046,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/us_portfolio\/10045\/revisions\/10046"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/media\/12910"}],"wp:attachment":[{"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/media?parent=10045"}],"wp:term":[{"taxonomy":"us_portfolio_category","embeddable":true,"href":"https:\/\/www.proefschriftmaken.nl\/en\/wp-json\/wp\/v2\/us_portfolio_category?post=10045"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}