Skupno modeliranje pri bolnikih s srčnim popuščanjem

Vavdi, Gregor (author), Kejžar, Nataša (mentor)

V magistrskem delu predstavljamo metodologijo skupnega modeliranja in z njo analiziramo podatke o bolnikih s srčnim popuščanjem. Podatki so zbrani v okviru projekta United4Health za 128 bolnikov v obdobju štirih let. Dnevno merjene (telemetrične) spremenljivke, ki jih uporabljamo za gradnjo modelov, so: diastolični krvni tlak, sistolični krvni tlak in srčni utrip. Na primeru diastoličnega tlaka predstavljamo izbiro različic modelov, ki smo jih ocenili, vrednotimo rezultate in interpretiramo razlike med njimi. Modele primerjamo s tistimi iz članka Njagi in sod. (2013), v katerem so naredili analizo pri enakem tipu podatkov. Predstavljamo in uporabljamo tudi Coxov model s časovnimi spremenljivkami in rezultate primerjamo s skupnimi modeli.

• Language: Slovenian
• Keywords: skupno modeliranje, skupni model, longitudinalni podatki, linearni mešani model, analiza preživetja, Coxov model s časovnimi spremenljivkami, dinamični diskriminantni indeks, slučajni vplivi, trenutna vrednost
• Work type: Master's thesis/paper
• Organization: FE - Faculty of Electrical Engineering
• Year: 2021
• PID: 20.500.12556/RUL-132835
• COBISS.SI-ID: 83615747
• Publication date in RUL: 04.11.2021

Secondary language

• Language: English
• Title: Joint modelling approach for chronic heart failure patients

Abstract

In the master's thesis, we present the methodology of joint modeling and analyze data on patients with chronic heart failure. Data were collected as part of the United4Health project for 128 patients over four years. Diastolic blood pressure, systolic blood pressure, and heart rate were selected for daily variables. In the case of diastolic pressure, we present the choice of several types of models that we have built due to the nature of the data and we interpreted the differences between them. We follow the article Njagi et al. (2013), in which they made the same analysis. We also use the extended Cox model and compare it with the joint modeling methodology.

• Keywords: joint modeling, joint model, longitudinal data, linear mix model, survival analysis, extended Cox model, random effect, current value, dinamic discriminant index