Razvoj modela za predviđanje operacijskoga mortaliteta bolesnika s rupturom aneurizama trbušne aorte

doctoral thesis

• Title (english): Data mining in development of prediction model for operative mortality in ruptured abdominal aortic aneurysms
• Author: Tomislav Meštrović
• Mentor: Zdenko Sonicki (mentor)
• Committee member: nije dostupno (član povjerenstva)
• Granter: University of Zagreb; School of Medicine; Zagreb
• Defense date and country: 2014-01-15, Croatia
• Scientific / art field, discipline and subdiscipline: BIOMEDICINE AND HEALTHCARE; Clinical Medical Sciences
• Universal decimal classification (UDC): 61 - Medical sciences

Abstract

Rupture aneurizama abdominalne aorte su praćene visokim operacijskim mortalitetom. Stoga je važno prepoznati čimbenike rizika i razviti model predviđanja operacijskog mortaliteta.
Cilj ove studije bio je razviti model predviđanja operacijskog mortaliteta udruživanjem statističkih metoda i dubinske analize podataka. Varijable vezane uz operacijski mortalitet su istražene retrospektivnom analizom povijesti bolesti bolesnika operiranih zbog rupturiranih aneurizama abdominalne aorte od 1996. do 2010. godine u Kliničkom Bolničkom Centru Zagreb.
Iz dobivenih podataka razvijeni su modeli predviđanja operacijskog mortaliteta s boljom diskriminacijskom moći na promatranom uzorku, od prethodno objavljenih, a ujedno su pronađene i naznake drugih, potencijalno važnih varijabli. Prijeoperacijski model sa samo dvije varijable ("Dob" i "Poremećaj svijesti") je pokazao dobru diskriminacijsku sposobnost na prospektivnoj validaciji, te ga smatramo vrijednim za daljnje korištenje u ustanovi, te kao polaznicu za razvitak predikcijskog sustava u našoj zemlji.
Integracija dubinske i statističke analize podataka predstavlja i metodološki doprinos analizi operacijskog mortaliteta rupturiranih aneurizama i u sličnim medicinskim domenama.
Rezultati ovog istraživanja mogli bi pomoći u kontroli kvalitete liječenja, standardizaciji izvješćivanja, racionalizaciji korištenja zdravstvenih resursa, a možda i u iznalaženju novih čimbenika operacijskog mortaliteta na koje se može utjecati.

Abstract (english)

Abdominal aortic aneurysm rupture is known to have high operative mortality. It is, therefore, imperative to unravel risk factors affecting mortality and to develop a model for outcome prediction.
The aim of this study was to develop a model for prediction of operative mortality in ruptured abdominal aortic aneurysms, by combining statistical and data mining methods. Variables potentially affecting mortality were investigated by retrospective analysis of medical records of patients operated due to the aortic aneurysm rupture between 1996. and 2010. in clinical Hospital Center Zagreb.
Two novel models of operative mortality have been developed, with better discriminatory power then the previously published ones on our patient sample. Hints of other potentially relevant variables have also been discovered. A simple preoperative prediction model, containing only two variables ("Age", "Loss of consciousness") was shown to have good discriminatory ability on prospective validation. We consider it to be a good model for clinical usage in our institution, as well as a good starting point for developing a national prediction model. Integration of data mining with medical statistics also represents a methodological advancement in the domain of ruptured aortic aneurysms and in similar fields.
Results of this study could improve the quality control process, reporting standardization, rational health resource utilisation, and, hopefully, unravel new, potentially correctible factors affecting operative mortality.

• Keywords: medicina
• Keywords (english): medicine
• Language: croatian
• URN:NBN: urn:nbn:hr:105:440251
• Study programme: Title: Biomedicine and Health Sciences; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, područje biomedicine i zdravstva (doktor/doktorica znanosti, područje biomedicine i zdravstva)
• Type of resource: Text
• Extent: 71 str.
• File origin: Born digital
• Access conditions: Open access
• Created on: 2023-05-05 07:47:41