National and Kapodistrian University of Athens
Παράγοντες κινδύνου πρώτου 24ώρου νοσηλείας που επηρεάζουν την τελική έκβαση ασθενών ΜΕΘ
Abstract
dc:descriptionIntroduction: In Intensive Care Units (ICUs) the ability to predict the outcome accurately and reliably, through special prognostic scoring systems, is very important as it has many applications in clinical practice. Numerous new variables are proposed in the literature that seems to influence the outcome of patients and could potentially be included in the new prognostic systems that will be developed in the future to increase their prognostic ability. Objectives: The primary aim of this study was the development and validation of three new mortality prognostic systems of ICU patients using data routinely collected during the first 24 hours of ICU admission. The three prognostic systems concern the three different types of outcome: 1) death in the ICU; 2) death in the hospital; and 3) death in the hospital only for patients who were discharged from the ICU. Additional goals of the study were: a) to compare the performance of the new systems to the most widely used conventional scoring models; and b) to identify the risk factors that affect the final outcome of ICU patients. Material and Methods: A prospective, observation cohort study was carried out in a medical/surgical, multidisciplinary, university ICU of a general hospital in Attica region in Greece. We studied 436 consecutive patients hospitalized in the ICU, from January 2012 to July 2013. For the data collection a specifically designed form was developed for recording the demographic and clinical characteristics of the patients, as well as the clinical follow-up of the first 24 hours in ICU and the assessment of the severity of the illness and the nursing workload, based on the following scoring systems: APACHE II (acute, physiology, age, chronic), SAPS III (Simplified Acute Physiology Score), SOFA score (Sepsis-Related Organ Failure Assessment), GCS (Glasgow Coma Score), TISS-28 (Therapeutic Intervention Scoring System) and NAS (Nursing Activity Score). For only the first two types of outcome, the initial sample was divided into two subgroups: the development set (N=400 patients) and the validation set (N=36 patients). From the development set we created the new final models, while the validation set was used for the evaluation of the performance of the former models for accurate prediction of death. For the development of the new models, in three cases, multivariate logistic regression modeling was used.Results: Three final prediction models of death in ICU patients were developed for the three types of outcome: 1) The new final model for predicting the ICU mortality is a novel, simple, flexible and handy, 12-variable tool and shows improved performance (ROC AUC=0.85, SMR=1.25), compared to the conventional scoring models (APACHE II: AUC=0.76, SMR=2.50, SAPS III: AUC=0.76, SMR=1.50) as well as higher predictive capability regarding ICU mortality (predicted mortality: 41.63±31.61, observed mortality: 41.67%). 2) The new final predictive model regarding the total hospital mortality consists of 11 variables. It shows improved discrimination (ROC AUC=0.80, SMR=0.85), compared to the conventional scoring models (APACHE II: AUC=0.78, SMR=1.13, SAPS III: AUC=0.77, SMR=0.94) as well as satisfactory predictive capability regarding ICU mortality (predicted mortality: 55.25±32.24, observed mortality: 47.22%). Regarding SOFA (ROC AUC=0.81, SMR=0.85), although it is slightly better than the new system, it is not so specific in assessing the severity of the illness and predicting the outcome, but in assessing the organic dysfunction. 3) The new final model for predicting hospital mortality for patients who have already exited the ICU is a very simple tool and consists of only 7 variables. It shows to fit better to our local population (ROC AUC=0.85, SMR=1.62) than the conventional scoring models (APACHE II: AUC=0.75, SMR=3.52, SAPS III: AUC=0.80, SMR=4.00) and it has a better predictive ability (predicted mortality: 20.67±22.14, observed mortality: 20.69%). However, these results are not very reliable, as this model was not evaluated by a valuation set due to the small sample size and therefore it needs further evaluation. Conclusions: The newly developed models are novel and quite simple risk-adjusted outcome prediction tools, as in two of them the rate of change of serum lactate concentration within the first day of ICU stay, is included for the first time. It seems that these new models are best suited to the local population of patients who are hospitalized in our ICU, as we have better performance and predictive capacity than APACHE II, SAPS III and SOFA. For the time being, the new models seem to be better than the conventional models, and reliable predictors of ICU mortality, despite the several limitations of the study. A wider validation using a larger sample size would be useful in order to ascertain their use in other ICUs.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Fika, Sofia
- Φίκα, Σοφία
Subjects
dc:subject × 16- Θνησιμότητα
- Πρόβλεψη
- Μονάδα εντατικής θεραπείας (ΜΕΘ)
- Βαρύτητα της ασθένειας
- Υπολογιστικά συστήματα
- Κλινική απόδοση
- Mortality
- Prediction
- Intensive care unit (ICU)
- Severity of illness
- Scoring systems
- Clinical performance
- Ιατρική και Επιστήμες Υγείας
- Επιστήμες Υγείας
- Medical and Health Sciences
- Health Sciences
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/43462