International classification of disease based injury severity score (ICISS): A data linkage study of hospital and death data in Victoria, Australia

Berecki‐Gisolf, Janneke; Fernando, Dasamal Tharanga; D'Elia, Angelo Dominic

doi:10.1016/j.injury.2022.01.011

Cited by 10 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, ICISS is more useful for population-based data sets than ISS [ 8 ]. The results from ICISS in our study were comparable to those from previous studies [ 26 , 30 ]. We also applied EMR-ISS to the NEDIS data set, which showed good performance in a previous study [ 11 ] but poor accuracy here.…”

Section: Discussionsupporting

confidence: 91%

Model for Predicting In-Hospital Mortality of Physical Trauma Patients Using Artificial Intelligence Techniques: Nationwide Population-Based Study in Korea

Lee¹,

Kang²,

Seo³

et al. 2022

J Med Internet Res

View full text Add to dashboard Cite

Background Physical trauma–related mortality places a heavy burden on society. Estimating the mortality risk in physical trauma patients is crucial to enhance treatment efficiency and reduce this burden. The most popular and accurate model is the Injury Severity Score (ISS), which is based on the Abbreviated Injury Scale (AIS), an anatomical injury severity scoring system. However, the AIS requires specialists to code the injury scale by reviewing a patient's medical record; therefore, applying the model to every hospital is impossible. Objective We aimed to develop an artificial intelligence (AI) model to predict in-hospital mortality in physical trauma patients using the International Classification of Disease 10th Revision (ICD-10), triage scale, procedure codes, and other clinical features. Methods We used the Korean National Emergency Department Information System (NEDIS) data set (N=778,111) compiled from over 400 hospitals between 2016 and 2019. To predict in-hospital mortality, we used the following as input features: ICD-10, patient age, gender, intentionality, injury mechanism, and emergent symptom, Alert/Verbal/Painful/Unresponsive (AVPU) scale, Korean Triage and Acuity Scale (KTAS), and procedure codes. We proposed the ensemble of deep neural networks (EDNN) via 5-fold cross-validation and compared them with other state-of-the-art machine learning models, including traditional prediction models. We further investigated the effect of the features. Results Our proposed EDNN with all features provided the highest area under the receiver operating characteristic (AUROC) curve of 0.9507, outperforming other state-of-the-art models, including the following traditional prediction models: Adaptive Boosting (AdaBoost; AUROC of 0.9433), Extreme Gradient Boosting (XGBoost; AUROC of 0.9331), ICD-based ISS (AUROC of 0.8699 for an inclusive model and AUROC of 0.8224 for an exclusive model), and KTAS (AUROC of 0.1841). In addition, using all features yielded a higher AUROC than any other partial features, namely, EDNN with the features of ICD-10 only (AUROC of 0.8964) and EDNN with the features excluding ICD-10 (AUROC of 0.9383). Conclusions Our proposed EDNN with all features outperforms other state-of-the-art models, including the traditional diagnostic code-based prediction model and triage scale.

show abstract

Section: Discussionsupporting

confidence: 91%

Model for Predicting In-Hospital Mortality of Physical Trauma Patients Using Artificial Intelligence Techniques: Nationwide Population-Based Study in Korea

Lee¹,

Kang²,

Seo³

et al. 2022

J Med Internet Res

View full text Add to dashboard Cite

show abstract

“…If injury-specific survival risk ratios (SRRs) change over time, this means that SRRs derived from earlier periods result in underestimates of expected survival, when applied to more recent data. In a recently derived ICISS measure for Victoria, serious injury rates, determined by a cut-off value of 0.941 21 , 22 , were much lower using the new measures for in-hospital death than using older, published Australian values 11 , 14 . In light of the observed improved survival of hospital-treated injury over the past 20 years, an amendment of the cut-off value of 0.941 should be considered.…”

Section: Discussionmentioning

confidence: 84%

“…Injury severity was captured as ICISS: the ICD-10 based Injury Severity Score, using recent, Victorian diagnosis-specific survival probabilities (DSPs), based on in-hospital death 14 . Age-specific DSP of the worst injury was used.…”

Section: Methodsmentioning

confidence: 99%

Trends in mortality outcomes of hospital-admitted injury in Victoria, Australia 2001–2021

Berecki‐Gisolf

Fernando

D'Elia

2023

Sci Rep

Self Cite

View full text Add to dashboard Cite

Due to advancements in trauma treatment methods, it is expected that survivability of hospital-admitted injuries gradually improves over time. However, measurement of trends in all-cause injury survivability is complicated by changes in case mix, demographics and hospital admission policy. The aim of this study is to determine trends in hospital-admitted injury survivability in Victoria, Australia, taking case-mix and patient demographics into account, and to explore the potential impact of changes in hospital admission practices. Injury admission records (ICD-10-AM codes S00-T75 and T79) between 1 July 2001 and 30 June 2021 were extracted from the Victorian Admitted Episodes Dataset. ICD-based Injury Severity Score (ICISS) calculated from Survival Risk Ratios for Victoria was used as an injury severity measure. Death-in-hospital was modelled as a function of financial year, adjusting for age group, sex and ICISS, as well as admission type and length of stay. There were 19,064 in-hospital deaths recorded in 2,362,991 injury-related hospital admissions in 2001/02–2020/21. Rates of in-hospital death decreased from 1.00% (866/86,998) in 2001/02 to 0.72% (1115/154,009) in 2020/21. ICISS was a good predictor of in-hospital death with an area-under-the-curve of 0.91. In-hospital death was associated with financial year (Odds Ratio 0.950 [95%CI 0.947, 0.952]), in logistic regression modelling adjusted for ICISS, age and sex. In stratified modelling, decreasing injury death trends were observed in each of the top 10 injury diagnoses (together constituting > 50% of cases). Admission type and length of stay were added to the model: these did not alter the effect of year on in-hospital death. In conclusion, a 28% reduction in rates of in-hospital deaths in Victoria was observed over the 20-year study period, in spite of aging of the injured population. This amounts to 1222 additional lives saved in 2020/21 alone. Survival Risk Ratios therefore change markedly over time. A better understanding of the drivers of positive change will help to further reduce the injury burden in Victoria.

show abstract

“…It is believed that the low incidence of known cases with comorbidities will not significantly influence the outcome prediction in the current study. Advantages of trauma scoring systems based on ICD-codes are widely known [49,50]. It was therefore assumed that application of ICD-codes into a machine learning model could provide a high-quality, easy to use and precise tool for trauma mortality prediction.…”

Section: Discussionmentioning

confidence: 99%

Use of neural network based on international classification ICD-10 in patients with head and neck injuries in Lublin Province, Poland, between 2006–2018, as a predictive value of the outcomes of injury sustained

Jojczuk¹,

Kamiński²,

Gajewski³

et al. 2023

Ann Agric Environ Med.

View full text Add to dashboard Cite

Introduction and objective. Head and neck injuries are a heterogeneous group in terms of both clinical course and prognosis. For years, there have been attempts to create an ideal tool to predict the outcomes and severity of injuries. The aim of this study was evaluation of the use of selected artificial intelligence methods for outcome predictions of head and neck injuries. Material sand Method. 6,824 consecutive cases of patients who sustained head and neck injuries, treated in hospitals in the Lublin Province between 2006-2018, whose data was provided by National Institute of Public Health / National Institute of Hygiene, were analyzed retrospectively. Patients were qualified using International Statistical Classification of Diseases and Related Health Problems (10th Revision). The multilayer perceptron (MLP) structure was utilized in numerical studies. Neural network training was achieved with the Broyden-Fletcher-Goldfarb-Shanno (BFGS) method. Results. In the designed network, the highest classification efficiency was obtained for the group of deaths (80.7%). The average value of correct classifications for all analyzed cases was 66%. The most important variable influencing the prognosis of an injured patient was diagnosis (weight 1.929). Gender and age were variables of less significance with weight 1.08 and 1.073, respectively. Conclusions. Designing a neural network was hindered due to the large amount of cases and linking of a large number of deaths with specific diagnosis (S06). With a predictive value of 80.7% for mortality, ANN can be a promising tool in the future; however, additional variables should be introduced into the algorithm to increase the predictive value of the network. Further studies, including other types of injuries and additional variables, are needed to introduce this method into clinical use.

show abstract

International classification of disease based injury severity score (ICISS): A data linkage study of hospital and death data in Victoria, Australia

Cited by 10 publications

References 22 publications

Model for Predicting In-Hospital Mortality of Physical Trauma Patients Using Artificial Intelligence Techniques: Nationwide Population-Based Study in Korea

Model for Predicting In-Hospital Mortality of Physical Trauma Patients Using Artificial Intelligence Techniques: Nationwide Population-Based Study in Korea

Trends in mortality outcomes of hospital-admitted injury in Victoria, Australia 2001–2021

Use of neural network based on international classification ICD-10 in patients with head and neck injuries in Lublin Province, Poland, between 2006–2018, as a predictive value of the outcomes of injury sustained

Contact Info

Product

Resources

About