Queensland Injury Surveillance Unit Research Articles

Lithium-ion battery related burns and emerging trends: a retrospective case series and data analysis of emergency presentations.

The rising prevalence of battery powered devices is driving a steady increase in lithium-ion battery-related burns. We present a case series of patients with lithium-ion battery-related burns and describe the clinical characteristics of this cohort. To further understand emerging trends, we compare our specialty inpatient case series with emergency department (ED) data collated by the Queensland Injury Surveillance Unit (QISU). This is a retrospective case series of all patients admitted to the Professor Stuart Pegg Adult Burns Centre for burns relating to lithium-ion batteries between January 2014 and October 2023. In addition, we provide a retrospective data analysis of ED presentations for lithium-ion battery-related burns or blast injuries collated by the QISU. Within the case series data, the most common injuries caused by lithium-ion batteries leading to burn unit admissions were due to e-scooters (57.1%). Burns relating to e-scooter batteries were more likely to involve a larger total body surface area and to be deeper in nature, than burns due to other products. The most common ED presentations were from energy storage devices such as powerpacks (43%). Lithium-ion battery-related burns are becoming more frequent. The majority of inpatient managed cases involved burns due to larger lithium-ion batteries used in e-scooters. This is the first Australian case series describing severe burns caused by this mechanism. Primary prevention through design, technological and behavioural strategies is required.

Application of a Machine Learning-Based Decision Support Tool to Improve an Injury Surveillance System Workflow.

Emergency department (ED)-based injury surveillance systems across many countries face resourcing challenges related to manual validation and coding of data. This study describes the evaluation of a machine learning (ML)-based decision support tool (DST) to assist injury surveillance departments in the validation, coding, and use of their data, comparing outcomes in coding time, and accuracy pre- and postimplementations. Manually coded injury surveillance data have been used to develop, train, and iteratively refine a ML-based classifier to enable semiautomated coding of injury narrative data. This paper describes a trial implementation of the ML-based DST in the Queensland Injury Surveillance Unit (QISU) workflow using a major pediatric hospital's ED data comparing outcomes in coding time and pre- and postimplementation accuracies. The study found a 10% reduction in manual coding time after the DST was introduced. The Kappa statistics analysis in both DST-assisted and -unassisted data shows increase in accuracy across three data fields, that is, injury intent (85.4% unassisted vs. 94.5% assisted), external cause (88.8% unassisted vs. 91.8% assisted), and injury factor (89.3% unassisted vs. 92.9% assisted). The classifier was also used to produce a timely report monitoring injury patterns during the novel coronavirus disease 2019 (COVID-19) pandemic. Hence, it has the potential for near real-time surveillance of emerging hazards to inform public health responses. The integration of the DST into the injury surveillance workflow shows benefits as it facilitates timely reporting and acts as a DST in the manual coding process.

A retrospective registry analysis of the transport‐related health burden of wheeled recreational devices in Queensland, Australia

ObjectiveEvaluate injury patterns from wheeled recreational devices (WRD) in the public space and explore risk factors for hospital admission. MethodA cross‐sectional analysis of WRD injury prevalence and risk factors for hospital admissions was conducted using data from the Queensland Injury Surveillance Unit (QISU) database for 2007 to 2017. Descriptive statistics and a log‐binomial regression model were used to calculate adjusted relative risk for hospital admission. ResultsMost WRD injury in the public space was related to stand‐alone WRD injury events such as falls, with few reported WRD users being hit by vehicles from 2007 to 2017. Stand‐alone WRD injury events had a higher independent risk of hospital admissions when injured in the head/neck/face (RR 2.08, 95%CI 1.6 to 2.8, p<0.001), and when the injury was a fracture (RR 2.57, 95%CI 2.1 to 3.3, p<0.001) or a brain injury (RR 3.19, 95%CI 2.5 to 4.1, p<0.001). ConclusionHead, brain and facial injuries and fractures are leading preventable factors for hospital admissions due to WRD injury. These types of injuries generate a preventable burden to the health system. Implications for public healthThe results support the need to consider legislation regarding mandatory helmet use for non‐motorised WRD when used on public roads and footpaths, while further research is conducted. This strategy could reduce the long‐term health outcomes associated with head, face and brain injury in young commuters.

Health implications of age and gender injury patterns of non-vehicle pedestrian trauma

IntroductionWalking is a healthy, environmentally sustainable, and economically profitable transportation strategy for citizens and governments. Despite these benefits, it is also important to understand, and to inform health professionals, policymakers, and public infrastructure planners about, the preventable injury potential of walking in the transportation system and the distribution of injury by gender and age. This is especially important for non-vehicle pedestrian trauma, given that is more prevalent and there is less information available about its injury outcomes in comparison to pedestrian injury caused by vehicle collisions. MethodA retrospective cohort study, and cross-sectional bivariate and multivariate analyses of non-vehicle pedestrian trauma in the traffic environment (NVPTE) were conducted. Relative risks (RR) of the incidence of injury and prevalence of injury outcomes by age and gender were calculated using data collected by the Queensland Injury Surveillance Unit (QISU) in Australia. ResultsPedestrians aged 60+ have a consistently higher risk of NVPTE incidence when compared with pedestrians aged <60. Older pedestrians have a higher risk of injury to the pelvis (RR 10.3 95% CI 3.81–27.5 p < 0.0001) and hip (RR 5.59 95% CI 3.53–8.85 p < 0.0001), and an adjusted higher risk of hospital admission (RR 1.5 95% CI 1.30–1.64 p < 0.0001). Pedestrian males <60 had more injuries to the head, sprains, strains and open wounds compared with females <60, while those aged 60+ had lower risk of fractures and higher risk of hospital admission. ConclusionPromoting walking for transport requires the incorporation of prevention strategies that consider key differences in injury outcomes. Clinical fall risk assessments and protective pedestrian infrastructure could play an important role in reducing the health and economic burden of pedestrian falls to the health system.

Intelligent human-machine approaches for assigning groups of injury codes to accident narratives

In injury surveillance, different aspects of an injury event are captured using injury codes such as the External-cause-of-injury (E-code), Major Injury Factor (MIF), and Intent. These are usually assigned by human coders based on accident narratives. Previous studies have examined automated and semi-automated filtering approaches that use machine learning (ML) models to assign single E-codes to accident narratives. In this study, our goal was to examine the effectiveness of these approaches for assigning groups of injury codes. This was done for three different types of injury codes (E-code, MIF, and Intent) using several ML models (Logistic Regression, Support Vector Machine, and Long-Short-Term-Memory based Recurrent-Neural-Network). Four filtering strategies were also tested which used the probability of prediction correctness assigned by the Logistic Regression model. These approaches were evaluated for a manually-coded dataset, provided by the Queensland Injury Surveillance Unit containing about half a million injury cases. The results showed very similar performance for the three ML models. The overall sensitivity of each model was quite high and almost identical for E-code (0.81–0.82), MIF (0.69–0.71), and Intent (0.96–0.97). However, the unweighted sensitivities were lower - E-code (0.67–0.75), MIF (0.59–0.62), and Intent (0.46–0.56), reflecting a general trend of each model to under-predict small categories. It was also observed that the probability of correctly assigning all three codes was low (0.58). Filtering approaches resulted in large improvements in sensitivity for smaller categories and the probability of predicting all three codes correctly.

Epidemiology of paediatric head and facial injuries in Queensland, Australia

IntroductionHead and facial injuries in paediatric patients are a major public health problem in Queensland, Australia, where control, minimisation, or prevention play key roles. Although many studies have been reported, few can be found concerning the epidemiology of paediatric maxillofacial injuries in Queensland, Australia. The aim of this study was to retrospectively assess and report the aetiology, patterns, distribution, and associated factors of head and facial injuries in children in Queensland, Australia.Material and methodsData were obtained from Queensland injury Surveillance Unit (QISU). A retrospective study included 10,723 injured children (from less than 1 year old to 12 years old) in the years 2012, 2013, 2014, 2015, and 2016. The analyses involved descriptive statistics and 2 test.ResultsThe number of patients injured varied among age groups. The nature and mechanism of injuries were different among age groups as well. Boys were more prone to maxillofacial injuries compared to females. The highest rate of injuries was in the 1–2-year age group for both genders. Contact with a static object was the main aetiology of maxillofacial injuries during leisure activity (playing). The home was the main site where the injuries occurred.ConclusionsPreventive measures should be strengthened both indoors and outdoors to reduce the incidence and severity of paediatric injuries in Queensland, Australia. In addition, children in the growing phase should be monitored periodically. Furthermore, increasing awareness can be attained by increasing safety education.

Patterns and features of methamphetamine-related presentations to emergency departments in QLD from 2005 to 2017.

This study explores the patterns and features of methamphetamine-related presentations to emergency departments (EDs) in Queensland. Despite an overall decrease in the use of methamphetamine in Australian, an increase in the use of the crystalized form of methamphetamine has been noted over recent years. A descriptive observational study was utilized to analyse emergency department (ED) injury surveillance data sourced from Queensland Injury Surveillance Unit (QISU) from 2005 to 2017. Data were analysed for presentations related to stimulants (n=564) with methamphetamine (n=250) included as a subcategory. Descriptive statistics were used to identify patterns and features of presentations related to methamphetamines. The relationship between demographic variables, service type variables, and drug type was assessed using chi-square and z-tests. Results included the following: 84.4% of methamphetamine-related presentations were allocated a triage score of 1, 2, or 3; 14.8% of all methamphetamine-related presentations required police involvement; 18% were brought in by ambulance; and 15.6% exhibited behaviour that was either, agitated, aggressive, or violent in nature. Methamphetamine-related presentations more frequently required police or ambulance services and more often included aggression or agitation. Methamphetamine-related presentations to ED have a high acuity and often require other emergency resources (police and ambulance). There is a need to develop policy for managing aggressive and agitated people presenting to EDs as a result of methamphetamine use and to further explore the experience of personnel (police and ambulance) managing persons under the influence of methamphetamine.

Semi-automated text mining strategies for identifying rare causes of injuries from emergency room triage data

Human coders, in many organizations conducting injury surveillance, routinely assign External-cause-of-injury codes (E-codes) to short narratives describing the incident, transcribed by triage nurses or others in hospital emergency rooms or other settings. Machine learning (ML) models trained on coded injury narratives can accurately assign E-codes to a large portion of the data, but tend to poorly predict cases falling into rare categories. In this study, we examined several ways of filtering out cases for human review that were likely to belong to rare categories from the predictions of Logistic Regression and Naïve Bayes classifiers for a manually-coded emergency department triage dataset of approximately 500,000 cases, collected between years 2002–2012, provided by the Queensland Injury Surveillance Unit. The ML models were trained using 90% of the data and the filtering approaches were evaluated on a prediction set comprised of the remaining cases. Cost analysis was also performed to compare the efficiency of each filtering method. The results showed that each filtering method greatly improved the ability to detect rare categories. Filtering using expert-designed causal linguistic rules combined with Logistic Regression prediction strength was found to be the most efficient approach. Several completely automated filtering approaches were also found to be effective.

Improving autocoding performance of rare categories in injury classification: Is more training data or filtering the solution?

Introduction: Classical Machine Learning (ML) models have been found to assign the external-cause-of-injury codes (E-codes) based on injury narratives with good overall accuracy but often struggle with rare categories, primarily due to lack of enough training cases and heavily skewed nature of injurdata. In this paper, we have: a) studied the effect of increasing the size of training data on the prediction performance of three classical ML models: Multinomial Naïve Bayes (MNB), Support Vector Machine (SVM) and Logistic Regression (LR), and b) studied the effect of filtering based on prediction strength of LR model when the model is trained on very-small (10,000 cases) and very-large (450,000 cases) training sets. MethodData from Queensland Injury Surveillance Unit from years 2002–2012, which was categorized into 20 broad E-codes was used for this study. Eleven randomly chosen training sets of size ranging from 10,000 to 450,000 cases were used to train the ML models, and the prediction performance was analyzed on a prediction set of 50,150 cases. Filtering approach was tested on LR models trained on smallest and largest training datasets. Sensitivity was used as the performance measure for individual categories. Weighted average sensitivity (WAvg) and Unweighted average sensitivity (UAvg) were used as the measures of overall performance. Filtering approach was also tested for estimating category counts and was compared with approaches of summing prediction probabilities and counting direct predictions by ML model. ResultsThe overall performance of all three ML models improved with increase in the size of training data. The overall sensitivities with maximum training size for LR and SVM models were similar (∼82%), and higher than MNB (76%). For all the ML models, the sensitivities of rare categories improved with increasing training data but they were considerably less than sensitivities of larger categories. With increasing training data size, LR and SVM exhibited diminishing improvement in UAvg whereas the improvement was relatively steady in case of MNB. Filtering based on prediction strength of LR model (and manual review of filtered cases) helped in improving the sensitivities of rare categories. A sizeable portion of cases still needed to be filtered even when the LR model was trained on very large training set. For estimating category counts, filtering approach provided best estimates for most E-codes and summing prediction probabilities approach provided better estimates for rare categories. ConclusionsIncreasing the size of training data alone cannot solve the problem of poor classification performance on rare categories by ML models. Filtering could be an effective strategy to improve classification performance of rare categories when large training data is not available.

Patterns of 'at-home' alcohol-related injury presentations to emergency departments.

This study aimed to establish the scale of alcohol-related injuries originating in the home. Despite recent media and public attention on alcohol-related injuries occurring at licensed venues, many occur in other locations including the home. A retrospective observational study. Emergency department surveillance data sourced from the Queensland Injury Surveillance Unit were interrogated for alcohol-related emergency department presentations from 2003-2012 (n=12,296). Descriptive analysis was undertaken to assess alcohol involvement in injury, and analysis of variance was used to determine the differences among group means and their associated presentations. The relationship between demographic variables and injury location was assessed using p value of <0·05 as statistically significant. Of all injuries that were positively identified as being alcohol related, 41·07% occurred at the 'other' location, 36·14% 'at home', 13·00% on the street and 9·78% at licensed premises. Of these, males (n=2635; 59%) represented a higher proportion than females (n=1807; 41%). Of injuries identified as domestic violence by spouse or partner (n=510), 59·5% occurred 'at home'. This is the first study to investigate alcohol-related injuries occurring at home. The home accounts for a greater proportion of injuries than the frequently assessed licensed premises location. Further research is required to validate these findings in a wider setting. A public health campaign is required to minimise harm associated with alcohol-related injuries in the home, and nurses are positioned to inform health policy makers around this issue. Furthermore, emergency department nurses are in a unique position to provide brief interventions around safe alcohol consumption and injury prevention.

The Extent of Consumer Product Involvement in Paediatric Injuries.

A challenge in utilising health sector injury data for Product Safety purposes is that clinically coded data have limited ability to inform regulators about product involvement in injury events, given data entry is bound by a predefined set of codes. Text narratives collected in emergency departments can potentially address this limitation by providing relevant product information with additional accompanying context. This study aims to identify and quantify consumer product involvement in paediatric injuries recorded in emergency department-based injury surveillance data. A total of 7743 paediatric injuries were randomly selected from Queensland Injury Surveillance Unit database and associated text narratives were manually reviewed to determine product involvement in the injury event. A Product Involvement Factor classification system was used to categorise these injury cases. Overall, 44% of all reviewed cases were associated with consumer products, with proximity factor (25%) being identified as the most common involvement of a product in an injury event. Only 6% were established as being directly due to the product. The study highlights the importance of utilising injury data to inform product safety initiatives where text narratives can be used to identify the type and involvement of products in injury cases.

Infant product‐related injuries: comparing specialised injury surveillance and routine emergency department data

Objective:To explore the potential for using a basic text search of routine emergency department data to identify product‐related injury in infants and to compare the patterns from routine ED data and specialised injury surveillance data. Methods:Data was sourced from the Emergency Department Information System (EDIS) and the Queensland Injury Surveillance Unit (QISU) for all injured infants between 2009 and 2011. A basic text search was developed to identify the top five infant products in QISU. Sensitivity, specificity, and positive predictive value were calculated and a refined search was used with EDIS. Results were manually reviewed to assess validity. Descriptive analysis was conducted to examine patterns between datasets. Results:The basic text search for all products showed high sensitivity and specificity, and most searches showed high positive predictive value. EDIS patterns were similar to QISU patterns with strikingly similar month‐of‐age injury peaks, admission proportions and types of injuries. Conclusions:This study demonstrated a capacity to identify a sample of valid cases of product‐related injuries for specified products using simple text searching of routine ED data. Implications:As the capacity for large datasets grows and the capability to reliably mine text improves, opportunities for expanded sources of injury surveillance data increase. This will ultimately assist stakeholders such as consumer product safety regulators and child safety advocates to appropriately target prevention initiatives.

Estimating under-reporting of road crash injuries to police using multiple linked data collections

The reliance on police data for the counting of road crash injuries can be problematic, as it is well known that not all road crash injuries are reported to police which under-estimates the overall burden of road crash injuries. The aim of this study was to use multiple linked data sources to estimate the extent of under-reporting of road crash injuries to police in the Australian state of Queensland. Data from the Queensland Road Crash Database (QRCD), the Queensland Hospital Admitted Patients Data Collection (QHAPDC), Emergency Department Information System (EDIS), and the Queensland Injury Surveillance Unit (QISU) for the year 2009 were linked. The completeness of road crash cases reported to police was examined via discordance rates between the police data (QRCD) and the hospital data collections. In addition, the potential bias of this discordance (under-reporting) was assessed based on gender, age, road user group, and regional location. Results showed that the level of under-reporting varied depending on the data set with which the police data was compared. When all hospital data collections are examined together the estimated population of road crash injuries was approximately 28,000, with around two-thirds not linking to any record in the police data. The results also showed that the under-reporting was more likely for motorcyclists, cyclists, males, young people, and injuries occurring in Remote and Inner Regional areas. These results have important implications for road safety research and policy in terms of: prioritising funding and resources; targeting road safety interventions into areas of higher risk; and estimating the burden of road crash injuries.

Alcohol‐related emergency department injury presentations in Queensland adolescents and young adults over a 13‐year period

The rate of alcohol-related emergency department (ED) presentations in young people has increased dramatically in recent decades. Injuries are the most common type of youth alcohol-related ED presentation, yet little is known about these injuries in young people. This paper describes the characteristics of alcohol-related ED injury presentations in young people over a 13-year period and determines if they differ by gender and/or age group (adolescents: 12-17 years; young adults: 18-24 years). The Queensland Injury Surveillance Unit (QISU) database collects injury surveillance data at triage in participating EDs throughout Queensland, Australia. A total of 4667 cases of alcohol-related injuries in young people (aged 12-24 years) were identified in the QISU database between January 1999 and December 2011, using an injury surveillance code and nursing triage text-based search strategy. Overall, young people accounted for 38% of all QISU alcohol-related ED injury presentations in patients aged 12 years or over. The majority of young adults presented with injuries due to violence and falls, whereas adolescents presented due to self-harm or intoxication without other injury. Males presented with injuries due to violence, whereas females presented with alcohol-related self-harm and intoxication. There is a need for more effective ways of identifying the degree of alcohol involvement in injuries among young people presenting to EDs.

Alcohol-related violence presenting to the emergency department: Is ‘glassing’ the big issue?

The study aims to describe the characteristics of patients presenting to EDs within Queensland, Australia with injuries because of assault with a glass implement ('glassing') and to set this within the broader context of presentations because of alcohol-related violence. This is an analysis of prospectively collected ED injury surveillance data collated by the Queensland Injury Surveillance Unit between 1999 and 2011. Cases of injury because of alcohol-related violence were identified and analysed using coded fields supplemented with qualitative data contained within the injury description text. Descriptive statistics were used to assess the characteristics of injury presentations because of alcohol-related violence. Violence included interpersonal violence and aggression (verbal aggression and object violence). A total of 4629 cases were studied. The study population was predominantly men (72%) and aged 18 to 24 (36%), with men in this age group comprising more than a quarter of the study population (28%). Nine per cent of alcohol-related assault injuries were a consequence of 'glassing'. The home was the most common location for alcohol-related violence (31%) and alcohol-related 'glassings' (33%). Overall, the most common glass object involved was a bottle (75%); however, within licensed venues an even mix of a drinking glass (44%) and glass bottle (45%) was identified. Contrary to public perception generated by media, 'glassing' incidents, particularly at licensed venues, constitute a relatively small proportion of all alcohol-related violence. The current study highlights the predominance of young men injured following alcohol-related violence, demonstrating a key focus area within the population for aiming prevention strategies.

Injuries leading to hospitalisation in the first year of life: analysis by trimester of age using coded data and textual description

To describe unintentional injuries to children aged less than one year, using coded and textual information, in three-month age bands to reflect their development over the year. Data from the Queensland Injury Surveillance Unit was used. The Unit collects demographic, clinical and circumstantial details about injured persons presenting to selected emergency departments across the State. Only injuries coded as unintentional in children admitted to hospital were included for this analysis. After editing, 1,082 children remained for analysis, 24 with transport-related injuries. Falls were the most common injury, but becoming proportionately less over the year, whereas burns and scalds and foreign body injuries increased. The proportion of injuries due to contact with persons or objects varied little, but poisonings were relatively more common in the first and fourth three-month periods. Descriptions indicated that family members were somehow causally involved in 16% of injuries. Our findings are in qualitative agreement with comparable previous studies. The pattern of injuries varies over the first year of life and is clearly linked to the child's increasing mobility. Injury patterns in the first year of life should be reported over shorter intervals. Preventive measures for young children need to be designed with their rapidly changing developmental stage in mind, using a variety of strategies, one of which could be opportunistic developmentally specific education of parents.

I have a bone to pick with you: the challenge of addressing dog-related injury in a rural community

BackgroundSarina, a rural community with a regional population of 12 000, is serviced by Mackay Safe Communities Network (MSCN). MSCN was established in 2000 to address high injury rates regionally....

Breaking the cycle of violence: detection of domestic violence in a regional emergency department

BackgroundDomestic violence is a serious issue in Australian society. Under reporting and poor management have affected how domestic violence victims are cared for within the health care system. Little research...

Bow Wow, Wolf Wolf: prevention and perception of dog bites

BackgroundHumans and dogs have developed a mutually beneficial relationship that has spanned thousands of years. The evolution of society however, has changed the nature of human–dog relationship. Dog bites can...

‘WHY SOMETIMES I'VE BELIEVED AS MANY AS SIX IMPOSSIBLE THINGS BEFORE BREAKFAST’: TRANSLATING CLINICAL EXPERIENCE INTO INJURY PREVENTION OUTCOMES

BackgroundClinicians in a variety of fields are at the receiving end of many injury incidents. Information relevant to preventing injury is often discussed, less often documented, and rarely acted upon...