Bring-Your-Own-Device Usage Trends in Australian Hospitals – A National Survey

Background: Healthcare is among the leading industries which drives the use of personal devices for work purposes (BYOD). However, allowing BYOD for healthcare workers comes at a cost, as it puts sensitive information assets such as patient data residing on personal devices at risk of potential data breaches. Objective: Previous review of the literature has highlighted the dearth of empirical studies in hospital settings regarding BYOD usage. As such, this paper aims to report BYOD usage trends in Australian hospitals through a national survey, first of its kind in Australia. Methods: An anonymous survey was conducted online among health IT personnel, asking them about their experiences about BYOD usage in their hospitals. 28 responses were collected based on public Australian hospitals, which included 21 hospital groups and 7 standalone hospitals, likely to represent more than 100 hospitals in total. Survey responses were quantitatively analysed through descriptive statistical analysis and cross tabulation. Results: BYOD is allowed in majority of the hospitals, and among all major staff groups, with doctors being the leading group. Participants ranked reasons for allowing BYOD, and most of them were related to improvement in clinical productivity, efficiency and mobility for clinical staff. Challenges were generally related to data security such as patient data breaches and compliance with data security laws, according to them. More than two thirds of hospitals had a cybersecurity officer employed, and CIOs were the most dominant group who held responsibility for managing BYOD within the hospital. Conclusion: This paper provides a starting point for better understanding of BYOD usage in a complex healthcare environment based on empirical evidence, one which highlights the security-usability conundrum, confirming previous literature themes.

Towards Automatic and Interpretable Assignments of Patients Presenting with Pain to the Emergency Department

Introduction: Pain is the most common symptom that patients present with to the emergency department. It is hard to identify patients who have presented in pain to the emergency department when compliance with structured pain assessment is low. An ability to identify patients presenting in pain allows further investigation of the quality of care provided. Background: Machine and deep learning techniques are commonly used for text analysis in healthcare. Applications such as the classification of diagnosis and unplanned readmissions from textual medical records have previously been described. In other work, conventional and deep-learning techniques have demonstrated high performance in identifying patients presenting to the emergency department in pain. However, these models have lacked interpretability. Methods: This paper proposes the use of machine learning techniques to identify patients who present in pain based upon their initial assessment using interpretable deep learning models. Results: The interpretable deep learning model of pain identification was shown to have more accuracy and precision than other machine and deep learning techniques. This technique has significant application to large datasets for the identification of the quality of care and real-time identification of patients presenting in pain to improve their care.

Visual Design and Anthropomorphism in a Mobile Pulmonary Rehabilitation Support Intervention

Pulmonary rehabilitation is a behavioral intervention that can improve symptom control and quality of life for patients with chronic obstructive pulmonary disease (COPD), but access, uptake and adherence are problematic. Our team has pursued the development of a mobile phone-based intervention (mobile pulmonary rehabilitation, mPR) with iterative design and a pilot study. The mPR intervention is delivered through two technologies: text messages (SMS) and a smartphone application. Our user-centered design analysis of pilot study data led to several insights. First, patients’ replies to the SMS suggested that messages were anthropomorphised and provided social support. Second, the smartphone application could help patients by clearly visualizing the exercise program, alternative exercises, and progress to date. We demonstrate the design iterations made to meet these requirements and we present feedback obtained from experts and from four COPD patients. We discuss implications for the design of mobile pulmonary rehabilitation interventions.

Caregivers’ Perspectives on Privacy in Aged Care Monitoring Devices

As people move into advanced old age, they may experience cognitive impairments and frailty, making it difficult for them to live without support from others. Caregivers might decide to use aged care monitoring devices (ACMDs) to support older adults under their care. However, these devices raise privacy concerns as they collect and share sensitive data from the older adult’s private life in order to provide monitoring capabilities. This study involved interviewing formal and informal caregivers who used/may use ACMDs to investigate their views on privacy. The study found that although caregivers consider protecting older adults’ privacy important, they may overlook privacy in order to gain benefits from ACMDs. We argue that ACMD developers should simplify privacy terms and conditions so that caregivers can make well-informed decisions when deciding to use the device. They also should consider providing users with flexible privacy settings so that users can decide what data to collect, whom to share it with and when.

Patient Flow Simulation Using Historically Informed Synthetic Data

Hospital overcrowding is a major problem for healthcare systems around the globe. In order to better estimate future demands and adequate resources for coping with such demands, statistical and computerised modelling can be applied. This can then allow healthcare administrators and decision makers to quantify the impacts of various “what-if” scenarios on hospital performance measures. This paper investigates the application of Discrete Event Simulation towards optimising Emergency Department resources while measuring overall length of stay and queuing time of emergency patients as a target performance measure. In particular, we explore strategies for generating historically informed synthetic data that helps the simulation model track patient flow through the target hospital over a future time frame. Using the developed simulation model, several resource configurations are tested using data from one of the busiest emergency departments in the state of Queensland as the baseline while quantifying the impacts of such changes on key patient flow metrics. It was found that adding a single bed (and associated resources) to the emergency department would result in a 23% decrease in average patient treatment delay.

Digital Innovations for Aged Care: Impacts in the COVID-19 Pandemic

This paper describes the extent to which remote interaction healthcare interventions supported by digital technology are currently being used, or have recently been newly developed for use, in the care of older people in Australia within the context of the existing Australian aged care system and in conjunction with the COVID-19 pandemic. We place emphasis on those interventions associated with primary care provision, and associated healthcare services such as allied health, rather than outreach from jurisdictional health services and acute care. The primary purpose of this study was to gain an indication of the extent and range of such interventions, and provide a pragmatic commentary on their usage. This has enabled the understanding of some characteristics for success, and drivers for rapid adoption of further digital technology interventions, in the aged care sector.

Automated Inter-Ictal Epileptiform Discharge Detection from Routine EEG

Epilepsy is the most common neurological disorder. The diagnosis commonly requires manual visual electroencephalogram (EEG) analysis which is time-consuming. Deep learning has shown promising performance in detecting interictal epileptiform discharges (IED) and may improve the quality of epilepsy monitoring. However, most of the datasets in the literature are small (n≤100) and collected from single clinical centre, limiting the generalization across different devices and settings. To better automate IED detection, we cross-evaluated a Resnet architecture on 2 sets of routine EEG recordings from patients with idiopathic generalized epilepsy collected at the Alfred Health Hospital and Royal Melbourne Hospital (RMH). We split these EEG recordings into 2s windows with or without IED and evaluated different model variants in terms of how well they classified these windows. The results from our experiment showed that the architecture generalized well across different datasets with an AUC score of 0.894 (95% CI, 0.881–0.907) when trained on Alfred’s dataset and tested on RMH’s dataset, and 0.857 (95% CI, 0.847–0.867) vice versa. In addition, we compared our best model variant with Persyst and observed that the model was comparable.

Improving the Physical Activity of Breast Cancer Survivors Through Fitness Trackers

Introduction: Increasing physical activity among posttreatment breast cancer survivors is essential, as greater physical activity reduces the relative risk of cancer-specific mortality. This trial examines how a fitness tracker-based intervention changes the physical activity behaviour of inactive posttreatment breast cancer survivors. Methods: Seventeen physically inactive posttreatment breast cancer survivors participated in a randomised cross-over controlled trial. Participants underwent a 12-week intervention of a fitness tracker combined with a behavioural counselling and goal-setting session and 12 weeks of normal activity (control). The primary outcome was the change in physical activity assessed by accelerometry over seven days. Results: The intervention achieved a mean increase of 4.5 min/day of moderate-vigorous physical activity, representative of a small-moderate effect (d = 0.34). Changes in time spent as a proportion of the day in light physical activity (-8.3%) and in sedentary behaviour (7.9%), were both significantly different to baseline (t (16) = 3.522, p < 0.01; t (16) = -3.162, p < 0.01). Conclusion: Interindividual differences in the change of patterns of physical activity behaviour suggest that only for some, fitness trackers can achieve a change in the level of moderate-vigorous physical activity.

Pandora’s Bot: Insights from the Syntax and Semantics of Suicide Notes

Conversation agents (chat-bots) are becoming ubiquitous in many domains of everyday life, including physical and mental health and wellbeing. With the high rate of suicide in Australia, chat-bot developers are facing the challenge of dealing with statements related to mental ill-health, depression and suicide. Advancements in natural language processing could allow for sensitive, considered responses, provided suicidal discourse can be accurately detected. Here suicide notes are examined for consistent linguistic syntax and semantic patterns used by individuals in mental health distress. Paper contains distressing content.

Utilising Electronic Health Record Data to Assess the Sepsis Inpatient Care Pathway: A Feasibility Study

The World Health Organisation has recently declared sepsis a global medical emergency. Obtaining quality data to establish the evidence on how clinicians recognise, diagnose, and treat sepsis is still a challenge. This feasibility study aimed to utilise routinely collected data from electronic health records (EHR) to assess the sepsis inpatient care pathway. We conducted a retrospective observational cohort study which included all patients admitted to a private teaching hospital between 2015 and 2018. De-identified patient demographic and clinical data were extracted and analysed. A total of 47 sepsis patients were identified based on diagnoses recorded and a review of clinical notes. A surgical procedure was conducted on more than half of these patients (n=25, 53%). Nearly two-thirds were given antibiotics (n=30, 64%), of which 87% (n=26) were administered within 2-hours of sepsis diagnosis. Eighteen patients were admitted to ICU and 13 of them were diagnosed as septic in ICU. We identified some aspects of EHR data that could be improved. Overall, routinely collected data from clinical information systems provides rich information to assess the sepsis patient care pathway.