Artificial Intelligence in Clinical Decision Support

Challenges for Evaluating AI and Practical Implications Artificial intelligence (AI) promises to transform clinical decision-making processes as it has the potential to harness the vast amounts of genomic, biomarker, and phenotype data that is being generated across the health system including from health records and delivery systems, to improve the safety and quality of care

Whole-genome sequencing – lessons for public health surveillance

Genomic sequencing has proven vital in detecting the source and variants of COVID-19, improving contact tracing, reducing disease transmission and ensuring health security for Australia. But genomic sequencing isn’t new and researchers from the NHMRC Centre of Digital Health (CRE) have been investigating it’s use in cases from public surveillance of Salmonella Typhimurium, predicting antibiotic

Study finds “serious problems with privacy” in mobile health apps

Patients should be informed Media release from BMJ An in-depth analysis of more than 20,000 health related mobile applications (mHealth apps) published by The BMJ today (16 June 2021) finds “serious problems with privacy and inconsistent privacy practices.” The researchers say the collection of personal user information is “a pervasive practice” and that patients “should

Digital scribes and AI – how it impacts on primary care consultations

Can co-designing artificial intelligence tools with general practitioners deliver better patient outcomes and what impact will it have on Doctors? And what about the healthcare system? We took it to the test in a study with general practitioners simulating an AI documentation assistant for use in patient consultations. While artificial intelligence is advancing rapidly across

Tortorella GL, Saurin TA, Fogliatto FS, Rosa VM, Tonetto LM, Magrabi F. Impacts of Healthcare 4.0 digital technologies on the resilience of hospitals. Technological Forecasting and Social Change. 2021;166:120666.

Healthcare 4.0 (H4.0) adapts principles and applications from the Industry 4.0 movement to healthcare, enabling real-time customization of care to patients and professionals. As such, H4.0 can potentially support resilient performance in healthcare systems, which refers to their adaptive capacity to cope with complexity. This paper explores the impact of ten H4.0 digital technologies on

Lyell D, Coiera E, Chen J, Shah P, Magrabi F. How machine learning is embedded to support clinician decision making: an analysis of FDA-approved medical devices. BMJ Health Care Inform. 2021 Apr;28(1):e100301. doi: 10.1136/bmjhci-2020-100301. PMID: 33853863; PMCID: PMC8054073.

Objective: To examine how and to what extent medical devices using machine learning (ML) support clinician decision making. Methods: We searched for medical devices that were (1) approved by the US Food and Drug Administration (FDA) up till February 2020; (2) intended for use by clinicians; (3) in clinical tasks or decisions and (4) used

Fernandez-Luque L, Kushniruk AW, Georgiou A, Basu J, Petersen C, Ronquillo C, Paton C, Nohr C, Kuziemsky C, Alhuwail D, Skiba D, Huesing E, Gabaron E, Borycki EM, Magrabi F, Denecke K, Topaz M, Al-Shorbaji N, Lacroix P, Cornet R, Iyengar S, Gogia SB, Kobayashi S, Deserno TM, Mettler T, Vimarlund V, Zhu X. Evidence-based health informatics as the foundation for the COVID-19 response: a joint call for action to transform hopes and hypes into realities. Methods of information in medicine. 2021;(accepted Jan 20).

Background: As a major public health crisis, the novel coronavirus disease 2019 (COVID-19) pandemic demonstrates the urgent need for safe, effective, and evidence-based implementations of digital health. The urgency stems from the frequent tendency to focus attention on seemingly high promising digital health interventions despite being poorly validated in times of crisis. Aim: In this

Albarqouni L, Moynihan R, Clark J, Scott AM, Duggan A, Del Mar C. Head of bed elevation to relieve gastroesophageal reflux symptoms: a systematic review. BMC Fam Pract. 2021 Jan 19;22(1):24. doi: 10.1186/s12875-021-01369-0. PMID: 33468060; PMCID: PMC7816499.

Background: Overuse of proton pump inhibitors (PPIs) – frequently used for relieving symptoms of gastroesophageal reflux disease (GORD) – raises long-term safety concerns, warranting evidence-based non-drug interventions. We conducted a systematic review to evaluate the effect of head-of-bed elevation on relieving symptoms of GORD in adults. Methods: We included controlled trials comparing the effect of

Yin K & AYS Lau. Field Methods for Patient Ergonomics. In Holden, R.J., & Valdez, R.S. (Eds.). (2021). The Patient Factor: Theories and Methods for Patient Ergonomics (1st ed.). CRC Press.

Patients are increasingly encouraged to take an active role in managing their health and health care. New technologies, cultural shifts, trends in healthcare delivery, and policies have brought to the forefront the “work” patients, families, and other non-professionals perform in pursuit of health. Volume 1 provides theoretical and methodological foundation for the emerging discipline of

Wang Y, Verspoor K, Baldwin T. (2020) Learning from Unlabelled Data for Clinical Semantic Textual Similarity. Proceedings of the 3rd Clinical Natural Language Processing Workshop at EMNLP2020.

Domain pretraining followed by task fine-tuning has become the standard paradigm for NLP tasks, but requires in-domain labelled data for task fine-tuning. To overcome this, we propose to utilise domain unlabelled data by assigning pseudo labels from a general model. We evaluate the approach on two clinical STS datasets, and achieve r= 0.80 on N2C2-STS.

Wang Y, Liu F, Verspoor K, Baldwin T (2020) Evaluating the Utility of Model Configurations and Data Augmentation on Clinical Semantic Textual Similarity. Proceedings of the Workshop on Biomedical Natural Language Processing (BioNLP) at ACL2020.

In this paper, we apply pre-trained language models to the Semantic Textual Similarity (STS) task, with a specific focus on the clinical domain. In low-resource setting of clinical STS, these large models tend to be impractical and prone to overfitting. Building on BERT, we study the impact of a number of model design choices, namely

Scott AM, Clark J, Mar CD, Glasziou P. Increased fluid intake to prevent urinary tract infections: systematic review and meta-analysis. Br J Gen Pract. 2020 Feb 27;70(692):e200-e207. doi: 10.3399/bjgp20X708125. PMID: 31988085; PMCID: PMC6988703.

Background: Approximately 15% of community-prescribed antibiotics are used in treating urinary tract infections (UTIs). Increase in antibiotic resistance necessitates considering alternatives. Aim: To assess the impact of increased fluid intake in individuals at risk for UTIs, for impact on UTI recurrence (primary outcome), antimicrobial use, and UTI symptoms (secondary outcomes). Design and setting: A systematic

Redfern J, Coorey G, Mulley J, Scaria A, Neubeck L, Hafiz N, Pitt C, Weir K, Forbes J, Parker S, Bampi F, Coenen A, Enright G, Wong A, Nguyen T, Harris M, Zwar N, Chow CK, Rodgers A, Heeley E, Panaretto K, Lau A, Hayman N, Usherwood T, Peiris D. A digital health intervention for cardiovascular disease management in primary care (CONNECT) randomized controlled trial. NPJ Digit Med. 2020 Sep 10;3:117. doi: 10.1038/s41746-020-00325-z. PMID: 32964140; PMCID: PMC7484809.

Digital health applications (apps) have the potential to improve health behaviors and outcomes. We aimed to examine the effectiveness of a consumer web-based app linked to primary care electronic health records (EHRs). CONNECT was a multicenter randomized controlled trial involving patients with or at risk of cardiovascular disease (CVD) recruited from primary care (Clinical Trial

Rahimi A, Baldwin T, Verspoor K. (2020) WikiUMLS: Aligning UMLS to Wikipedia via Cross-lingual Neural Ranking. COLING 2020.

We present our work on aligning the Unified Medical Language System (UMLS) to Wikipedia, to facilitate manual alignment of the two resources. We propose a cross-lingual neural reranking model to match a UMLS concept with a Wikipedia page, which achieves a recall@1of 72%, a substantial improvement of 20% over word- and char-level BM25, enabling manual

Pederson M, Verspoor K, Jenkinson M, Law M, Abbott DF, Jackson GD (2020). Artificial Intelligence for clinical decision support in neurology. Brain Communications. doi: 10.1093/braincomms/fcaa096

Artificial intelligence is one of the most exciting methodological shifts in our era. It holds the potential to transform healthcare as we know it, to a system where humans and machines work together to provide better treatment for our patients. It is now clear that cutting edge artificial intelligence models in conjunction with high-quality clinical