Josh Andrews

Josh Andrews

Auto-generated profile, based on publicly available information

Josh Andrews

Faculty of Health and Medical Sciences The University of Adelaide Adelaide South Australia Australia

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 June 2026 in Internal Medicine Journal
Verified
Journal Article
Open access
PDF file
JAJoshua Andrews
Joshua Andrews
SKSanjat Kanjilal
Sanjat Kanjilal
CKCamile Kotton
Camile Kotton
BSBrandon Stretton
Brandon Stretton
ALAlasdair Leslie
Alasdair Leslie
Background: Large language models (LLMs) show potential to support antimicrobial prescribing but require simulation‐based, institution‐specific safety evaluation prior to any consideration of clinical use. In Australia, antimicrobial prescribing represents a high‐risk domain for digital decision‐support systems due to patient safety and antimicrobial resistance implications. Aim: To characterise prescribing accuracy, error phenotypes and antimicrobial stewardship risk associated with a LLM that was provided with publicly available surgical prophylaxis guidelines during inference (without fine‐tuning or model modification) across 20 simulated surgical scenarios. Methods: Twenty simulated surgical scenarios were tested using a LLM that was prompt‐conditioned with publicly available guideline text during inference, without any fine‐tuning or modification of model weights. For each case, the model generated recommendations for agent, dose, timing, re‐dosing and guideline citation. Outputs were independently assessed by two local clinicians familiar with the guideline, with accuracy scored across five domains and harm classified using a modified National Coordinating Council for Medical Error Reporting and Prevention (NCC MERP) Index. Results: Clinically significant antimicrobial prescribing risk was identified in 10% of simulated scenarios (2/20), recognising wide confidence intervals due to the small sample size. These included omission of required anaerobic coverage and failure to redose prophylaxis in prolonged procedures. Overall guideline concordance was 4/5, with perfect dose accuracy but lower performance for timing (70%) and guideline citation (45%). Conclusions: This study demonstrates the feasibility of constructing institutionally governed, guideline‐based AI systems while identifying stewardship‐relevant safety risks that currently preclude clinical use without further validation.

25 July 2026 in Internal Medicine Journal
Verified
Journal Article
Closed access
JAJoshua Andrews
Joshua Andrews
SKSanjat Kanjilal
Sanjat Kanjilal
CKCamile Kotton
Camile Kotton
BSBrandon Stretton
Brandon Stretton
ALAlasdair Leslie
Alasdair Leslie

1 September 2026 in South Australian Journal of Medicine
Verified
Journal Article
Closed access
JAJoshua Andrews
Joshua Andrews
Background: Acute post-operative pain is common and strongly influences opioid consumption, recovery and early readmission. Whether routinely collected ward pain data contains identifiable phenotypes remains uncertain, particularly across diverse surgical populations. Objectives: To derive pain phenotypes and evaluate their associations with inpatient opioid use, opioid quantity dispensed at discharge, length of stay and 30-day readmissions. Methods: We conducted a multicentre retrospective observational cohort study of 40,049 adult surgical admissions across five Australian hospitals. Our sample was limited to those who had undergone a single operation for which a post-operative pain score was recorded. Seven numeric features of the pain rating scale were standardised and clustered using k-means (k = 3) to identify pain phenotypes. Associations were examined using multivariable linear regression models (log-transformed opioid dose [in oral morphine milligram equivalents, MME] and length of stay) and logistic regression for 30-day readmission, adjusting for demographic, clinical and operative covariates. Results: Three phenotypes were identified: Phenotype 1, intermittent low average pain (41.8%); Phenotype 2, consistently minimal pain (38.0%); and Phenotype 3, persistently high pain (20.3%). Compared with Phenotype 1, Phenotypes 2 and 3 showed 85% lower and 171% higher adjusted opioid consumption, respectively. Further, discharge MME was 24% lower and 108% higher (all P < 0.01), respectively. Thirty-day readmission was reduced only in Phenotype 2 (adjusted OR 0.66; 95% CI 0.61–0.72). Phenotype 3 showed no difference in readmission risk compared with Phenotype 1. Conclusion: Using only routinely recorded ward pain scores, this simple phenotyping approach identified patient groups who had distinct opioid dose requirements and early surgical outcomes. These findings suggest that routinely collected pain data from electronic medical records could be used prospectively to flag high-risk patients for targeted analgesic review, without requiring additional data collection. Keywords: post-operative pain · opioid analgesics · pain phenotype · length of stay · readmission

24 September 2026 in Anaesthesia and Intensive Care
Verified
Journal Article
Closed access
JAJosh Andrews
Josh Andrews
ALAlasdair Leslie
Alasdair Leslie
VTVenkatesan Thiruvenkatarajan
Venkatesan Thiruvenkatarajan
BSBrandon Stretton
Brandon Stretton

22 July 2026 in Journal of Hand Surgery (European Volume)
Verified
Journal Article
Closed access
JAJoshua Andrews
Joshua Andrews
SBSimon Bellringer
Simon Bellringer
MAMelanie Amarasooriya
Melanie Amarasooriya
GBGregory Bain
Gregory Bain
Introduction: Scapholunate instability is the most common pattern of carpal instability, yet the relative contributions of the intrinsic and extrinsic ligaments remain debated. This systematic review evaluated cadaveric studies of wrist biomechanics assessing kinematic changes following carpal ligament sectioning to clarify the stabilizing roles of individual ligaments. Methods: The EMBASE, MEDLINE, Cochrane and Scopus databases were searched from inception to June 2025. Cadaveric studies reporting scapholunate gap (SLG), scapholunate angle (SLA), radiolunate angle (RLA) or dorsal intercalated segment instability (DISI) after ligament sectioning were included. Data were extracted for cumulative and sequential sectioning patterns and synthesized qualitatively. Results: Thirty-three studies involving 406 wrists were included. With an intact scapholunate interosseous ligament (SLIL), sectioning single or multiple extrinsic carpal ligaments produced minimal changes in SLG, SLA, or RLA. Scapholunate interosseous ligament sectioning increased SLG (mean +1.7 mm), SLA (+8°) and RLA (+3°), with clenched-fist loading further elevating SLG (+0.7 mm) and SLA (+4°). The dorsal intercarpal ligament (DIC) was the dominant secondary stabilizer. Combined SLIL–DIC injury produced the greatest increases across all measures, and DISI occurred only when the SLIL and either the DIC or the scaphotrapeziotrapezoid complex was sectioned. Conclusion: Cadaveric evidence confirms the SLIL as the primary stabilizer of the scapholunate joint and identifies the DIC as the key secondary stabilizer. These biomechanical findings support early diagnosis, targeted reconstruction and the need to restore both primary and secondary stabilizers to prevent progressive scapholunate instability.

12 September 2025 in Australian Journal of Rural Health
Verified
Journal Article
Closed access
JAJosh Andrews
Josh Andrews
ALAlasdair Leslie
Alasdair Leslie
CGChristina Gao
Christina Gao
DLD‐Yin Lin
D‐Yin Lin
BSBrandon Stretton
Brandon Stretton
SBStephen Bacchi
Stephen Bacchi