Evaluating ChatGPT text mining of clinical records for companion animal obesity monitoring.

BACKGROUND: Veterinary clinical narratives remain a largely untapped resource for addressing complex diseases. Here we compare the ability of a large language model (ChatGPT) and a previously developed regular expression (RegexT) to identify overweight body condition scores (BCS) in veterinary narratives pertaining to companion animals. METHODS: BCS values were extracted from 4415 anonymised clinical narratives using either RegexT or by appending the narrative to a prompt sent to ChatGPT, prompting the model to return the BCS information. Data were manually reviewed for comparison. RESULTS: The precision of RegexT was higher (100%, 95% confidence interval [CI] 94.81%-100%) than that of ChatGPT (89.3%, 95% CI 82.75%-93.64%). However, the recall of ChatGPT (100%, 95% CI 96.18%-100%) was considerably higher than that of RegexT (72.6%, 95% CI 63.92%-79.94%). LIMITATIONS: Prior anonymisation and subtle prompt engineering are needed to improve ChatGPT output. CONCLUSIONS: Large language models create diverse opportunities and, while complex, present an intuitive interface to information. However, they require careful implementation to avoid unpredictable errors.

A mixed-methods approach utilising electronic health records to examine antimicrobial prescription surrounding gastrointestinal clinical presentations in dogs and cats.

Introduction: Systemically-administered antimicrobials are often prescribed in canine and feline gastrointestinal clinical presentations. Responsible use of antimicrobials, particularly those considered Highest Priority Critically Important Antimicrobials (HPCIAs) is vital to tackle antimicrobial resistance. Although practice-level prescription guidance is available, further strategies based on a greater understanding of antimicrobial prescription at the population-level are needed. Here, we used a mixed-methods approach, harnessing veterinary electronic health records (EHRs) to characterise the use of antimicrobials in canine and feline gastrointestinal presentations, and to explore justification and reasoning around antimicrobial prescribing, particularly of HPCIAs. Methods: This observational study used 23,337 EHRs complemented with veterinary practitioner-completed questionnaires, from canine and feline gastrointestinal consultations from 225 volunteer UK veterinary practices between April 2014 and September 2018. Results: A total of 83.4% (95% confidence interval (CI) 82.6-84.3) gastrointestinal presentations were reported as mild, with non-haemorrhagic diarrhoea and vomiting the most frequently reported clinical signs. Systemically-administered antimicrobials occurred in 28.6% of canine (95% CI 26.9-30.3) and 22.4% of feline (95% CI 20.4-24.4) gastrointestinal consultations, with HPCIA prescription occurring more frequently in cats. Results of multivariable analysis showed the presence of non-haemorrhagic diarrhoea (canine Odds Ratio (OR) 2.1, 95% CI 1.9-2.3; feline OR 1.8, 95% CI 1.5-2.1), haemorrhagic diarrhoea (canine OR 4.2, 95% CI 3.8-4.7; feline OR 3.1, 95% CI 2.4-3.8), and moderate/severe presentations (canine OR 1.9, 95% CI 1.7-2.8; feline OR 2.0, 95% CI 1.7-2.5) were positively associated with receiving a systemically-administered antimicrobial. Thematic analysis of clinical narrative content of 516 gastrointestinal consultations where HPCIAs were prescribed allowed the identification of ten factors underpinning reasoning or decision-making for HPCIA prescription: perceived animal/owner compliance; owner's expectations; perceived risk of infection; clinical signs; recent clinical history; perceived positive previous response to antimicrobial therapy; geriatric patients and euthanasia; concomitant conditions; diagnostic testing and the behavioral trend to trial antimicrobial therapy empirically in gastrointestinal cases. No explicit justification for HPCIA prescription was recorded in 77% of cases. Discussion: Improving recorded justification represents a clear target for stewardship programmes. By utilising a complementary mixed-methods approach to EHRs, this study unlocks previously untapped data recorded within EHRs. These results can help inform targeted interventions, contributing towards enhanced antimicrobial stewardship.