Building the workforce of tomorrow: The weighting of rural exposure in standardised curriculum vitae scoring criteria for entrance into Australian specialty training programs.

OBJECTIVE: To determine the weighting of rural exposure within publicly available standardised curriculum vitae (CV) scoring criteria for trainee medical officer's applying into medical and surgical specialty training programs in Australia and New Zealand. METHODS/DESIGN: An observational analysis of rural exposure point allocations within publicly available standardised CV scoring criteria for entrance into specialty training programs. SETTING: All Australian and New Zealand medical and surgical specialties training programs outlined by the Australian Health Practitioner Regulation Agency (AHPRA) who publish publicly available standardised CV scoring criteria for entrance into specialty training were included. RESULTS: Of the 14 specialty training programs that publish publicly available standardised CV scoring criteria, 8/14 allocate points towards rural exposure. While the allocation of points within this scoring domain varies between the eight training programs, the mean weighting of rural exposure is 13.7%. CONCLUSIONS: The relative weighting of rural exposure varies between the eight specialty training programs who include rural exposure as a CV scoring criteria. The deliberate and strategic construction of CV scoring criteria and inclusion of rural exposure points is important to continue developing the Australian rural specialist workforce. Future development of standardised CV scoring criteria should continue to consider point allocation towards rural exposure and related activities to ensure that the requirements of rural Australian healthcare needs are met across medical and surgical specialties.

Subcutaneous sodium valproate in palliative care: A systematic review.

BACKGROUND: Seizures are an important palliative symptom, the management of which can be complicated by patients' capacity to swallow oral medications. In this setting, and the wish to avoid intravenous access, subcutaneous infusions may be employed. Options for antiseizure medications that can be provided subcutaneously may be limited. Subcutaneous sodium valproate may be an additional management strategy. AIM: To evaluate the published experience of subcutaneous valproate use in palliative care, namely with respect to effectiveness and tolerability. DESIGN: A systematic review was registered (PROSPERO CRD42023453427), conducted and reported according to PRISMA reporting guidelines. DATA SOURCES: The databases PubMed, EMBASE and Scopus were searched for publications until August 11, 2023. RESULTS: The searches returned 429 results, of which six fulfilled inclusion criteria. Case series were the most common study design, and most studies included <10 individuals who received subcutaneous sodium valproate. There were three studies that presented results on the utility of subcutaneous sodium valproate for seizure control, which described it to be an effective strategy. One study also described it as an effective treatment for neuropathic pain. The doses were often based on presumed 1:1 oral to subcutaneous conversion ratios. Only one study described a local site adverse reaction, which resolved with a change of administration site. CONCLUSIONS: There are limited data on the use of subcutaneous sodium valproate in palliative care. However, palliative symptoms for which subcutaneous sodium valproate have been used successfully are seizures and neuropathic pain. The available data have described few adverse effects, supporting its use with an appropriate degree of caution.

Improving the performance of machine learning penicillin adverse drug reaction classification with synthetic data and transfer learning.

BACKGROUND: Machine learning may assist with the identification of potentially inappropriate penicillin allergy labels. Strategies to improve the performance of existing models for this task include the use of additional training data, synthetic data and transfer learning. AIMS: The aims of this study were to investigate the use of additional training data and novel machine learning strategies, namely synthetic data and transfer learning, to improve the performance of penicillin adverse drug reaction (ADR) machine learning classification. METHODS: Machine learning natural language processing was applied to free-text penicillin ADR data extracted from a public health system electronic health record (EHR). The models were developed by training on various labelled data sets. ADR entries were split into training and testing data sets and used to develop and test a variety of machine learning models. The effect of training on additional data and synthetic data versus the use of transfer learning was analysed. RESULTS: Following the application of these techniques, the area under the receiver operator curve of best-performing models for the classification of penicillin allergy (vs intolerance) and high-risk allergy (vs low-risk allergy) improved to 0.984 (using the artificial neural network model) and 0.995 (with the transfer learning approach) respectively. CONCLUSIONS: Machine learning models demonstrate high levels of accuracy in the classification and risk stratification of penicillin ADR labels using the reaction documented in the EHR. The model can be further optimised by incorporating additional training data and using transfer learning. Practical applications include automating case detection for penicillin allergy delabelling programmes.

Perceived and actual value of Student-led Objective Structured Clinical Examinations.

INTRODUCTION: Student-led Objective Structured Clinical Examinations (OSCEs) provide formative learning opportunities prior to Faculty-led OSCEs. It is important to undertake quality assurance measurements of peer-led assessments because, if they are found to be unreliable and invalid, they may have detrimental impacts. The objectives of this study were to explore and evaluate Student-led OSCEs hosted by fifth-year medical students. METHODS: Student-led OSCE results were analysed to examine reliability (Cronbach's alpha). The relationship between Student-led and Faculty-led OSCEs was evaluated using linear regression. Qualitative data were acquired by survey and semi-structured interviews and were analysed using an inductive content analysis approach. RESULTS: In total, 85 (94%) of 91 eligible students consented to study participation. Student-led OSCEs had a low-moderate reliability [Cronbach alpha = 0.47 (primary care) and 0.61 (human reproduction/paediatrics) (HRH)]. A statistically significant, positive relationship between Student-led and Faculty-led OSCE results was observed. Faculty-led OSCE grades increased by 0.49 (95% CI: 0.18, 0.80) to 1.09 (95% CI: 0.67, 1.52), for each percentage increase in Student-led OSCE result. Student-led OSCE participants highly valued the authentic peer-assessed experience. Reported benefits included a reduction of perceived stress and anxiety prior to Faculty-led OSCEs, recognition of learning gaps, contribution to overall clinical competency and facilitation of collaboration between peers. DISCUSSION: Student-led OSCEs are moderately reliable and can predict Faculty-led OSCE performance. This form of near-peer assessment encourages the metacognitive process of reflective practice and can be effectively implemented to direct further study. Faculties should collaborate with their student bodies to facilitate Student-led OSCEs and offer assistance to improve the quality, and benefits, of these endeavours.

Current safeguards, risk mitigation, and transparency measures of large language models against the generation of health disinformation: repeated cross sectional analysis.

OBJECTIVES: To evaluate the effectiveness of safeguards to prevent large language models (LLMs) from being misused to generate health disinformation, and to evaluate the transparency of artificial intelligence (AI) developers regarding their risk mitigation processes against observed vulnerabilities. DESIGN: Repeated cross sectional analysis. SETTING: Publicly accessible LLMs. METHODS: In a repeated cross sectional analysis, four LLMs (via chatbots/assistant interfaces) were evaluated: OpenAI's GPT-4 (via ChatGPT and Microsoft's Copilot), Google's PaLM 2 and newly released Gemini Pro (via Bard), Anthropic's Claude 2 (via Poe), and Meta's Llama 2 (via HuggingChat). In September 2023, these LLMs were prompted to generate health disinformation on two topics: sunscreen as a cause of skin cancer and the alkaline diet as a cancer cure. Jailbreaking techniques (ie, attempts to bypass safeguards) were evaluated if required. For LLMs with observed safeguarding vulnerabilities, the processes for reporting outputs of concern were audited. 12 weeks after initial investigations, the disinformation generation capabilities of the LLMs were re-evaluated to assess any subsequent improvements in safeguards. MAIN OUTCOME MEASURES: The main outcome measures were whether safeguards prevented the generation of health disinformation, and the transparency of risk mitigation processes against health disinformation. RESULTS: Claude 2 (via Poe) declined 130 prompts submitted across the two study timepoints requesting the generation of content claiming that sunscreen causes skin cancer or that the alkaline diet is a cure for cancer, even with jailbreaking attempts. GPT-4 (via Copilot) initially refused to generate health disinformation, even with jailbreaking attempts-although this was not the case at 12 weeks. In contrast, GPT-4 (via ChatGPT), PaLM 2/Gemini Pro (via Bard), and Llama 2 (via HuggingChat) consistently generated health disinformation blogs. In September 2023 evaluations, these LLMs facilitated the generation of 113 unique cancer disinformation blogs, totalling more than 40 000 words, without requiring jailbreaking attempts. The refusal rate across the evaluation timepoints for these LLMs was only 5% (7 of 150), and as prompted the LLM generated blogs incorporated attention grabbing titles, authentic looking (fake or fictional) references, fabricated testimonials from patients and clinicians, and they targeted diverse demographic groups. Although each LLM evaluated had mechanisms to report observed outputs of concern, the developers did not respond when observations of vulnerabilities were reported. CONCLUSIONS: This study found that although effective safeguards are feasible to prevent LLMs from being misused to generate health disinformation, they were inconsistently implemented. Furthermore, effective processes for reporting safeguard problems were lacking. Enhanced regulation, transparency, and routine auditing are required to help prevent LLMs from contributing to the mass generation of health disinformation.

Current management options for the treatment of refractory postoperative cystoid macular edema: A systematic review.

Post-operative cystoid macular edema (PCME) is an important complication following intraocular surgery that often resolves spontaneously without treatment. In some cases, PCME may persist despite initial medical therapy, which can adversely impact visual outcomes. Various topical, intraocular and systemic treatments exist for the prevention and management of cystoid macular edema; however, there is no consensus on treatment of refractory cases in the postoperative setting. In accordance with the PRISMA guidelines, we systematically reviewed 68 articles describing management options and their outcomes for treatment-resistant cases of PCME. The most commonly reported treatments included steroid (39 studies) and biological-based (17 studies) therapies. We provide an overview of the treatment options for refractory PCME.

ChatGPT-Based Learning: Generative Artificial Intelligence in Medical Education.

Large language models like ChatGPT are a type of machine learning model that can offer a positive paradigm shift in case-based/problem-based learning (CBL/PBL). ChatGPT may be able to augment the existing paradigm to work in conjunction with the clinical-teacher in PBL/CBL case generation. It can develop realistic patient cases that could be revised by clinical teachers to ensure accuracy and relevance. Further, it can be directed to include specific case content in order to facilitate the constructive alignment of the case with the broader learning objectives of the curriculum. There is also the possibility of improving engagement by 'gamifying' CBL/PBL. Supplementary Information: The online version contains supplementary material available at 10.1007/s40670-023-01934-5.

Renal disease is not associated with delays in hyperacute stroke management in South Australia.

OBJECTIVE: The aims of the present study were to determine how renal disease is associated with the time to receive hyperacute stroke care. METHODS: The present study involved a 5-year cohort of all patients admitted to stroke units in South Australia. RESULTS: In those with pre-existing renal disease there were no significant differences in the time taken to receive a scan, thrombolysis or endovascular thrombectomy. CONCLUSIONS: The present study shows that in protocolised settings there were no significant delays in hyperacute stroke management for patients with renal disease.

Prognostic value of left ventricular systolic function before vascular surgery: a systematic review.

BACKGROUND: Vascular surgery carries a high risk of post-operative cardiac complications. Recent studies have shown an association between asymptomatic left ventricular systolic dysfunction and increased risk of major adverse cardiovascular events (MACE). This systematic review aims to evaluate the prognostic value of left ventricular function as determined by left ventricular ejection fraction (LVEF) measured by resting echocardiography before vascular surgery. METHODS: This review conformed to PRISMA and MOOSE guidelines. PubMed, OVID Medline and Cochrane databases were searched from inception to 27 October 2022. Eligible studies assessed vascular surgery patients, with multivariable-adjusted or propensity-matched observational studies measuring LVEF via resting echocardiography and providing risk estimates for outcomes. The primary outcomes measures were all-cause mortality and congestive heart failure at 30 days. Secondary outcome included the composite outcome MACE. RESULTS: Ten observational studies were included (4872 vascular surgery patients). Studies varied widely in degree of left ventricular systolic dysfunction, symptom status, and outcome reporting, precluding reliable meta-analysis. Available data demonstrated a trend towards increased incidence of all-cause mortality, congestive heart failure and MACE in patients with pre-operative LVEF <50%. Methodological quality of the included studies was found to be of moderate quality according to the Newcastle Ottawa Checklist. CONCLUSION: The evidence surrounding the prognostic value of LVEF measurement before vascular surgery is currently weak and inconclusive. Larger scale, prospective studies are required to further refine cardiac risk prediction before vascular surgery.

Vital sign measurements demonstrate terminal digit bias and boundary effects.

OBJECTIVE: The measurement and recording of vital signs may be impacted by biases, including preferences for even and round numbers. However, other biases, such as variation due to defined numerical boundaries (also known as boundary effects), may be present in vital signs data and have not yet been investigated in a medical setting. We aimed to assess vital signs data for such biases. These parameters are clinically significant as they influence care escalation. METHODS: Vital signs data (heart rate, respiratory rate, oxygen saturation and systolic blood pressure) were collected from a tertiary hospital electronic medical record over a 2-year period. These data were analysed using polynomial regression with additional terms to assess for underreporting of out-of-range observations and overreporting numbers with terminal digits of 0 (round numbers), 2 (even numbers) and 5. RESULTS: It was found that heart rate, oxygen saturation and systolic blood pressure demonstrated 'boundary effects', with values inside the 'normal' range disproportionately more likely to be recorded. Even number bias was observed in systolic heart rate, respiratory rate and blood pressure. Preference for multiples of 5 was observed for heart rate and blood pressure. Independent overrepresentation of multiples of 10 was demonstrated in heart rate data. CONCLUSION: Although often considered objective, vital signs data are affected by bias. These biases may impact the care patients receive. Additionally, it may have implications for creating and training machine learning models that utilise vital signs data.