UNASSIGNED: Teaching medical students the skills required to acquire, interpret, apply, and communicate clinical information is an integral part of medical education. A crucial aspect of this process involves providing students with feedback regarding the quality of their free-text clinical notes.
UNASSIGNED: The goal of this study was to assess the ability of ChatGPT 3.5, a large language model, to score medical students\' free-text history and physical notes.
UNASSIGNED: This is a single-institution, retrospective study. Standardized patients learned a prespecified clinical case and, acting as the patient, interacted with medical students. Each student wrote a free-text history and physical note of their interaction. The students\' notes were scored independently by the standardized patients and ChatGPT using a prespecified scoring rubric that consisted of 85 case elements. The measure of accuracy was percent correct.
UNASSIGNED: The study population consisted of 168 first-year medical students. There was a total of 14,280 scores. The ChatGPT incorrect scoring rate was 1.0%, and the standardized patient incorrect scoring rate was 7.2%. The ChatGPT error rate was 86%, lower than the standardized patient error rate. The ChatGPT mean incorrect scoring rate of 12 (SD 11) was significantly lower than the standardized patient mean incorrect scoring rate of 85 (SD 74; P=.002).
UNASSIGNED: ChatGPT demonstrated a significantly lower error rate compared to standardized patients. This is the first study to assess the ability of a generative pretrained transformer (GPT) program to score medical students\' standardized patient-based free-text clinical notes. It is expected that, in the near future, large language models will provide real-time feedback to practicing physicians regarding their free-text notes. GPT artificial intelligence programs represent an important advance in medical education and medical practice.

Intracerebral hemorrhage (ICH) is a critical condition characterized by a high prevalence, substantial mortality rates, and unpredictable clinical outcomes, which results in a serious threat to human health. Improving the timeliness and accuracy of prognosis assessment is crucial to minimizing mortality and long-term disability associated with ICH. Due to the complexity of ICH, the diagnosis of ICH in clinical practice heavily relies on the professional expertise and clinical experience of physicians. Traditional prognostic methods largely depend on the specialized knowledge and subjective judgment of healthcare professionals. Meanwhile, existing artificial intelligence (AI) methodologies, which predominantly utilize features derived from computed tomography (CT) scans, fall short of capturing the multifaceted nature of ICH. Although existing methods are capable of integrating clinical information and CT images for prognosis, the effectiveness of this fusion process still requires improvement. To surmount these limitations, the present study introduces a novel AI framework, termed the ICH Network (ICH-Net), which employs a joint-attention cross-modal network to synergize clinical textual data with CT imaging features. The architecture of ICH-Net consists of three integral components: the Feature Extraction Module, which processes and abstracts salient characteristics from the clinical and imaging data, the Feature Fusion Module, which amalgamates the diverse data streams, and the Classification Module, which interprets the fused features to deliver prognostic predictions. Our evaluation, conducted through a rigorous five-fold cross-validation process, demonstrates that ICH-Net achieves a commendable accuracy of up to 87.77%, outperforming other state-of-the-art methods detailed within our research. This evidence underscores the potential of ICH-Net as a formidable tool in prognosticating ICH, promising a significant advancement in clinical decision-making and patient care.

BACKGROUND: Large language models (LLMs) have shown remarkable capabilities in natural language processing (NLP), especially in domains where labeled data are scarce or expensive, such as the clinical domain. However, to unlock the clinical knowledge hidden in these LLMs, we need to design effective prompts that can guide them to perform specific clinical NLP tasks without any task-specific training data. This is known as in-context learning, which is an art and science that requires understanding the strengths and weaknesses of different LLMs and prompt engineering approaches.
OBJECTIVE: The objective of this study is to assess the effectiveness of various prompt engineering techniques, including 2 newly introduced types-heuristic and ensemble prompts, for zero-shot and few-shot clinical information extraction using pretrained language models.
METHODS: This comprehensive experimental study evaluated different prompt types (simple prefix, simple cloze, chain of thought, anticipatory, heuristic, and ensemble) across 5 clinical NLP tasks: clinical sense disambiguation, biomedical evidence extraction, coreference resolution, medication status extraction, and medication attribute extraction. The performance of these prompts was assessed using 3 state-of-the-art language models: GPT-3.5 (OpenAI), Gemini (Google), and LLaMA-2 (Meta). The study contrasted zero-shot with few-shot prompting and explored the effectiveness of ensemble approaches.
RESULTS: The study revealed that task-specific prompt tailoring is vital for the high performance of LLMs for zero-shot clinical NLP. In clinical sense disambiguation, GPT-3.5 achieved an accuracy of 0.96 with heuristic prompts and 0.94 in biomedical evidence extraction. Heuristic prompts, alongside chain of thought prompts, were highly effective across tasks. Few-shot prompting improved performance in complex scenarios, and ensemble approaches capitalized on multiple prompt strengths. GPT-3.5 consistently outperformed Gemini and LLaMA-2 across tasks and prompt types.
CONCLUSIONS: This study provides a rigorous evaluation of prompt engineering methodologies and introduces innovative techniques for clinical information extraction, demonstrating the potential of in-context learning in the clinical domain. These findings offer clear guidelines for future prompt-based clinical NLP research, facilitating engagement by non-NLP experts in clinical NLP advancements. To the best of our knowledge, this is one of the first works on the empirical evaluation of different prompt engineering approaches for clinical NLP in this era of generative artificial intelligence, and we hope that it will inspire and inform future research in this area.

BACKGROUND: The referral is the basis for radiologists\' assessment of modality, protocol and urgency, and insufficient information may threaten patient safety. The aim of this study was to assess the completeness of referrals for lower extremity venous duplex ultrasonography (LEVDUS) and computed tomography pulmonary angiography (CTPA), and to investigate associations between the provided clinical information including risk factors, symptoms and lab results in the referrals and positive findings of deep vein thrombosis (DVT) and pulmonary embolism (PE), respectively.
METHODS: Referrals for LEVDUS (801) and CTPA (800) performed from 2016 to 2019 were obtained. Three categories of clinical information from the referrals were recorded: symptoms, risk factors and laboratory results, as well as positive imaging findings of venous thromboembolism (VTE). Referral completeness was rated from zero to three according to how many categories of clinical information the referral provided.
RESULTS: Information from all three clinical information categories was provided in 15% and 25% of referrals for LEVDUS and CTPA, respectively, while 2% and 10% of referrals did not contain any clinical information. Symptoms were provided most often (85% for LEVDUS and 94% for CTPA). Provided information about risk factors was significantly associated with positive findings for LEVDUS, (p = 0.02) and CTPA (p < 0.001).
CONCLUSIONS: A great majority of referrals failed to provide one or more categories of clinical information. Risk factors were associated with a positive finding of VTE on LEVDUS and CTPA.
CONCLUSIONS: Improving clinical information in referrals may improve justification, patient safety and quality of radiology services.

Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN) are potentially life-threatening severe cutaneous adverse drug reactions. These diseases are rare, and their onset is difficult to predict because of their idiosyncratic reactivity. The Japan Severe Adverse Reactions Research Group, led by the National Institute of Health Sciences, has operated a nationwide to collect clinical information and genomic samples from patients with SJS/TEN since 2006. This study evaluated the associations of clinical symptoms with sequelae and specific causative drugs/drug groups in Japanese patients with SJS/TEN to identify clinical clues for SJS/TEN treatment and prognosis. Acetaminophen, antibiotics, and carbocisteine were linked to high frequencies of severe ocular symptoms and ocular sequelae (p < 0.05). For erythema and erosion areas, antipyretic analgesics had higher rates of skin symptom affecting <10% of the skin than the other drugs, suggesting narrower lesions (p < 0.004). Hepatic dysfunction, was common in both SJS and TEN, and antiepileptic drugs carried higher risks of hepatic dysfunction than the other drug groups (p = 0.0032). This study revealed that the clinical manifestations of SJS/TEN vary according to the causative drugs.

BACKGROUND: Yaoundé Central Hospital (YCH), located in the capital of Cameroon, is one of the leading referral hospitals in Cameroon. The hospital has several departments, including the Department of Gynecology-Obstetrics (hereinafter referred to as \"the Maternity\"). This clinical department has faced numerous problems with clinical information management, including the lack of high-quality and reliable clinical information, lack of access to this information, and poor use of this information.
OBJECTIVE: We aim to improve the management of clinical information generated at the Maternity at YCH and to describe the challenges, success factors, and lessons learned during its implementation and use.
METHODS: Based on an open-source hospital information system (HIS), this intervention implemented a clinical information system (CIS) at the Maternity at YCH and was carried out using the HERMES model-the first part aimed to cover outpatient consultations, billing, and cash management of the Maternity. Geneva University Hospitals supported this project, and several outcomes were measured at the end. The following outcomes were assessed: project management, technical and organizational aspects, leadership, change management, user training, and system use.
UNASSIGNED: The first part of the project was completed, and the CIS was deployed in the Maternity at YCH. The main technical activities were adapting the open-source HIS to manage outpatient consultations and develop integrated billing and cash management software. In addition to technical aspects, we implemented several other activities. They consisted of the implementation of appropriate project governance or management, improvement of the organizational processes at the Maternity, promotion of the local digital health leadership and performance of change management, and implementation of the training and support of users. Despite barriers encountered during the project, the 6-month evaluation showed that the CIS was effectively used during the first 6 months.
CONCLUSIONS: Implementation of the HIS or CIS is feasible in a resource-limited setting such as Cameroon. The CIS was implemented based on good practices at the Maternity at YCH. This project had successes but also many challenges. Beyond project management and technical and financial aspects, the other main problems of implementing health information systems or HISs in Africa lie in digital health leadership, governance, and change management. This digital health leadership, governance, and change management should prioritize data as a tool for improving productivity and managing health institutions, and promote a data culture among health professionals to support a change in mindset and the acquisition of information management skills. Moreover, in countries with a highly centralized political system like ours, a high-level strategic and political anchor for such projects is often necessary to guarantee their success.

Radiological reporting errors have a direct negative impact on patient treatment. The purpose of this study was to investigate the contribution of clinical information (CI) in radiological reporting of oncological imaging and the dependence on the radiologists’ experience level (EL). Sixty-four patients with several types of carcinomas and twenty patients without tumors were enrolled. Computed tomography datasets acquired in primary or follow-up staging were independently analyzed by three radiologists (R) with different EL (R1: 15 years; R2: 10 years, R3: 1 year). Reading was initially performed without and 3 months later with CI. Overall, diagnostic accuracy and sensitivity for primary tumor detection increased significantly when receiving CI from 77% to 87%; p = 0.01 and 73% to 83%; p = 0.01, respectively. All radiologists benefitted from CI; R1: 85% vs. 92%, p = 0.15; R2: 77% vs. 83%, p = 0.33; R3: 70% vs. 86%, p = 0.02. Overall, diagnostic accuracy and sensitivity for detecting lymphogenous metastases increased from 80% to 85% (p = 0.13) and 42% to 56% (p = 0.13), for detection of hematogenous metastases from 85% to 86% (p = 0.61) and 46% to 60% (p = 0.15). Specificity remained stable (>90%). Thus, CI in oncological imaging seems to be essential for correct radiological reporting, especially for residents, and should be available for the radiologist whenever possible.

BACKGROUND: After the Great East Japan Earthquake in 2011, backup systems for clinical information were launched in Japan. The system in Miyagi Prefecture called the Miyagi Medical and Welfare Information Network (MMWIN) is used as a health information exchange network to share clinical information among various medical facilities for patients who have opted in. Hospitals and clinics specializing in chronic renal failure require patients\' data and records during hemodialysis to facilitate communication in daily clinical activity and preparedness for disasters.
OBJECTIVE: This study aimed to facilitate the sharing of clinical data of patients undergoing hemodialysis among different hemodialysis facilities.
METHODS: We introduced a document-sharing system to make hemodialysis reports available on the MMWIN. We also recruited hospitals and clinics to share the hemodialysis reports of their patients and promoted the development of a network between emergency and dialysis clinics.
RESULTS: In addition to basic patient information as well as information on diagnosis, prescription, laboratory data, hospitalization, allergy, and image data from different facilities, specific information about hemodialysis is available, as well as a backup of indispensable information in preparation for disasters. As of June 1, 2021, 12 clinics and 10 hospitals of 68 dialysis facilities in Miyagi participated in the MMWIN. The number of patients who underwent hemodialysis in Miyagi increased by more than 40%.
CONCLUSIONS: Our backup system successfully developed a network of hemodialysis facilities. We have accumulated data that are beneficial to prevent the fragmentation of patient information and would be helpful in transferring patients efficiently during unpredictable disasters.

To evaluate the performance of a machine learning method to differentiate malignant from benign soft tissue tumors based on textural features on multiparametric magnetic resonance imaging (mpMRI).
We enrolled 163 patients with soft tissue tumors whose diagnosis was pathologically proven (71 malignant, 92 benign). All patients underwent mpMRI. Twelve histographic and textural parameters were assessed on T1-weighted imaging (T1WI), T2-weighted imaging, diffusion-weighted imaging, apparent diffusion coefficient maps, and contrast-enhanced T1WI imaging. We compared mean signals of all sequences from the malignant and benign tumors using Welch\'s t-test. Prediction models were developed via a machine learning technique (support vector machine) using textural features of each sequence, clinical information (sex + age + tumor size), and the combined model incorporating all features. Areas under the receiver operating characteristic curves (AUCs) of these models were calculated using fivefold cross validation.
The diagnostic ability of clinical information model (AUC 0.85) was not inferior to the model with textural features of each sequence (AUC 0.79-0.84). The combined model showed the highest diagnostic ability (AUC 0.89). The AUC of the combined model (0.89) was comparable to those of two board-certified radiologists (0.89 and 0.87).
Machine learning methods based on textural features on mpMRI and clinical information offer adequate diagnostic performance to differentiate between malignant and benign soft tissue tumors.

OBJECTIVE: Alzheimer\'s is the most common irreversible neurodegenerative disease. Its symptoms range from memory impairments to degradation of multiple cognitive abilities and ultimately death. Mild cognitive impairment (MCI) is the earliest detectable stage that happens between normal aging and early dementia, and even though MCI subjects have a chance of changing back to cognitively normal or even staying the same, there is a risk that their condition progresses to Alzheimer\'s disease (AD) annually. Therefore predicting AD among MCI subjects is pivotal for starting treatments at an opportune time in case of progression, and if staying stable is the case, the need for consistent medical observations would eliminate. Thus, we aim to diagnose possible conversion from MCI to AD by exploiting a class of deep learning (DL) methods called convolutional neural network (CNN).
METHODS: We proposed a three-dimensional CNN (3D-CNN) to combine and analyze resting-state functional magnetic resonance imaging (rs-fMRI), clinical assessment results, and demographic information to predict conversion from MCI to AD in an average 5-years interval. Initially, a 3D-CNN was developed based on fMRI single volumes of 266 samples from 81 subjects; then, we used neuron layers to combine clinical data with fMRI to improve the results.
RESULTS: At first, the CNN model demonstrated an AUC of 87.67% and an accuracy of 85.7%, then after combining clinical and rs-fMRI features, we observed the following improved scores: an AUC of 91.72%, an accuracy of 87.6%, a sensitivity of 75.58% and a specificity of 92.57%.
CONCLUSIONS: Our developed algorithm managed to predict prognosis from MCI to AD with high levels of accuracy, proving the potential of DL approaches in solving the matter and the efficiency of integrating clinical information with imaging according to the proposed method.