Can large language models provide secondary reliable opinion on treatment options for dermatological diseases?

OBJECTIVE: To investigate the consistency and reliability of medication recommendations provided by ChatGPT for common dermatological conditions, highlighting the potential for ChatGPT to offer second opinions in patient treatment while also delineating possible limitations. MATERIALS AND METHODS: In this mixed-methods study, we used survey questions in April 2023 for drug recommendations generated by ChatGPT with data from secondary databases, that is, Taiwan's National Health Insurance Research Database and an US medical center database, and validated by dermatologists. The methodology included preprocessing queries, executing them multiple times, and evaluating ChatGPT responses against the databases and dermatologists. The ChatGPT-generated responses were analyzed statistically in a disease-drug matrix, considering disease-medication associations (Q-value) and expert evaluation. RESULTS: ChatGPT achieved a high 98.87% dermatologist approval rate for common dermatological medication recommendations. We evaluated its drug suggestions using the Q-value, showing that human expert validation agreement surpassed Q-value cutoff-based agreement. Varying cutoff values for disease-medication associations, a cutoff of 3 achieved 95.14% accurate prescriptions, 5 yielded 85.42%, and 10 resulted in 72.92%. While ChatGPT offered accurate drug advice, it occasionally included incorrect ATC codes, leading to issues like incorrect drug use and type, nonexistent codes, repeated errors, and incomplete medication codes. CONCLUSION: ChatGPT provides medication recommendations as a second opinion in dermatology treatment, but its reliability and comprehensiveness need refinement for greater accuracy. In the future, integrating a medical domain-specific knowledge base for training and ongoing optimization will enhance the precision of ChatGPT's results.

Adjustable Cuffless Smartphone Attachment (ACSA+) for Estimation of Blood Pressure Trends: A Pilot Study.

Among the elderly, hypertension remains one of the prevalent health conditions, which requires monitoring and intervention strategies. Nevertheless, regular reporting of blood pressure (BP) from these individuals still poses multiple challenges. However, most people own cell phone and are engaged in phone conversations daily. Here, we propose an adjustable cuffless smartphone attachment (ACSA+) equipped with a PPG sensor for the estimation of BP during phone conversations. ACSA+ can be easily attached to the back of any modern cell phone. ACSA+ will help to continuously collect BP data and store it as a trend line.

Artificial Intelligence Approach for Severe Dengue Early Warning System.

Dengue fever is a viral infectious disease transmitted through mosquito bites, and has symptoms ranging from mild flu-like symptoms to deadly complications. Dengue fever is one of the global burden diseases which annually have 50-100 million cases with 500,000 cases of severe dengue fever, of which 22,000 deaths occur mostly in children. Despite the discovery of vaccines, vector control is still the main approach for prevention efforts. Early detection and accessibility to medical care can reduce severe Dengue mortality rate from 50% to 2%. In the previous study, both statistical and machine learning methods have the potential for predicting a Dengue outbreak, but the study is still fragmented and limited on implementing the generated model into an early warning system application. In this study, we developed an artificial intelligence model with spatiotemporal to predict Dengue outbreak and Dengue incidence case which is ready to be implemented into an early warning system application. Indonesia, especially Semarang City, has experienced an endemic Dengue. We used Semarang City spatiotemporal, meteorological, climatological, and Dengue surveillance epidemiology data from January 2014 to December 2021 in 16 districts of Semarang City. We reviewed 7208 samples from 16 districts and 1 city per week during 8 years. The entire dataset was divided into training (80%) and testing (20%) to develop a prediction model. We used machine learning and Long Short Term Memory (LSTM) to predict Dengue outbreak 1 week before the event for each district. and machine learning to predict Dengue incident cases 1 week before the event for each district. Accuracy, area under the receiver operating characteristic curve (AUROC), precision, recall, and F1 score were considered to evaluate the Dengue outbreak prediction model. The Dengue incidence cases prediction model will evaluate using Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-Squared (R2). Extra Trees Classifier model shown outperform in Dengue outbreak prediction, with accuracy 0.8925, AUROC 0. 9529, Recall 0.6117, precision 0.8880, and F1 score 0.7238. CatBoost Regressor model is shown to outperform in Dengue incidence cases prediction, with R2 0.5621, MAE 0.6304, MSE 1.1997, and RMSE 1.0891. The study proves that Artificial Intelligence (AI) with a spatiotemporal approach can give higher performance in Dengue outbreak and incidence cases prediction. Utilization of AI approaches that are sensitive with spatiotemporal feasibility to implement in Dengue early warning system application may contribute to increase the policy makers and community attention to do accurate community-based vector control.

Machine-Learning Based Risk Assessment for Cancer Therapy-Related Cardiac Adverse Events Among Breast Cancer Patients.

The study aims to develop machine-learning models to predict cardiac adverse events in female breast cancer patients who receive adjuvant therapy. We selected breast cancer patients from a retrospective dataset of the Taipei Medical University Clinical Research Database and Taiwan Cancer Registry between January 2004 and December 2020. Patients were monitored at the date of prescribed chemo- and/or -target therapies until cardiac adverse events occurred during a year. Variables were used, including demographics, comorbidities, medications, and lab values. Logistics regression (LR) and artificial neural network (ANN) were used. The performance of the algorithms was measured by the area under the receiver operating characteristic curve (AUC). In total, 1321 patients (an equal 15039 visits) were included. The best performance of the artificial neural network (ANN) model was achieved with the AUC, precision, recall, and F1-score of 0.89, 0.14, 0.82, and 0.2, respectively. The most important features were a pre-existing cardiac disease, tumor size, estrogen receptor (ER), human epidermal growth factor receptor 2 (HER2), cancer stage, and age at index date. Further research is necessary to determine the feasibility of applying the algorithm in the clinical setting and explore whether this tool could improve care and outcomes.

Speech Emotion Recognition Applied to Real-World Medical Consultation.

Since 2020, the COVID-19 epidemic has changed our lives in healthcare behaviors. Forced to wear masks influenced doctor-patient interaction perceptions truly, thus, to build a satisfying relationship is not just empathize with facial expressions. The voice becomes more important for the sake of conquering the burden of masks. Hence, verbal and non-verbal communication will be crucial criteria for doctor-patient interaction during medical consultations and other conversations. In these years, speech emotion recognition has been a popular research domain. In spite of abundant work conducted, nonverbal emotion recognition in medical scenarios is still required to reveal. In this study, we investigate YAMNet transfer learning on Chinese Mandarin speech corpus NTHU-NTUA Chinese Interactive Emotion Corpus (NNIME) and use real-world dermatology clinic recording to test the generalization capability. The results showed that the accuracy validated on NNIME data was 0.59 for activation prediction and 0.57 for valence. Furthermore, the validation accuracy on the doctor-patient dataset was 0.24 for activation and 0.58 for valence, respectively.

Facial Mimicry and Doctor-Patient Satisfaction: The Feasibility of Artificial Empathy in a Clinical Video Data.

Good nonverbal communication between doctor and patient is essential for achieving a successful and therapeutic doctor-patient relationship. Increasing evidence has shown that nonverbal communication mimicry, particularly facial mimicry, where one mirrors another's facial expressions, is linked to empathy and emotion recognition. Empathy is also the key driver of patient satisfaction. This study explores how facial expressions and facial mimicry influence doctor-patient satisfaction during a clinical encounter. We used a facial emotion recognition-based artificial empathy model to analyze 315 recorded clinical video data of doctors and patients in a dermatology outpatient clinic. The results show a significant negative correlation between patients' emotions of sadness and neutral and doctor satisfaction, but no correlation between the duration of doctors mimicking patient emotions and patient satisfaction. These findings provide valuable insights into the future design of systems that can further enhance clinician awareness to maintain communication skills in the search for better doctor-patient satisfaction.

Crowdsource authoring as a tool for enhancing the quality of competency assessments in healthcare professions.

The current Objective Structured Clinical Examination (OSCE) is complex, costly, and difficult to provide high-quality assessments. This pilot study employed a focus group and debugging stage to test the Crowdsource Authoring Assessment Tool (CAAT) for the creation and sharing of assessment tools used in editing and customizing, to match specific users' needs, and to provide higher-quality checklists. Competency assessment international experts (n = 50) were asked to 1) participate in and experience the CAAT system when editing their own checklist, 2) edit a urinary catheterization checklist using CAAT, and 3) complete a Technology Acceptance Model (TAM) questionnaire consisting of 14 items to evaluate its four domains. The study occurred between October 2018 and May 2019. The median time for developing a new checklist using the CAAT was 65.76 minutes whereas the traditional method required 167.90 minutes. The CAAT system enabled quicker checklist creation and editing regardless of the experience and native language of participants. Participants also expressed the CAAT enhanced checklist development with 96% of them willing to recommend this tool to others. The use of a crowdsource authoring tool as revealed by this study has efficiently reduced the time to almost a third it would take when using the traditional method. In addition, it allows collaborations to partake on a simple platform which also promotes contributions in checklist creation, editing, and rating.

Using alert dwell time to filter universal clinical alerts: A machine learning approach.

BACKGROUND: Alerts in computerized physician order entry (CPOE) systems can improve patient safety. However, alerts in rule-based systems cannot be customized based on individual patient or user characteristics. This limitation can lead to the presentation of irrelevant alerts and subsequent alert fatigue. OBJECTIVE: We used machine learning approaches with alert dwell time to filter out irrelevant alerts for physicians based on contextual factors. METHODS: We utilized five machine learning algorithms and a total of 1,120 features grouped into six categories: alert, demographic, environment, diagnosis, prescription, and laboratory results. The output of the models was the alert dwell time within a specified time window to determine the optimal range by the sensitivity analysis. RESULTS: We used 813,026 records (19 categories) from the hospital's outpatient clinic data from 2020 to 2021. The sensitivity analysis showed that a time window with a range of 0.3-4.0 s had the best performance, with an area under the receiver operating characteristic (AUROC) curve of 0.73 and an area under the precision-recall curve (AUPRC) of 0.97. The model built with alert and demographic feature groups showed the best performance, with an AUROC of 0.73. The most significant individual feature groups were alert and demographic, with AUROCs of 0.66 and 0.62, respectively. CONCLUSION: Our study found that alerts and user and patient demographic features are more crucial than clinical features when constructing universal context-aware alerts. Using alert dwell time in combination with a time window is an effective way to determine the trigger status of an alert. The findings of this study can provide useful insights for researchers working on specific and universal context-aware alerts.

Incidence and Mortality of Malignant Brain Tumors after 20 Years of Mobile Use.

(1) Objective: This population-based study was performed to examine the trends of incidence and deaths due to malignant neoplasm of the brain (MNB) in association with mobile phone usage for a period of 20 years (January 2000-December 2019) in Taiwan. (2) Methods: Pearson correlation, regression analysis, and joinpoint regression analysis were used to examine the trends of incidence of MNB and deaths due to MNB in association with mobile phone usage. (3) Results: The findings indicate a trend of increase in the number of mobile phone users over the study period, accompanied by a slight rise in the incidence and death rates of MNB. The compound annual growth rates further support these observations, highlighting consistent growth in mobile phone users and a corresponding increase in MNB incidences and deaths. (4) Conclusions: The results suggest a weaker association between the growing number of mobile phone users and the rising rates of MNB, and no significant correlation was observed between MNB incidences and deaths and mobile phone usage. Ultimately, it is important to acknowledge that conclusive results cannot be drawn at this stage and further investigation is required by considering various other confounding factors and potential risks to obtain more definitive findings and a clearer picture.

Machine Learning Approaches for Predicting Psoriatic Arthritis Risk Using Electronic Medical Records: Population-Based Study.

BACKGROUND: Psoriasis (PsO) is a chronic, systemic, immune-mediated disease with multiorgan involvement. Psoriatic arthritis (PsA) is an inflammatory arthritis that is present in 6%-42% of patients with PsO. Approximately 15% of patients with PsO have undiagnosed PsA. Predicting patients with a risk of PsA is crucial for providing them with early examination and treatment that can prevent irreversible disease progression and function loss. OBJECTIVE: The aim of this study was to develop and validate a prediction model for PsA based on chronological large-scale and multidimensional electronic medical records using a machine learning algorithm. METHODS: This case-control study used Taiwan's National Health Insurance Research Database from January 1, 1999, to December 31, 2013. The original data set was split into training and holdout data sets in an 80:20 ratio. A convolutional neural network was used to develop a prediction model. This model used 2.5-year diagnostic and medical records (inpatient and outpatient) with temporal-sequential information to predict the risk of PsA for a given patient within the next 6 months. The model was developed and cross-validated using the training data and was tested using the holdout data. An occlusion sensitivity analysis was performed to identify the important features of the model. RESULTS: The prediction model included a total of 443 patients with PsA with earlier diagnosis of PsO and 1772 patients with PsO without PsA for the control group. The 6-month PsA risk prediction model that uses sequential diagnostic and drug prescription information as a temporal phenomic map yielded an area under the receiver operating characteristic curve of 0.70 (95% CI 0.559-0.833), a mean sensitivity of 0.80 (SD 0.11), a mean specificity of 0.60 (SD 0.04), and a mean negative predictive value of 0.93 (SD 0.04). CONCLUSIONS: The findings of this study suggest that the risk prediction model can identify patients with PsO at a high risk of PsA. This model may help health care professionals to prioritize treatment for target high-risk populations and prevent irreversible disease progression and functional loss.

Emotion recognition in doctor-patient interactions from real-world clinical video database: Initial development of artificial empathy.

BACKGROUND AND OBJECTIVE: The promising use of artificial intelligence (AI) to emulate human empathy may help a physician engage with a more empathic doctor-patient relationship. This study demonstrates the application of artificial empathy based on facial emotion recognition to evaluate doctor-patient relationships in clinical practice. METHODS: A prospective study used recorded video data of doctor-patient clinical encounters in dermatology outpatient clinics, Taipei Municipal Wanfang Hospital, and Taipei Medical University Hospital collected from March to December 2019. Two cameras recorded the facial expressions of four doctors and 348 adult patients during regular clinical practice. Facial emotion recognition was used to analyze the basic emotions of doctors and patients with a temporal resolution of 1 second. In addition, a physician-patient satisfaction questionnaire was administered after each clinical session, and two standard patients gave impartial feedback to avoid bias. RESULTS: Data from 326 clinical session videos showed that (1) Doctors expressed more emotions than patients (t [326] > = 2.998, p < = 0.003), including anger, happiness, disgust, and sadness; the only emotion that patients showed more than doctors was surprise (t [326] = -4.428, p < .001) (p < .001). (2) Patients felt happier during the latter half of the session (t [326] = -2.860, p = .005), indicating a good doctor-patient relationship. CONCLUSIONS: Artificial empathy can offer objective observations on how doctors' and patients' emotions change. With the ability to detect emotions in 3/4 view and profile images, artificial empathy could be an accessible evaluation tool to study doctor-patient relationships in practical clinical settings.

Deep Learning for the Diagnosis of Esophageal Cancer in Endoscopic Images: A Systematic Review and Meta-Analysis.

Esophageal cancer, one of the most common cancers with a poor prognosis, is the sixth leading cause of cancer-related mortality worldwide. Early and accurate diagnosis of esophageal cancer, thus, plays a vital role in choosing the appropriate treatment plan for patients and increasing their survival rate. However, an accurate diagnosis of esophageal cancer requires substantial expertise and experience. Nowadays, the deep learning (DL) model for the diagnosis of esophageal cancer has shown promising performance. Therefore, we conducted an updated meta-analysis to determine the diagnostic accuracy of the DL model for the diagnosis of esophageal cancer. A search of PubMed, EMBASE, Scopus, and Web of Science, between 1 January 2012 and 1 August 2022, was conducted to identify potential studies evaluating the diagnostic performance of the DL model for esophageal cancer using endoscopic images. The study was performed in accordance with PRISMA guidelines. Two reviewers independently assessed potential studies for inclusion and extracted data from retrieved studies. Methodological quality was assessed by using the QUADAS-2 guidelines. The pooled accuracy, sensitivity, specificity, positive and negative predictive value, and the area under the receiver operating curve (AUROC) were calculated using a random effect model. A total of 28 potential studies involving a total of 703,006 images were included. The pooled accuracy, sensitivity, specificity, and positive and negative predictive value of DL for the diagnosis of esophageal cancer were 92.90%, 93.80%, 91.73%, 93.62%, and 91.97%, respectively. The pooled AUROC of DL for the diagnosis of esophageal cancer was 0.96. Furthermore, there was no publication bias among the studies. The findings of our study show that the DL model has great potential to accurately and quickly diagnose esophageal cancer. However, most studies developed their model using endoscopic data from the Asian population. Therefore, we recommend further validation through studies of other populations as well.

Proton Pump Inhibitors Use and the Risk of Pancreatic Cancer: Evidence from Eleven Epidemiological Studies, Comprising 1.5 Million Individuals.

Previous epidemiological studies have shown that proton pump inhibitor (PPI) may modify the risk of pancreatic cancer. We conducted an updated systematic review and meta-analysis of observational studies assessing the effect of PPI on pancreatic cancer. PubMed, Embase, Scopus, and Web of Science were searched for studies published between 1 January 2000, and 1 May 2022. We only included studies that assessed exposure to PPI, reported pancreatic cancer outcomes, and provided effect sizes (hazard ratio or odds ratio) with 95% confidence intervals (CIs). We calculated an adjusted pooled risk ratio (RR) with 95%CIs using the random-effects model. Eleven studies (eight case-control and three cohorts) that reported 51,629 cases of pancreatic cancer were included. PPI was significantly associated with a 63% increased risk of pancreatic cancer (RRadj. 1.63, 95%CI: 1.19-2.22, p = 0.002). Subgroup analysis showed that the pooled RR for rabeprazole and lansoprazole was 4.08 (95%CI: 0.61-26.92) and 2.25 (95%CI: 0.83-6.07), respectively. Moreover, the risk of pancreatic cancer was established for both the Asian (RRadj. 1.37, 95%CI: 0.98-1.81) and Western populations (RRadj.2.76, 95%CI: 0.79-9.56). The findings of this updated meta-analysis demonstrate that the use of PPI was associated with an increased risk of pancreatic cancer. Future studies are needed to improve the quality of evidence through better verification of PPI status (e.g., patient selection, duration, and dosages), adjusting for possible confounders, and ensuring long-term follow-up.

Temporal Phenomics - A Powerful Approach Using AI to Achieve "Earlier Medicine".

The resurgence of machine learning AI has triggered the importance of collecting "personal big data" over a long period of time from wearable devices and EHRs. Collecting data from this large number of variables over a significant period of time has further induced the study on "Temporal Phenomics", which can be a powerful approach to achieve pre-emptive and "earlier medicine". The paper presents a methodology to make studying "Temporal Phenomics" more feasible and convenient without limitations on the number of variables and the length of time periods.

SlimMe, a Chatbot With Artificial Empathy for Personal Weight Management: System Design and Finding.

As the obesity rate continues to increase persistently, there is an urgent need to develop an effective weight loss management strategy. Nowadays, the development of artificial intelligence (AI) and cognitive technologies coupled with the rapid spread of messaging platforms and mobile technology with easier access to internet technology offers professional dietitians an opportunity to provide extensive monitoring support to their clients through a chatbot with artificial empathy. This study aimed to design a chatbot with artificial empathic motivational support for weight loss called "SlimMe" and investigate how people react to a diet bot. The SlimMe infrastructure was built using Dialogflow as the natural language processing (NLP) platform and LINE mobile messenger as the messaging platform. We proposed a text-based emotion analysis to simulate artificial empathy responses to recognize the user's emotion. A preliminary evaluation was performed to investigate the early-stage user experience after a 7-day simulation trial. The result revealed that having an artificially empathic diet bot for weight loss management is a fun and exciting experience. The use of emoticons, stickers, and GIF images makes the chatbot response more interactive. Moreover, the motivational support and persuasive messaging features enable the bot to express more empathic and engaging responses to the user. In total, there were 1,007 bot responses from 892 user input messages. Of these, 67.38% (601/1,007) of the chatbot-generated responses were accurate to a relevant user request, 21.19% (189/1,007) inaccurate responses to a relevant request, and 10.31% (92/1,007) accurate responses to an irrelevant request. Only 1.12% (10/1,007) of the chatbot does not answer. We present the design of an artificially empathic diet bot as a friendly assistant to help users estimate their calorie intake and calories burned in a more interactive and engaging way. To our knowledge, this is the first chatbot designed with artificial empathy features, and it looks very promising in promoting long-term weight management. More user interactions and further data training and validation enhancement will improve the bot's in-built knowledge base and emotional intelligence base.

Proton Pump Inhibitor Use and Risk of Gastric Cancer: Current Evidence from Epidemiological Studies and Critical Appraisal.

Proton pump inhibitors (PPIs) are used for maintaining or improving gastric problems. Evidence from observational studies indicates that PPI therapy is associated with an increased risk of gastric cancer. However, the evidence for PPIs increasing the risk of gastric cancer is still being debated. Therefore, we aimed to investigate whether long-term PPI use is associated with an increased risk of gastric cancer. We systematically searched the relevant literature in electronic databases, including PubMed, EMBASE, Scopus, and Web of Science. The search and collection of eligible studies was between 1 January 2000 and 1 July 2021. Two independent authors were responsible for the study selection process, and they considered only observational studies that compared the risk of gastric cancer with PPI treatment. We extracted relevant information from selected studies, and assessed the quality using the Newcastle−Ottawa scale (NOS). Finally, we calculated overall risk ratios (RRs) with 95% confidence intervals (CIs) of gastric cancer in the group receiving PPI therapy and the control group. Thirteen observational studies, comprising 10,557 gastric cancer participants, were included. Compared with patients who did not take PPIs, the pooled RR for developing gastric cancer in patients receiving PPIs was 1.80 (95% CI, 1.46−2.22, p < 0.001). The overall risk of gastric cancer also increased in patients with gastroesophageal reflux disease (GERD), H. pylori treatment, and various adjusted factors. The findings were also consistent across several sensitivity analyses. PPI use is associated with an increased risk of gastric cancer in patients compared with those with no PPI treatment. The findings of this updated study could be used in making clinical decisions between physicians and patients about the initiation and continuation of PPI therapy, especially in patients at high risk of gastric cancer. Additionally, large randomized controlled trials are needed to determine whether PPIs are associated with a higher risk of gastric cancer.