Journal of Medical Systems最新文献

Decision-making in chronic diseases guided by clinical decision support systems that use models including multiple variables based on artificial intelligence requires scientific validation in different populations to optimize the use of limited human, financial, and clinical resources in healthcare systems worldwide. This cohort study evaluated three machine learning algorithms-XGBoost, Elastic Net logistic regression, and an Artificial Neural Network-to develop a prediction model for three outcomes: mortality, hospitalization, and emergency department visits. The objective was to build a clinical decision support system for patients with noncommunicable diseases treated at the Alma Mater Hospital complex in Medellín, Colombia. We collected 4845 electronic medical record entries from 5000 patients included in the study. The median age was 71.83 years, with 63.8% women and 29.7% receiving home care. The most prevalent medical conditions were diabetes (52.9%), hypertension (67.2%), dyslipidemia (57.3%), and COPD (19.4%). For mortality prediction, the Elastic Net logistic regression model achieved an AUCROC of 0.883 (95% CI: 0.848-0.917), the XGBoost model reached an AUCROC of 0.896 (95% CI: 0.865-0.927), and the Neural Network achieved 0.886 (95% CI: 0.853-0.916). For hospitalization, the Elastic Net model had an AUCROC of 0.952 (95% CI: 0.937-0.965), the XGBoost model achieved 0.963 (95% CI: 0.952-0.974), and the Neural Network scored 0.932 (95% CI: 0.915-0.948). For emergency department visits, the AUCROC values were 0.980 (95% CI: 0.971-0.987) for Elastic Net, 0.977 (95% CI: 0.967-0.986) for XGBoost, and 0.976 (95% CI: 0.968-0.982) for the neural network. A dashboard was developed to interact with an ensemble risk categorization segmenting patient risk in the cohort to aid in clinical decision-making. A clinical decision support system based on artificial intelligence using electronic medical records possibly can help segmenting the risk in populations with Noncommunicable Diseases for effective decision-making.

Purpose: Mobile health plays an important role in providing individualized information about the health status of patients. Limited information exists on intensive care unit (ICU) patients with the risk of suffering from the post-intensive care syndrome (PICS), summarizing long-term physical, mental and cognitive impairment. This web-based survey study aims to identify specific needs of former ICU patients for utilizing a newly developed, so called Post-Intensive Care Outcome Surveillance (PICOS) app to collect relevant PICS-related parameters.

Methods: A prototype app was developed following interaction principles for interactive systems of usability engineering. Patients from four different German hospitals were asked about demographics, interaction with technology and their perception of the prototype regarding hedonic motivation, perceived ease of use and performance expectancy.

Results: 123 patients participated in the survey; the majority owned and used smartphones. Nearly half of respondents would seek help from family members or caregivers using the app. There was a difference in affinity for technology for participants who own a smartphone and those who do not, t(116) = - 0.97, p = .335, and no significant difference in affinity for technology whether the participants would like support when using the app or not, t(97) = 1.81, p = .073. The average hedonic motivation for using the app was M = 4.44 (SD = 1.304).

Conclusion: This app prototype was perceived as both beneficial and easy to use, indicating its success among former ICU patients. Due to aging and ongoing health impairments, every second patient would need assistance with the initial use of the app.

Medical devices significantly enhance healthcare by integrating advanced technology to improve patient outcomes. Ensuring their safety and reliability requires a delicate balance between innovation and rigorous oversight, managed through the collaborative efforts of standards development organizations, standards accrediting organizations, and regulatory agencies such as the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA). This article explores the historical evolution of medical device regulation, the role of standards organizations, and the impact of regulatory practices on device safety. Highlighting the critical need for stringent regulations, informed by instances where medical devices caused patient harm, we discuss the processes and collaborations between various international standards and regulatory frameworks that ensure device safety and effectiveness. This comprehensive review addresses the complexities of regulatory compliance and standardization, aiming to bridge the knowledge gap among healthcare providers and enhance the implementation of safety standards in medical technology.

Predicting health-related outcomes can help with proactive healthcare planning and resource management. This is especially important on the older population, an age group growing in the coming decades. Considering longitudinal rather than cross-sectional information from primary care electronic health records (EHRs) can contribute to more informed predictions. In this work, we developed prediction models using longitudinal EHRs to inform resource allocation. In this study, we developed deep-learning-based prognostic models to predict 1-year and 5-year all-cause mortality, nursing home admission, and home care need in people over 65 years old using all the longitudinal information from EHRs. The models included attention mechanisms to increase their transparency. EHRs were drawn from SIDIAP (primary care, Catalonia (Spain)) from 2010-2019. Performance on the test set was compared to that from baseline models using cross-sectional one-year history only. Data from 1,456,052 individuals over 65 years old were considered. Cohen's kappa obtained using longitudinal data was 3.4-fold (1-year all-cause mortality), 10.3-fold (5-year all-cause mortality), 1.1-fold (5-year nursing home admission), and 1.2-fold (5-year home care need) higher than that obtained by the one-year history baseline models. Our models performed better than those not considering longitudinal data, especially when predicting further into the future. However, nursing home admission and home care need in the long term were harder to predict, suggesting their dependence on more abrupt changes. The attention maps helped to understand the predictions, enhancing model transparency. These prediction models can contribute to improve resource allocation in the general population of aging adults.

Large language models (LLMs) have been utilized to automate tasks like writing discharge summaries and operative reports in neurosurgery. The present study evaluates their ability to identify current procedural terminology (CPT) codes from operative reports. Three LLMs (ChatGPT 4.0, AtlasGPT and Gemini) were evaluated in their ability to provide CPT codes for diagnostic or interventional procedures in endovascular neurosurgery at a single institution. Responses were classified as correct, partially correct or incorrect, and the percentage of correctly identified CPT codes were calculated. The Chi-Square test and Kruskal Wallis test were used to compare responses across LLMs. A total of 30 operative notes were used in the present study. AtlasGPT identified CPT codes for 98.3% procedures with partially correct responses, while ChatGPT and Gemini provided partially correct responses for 86.7% and 30% procedures, respectively (P < 0.001). AtlasGPT identified CPT codes correctly in an average of 35.3% of procedures, followed by ChatGPT (35.1%) and Gemini (8.9%) (P < 0.001). A pairwise comparison among three LLMs revealed that AtlasGPT and ChatGPT outperformed Gemini. Untrained LLMs have the ability to identify partially correct CPT codes in endovascular neurosurgery. Training these models could further enhance their ability to identify CPT codes and minimize healthcare expenditure.

Effective referral triaging enhances patient service outcomes, experience and access to care especially for specialized procedures. This study presents the development and implementation of an automated triaging system to predict patients who would benefit from Spinal Cord Stimulation (SCS) procedure for their pain management. The proposed triage system aims to improve the triage process by reducing unnecessary appointments before SCS assessment, ensuring appropriate pain management care. It compares various machine learning techniques for the prediction while addressing the class imbalance and overlap challenges inherent in the data. Both data-level and algorithm-level approaches were explored. Two years of patient data was collected including patient characteristics, diagnosis history, pain symptoms, appointment history, medication history, and concepts from clinical notes extracted using Natural Language Processing. EasyEnsemble with Ada Boosting method, an algorithm-level approach, showed the most promising results. The tenfold validation indicated the average area under curve of 0.82, true positive rate (TPR) of 77.3%, and true negative rate (TNR) of 73.0%. The probability threshold was adjusted to 0.575 to meet practice expectation of 15% or less on false positive rate (FPR). The implementation pipeline for the selected model was designed to be applicable to real clinical settings. The one-year implementation results showed TPR of 64.7% and TNR of 87.2%, which reduced FPR by 12.8% while reduced TPR by 12.6%. The trade-off was acceptable to practice. The proposed triage system demonstrated promising accuracy, leading to the enhancement of scheduling systems, patient care, and the reduction of unnecessary appointments in a pain medicine setting.

Low medication adherence poses a great risk of poor treatment outcomes among patients with chronic diseases. Recently, mobile applications (apps) have been recognized as effective interventions, enabling patients to adhere to their prescriptions. This study aimed to establish the effectiveness of mobile app interventions for medication adherence, affecting features, and dropout rates by focusing on previous randomized controlled trials (RCTs). This study conducted a systematic review and meta-analysis of mobile app interventions targeting medication adherence in patients with chronic diseases. Electronic searches of eight databases were conducted on April 21, 2023, for studies published between 2013 and 2023. Comprehensive meta-analysis software was used to estimate the standardized mean difference (SMD) of pooled outcomes, odds ratios (ORs), and confidence intervals (CIs). Subgroup analysis was applied to investigate and compare the effectiveness of the interventional strategies and their features. The risk of bias of the included RCTs was evaluated by applying the risk of bias tool. Publication bias was examined using the fail-safe N method. Twenty-six studies with 5,174 participants were included (experimental group 2603, control group 2571). The meta-analysis findings showed a positive impact of mobile apps on improving medication adherence (OR = 2.371, SMD = 0.279). The subgroup analysis results revealed greater effectiveness of interventions using interactive strategies (OR = 2.652, SMD = 0.283), advanced reminders (OR = 1.849, SMD = 0.455), data-sharing (OR = 2.404, SMD = 0.346), and pill dispensers (OR = 2.453). The current study found that mobile interventions had significant effects on improving medication adherence. Subgroup analysis showed that the roles of stakeholders in health providers' interactions with patients and developers' understanding of patients and disease characteristics are critical. Future studies should incorporate advanced technology reflecting acceptability and the needs of the target population.

Anesthetic gases contribute to global warming. We described a two-year performance improvement project to examine the association of individualized provider dashboard feedback of anesthetic gas carbon dioxide equivalent (CDE₂₀) production and median perioperative fresh gas flows (FGF) during general anesthetics during perioperative management. Using a custom structured query language (SQL) query, hourly CDE₂₀ for each anesthetic gas and median FGF were determined. During the first year, practitioners were not given any feedback on their use of anesthetic gases. During the second year of the study protocol, a commercially available business intelligence platform was used to deliver individualized monthly dashboard of these parameters to each practitioner. Continuous values are expressed as median [first quartile, third quartile]. During the study period, 53,294 patients managed by 79 anesthesiologists were available for analysis. Bivariate analysis revealed an overall decrease in median FGF from 2.0 [1.9, 3.0] liters/minute (l/min) to 1.9 [1.7, 2.0] l/min (p < 0.001). There was a significant decrease in the overall total CDE₂₀ from 5.10 [0,12.3] to 3.59 [0,8.78] kg/hr (p < 0.001). Multivariate analysis demonstrated an initial decrease in monthly practitioner total CDE₂₀ production with the intervention (odds ratio (OR) 0.875 95% confidence interval (CI) 0.809-0.996, p < 0.001) and a faster decrease rate in monthly total CDE₂₀ (OR 0.986, 95% CI 0.976-0,996, p < 0.001). Dashboard distribution initially decreased isoflurane (intervention OR 0.97 95% CI 0.96-0.99, p = 0.001) and N₂O (OR 0.82 95% CI 0.73-0.94, p = 0.003) CDE₂₀ production and was associated with a steeper declining rate of isoflurane (OR 0.87, CI 0.79-0.94, p < 0.001) and desflurane (OR 0.9, 0.84-0.97, p = 0.005) CDE₂₀ production. The intervention did not have a significant effect on the monthly rate of decline of sevoflurane or N₂O CDE₂₀. The average practitioner FGF decreased by 0.3 l/m (95% confidence interval (CI): -0,011, -0.5, p = 0.002) with dashboard distributions. Dashboard distribution may be an effective tool to decrease FGF as well as components of anesthetic greenhouse gas emissions.

With the rise of AI platforms, patients increasingly use them for information, relying on advanced language models like ChatGPT for answers and advice. However, the effectiveness of ChatGPT in educating thyroid cancer patients remains unclear. We designed 50 questions covering key areas of thyroid cancer management and generated corresponding responses under four different prompt strategies. These answers were evaluated based on four dimensions: accuracy, comprehensiveness, human care, and satisfaction. Additionally, the readability of the responses was assessed using the Flesch-Kincaid grade level, Gunning Fog Index, Simple Measure of Gobbledygook, and Fry readability score. We also statistically analyzed the references in the responses generated by ChatGPT. The type of prompt significantly influences the quality of ChatGPT's responses. Notably, the "statistics and references" prompt yields the highest quality outcomes. Prompts tailored to a "6th-grade level" generated the most easily understandable text, whereas responses without specific prompts were the most complex. Additionally, the "statistics and references" prompt produced the longest responses while the "6th-grade level" prompt resulted in the shortest. Notably, 87.84% of citations referenced published medical literature, but 12.82% contained misinformation or errors. ChatGPT demonstrates considerable potential for enhancing the readability and quality of thyroid cancer patient education materials. By adjusting prompt strategies, ChatGPT can generate responses that cater to diverse patient needs, improving their understanding and management of the disease. However, AI-generated content must be carefully supervised to ensure that the information it provides is accurate.

Generative Artificial Intelligence (Gen AI) has transformative potential in healthcare to enhance patient care, personalize treatment options, train healthcare professionals, and advance medical research. This paper examines various clinical and non-clinical applications of Gen AI. In clinical settings, Gen AI supports the creation of customized treatment plans, generation of synthetic data, analysis of medical images, nursing workflow management, risk prediction, pandemic preparedness, and population health management. By automating administrative tasks such as medical documentations, Gen AI has the potential to reduce clinician burnout, freeing more time for direct patient care. Furthermore, application of Gen AI may enhance surgical outcomes by providing real-time feedback and automation of certain tasks in operating rooms. The generation of synthetic data opens new avenues for model training for diseases and simulation, enhancing research capabilities and improving predictive accuracy. In non-clinical contexts, Gen AI improves medical education, public relations, revenue cycle management, healthcare marketing etc. Its capacity for continuous learning and adaptation enables it to drive ongoing improvements in clinical and operational efficiencies, making healthcare delivery more proactive, predictive, and precise.