-
JMIR Research Protocols Jun 2024Artificial intelligence (AI) medical devices have the potential to transform existing clinical workflows and ultimately improve patient outcomes. AI medical devices have...
BACKGROUND
Artificial intelligence (AI) medical devices have the potential to transform existing clinical workflows and ultimately improve patient outcomes. AI medical devices have shown potential for a range of clinical tasks such as diagnostics, prognostics, and therapeutic decision-making such as drug dosing. There is, however, an urgent need to ensure that these technologies remain safe for all populations. Recent literature demonstrates the need for rigorous performance error analysis to identify issues such as algorithmic encoding of spurious correlations (eg, protected characteristics) or specific failure modes that may lead to patient harm. Guidelines for reporting on studies that evaluate AI medical devices require the mention of performance error analysis; however, there is still a lack of understanding around how performance errors should be analyzed in clinical studies, and what harms authors should aim to detect and report.
OBJECTIVE
This systematic review will assess the frequency and severity of AI errors and adverse events (AEs) in randomized controlled trials (RCTs) investigating AI medical devices as interventions in clinical settings. The review will also explore how performance errors are analyzed including whether the analysis includes the investigation of subgroup-level outcomes.
METHODS
This systematic review will identify and select RCTs assessing AI medical devices. Search strategies will be deployed in MEDLINE (Ovid), Embase (Ovid), Cochrane CENTRAL, and clinical trial registries to identify relevant papers. RCTs identified in bibliographic databases will be cross-referenced with clinical trial registries. The primary outcomes of interest are the frequency and severity of AI errors, patient harms, and reported AEs. Quality assessment of RCTs will be based on version 2 of the Cochrane risk-of-bias tool (RoB2). Data analysis will include a comparison of error rates and patient harms between study arms, and a meta-analysis of the rates of patient harm in control versus intervention arms will be conducted if appropriate.
RESULTS
The project was registered on PROSPERO in February 2023. Preliminary searches have been completed and the search strategy has been designed in consultation with an information specialist and methodologist. Title and abstract screening started in September 2023. Full-text screening is ongoing and data collection and analysis began in April 2024.
CONCLUSIONS
Evaluations of AI medical devices have shown promising results; however, reporting of studies has been variable. Detection, analysis, and reporting of performance errors and patient harms is vital to robustly assess the safety of AI medical devices in RCTs. Scoping searches have illustrated that the reporting of harms is variable, often with no mention of AEs. The findings of this systematic review will identify the frequency and severity of AI performance errors and patient harms and generate insights into how errors should be analyzed to account for both overall and subgroup performance.
TRIAL REGISTRATION
PROSPERO CRD42023387747; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=387747.
INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)
PRR1-10.2196/51614.
PubMed: 38941147
DOI: 10.2196/51614 -
JAMA Network Open Jun 2024While adults aged 80 years and older account for 70% of hip fractures in the US, performance of fracture risk assessment tools in this population is uncertain.
IMPORTANCE
While adults aged 80 years and older account for 70% of hip fractures in the US, performance of fracture risk assessment tools in this population is uncertain.
OBJECTIVE
To compare performance of the Fracture Risk Assessment Tool (FRAX), Garvan Fracture Risk Calculator, and femoral neck bone mineral density (FNBMD) alone in 5-year hip fracture prediction.
DESIGN, SETTING AND PARTICIPANTS
Prognostic analysis of 3 prospective cohort studies including participants attending an index examination (1997 to 2016) at age 80 years or older. Data were analyzed from March 2023 to April 2024.
MAIN OUTCOMES AND MEASURES
Participants contacted every 4 or 6 months after index examination to ascertain incident hip fractures and vital status. Predicted 5-year hip fracture probabilities calculated using FRAX and Garvan models incorporating FNBMD and FNBMD alone. Model discrimination assessed by area under receiver operating characteristic curve (AUC). Model calibration assessed by comparing observed vs predicted hip fracture probabilities within predicted risk quintiles.
RESULTS
A total of 8890 participants were included, with a mean (SD) age at index examination of 82.6 (2.7) years; 4906 participants (55.2%) were women, 866 (9.7%) were Black, 7836 (88.1%) were White, and 188 (2.1%) were other races and ethnicities. During 5-year follow-up, 321 women (6.5%) and 123 men (3.1%) experienced a hip fracture; 818 women (16.7%) and 921 men (23.1%) died before hip fracture. Among women, AUC was 0.69 (95% CI, 0.67-0.72) for FRAX, 0.69 (95% CI, 0.66-0.72) for Garvan, and 0.72 (95% CI, 0.69-0.75) for FNBMD alone (FNBMD superior to FRAX, P = .01; and Garvan, P = .01). Among men, AUC was 0.71 (95% CI, 0.66-0.75) for FRAX, 0.76 (95% CI, 0.72-0.81) for Garvan, and 0.77 (95% CI, 0.72-0.81) for FNBMD alone (P < .001 Garvan and FNBMD alone superior to FRAX). Among both sexes, Garvan greatly overestimated hip fracture risk among individuals in upper quintiles of predicted risk, while FRAX modestly underestimated risk among those in intermediate quintiles of predicted risk.
CONCLUSIONS AND RELEVANCE
In this prognostic study of adults aged 80 years and older, FRAX and Garvan tools incorporating FNBMD compared with FNBMD alone did not improve 5-year hip fracture discrimination. FRAX modestly underpredicted observed hip fracture probability in intermediate-risk individuals. Garvan markedly overpredicted observed hip fracture probability in high-risk individuals. Until better prediction tools are available, clinicians should prioritize consideration of hip BMD, life expectancy, and patient preferences in decision-making regarding drug treatment initiation for hip fracture prevention in late-life adults.
PubMed: 38941095
DOI: 10.1001/jamanetworkopen.2024.18612 -
JAMA Health Forum Jun 2024
Topics: Artificial Intelligence; Humans; Patient Safety; Hospitals
PubMed: 38941085
DOI: 10.1001/jamahealthforum.2024.1369 -
Emergency Radiology Jun 2024Traumatic upper extremity injuries are a common cause of emergency department visits, comprising between 10-30% of traumatic injury visits. Timely and accurate...
Traumatic upper extremity injuries are a common cause of emergency department visits, comprising between 10-30% of traumatic injury visits. Timely and accurate evaluation is important to prevent severe complications such as permanent deformities, ischemia, or even death. Computed tomography (CT) and CT angiography (CTA) are the favored non-invasive imaging techniques for assessing upper extremity trauma, playing a crucial role in both the treatment planning and decision-making processes for such injuries. In CT postprocessing, a novel 3D rendering method, cinematic rendering (CR), employs sophisticated lighting models that simulate the interaction of multiple photons with the volumetric dataset. This technique produces images with realistic shadows and improved surface detail, surpassing the capabilities of volume rendering (VR) or maximal intensity projection (MIP). Considering the benefits of CR, we demonstrate its use and ability to achieve photorealistic anatomic visualization in a series of 11 cases where patients presented with traumatic upper extremity injuries, including bone, vascular, and skin/soft tissue injuries, adding to diagnostic confidence and intervention planning.
PubMed: 38941025
DOI: 10.1007/s10140-024-02259-5 -
Journal of Nephrology Jun 2024Implementing Advance Care Planning (ACP) for patients with End-Stage Kidney Disease (ESKD), particularly in the context of hemodialysis, presents significant challenges.... (Review)
Review
Implementing Advance Care Planning (ACP) for patients with End-Stage Kidney Disease (ESKD), particularly in the context of hemodialysis, presents significant challenges. Despite existing legal frameworks, disparities in advance care planning practices are evident across Europe. The present perspective introduces a multidisciplinary model, initiated in 2019. This model incorporates a specialized team comprising a nephrologist, a psychologist, a palliative care specialist, and an anesthesiologist/intensivist. Through this collaborative approach, we aimed to comprehensively address the intricate medical, emotional, and psychological dimensions in advance care planning. In this point of view, we discuss the strengths of our model, its potential for European Nephrology, and advocate for guidelines to enhance advance care planning implementation within the nephrology community.
PubMed: 38941001
DOI: 10.1007/s40620-024-02002-w -
Child's Nervous System : ChNS :... Jun 2024The purpose of this study was to evaluate the surgical complications of patients treated for nonsyndromic sagittal craniosynostosis and the necessity for reoperations...
OBJECTIVE
The purpose of this study was to evaluate the surgical complications of patients treated for nonsyndromic sagittal craniosynostosis and the necessity for reoperations due to craniocerebral disproportion.
MATERIALS AND METHODS
The patient cohort of this study consisted of patients (N = 82) who were treated in the Oulu University Hospital using the open vault cranioplasty with a modified H-technique between the years 2008 to 2022. There were 69 males (84.1%) and 13 females (15.9%). The mean age at the primary operation was 6.1 months. Mean follow-up time was 9.0 years.
RESULTS
There were no major complications related to the procedures. Two patients (2.4%) had a minor dural lesion. There were no postoperative wound infections. Of the 82 patients, seven patients with primary craniosynostosis (13.0%) developed symptomatic craniocerebral disproportion requiring reoperation to increase intracranial volume. In all these patients, invasive intracranial pressure (ICP) monitoring was performed prior to decision-making. In the majority of cases, the aesthetical outcome was considered good or excellent.
CONCLUSION
The operative method used was feasible and safe. Thirteen percent of patients who were followed over 5 years required major surgery due to development of craniocerebral disproportion later in life.
PubMed: 38940955
DOI: 10.1007/s00381-024-06519-0 -
Journal of Imaging Informatics in... Jun 2024Postoperative complications of radical gastrectomy seriously affect postoperative recovery and require accurate risk prediction. Therefore, this study aimed to develop a...
Postoperative complications of radical gastrectomy seriously affect postoperative recovery and require accurate risk prediction. Therefore, this study aimed to develop a prediction model specifically tailored to guide perioperative clinical decision-making for postoperative complications in patients with gastric cancer. A retrospective analysis was conducted on patients who underwent radical gastrectomy at the First Affiliated Hospital of Nanjing Medical University between April 2022 and June 2023. A total of 166 patients were enrolled. Patient demographic characteristics, laboratory examination results, and surgical pathological features were recorded. Preoperative abdominal CT scans were used to segment the visceral fat region of the patients through 3Dslicer, a 3D Convolutional Neural Network (3D-CNN) to extract image features and the LASSO regression model was employed for feature selection. Moreover, an ensemble learning strategy was adopted to train the features and predict postoperative complications of gastric cancer. The prediction performance of the LGBM (Light Gradient Boosting Machine), XGB (XGBoost), RF (Random Forest), and GBDT (Gradient Boosting Decision Tree) models was evaluated through fivefold cross-validation. This study successfully constructed a model for predicting early complications following radical gastrectomy based on the optimal algorithm, LGBM. The LGBM model yielded an AUC value of 0.9232 and an accuracy of 87.28% (95% CI, 75.61-98.95%), surpassing the performance of other models. Through ensemble learning and integration of perioperative clinical data and visceral fat radiomics, a predictive LGBM model was established. This model has the potential to facilitate individualized clinical decision-making and the early recovery of patients with gastric cancer post-surgery.
PubMed: 38940888
DOI: 10.1007/s10278-024-01172-0 -
Alternative Therapies in Health and... Jun 2024This meta-analysis evaluates the diagnostic value of echocardiography for Acute Heart Failure (AHF) and its utility in urgent clinical situations, emphasizing its...
OBJECTIVE
This meta-analysis evaluates the diagnostic value of echocardiography for Acute Heart Failure (AHF) and its utility in urgent clinical situations, emphasizing its significance for accurate and timely diagnosis in critical care.
METHODS
Relevant studies from databases like PubMed and Embase were selected using terms such as 'Ultrasound' and 'acute heart failure'. Inclusion criteria focused on studies evaluating echocardiographic diagnosis in adult patients presenting with symptoms suggestive of AHF. Quality assessment was performed using RevMan 5.3 and QUADAS. Key metrics like sensitivity, specificity, and likelihood ratios were analyzed using STATA 15.1. The types of echocardiography assessed included transthoracic and focused cardiac ultrasound.
RESULTS
Eighteen articles were included, indicating echocardiography's high sensitivity (0.92) and specificity (0.96) in diagnosing AHF. The combined positive likelihood ratio of 23.2 suggests that patients with AHF are over 23 times more likely to have a positive echocardiography result than those without AHF, greatly influencing clinical decision-making toward confirming the diagnosis. The AUC of the SROC curve was 0.98, indicating excellent overall accuracy.
CONCLUSION
Echocardiography is highly accurate in diagnosing AHF, underscored by its critical role in early treatment decisions and potential integration into standard care protocols, thereby preventing adverse outcomes and improving patient management.
PubMed: 38940797
DOI: No ID Found -
Alternative Therapies in Health and... Jun 2024The objective of this study is to develop a prediction model for the pathological upgrading of low-grade dysplasia (LGD) in gastric mucosa. The study aims to compare the...
OBJECTIVE
The objective of this study is to develop a prediction model for the pathological upgrading of low-grade dysplasia (LGD) in gastric mucosa. The study aims to compare the performance of a traditional model based on clinical and endoscopic factors with an enhanced model that incorporates AMACR staining of biopsy tissues.
METHODS
The study utilized a training dataset of 405 LGD cases to establish and compare the traditional and enhanced prediction models. Factors associated with upgrading were identified, and the traditional model was based on these factors. The enhanced model incorporated AMACR staining. The models' performances were evaluated using the area under the curve (AUC), bootstrap resampling, and decision curve analysis. External validation was performed using 171 LGD cases. Statistical techniques such as logistic regression and resampling methods were employed to assess the models' predictive abilities and robustness.
RESULTS
In the training dataset, the traditional model achieved an AUC of 0.824 (95% confidence interval [CI]: 0.783-0.865) for predicting pathological upgrading. However, the enhanced model, which incorporated AMACR staining, exhibited a significantly improved performance with an AUC of 0.878 (95% CI: 0.843-0.913). This increase in AUC by 0.054 (95% CI: 0.015-0.093) demonstrates a statistically significant enhancement provided by the inclusion of AMACR staining in the prediction model for pathological upgrading of LGD lesions in gastric mucosa.
CONCLUSION
The findings of this study highlight the practical implications of the enhanced prediction model incorporating AMACR staining for low-grade gastric mucosal dysplasia (LGD). The significantly improved performance of the enhanced model in predicting pathological upgrading emphasizes its potential to revolutionize the management and treatment strategies for patients with LGD. By providing a more accurate prediction of upgrading, the enhanced model enables early intervention and timely decision-making, leading to improved outcomes and prognosis for patients. The incorporation of AMACR staining in the prediction model holds promise for enhancing diagnostic strategies and reducing the incidence of postoperative pathological upgrading. This research underscores the importance of leveraging advanced techniques to improve the early detection rate of gastric cancer and ultimately benefit patient care.
PubMed: 38940773
DOI: No ID Found -
Journal of Cognitive Neuroscience Jun 2024In value-based decisions, there are frequently multiple attributes, such as cost, quality, or quantity, that contribute to the overall goodness of an option. Because one...
In value-based decisions, there are frequently multiple attributes, such as cost, quality, or quantity, that contribute to the overall goodness of an option. Because one option may not be better in all attributes at once, the decision process should include a means of weighing relevant attributes. Most decision-making models solve this problem by computing an integrated value, or utility, for each option from a weighted combination of attributes. However, behavioral anomalies in decision-making, such as context effects, indicate that other attribute-specific computations might be taking place. Here, we tested whether rhesus macaques show evidence of attribute-specific processing in a value-based decision-making task. Monkeys made a series of decisions involving choice options comprising a sweetness and probability attribute. Each attribute was represented by a separate bar with one of two mappings between bar size and the magnitude of the attribute (i.e., bigger = better or bigger = worse). We found that translating across different mappings produced selective impairments in decision-making. Choices were less accurate and preferences were more variable when like attributes differed in mapping, suggesting that preventing monkeys from easily making direct attribute comparisons resulted in less accurate choice behavior. This was not the case when mappings of unalike attributes within the same option were different. Likewise, gaze patterns favored transitions between like attributes over transitions between unalike attributes of the same option, so that like attributes were sampled sequentially to support within-attribute comparisons. Together, these data demonstrate that value-based decisions rely, at least in part, on directly comparing like attributes of multiattribute options.
PubMed: 38940740
DOI: 10.1162/jocn_a_02208