Prediction of subsequent fragility fractures: application of machine learning.

BACKGROUND: Machine learning (ML) has shown exceptional promise in various domains of medical research. However, its application in predicting subsequent fragility fractures is still largely unknown. In this study, we aim to evaluate the predictive power of different ML algorithms in this area and identify key features associated with the risk of subsequent fragility fractures in osteoporotic patients. METHODS: We retrospectively analyzed data from patients presented with fragility fractures at our Fracture Liaison Service, categorizing them into index fragility fracture (n = 905) and subsequent fragility fracture groups (n = 195). We independently trained ML models using 27 features for both male and female cohorts. The algorithms tested include Random Forest, XGBoost, CatBoost, Logistic Regression, LightGBM, AdaBoost, Multi-Layer Perceptron, and Support Vector Machine. Model performance was evaluated through 10-fold cross-validation. RESULTS: The CatBoost model outperformed other models, achieving 87% accuracy and an AUC of 0.951 for females, and 93.4% accuracy with an AUC of 0.990 for males. The most significant predictors for females included age, serum C-reactive protein (CRP), 25(OH)D, creatinine, blood urea nitrogen (BUN), parathyroid hormone (PTH), femoral neck Z-score, menopause age, number of pregnancies, phosphorus, calcium, and body mass index (BMI); for males, the predictors were serum CRP, femoral neck T-score, PTH, hip T-score, BMI, BUN, creatinine, alkaline phosphatase, and spinal Z-score. CONCLUSION: ML models, especially CatBoost, offer a valuable approach for predicting subsequent fragility fractures in osteoporotic patients. These models hold the potential to enhance clinical decision-making by supporting the development of personalized preventative strategies.

Machine learning-based mortality prediction models for smoker COVID-19 patients.

BACKGROUND: The large number of SARS-Cov-2 cases during the COVID-19 global pandemic has burdened healthcare systems and created a shortage of resources and services. In recent years, mortality prediction models have shown a potential in alleviating this issue; however, these models are susceptible to biases in specific subpopulations with different risks of mortality, such as patients with prior history of smoking. The current study aims to develop a machine learning-based mortality prediction model for COVID-19 patients that have a history of smoking in the Iranian population. METHODS: A retrospective study was conducted across six medical centers between 18 and 2020 and 15 March 2022, comprised of 678 CT scans and laboratory-confirmed COVID-19 patients that had a history of smoking. Multiple machine learning models were developed using 10-fold cross-validation. The target variable was in-hospital mortality and input features included patient demographics, levels of care, vital signs, medications, and comorbidities. Two sets of models were developed for at-admission and post-admission predictions. Subsequently, the top five prediction models were selected from at-admission models and post-admission models and their probabilities were calibrated. RESULTS: The in-hospital mortality rate for smoker COVID-19 patients was 20.1%. For "at admission" models, the best-calibrated model was XGBoost which yielded an accuracy of 87.5% and F1 score of 86.2%. For the "post-admission" models, XGBoost also outperformed the rest with an accuracy of 90.5% and F1 score of 89.9%. Active smoking was among the most important features in patients' mortality prediction. CONCLUSION: Our machine learning-based mortality prediction models have the potential to be adapted for improving the management of smoker COVID-19 patients and predicting patients' chance of survival.

Establishment of Iran Musculoskeletal Tumor Registry: A Study Protocol and Lessons Learned from Implementation and the Pilot Phase.

Background: The number of available musculoskeletal tumor registries is relatively small. We developed a registry system focused on the clinical aspects of musculoskeletal tumors to improve quality of care indexes through the development of updated national protocols. In this study, we describe our protocol, challenges, and the data collected during the implementation of the registry system in a single-specialty orthopedic center in Iran. Methods: Three main malignant bone tumors, including osteosarcoma, Ewing sarcoma, and chondrosarcoma, were included in the registry. After establishing a steering committee, we defined the minimum data set based on a literature review and suggestions from an expert panel. Accordingly, the data collection forms and the web-based software were developed. The collected information was categorized into 9 classes, including demographics, socioeconomic data, signs and symptoms, past medical history, family history, laboratory tests, tumor characteristics, primary treatment, and follow-up. Data collection was performed both retrospectively and prospectively. Results: Until September 21, 2022, a total of 71 patients were registered (21 patients prospectively and 50 patients retrospectively) and consisted of 36 (50.7%) cases of osteosarcoma, 13 (18.3%) cases of Ewing sarcoma, and 22 (31%) cases of chondrosarcoma. The implementation of the registry demonstrated promising data regarding the tumor characteristics, delay patterns, and socioeconomic status of the patients. Conclusion: The main lessons learned were to develop a monitoring system to make sure that the new staff is adequately trained for the registration process as well as avoid the inclusion of time-consuming useless data in the minimum data set.

The Persian Version of the Mobile Application Rating Scale (MARS-Fa): Translation and Validation Study.

BACKGROUND: Approximately 110 million Farsi speakers worldwide have access to a growing mobile app market. Despite restrictions and international sanctions, Iran's internal mobile health app market is growing, especially for Android-based apps. However, there is a need for guidelines for developing health apps that meet international quality standards. There are also no tools in Farsi that assess health app quality. Developers and researchers who operate in Farsi could benefit from such quality assessment tools to improve their outputs. OBJECTIVE: This study aims to translate and culturally adapt the Mobile Application Rating Scale in Farsi (MARS-Fa). This study also evaluates the validity and reliability of the newly developed MARS-Fa tool. METHODS: We used a well-established method to translate and back translate the MARS-Fa tool with a group of Iranian and international experts in Health Information Technology and Psychology. The final translated version of the tool was tested on a sample of 92 apps addressing smartphone addiction. Two trained reviewers completed an independent assessment of each app in Farsi and English. We reported reliability and construct validity estimates for the objective scales (engagement, functionality, aesthetics, and information quality). Reliability was based on the evaluation of intraclass correlation coefficients, Cronbach α and Spearman-Brown split-half reliability indicators (for internal consistency), as well as Pearson correlations for test-retest reliability. Construct validity included convergent and discriminant validity (through item-total correlations within the objective scales) and concurrent validity using Pearson correlations between the objective and subjective scores. RESULTS: After completing the translation and cultural adaptation, the MARS-Fa tool was used to assess the selected apps for smartphone addiction. The MARS-Fa total scale showed good interrater reliability (intraclass correlation coefficient=0.83, 95% CI 0.74-0.89) and good internal consistency (Cronbach α=.84); Spearman-Brown split-half reliability for both raters was 0.79 to 0.93. The instrument showed excellent test-retest reliability (r=0.94). The correlations among the MARS-Fa subdomains and the total score were all significant and above r=0.40, suggesting good convergent and discriminant validity. The MARS-Fa was positively and significantly correlated with subjective quality (r=0.90, P<.001), and so were the objective subdomains of engagement (r=0.85, P<.001), information quality (r=0.80, P<.001), aesthetics (r=0.79, P<.001), and functionality (r=0.57, P<.001), indicating concurrent validity. CONCLUSIONS: The MARS-Fa is a reliable and valid instrument to assess mobile health apps. This instrument could be adopted by Farsi-speaking researchers and developers who want to evaluate the quality of mobile apps. While we tested the tool with a sample of apps addressing smartphone addiction, the MARS-Fa could assess other domains or issues since the Mobile App Rating Scale has been used to rate apps in different contexts and languages.

Telemedicine in the emergency department: an overview of systematic reviews.

AIM: There is both favorable and controversial evidence on the application of telemedicine in the emergency department (ED), which has created uncertainty regarding the effectiveness of these systems. We performed a systematic review of the literature on systematic reviews to provide an overview of the benefits and challenges to the application of telemedicine systems for the ED. SUBJECT AND METHODS: PubMed, Web of Science, Scopus, Cochrane Library, and Google Scholar databases were explored for systematic reviews of telemedicine applications for the ED. Each review was critically appraised by two authors for data items to be extracted and evaluated. The most highly recommended technology, feasibility, benefits, and challenges to the application of telemedicine systems were studied and reported. RESULTS: We identified 18 studies of varying methodological quality and summarized their key findings. Form these 18 studies, 12 papers yielded a high risk of bias in their investigation. Nine papers concluded that real-time video conferencing was the best method of delivery, eight papers found cost reduction as an outcome of implementing these systems, and six studies found technical and infrastructure issues as a challenge when implementing telemedicine for EDs. CONCLUSION: There is strong evidence suggesting that the use of telemedicine positively impacts patient care. However, there are many challenges in implementing telemedicine that may impede the process or even impact patient safety. In conclusion, despite the high potential of telemedicine systems, there is still a need for better quality of evidence in order to confirm their feasibility in the ED.