No abstract

Background: The use of online medical forums is on the rise globally. Data scraping is a method of extracting website content using an automated computer program. We scraped users' questions regarding back and neck pain (BNP) from popular Israeli online medical forums. We aimed to identify the sort of questions being asked about BNP, and to analyse explicit themes that characterize their questions. Methods: Six leading Israeli BNP forums were identified. In phase 1, Python scripts scraped 12,418 questions into a data set. In phase 2-five themes were identified: Surgery (n = 2,957); health care professions (n = 2,361); Sports (n = 2,304); drugs (n = 1,419) and interpretation of imaging (n = 845). Phase 3-included the categorization of explicit fear-related words by the authors. Phase 4-analysis of explicit fear-related themes yielded 402 questions. Results: Gender was identified for 394 users, and age was identified for 181 users. A total of 248 users (61.6%) were women and 146 men (36.3%). Mean age 36.3 ± 16.15 for women and 35.5 ± 16.1 for men. The most commonly expressed fears were related to: invasive procedures, 30.9% (131 questions); fear of serious condition or misdiagnosis, 17.0% (72 questions); General concerns, 13.7% (58 questions); fear of worsening or relapse, 12.3% (52 questions); adverse effects of oral drugs or radiation, 10.8% (46 questions) and concerns related to lifestyle, 9.7% (41 questions). Conclusions: Web scraping is a feasible strategy with which to explore medical forums and the above-mentioned themes, all of which are of potential clinical significance. Significance: Using automated algorithms, a total of 12,369 questions from online back and neck medical forums were scraped and analysed. Secondary analysis categorized fear-related themes that were mentioned by users. Identifying and addressing patients' fear has potential to improve communication and therapeutic outcome. For example, questions regarding surgery were typically asked after the option was mentioned by a physician. This insight should encourage physicians to devote extra time explaining the possible implications of surgery, should they consider it as an option.

Background Ciprofloxacin is a widely used antibiotic that has lost efficiency due to extensive resistance. We developed machine learning (ML) models that predict the probability of ciprofloxacin resistance in hospitalized patients. Methods Data were collected from electronic records of hospitalized patients with positive bacterial cultures, during 2016-2019. Susceptibility results to ciprofloxacin (n = 10,053 cultures) were obtained for Escherichia coli, Klebsiella pneumoniae, Morganella morganii, Pseudomonas aeruginosa, Proteus mirabilis and Staphylococcus aureus. An ensemble model, combining several base models, was developed to predict ciprofloxacin resistant cultures, either with (gnostic) or without (agnostic) information on the infecting bacterial species. Results The ensemble models’ predictions are well-calibrated, and yield ROC-AUCs (area under the receiver operating characteristic curve) of 0.737 (95%CI 0.715–0.758) and 0.837 (95%CI 0.821–0.854) on independent test-sets for the agnostic and gnostic datasets, respectively. Shapley additive explanations analysis identifies that influential variables are related to resistance of previous infections, where patients arrived from (hospital, nursing home, etc.), and recent resistance frequencies in the hospital. A decision curve analysis reveals that implementing our models can be beneficial in a wide range of cost-benefits considerations of ciprofloxacin administration. Conclusions This study develops ML models to predict ciprofloxacin resistance in hospitalized patients. The models achieve high predictive ability, are well calibrated, have substantial net-benefit across a wide range of conditions, and rely on predictors consistent with the literature. This is a further step on the way to inclusion of ML decision support systems into clinical practice.

No abstract

Background: Ciprofloxacin is a widely used antibiotic that has lost efficiency due to extensive resistance. We developed machine learning (ML) models that predict the probability of ciprofloxacin resistance in hospitalized patients. Methods: Data were collected from electronic records of hospitalized patients with positive bacterial cultures, during 2016-2019. Susceptibility results to ciprofloxacin (n=10,053 cultures) were obtained for E. coli, K. pneumoniae, M. morganii, P.aeruginosa, P. mirabilis and S. aureus. An ensemble model, combining several base models, was developed to predict ciprofloxacin resistant cultures, either with (gnostic) or without (agnostic) information on the infecting bacterial species. Results: The ensemble models' predictions were well-calibrated, and yielded ROC-AUCs (area under the receiver operating characteristic curve) of 0.763 (95%CI 0.634-0.785) and 0.849 (95%CI 0.799-0.921) on independent test-sets for the agnostic and gnostic datasets, respectively. Shapley additive explanations analysis identified that influential variables were related to resistance of previous infections, where patients arrived from (hospital, nursing home, etc.), sex, and recent resistance frequencies in the hospital. A decision curve analysis revealed that implementing our models can be beneficial in a wide range of cost-benefits considerations of ciprofloxacin administration. Conclusions: This study develops ML models to predict ciprofloxacin resistance in hospitalized patients. The models achieved high predictive ability, were well calibrated, had substantial net-benefit across a wide range of conditions, and relied on predictors consistent with the literature. This is a further step on the way to inclusion of ML decision support systems into clinical practice.