Although the number of hospitals performing cardiac surgery has increased rapidly in China, information regarding the trends in coronary artery bypass grafting (CABG) outcomes remains unknown. We used data from the Chinese Cardiac Surgery Registry, the largest registry system that accounts for nearly 50% of total annual CABG volume in China, to assess trends of in-hospital mortality and major complication rates for patients receiving isolated CABG in 102 urban teaching hospitals in China from 25 January 2004 through 31 December 2013 (except 2006 and 2009). Using a mixed effects model, we estimated annual trends in each of these two outcomes overall and by age groups (18–64 and 65 years or older), adjusted for patient characteristics. We also assessed the trends in pre-operative, post-operative, and total length of stay (LOS). The study included 40 652 patients across 102 hospitals. Between 2004 and 2013, patients' mean age decreased from 62.7 to 61.4 years, in-hospital mortality decreased from 2.8% to 1.6% (difference, 1.3%, 95% CI: 0.70–1.85), and major complication rates decreased from 7.8% to 3.8% (difference, 4.0%; 95% CI: 3.05–4.90). The reduction in mortality and major complication rates were consistent across age groups. Between 2004 and 2013, the median (inter-quartile range) pre-operative LOS remained unchanged, post-operative LOS declined from 12.0 (8.0) to 10.0 (7.0) days, and total LOS declined from 22.0 (13.0) to 20.0 (12.0) days. Isolated CABG-related in-hospital mortality, major complication rates, and LOS have improved in urban teaching hospitals in China over the last decade.

Suicidal and self-injurious incidents in correctional settings deplete the institutional and healthcare resources, create disorder and stress for staff and other inmates. Traditional statistical analyses provide some guidance, but they can only be applied to structured data that are often difficult to collect and their recommendations are often expensive to act upon. This study aims to extract information from medical and mental health progress notes using AI algorithms to make actionable predictions of suicidal and self-injurious events to improve the efficiency of triage for health care services and prevent suicidal and injurious events from happening at California's Orange County Jails. The results showed that the notes data contain more information with respect to suicidal or injurious behaviors than the structured data available in the EHR database at the Orange County Jails. Using the notes data alone (under-sampled to 50%) in a Transformer Encoder model produced an AUC-ROC of 0.862, a Sensitivity of 0.816, and a Specificity of 0.738. Incorporating the information extracted from the notes data into traditional Machine Learning models as a feature alongside structured data (under-sampled to 50%) yielded better performance in terms of Sensitivity (AUC-ROC: 0.77, Sensitivity: 0.89, Specificity: 0.65). In addition, under-sampling is an effective approach to mitigating the impact of the extremely imbalanced classes.

Objectives The objective was to explore the differences in medication use pattern of lipid-lowering drug (LLD) and antiplatelet agents among post-percutaneous coronary intervention patients with acute coronary syndrome aged <65 in Hong Kong (HK) and the USA. Design Retrospective study. Setting This study used deidentified claims data from Clinformatics Data Mart database (OptumInsight, Eden Prairie, Minnesota, USA) and electronic health records from HK Hospital Authority Clinical Data Analysis and Reporting System database. Participants We used 1 year prescription records of LLDs and antiplatelet agents among 1013 USA patients and 270 HK Chinese patients in 2011–2013. Primary and secondary outcome measures Continuity was investigated on the assumption that one defined daily dose represented 1 day treatment. Medication possession ratio method was used to evaluate the adherence. Multivariate-adjusted logistic regressions were constructed to compare the good continuity and adherence levels in the merged database with the cutoffs set at 80%, and Cox proportional hazard models were built using the time to discontinuation as the dependent variable, to assess the persistence level. Results HK Chinese patients were less adherent (67.41% vs 84.60%, adjusted odds ratio (AOR) for Americans over Chinese=2.23 (95% CI=1.60 to 3.12), p<0.001) to antiplatelet agents compared with American patients but better adherent to statins (90.00% vs 78.18%, AOR=0.37 (0.23 to 0.58), p<0.001). The discontinuation with statins was more common in American patients (13.33% vs 34.25%, adjusted hazard ratio (AHR)=2.95 (2.05 to 4.24), p<0.001). Low-to-moderate potency statins and clopidogrel were favoured by our HK local physicians, while American patients received higher doses of statins and prasugrel. Conclusions We seemed to find HK physicians tended to prescribe cheaper and lower doses of statins and antiplatelet agents when compared with the privately insured patients in the USA, though the adherence and persistence levels of HK patients with statins were relatively good.