Background and objectiveThe volume of healthcare data is growing rapidly with the adoption of health information technology. We focus on automated ICD9 code assignment from discharge summary content and methods for evaluating such assignments.MethodsWe study ICD9 diagnosis codes and discharge summaries from the publicly available Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC II) repository. We experiment with two coding approaches: one that treats each ICD9 code independently of each other (flat classifier), and one that leverages the hierarchical nature of ICD9 codes into its modeling (hierarchy-based classifier). We propose novel evaluation metrics, which reflect the distances among gold-standard and predicted codes and their locations in the ICD9 tree. Experimental setup, code for modeling, and evaluation scripts are made available to the research community.ResultsThe hierarchy-based classifier outperforms the flat classifier with F-measures of 39.5% and 27.6%, respectively, when trained on 20 533 documents and tested on 2282 documents. While recall is improved at the expense of precision, our novel evaluation metrics show a more refined assessment: for instance, the hierarchy-based classifier identifies the correct sub-tree of gold-standard codes more often than the flat classifier. Error analysis reveals that gold-standard codes are not perfect, and as such the recall and precision are likely underestimated.ConclusionsHierarchy-based classification yields better ICD9 coding than flat classification for MIMIC patients. Automated ICD9 coding is an example of a task for which data and tools can be shared and for which the research community can work together to build on shared models and advance the state of the art.

MotivationCatheter-associated urinary tract infections (CAUTI) are a common and serious healthcare-associated infection. Despite many efforts to reduce the occurrence of CAUTI, there remains a gap in the literature about CAUTI risk factors, especially pertaining to the effect of catheter dwell-time on CAUTI development and patient comorbidities.ObjectiveTo examine how the risk for CAUTI changes over time. Additionally, to assess whether time from catheter insertion to CAUTI event varied according to risk factors such as age, sex, patient type (surgical vs medical) and comorbidities.DesignRetrospective cohort study of all patients who were catheterised from 2012 to 2016, including those who did and did not develop CAUTIs. Both paediatric and adult patients were included. Indwelling urinary catheterisation is the exposure variable. The variable is interval, as all participants were exposed but for different lengths of time.SettingUrban academic health system of over 2500 beds. The system encompasses two large academic medical centres, two community hospitals and a paediatric hospital.ResultsThe study population was 47 926 patients who had 61 047 catheterisations, of which 861 (1.41%) resulted in a CAUTI. CAUTI rates were found to increase non-linearly for each additional day of catheterisation; CAUTI-free survival was 97.3% (CI: 97.1 to 97.6) at 10 days, 88.2% (CI: 86.9 to 89.5) at 30 days and 71.8% (CI: 66.3 to 77.8) at 60 days. This translated to an instantaneous HR of. 49%–1.65% in the 10–60 day time range. Paraplegia, cerebrovascular disease and female sex were found to statistically increase the chances of a CAUTI.ConclusionsUsing a very large data set, we demonstrated the incremental risk of CAUTI associated with each additional day of catheterisation, as well as the risk factors that increase the hazard for CAUTI. Special attention should be given to patients carrying these risk factors, for example, females or those with mobility issues.

We present the Unsupervised Phenome Model (UPhenome), a probabilistic graphical model for large-scale discovery of computational models of disease, or phenotypes. We tackle this challenge through the joint modeling of a large set of diseases and a large set of clinical observations. The observations are drawn directly from heterogeneous patient record data (notes, laboratory tests, medications, and diagnosis codes), and the diseases are modeled in an unsupervised fashion. We apply UPhenome to two qualitatively different mixtures of patients and diseases: records of extremely sick patients in the intensive care unit with constant monitoring, and records of outpatients regularly followed by care providers over multiple years. We demonstrate that the UPhenome model can learn from these different care settings, without any additional adaptation. Our experiments show that (i) the learned phenotypes combine the heterogeneous data types more coherently than baseline LDA-based phenotypes; (ii) they each represent single diseases rather than a mix of diseases more often than the baseline ones; and (iii) when applied to unseen patient records, they are correlated with the patients' ground-truth disorders. Code for training, inference, and quantitative evaluation is made available to the research community.

Engaging patients in admission medication reconciliation using an electronic home medication review tool may improve medication safety during hospitalization.