Using Concept Unique Identifiers to Filter Electronic Health Records for Delirium Cases.

Delirium, an acute mental status change associated with inattention, confusion, hypervigilance, or somnolence due to a medical cause, is considered a medical emergency. Unfortunately, screening and diagnosis of delirium in acute care are often inadequate. It is estimated that 60% of delirium cases are not identified, and in claims data, they are underreported. Using information technology, we investigated whether concept unique identifiers from the Unified Language Medical System Metathesaurus could be used as a method to filter electronic health records for possible delirium cases. This article provides the reader with an overview of delirium, the Unified Language Medical System Metathesaurus, and our method for retrospectively filtering electronic health records for delirium cases from our clinical research database. Using a retrospective observational approach, we randomly selected 150 electronic health records with narrative notes containing a delirium concept unique identifier. One hundred records were used for training and 50 were used for validation and interrater reliability. Our results validate electronic health record-selected concept unique identifiers and provide insights into their use. Refinement and application of this method on a larger scale can provide an initial filter for identifying patients with delirium from the electronic health record.

Application of machine learning to the prediction of postoperative sepsis after appendectomy.

BACKGROUND: We applied various machine learning algorithms to a large national dataset to model the risk of postoperative sepsis after appendectomy to evaluate utility of such methods and identify factors associated with postoperative sepsis in these patients. METHODS: The National Surgery Quality Improvement Program database was used to identify patients undergoing appendectomy between 2005 and 2017. Logistic regression, support vector machines, random forest decision trees, and extreme gradient boosting machines were used to model the occurrence of postoperative sepsis. RESULTS: In the study, 223,214 appendectomies were identified; 2,143 (0.96%) were indicated as having postoperative sepsis. Logistic regression (area under the curve 0.70; 95% confidence interval, 0.68-0.73), random forest decision trees (area under the curve 0.70; 95% confidence interval, 0.68-0.73), and extreme gradient boosting (area under the curve 0.70; 95% confidence interval, 0.68-0.73) afforded similar performance, while support vector machines (area under the curve 0.51; 95% confidence interval, 0.50-0.52) had worse performance. Variable importance analyses identified preoperative congestive heart failure, transfusion, and acute renal failure as predictors of postoperative sepsis. CONCLUSION: Machine learning methods can be used to predict the development of sepsis after appendectomy with moderate accuracy. Such predictive modeling has potential to ultimately allow for preoperative recognition of patients at risk for developing postoperative sepsis after appendectomy thus facilitating early intervention and reducing morbidity.

Obstructive sleep apnea co-morbidity in patients with fibromyalgia: a single-center retrospective analysis and literature review.

Background: Fibromyalgia (FM) is a chronic medical condition characterized by widespread pain, sleep disturbance, and cognitive dysfunction. Sleep disorders are thought to play a prominent role in the etiology and symptomatic management of FM, specifically obstructive sleep apnea (OSA). In order to provide collaborative care, we need a better understanding of any overlapping presentation of FM and OSA. We conducted a site-wide review of patients from 2012-2016 to identify FM patients diagnosed with OSA. Methods: Charts were reviewed in patients aged 18 and above from 2012-2016 using ICD codes from a clinical data repository. Intersection of patients with a diagnosis of FM and OSA in clinics of psychiatry, sleep, rheumatology, and other outpatient clinics was compared. Polysomnography order patterns for FM patients were investigated. Results: Co-morbidity was highest in the sleep clinic (85.8%) compared to psychiatry (42.0%), rheumatology (18.7%), and other outpatient clinics (3.6%) (p<0.001). In the rheumatology and other outpatient clinics, 93.5% and 96% of patients respectively, had no polysomnography ordered. Pairwise comparison of co-morbidity in clinics: sleep vs psychiatry, sleep vs rheumatology, sleep vs other clinics, psychiatry vs rheumatology, psychiatry vs other clinics, and rheumatology vs other clinics were statistically significant after applying a Sidak adjustment to the p-values (all p<0.001). Conclusion: Our analysis suggests that there could be a correlation between FM and OSA, and referral to sleep studies is recommended in the management of patients with FM. The varying prevalence of FM patients with co-morbid OSA in sleep clinics when compared to other outpatient clinics suggests a discrepancy in the identification of FM patients with OSA. When properly screened, OSA co-morbidity has the potential to be higher in other outpatient clinics.

Development and application of a high throughput natural language processing architecture to convert all clinical documents in a clinical data warehouse into standardized medical vocabularies.

OBJECTIVE: Natural language processing (NLP) engines such as the clinical Text Analysis and Knowledge Extraction System are a solution for processing notes for research, but optimizing their performance for a clinical data warehouse remains a challenge. We aim to develop a high throughput NLP architecture using the clinical Text Analysis and Knowledge Extraction System and present a predictive model use case. MATERIALS AND METHODS: The CDW was comprised of 1 103 038 patients across 10 years. The architecture was constructed using the Hadoop data repository for source data and 3 large-scale symmetric processing servers for NLP. Each named entity mention in a clinical document was mapped to the Unified Medical Language System concept unique identifier (CUI). RESULTS: The NLP architecture processed 83 867 802 clinical documents in 13.33 days and produced 37 721 886 606 CUIs across 8 standardized medical vocabularies. Performance of the architecture exceeded 500 000 documents per hour across 30 parallel instances of the clinical Text Analysis and Knowledge Extraction System including 10 instances dedicated to documents greater than 20 000 bytes. In a use-case example for predicting 30-day hospital readmission, a CUI-based model had similar discrimination to n-grams with an area under the curve receiver operating characteristic of 0.75 (95% CI, 0.74-0.76). DISCUSSION AND CONCLUSION: Our health system's high throughput NLP architecture may serve as a benchmark for large-scale clinical research using a CUI-based approach.

Acupuncture points used in the prophylaxis against recurrent uncomplicated cystitis, patterns identified and their possible relationship to physiological measurements.

OBJECTIVE: To explore the correlation between single acupoints used and the recurrence rate of cystitis among cystitis-prone women receiving acupuncture as a prophylactic treatment. METHODS: In all, 58 cystitis-prone women were included in the analysis. Customised acupuncture treatments were given twice a week, over 4 weeks. The main effect parameter was the number of cystitis episodes during the 6-month observation time. Residual urine was measured at baseline, 2, 4 and 6 months using portable ultrasound equipment. Sympathetic and vagotone nerve activities were measured by using skin conductance and respiratory sinus arrhythmia, respectively. RESULTS: The main acupoints used for patients with Kidney (Shen) qi/yang deficiency were Shenshu (BL23), Taixi (KI3), Zhongji (CV3), Sanyinjiao (SP6) and Pangguangshu (BL28), compared with Taichong (LR3), CV3, BL28, Yinlingquan (SP9) and SP6 for Liver (Gan) qi stagnation, and SP6, CV3, BL28, Zusanli (ST36) and SP9 for Spleen (Pi) qi/yang deficiency patients. The combination BL23 and KI3 were used in 16 women, 13 of which were Kidney pattern related patients. When used, the number of symptomatic episodes were reduced to a third compared with what occurred in the 42 women where this combination was not used (3/16 vs. 28/42, P<0.05). BL23 application correlated to a significant reduction in residual urine measured a few days after treatment. Patients with the pattern of Spleen qi/yang deficiency had an initial increase in residual urine after treatments. CONCLUSION: Treating Kidney pattern related patients with the combination of BL23 and KI3 resulted in far better outcome than other points/combination of points for other Chinese medicine diagnoses. The acupoint SP6 may be less indicated than previously assumed when treating cystitis-prone women prophylactically.