Christine Yu

Defining a Patient Population With Cirrhosis: An Automated Algorithm With Natural Language Processing.

Objectives: The objective of this study was to use natural language processing (NLP) as a supplement to International Classification of Diseases, Ninth Revision (ICD-9) and laboratory values in an automated algorithm to better define and risk-stratify patients with cirrhosis. Background: Identification of patients with cirrhosis by manual data collection is time-intensive and laborious, whereas using ICD-9 codes can be inaccurate. NLP, a novel computerized approach to analyzing electronic free text, has been used to automatically identify patient cohorts with gastrointestinal pathologies such as inflammatory bowel disease. This methodology has not yet been used in cirrhosis. Study Design: This retrospective cohort study was conducted at the University of California, Los Angeles Health, an academic medical center. A total of 5343 University of California, Los Angeles primary care patients with ICD-9 codes for chronic liver disease were identified during March 2013 to January 2015. An algorithm incorporating NLP of radiology reports, ICD-9 codes, and laboratory data determined whether these patients had cirrhosis. Of the 5343 patients, 168 patient charts were manually reviewed at random as a gold standard comparison. Positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity of the algorithm and each of its steps were calculated. Results: The algorithm’s PPV, NPV, sensitivity, and specificity were 91.78%, 96.84%, 95.71%, and 93.88%, respectively. The NLP portion was the most important component of the algorithm with PPV, NPV, sensitivity, and specificity of 98.44%, 93.27%, 90.00%, and 98.98%, respectively. Conclusions: NLP is a powerful tool that can be combined with administrative and laboratory data to identify patients with cirrhosis within a population.

A Multi-Level Fit-Based Quality Improvement Initiative to Improve Colorectal Cancer Screening in a Managed Care Population

Introduction: Colorectal cancer (CRC) is a common but largely preventable disease with suboptimal screening rates despite national guidelines to screen individuals age 50–75. Single‐component interventions aimed to improve screening uptake only modestly improve rates; data suggest that multi‐modal approaches may be more effective. Methods: We designed, implemented, and evaluated the impact of a multi‐modal intervention on CRC screening uptake among unscreened patients in a large managed care population. Patient‐level components included a mailed letter with education about screening options and pre‐colonoscopy telephone counseling. For providers, we facilitated communication of screening test results and work‐flow for abnormal results. System‐level modifications included establishment of a patient navigator, expedited work‐up for abnormal results, and stream‐lined colonoscopy scheduling. We measured the rate of screening uptake overall, screening uptake by modality, change in the proportion of the population screened, and positive fecal immunochemical test (FIT) follow‐up rates in the 1‐year study period. Results: There were 5093 patients in the intervention cohort. Of these, 33.2% participated in FIT or colonoscopy screening within 1 year of the mailing. A total of 1078 (21.2%) participants completed a FIT and 611 (12.0%) completed a screening colonoscopy. The screening rate in the managed care population increased from 65.1 to 76.6%. Fifty‐nine patients (5.5%) had a positive FIT, of which 30 (50.8%) completed a diagnostic colonoscopy. Conclusion: Multi‐modal interventions can result in substantial improvement in CRC screening uptake in large and diverse managed care populations. Translational Impact: Health systems should shift their focus from single‐level to multi‐level interventions when addressing barriers to CRC screening.

Barriers to Follow-Up Colonoscopies for Patients With Positive Results From Fecal Immunochemical Tests During Colorectal Cancer Screening

Background & Aims Colorectal cancer is common yet largely preventable. The fecal immunochemical test (FIT) is a highly recommended screening method, but patients with positive results must receive a follow‐up colonoscopy to determine if they have precancerous or cancerous lesions. We characterized colonoscopic follow‐up evaluations and reasons for lack of follow‐up in a Veterans Affairs (VA) cohort. Methods We conducted a retrospective cross‐sectional analysis of patients 50 to 75 years old with a positive FIT result from January 1, 2014, through May 31, 2016, in a network of 12 VAs sites in southern California. We determined the proportion of patients who received a follow‐up colonoscopy, median time to colonoscopy, and colonoscopy findings. For patients who did not undergo colonoscopy, we determined the documented reason for lack of colonoscopy and factors associated with declining the colonoscopy examination. Results Of the 10,635 FITs performed, 916 (8.6%) produced positive results; 569 of these (62.1%) were followed by colonoscopy. The median time to colonoscopy after a positive FIT result was 83 days (interquartile range, 54–145 d), which did not vary between veterans who received a colonoscopy at a VA facility (81 d; interquartile range, 52–143 d) vs a non‐VA site (87 d; interquartile range, 60–154 d) (P = .2). For the 347 veterans (37.9%) who did not undergo follow‐up colonoscopy, the reasons were patient‐related (49.3%), provider‐related (16.4%), system‐related (12.1%), or multifactorial (22.2%). Overall, patient decline of colonoscopy (35.2%) was the most common reason. Conclusions In a cohort of veterans with positive results from FITs during CRC screening, reasons for lack of follow‐up colonoscopy varied and included patient, provider, and system factors. These findings can be used to reduce barriers to follow‐up colonoscopy and to address system‐level challenges in scheduling and attrition for colonoscopy.