Care coordination for patients with cirrhosis: a "win-win" solution for patients, caregivers, providers, and healthcare expenditures

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.

The 3-month readmission rate remains unacceptably high in a large North American cohort of patients with cirrhosis

In smaller single-center studies, patients with cirrhosis are at a high readmission risk, but a multicenter perspective study is lacking. We evaluated the determinants of 3-month readmissions among inpatients with cirrhosis using the prospective 14-center North American Consortium for the Study of End-Stage Liver Disease cohort. Patients with cirrhosis hospitalized for nonelective indications provided consent and were followed for 3 months postdischarge. The number of 3-month readmissions and their determinants on index admission and discharge were calculated. We used multivariable logistic regression for all readmissions and for hepatic encephalopathy (HE), renal/metabolic, and infection-related readmissions. A score was developed using admission/discharge variables for the total sample, which was validated on a random half of the total population. Of the 1353 patients enrolled, 1177 were eligible on discharge and 1013 had 3-month outcomes. Readmissions occurred in 53% (n = 535; 316 with one, 219 with two or more), with consistent rates across sites. The leading causes were liver-related (n = 333; HE, renal/metabolic, and infections). Patients with cirrhosis and with worse Model for End-Stage Liver Disease score or diabetes, those taking prophylactic antibiotics, and those with prior HE were more likely to be readmitted. The admission model included Model for End-Stage Liver Disease and diabetes (c-statistic = 0.64, after split-validation 0.65). The discharge model included Model for End-Stage Liver Disease, proton pump inhibitor use, and lower length of stay (c-statistic = 0.65, after split-validation 0.70). Thirty percent of readmissions could not be predicted. Patients with liver-related readmissions consistently had index-stay nosocomial infections as a predictor for HE, renal/metabolic, and infection-associated readmissions (odds ratio = 1.9-3.0).

CONCLUSIONS

Three-month readmissions occurred in about half of discharged patients with cirrhosis, which were associated with cirrhosis severity, diabetes, and nosocomial infections; close monitoring of patients with advanced cirrhosis and prevention of nosocomial infections could reduce this burden. (Hepatology 2016;64:200-208).