Access to specialty care has been associated with improved survival in patients with liver disease but universal access is not always feasible. Methods of care delivery using virtual modalities including the SCAN-ECHO (Specialty Access Network-Extension of Community Healthcare Outcome) program were implemented by the Veterans Health Administration (VHA) to address this need but limited data are available on patient outcomes. We sought to evaluate the efficacy of a SCAN-ECHO visit within the context of a regional cohort of patients with liver disease in the VHA (n = 62,237) following implementation in the Ann Arbor SCAN-ECHO Liver Clinic from June 1, 2011, to March 31, 2015. The effect of a SCAN-ECHO visit on all-cause mortality was compared with patients with no liver clinic visit. To adjust for the differences among patients who had a SCAN-ECHO visit versus those with no visit, propensity score matching was performed on condition factors that affect the likelihood of a SCAN-ECHO visit: demographics, geographic location, liver disease diagnosis, severity, and comorbidities. During the study period, 513 patients who had a liver SCAN-ECHO visit were found within the cohort. Patients who had completed a virtual SCAN-ECHO visit were more likely younger, rural, with more significant liver disease, and evidence for cirrhosis. Propensity-adjusted mortality rates using the Cox Proportional Hazard Model showed that a SCAN-ECHO visit was associated with a hazard ratio of 0.54 (95% confidence interval 0.36-0.81, P = 0.003) compared with no visit. Conclusion: Improved survival in patients using SCAN-ECHO suggests that this approach may be an effective method to improve access for selected patients with liver disease, particularly in rural and underserved populations where access to specialty care is limited.

Key Points Question Can deep learning recurrent neural network (RNN) models using raw longitudinal data extracted directly from electronic health records outperform conventional regression models in predicting the risk of developing hepatocellular carcinoma (HCC)? Findings This prognostic study included 48 151 patients with hepatitis C virus (HCV)–related cirrhosis in the national Veterans Health Administration who had at least 3 years of follow-up after the diagnosis of cirrhosis. Deep learning RNN models outperformed conventional linear regression models and could be used to identify patients with HCV-related cirrhosis at high risk of developing HCC. Meaning The findings of this study suggest that RNN models could have multiple applications in clinical practice and could be applied to HCC outreach and surveillance strategies.

BackgroundMachine learning (ML) algorithms provide effective ways to build prediction models using longitudinal information given their capacity to incorporate numerous predictor variables without compromising the accuracy of the risk prediction. Clinical risk prediction models in chronic hepatitis C virus (CHC) can be challenging due to non-linear nature of disease progression. We developed and compared two ML algorithms to predict cirrhosis development in a large CHC-infected cohort using longitudinal data.Methods and findingsWe used national Veterans Health Administration (VHA) data to identify CHC patients in care between 2000–2016. The primary outcome was cirrhosis development ascertained by two consecutive aspartate aminotransferase (AST)-to-platelet ratio indexes (APRIs) > 2 after time zero given the infrequency of liver biopsy in clinical practice and that APRI is a validated non-invasive biomarker of fibrosis in CHC. We excluded those with initial APRI > 2 or pre-existing diagnosis of cirrhosis, hepatocellular carcinoma or hepatic decompensation. Enrollment was defined as the date of the first APRI. Time zero was defined as 2 years after enrollment. Cross-sectional (CS) models used predictors at or closest before time zero as a comparison. Longitudinal models used CS predictors plus longitudinal summary variables (maximum, minimum, maximum of slope, minimum of slope and total variation) between enrollment and time zero. Covariates included demographics, labs, and body mass index. Model performance was evaluated using concordance and area under the receiver operating curve (AuROC). A total of 72,683 individuals with CHC were analyzed with the cohort having a mean age of 52.8, 96.8% male and 53% white. There are 11,616 individuals (16%) who met the primary outcome over a mean follow-up of 7 years. We found superior predictive performance for the longitudinal Cox model compared to the CS Cox model (concordance 0.764 vs 0.746), and for the longitudinal boosted-survival-tree model compared to the linear Cox model (concordance 0.774 vs 0.764). The accuracy of the longitudinal models at 1,3,5 years after time zero also showed superior performance compared to the CS model, based on AuROC.ConclusionsBoosted-survival-tree based models using longitudinal information are statistically superior to cross-sectional or linear models for predicting development of cirrhosis in CHC, though all four models were highly accurate. Similar statistical methods could be applied to predict outcomes in other non-linear chronic disease states.

Access to ambulatory GI care was associated with improved 5-year survival for patients with liver disease. Innovative care coordination techniques may prove beneficial in extending access to care to liver disease patients.