Development and Validation of an Algorithm to Identify Nonalcoholic Fatty Liver Disease in the Electronic Medical Record

Rise in First-Time ERCP for Benign Indications >1 Year After Cholecystectomy Is Associated With Worse Outcomes

BACKGROUND & AIMS

Greater availability of less invasive biliary imaging to rule out choledocholithiasis should reduce the need for diagnostic endoscopic retrograde cholangiopancreatography (ERCP) in patients who have a remote history of cholecystectomy. The primary aims were to determine the incidence, characteristics, and outcomes of individuals who undergo first-time ERCP >1 year after cholecystectomy (late-ERCP).

METHODS

Data from a commercial insurance claim database (Optum Clinformatics) identified 583,712 adults who underwent cholecystectomy, 4274 of whom underwent late-ERCP, defined as first-time ERCP for nonmalignant indications >1 year after cholecystectomy. Outcomes were exposure and temporal trends in late-ERCP, biliary imaging utilization, and post-ERCP outcomes. Multivariable logistic regression was used to examine patient characteristics associated with undergoing late-ERCP.

RESULTS

Despite a temporal increase in the use of noninvasive biliary imaging (35.9% in 2004 to 65.6% in 2021; P < .001), the rate of late-ERCP increased 8-fold (0.5-4.2/1000 person-years from 2005 to 2021; P < .001). Although only 44% of patients who underwent late-ERCP had gallstone removal, there were high rates of post-ERCP pancreatitis (7.1%), hospitalization (13.1%), and new chronic opioid use (9.7%). Factors associated with late-ERCP included concomitant disorder of gut-brain interaction (odds ratio [OR], 6.48; 95% confidence interval [CI], 5.88-6.91) and metabolic dysfunction steatotic liver disease (OR, 3.27; 95% CI, 2.79-3.55) along with use of anxiolytic (OR, 3.45; 95% CI, 3.19-3.58), antispasmodic (OR, 1.60; 95% CI, 1.53-1.72), and chronic opioids (OR, 6.24; 95% CI, 5.79-6.52).

CONCLUSIONS

The rate of late-ERCP postcholecystectomy is increasing significantly, particularly in patients with comorbidities associated with disorder of gut-brain interaction and mimickers of choledocholithiasis. Late-ERCPs are associated with disproportionately higher rates of adverse events, including initiation of chronic opioid use.

Identification of hepatic steatosis among persons with and without HIV using natural language processing

BACKGROUND

Steatotic liver disease (SLD) is a growing phenomenon, and our understanding of its determinants has been limited by our ability to identify it clinically. Natural language processing (NLP) can potentially identify hepatic steatosis systematically within large clinical repositories of imaging reports. We validated the performance of an NLP algorithm for the identification of SLD in clinical imaging reports and applied this tool to a large population of people with and without HIV.

METHODS

Patients were included in the analysis if they enrolled in the Veterans Aging Cohort Study between 2001 and 2017, had an imaging report inclusive of the liver, and had ≥2 years of observation before the imaging study. SLD was considered present when reports contained the terms "fatty," "steatosis," "steatotic," or "steatohepatitis." The performance of the SLD NLP algorithm was compared to a clinical review of 800 reports. We then applied the NLP algorithm to the first eligible imaging study and compared patient characteristics by SLD and HIV status.

RESULTS

NLP achieved 100% sensitivity and 88.5% positive predictive value for the identification of SLD. When applied to 26,706 eligible Veterans Aging Cohort Study patient imaging reports, SLD was identified in 72.2% and did not significantly differ by HIV status. SLD was associated with a higher prevalence of metabolic comorbidities, alcohol use disorder, and hepatitis B and C, but not HIV infection.

CONCLUSIONS

While limited to those undergoing radiologic study, the NLP algorithm accurately identified SLD in people with and without HIV and offers a valuable tool to evaluate the determinants and consequences of hepatic steatosis.

Where you live matters: Area deprivation predicts poor survival and liver transplant waitlisting

Social determinants of health (SDOH) are important predictors of poor clinical outcomes in chronic diseases, but their associations among the general cirrhosis population and liver transplantation (LT) are limited. We conducted a retrospective, multiinstitutional analysis of adult (≥18-years-old) patients with cirrhosis in metropolitan Chicago to determine the associations of poor neighborhood-level SDOH on decompensation complications, mortality, and LT waitlisting. Area deprivation index and covariates extracted from the American Census Survey were aspects of SDOH that were investigated. Among 15 101 patients with cirrhosis, the mean age was 57.2 years; 6414 (42.5%) were women, 6589 (43.6%) were non-Hispanic White, 3652 (24.2%) were non-Hispanic Black, and 2662 (17.6%) were Hispanic. Each quintile increase in area deprivation was associated with poor outcomes in decompensation (sHR [subdistribution hazard ratio] 1.07; 95% CI 1.05-1.10; P < .001), waitlisting (sHR 0.72; 95% CI 0.67-0.76; P < .001), and all-cause mortality (sHR 1.09; 95% CI 1.06-1.12; P < .001). Domains of SDOH associated with a lower likelihood of waitlisting and survival included low income, low education, poor household conditions, and social support (P < .001). Overall, patients with cirrhosis residing in poor neighborhood-level SDOH had higher decompensation, and mortality, and were less likely to be waitlisted for LT. Further exploration of structural barriers toward LT or optimizing health outcomes is warranted.

A natural language processing algorithm accurately classifies steatotic liver disease pathology to estimate the risk of cirrhosis

BACKGROUND

Histopathology remains the gold standard for diagnosing and staging metabolic dysfunction-associated steatotic liver disease (MASLD). The feasibility of studying MASLD progression in electronic medical records based on histological features is limited by the free-text nature of pathology reports. Here we introduce a natural language processing (NLP) algorithm to automatically score MASLD histology features.

METHODS

From the Mass General Brigham health care system electronic medical record, we identified all patients (1987-2021) with steatosis on index liver biopsy after excluding excess alcohol use and other etiologies of liver disease. An NLP algorithm was constructed in Python to detect steatosis, lobular inflammation, ballooning, and fibrosis stage from pathology free-text and manually validated in >1200 pathology reports. Patients were followed from the index biopsy to incident decompensated liver disease accounting for covariates.

RESULTS

The NLP algorithm demonstrated positive and negative predictive values from 93.5% to 100% for all histologic concepts. Among 3134 patients with biopsy-confirmed MASLD followed for 20,604 person-years, rates of the composite endpoint increased monotonically with worsening index fibrosis stage (p for linear trend <0.005). Compared to simple steatosis (incidence rate, 15.06/1000 person-years), the multivariable-adjusted HRs for cirrhosis were 1.04 (0.72-1.5) for metabolic dysfunction-associated steatohepatitis (MASH)/F0, 1.19 (0.92-1.54) for MASH/F1, 1.89 (1.41-2.52) for MASH/F2, and 4.21 (3.26-5.43) for MASH/F3.

CONCLUSIONS

The NLP algorithm accurately scores histological features of MASLD from pathology free-text. This algorithm enabled the construction of a large and high-quality MASLD cohort across a multihospital health care system and disclosed an accelerating risk for cirrhosis based on the index MASLD fibrosis stage.

NAFLD and NASH are obesity-independent risk factors in COVID-19: Matched real-world results from the large PINC AI™ Healthcare Database

BACKGROUND & AIMS

Non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) are potential risk factors for severe pneumonia and other infections. Available data on the role of NAFLD/NASH in worsening outcomes for COVID-19 are controversial and might be confounded by comorbidities.

METHODS

We used the PINC AI™ Healthcare Data Special Release (PHD-SR) to identify patients with COVID-19 (ICD-10) at approximately 900 hospitals in the United States. We performed exact matching (age, gender, and ethnicity) for patients with or without NAFLD/NASH, adjusting for demographics (admission type, region) and comorbidities (e.g., obesity, diabetes) through inverse probability of treatment weighting and then analysed hospitalisation-related outcomes.

RESULTS

Among 513 623 patients with SARS-CoV-2 (COVID-19), we identified 14 667 with NAFLD/NASH who could be matched to 14 667 controls. Mean age was 57.6 (±14.9) years, 50.8% were females and 43.7% were non-Hispanic whites. After matching, baseline characteristics (e.g., age, ethnicity, and gender) and comorbidities (e.g., hypertension, obesity, diabetes, and cardiovascular disease) were well balanced (standard difference (SD) <.10), except for cirrhosis and malignancies. Patients with COVID-19 and NAFLD/NASH had higher FIB-4 scores, a significantly longer hospital length of stay (LOS) and intensive care LOS than controls (9.4 vs. 8.3 days, and 10.4 vs. 9.3, respectively), even after adjusting for cirrhosis and malignancies. Patients with COVID-19 and NAFLD/NASH also had significantly higher risk of needing invasive mandatory ventilation (IMV) (odds ratio 1.0727; 95% CI 1.0095-1.1400). Other outcomes were similar in both groups.

CONCLUSIONS

In this large real-world cohort of patients hospitalised for COVID-19 in the United States, NAFLD/NASH were obesity-independent risk factors for complicated disease courses.

Artificial Intelligence in Liver Diseases: Recent Advances

Liver diseases cause a significant burden on public health worldwide. In spite of great advances during recent years, there are still many challenges in the diagnosis and treatment of liver diseases. During recent years, artificial intelligence (AI) has been widely used for the diagnosis, risk stratification, and prognostic prediction of various diseases based on clinical datasets and medical images. Accumulative studies have shown its performance for diagnosing patients with nonalcoholic fatty liver disease and liver fibrosis and assessing their severity, and for predicting treatment response and recurrence of hepatocellular carcinoma, outcomes of liver transplantation recipients, and risk of drug-induced liver injury. Herein, we aim to comprehensively summarize the current evidence regarding diagnostic, prognostic, and/or therapeutic role of AI in these common liver diseases.

Increasing prevalence of cirrhosis among insured adults in the United States, 2012-2018

BACKGROUND

Liver cirrhosis is a chronic disease that is known as a "silent killer" and its true prevalence is difficult to describe. It is imperative to accurately characterize the prevalence of cirrhosis because of its increasing healthcare burden.

METHODS

In this retrospective cohort study, trends in cirrhosis prevalence were evaluated using administrative data from one of the largest national health insurance providers in the US. (2011-2018). Enrolled adult (≥18-years-old) patients with cirrhosis defined by ICD-9 and ICD-10 were included in the study. The primary outcome measured in the study was the prevalence of cirrhosis 2011-2018.

RESULTS

Among the 371,482 patients with cirrhosis, the mean age was 62.2 (±13.7) years; 53.3% had commercial insurance and 46.4% had Medicare Advantage. The most frequent cirrhosis etiologies were alcohol-related (26.0%), NASH (20.9%) and HCV (20.0%). Mean time of follow-up was 725 (±732.3) days. The observed cirrhosis prevalence was 0.71% in 2018, a 2-fold increase from 2012 (0.34%). The highest prevalence observed was among patients with Medicare Advantage insurance (1.67%) in 2018. Prevalence increased in each US. state, with Southern states having the most rapid rise (2.3-fold). The most significant increases were observed in patients with NASH (3.9-fold) and alcohol-related (2-fold) cirrhosis.

CONCLUSION

Between 2012-2018, the prevalence of liver cirrhosis doubled among insured patients. Alcohol-related and NASH cirrhosis were the most significant contributors to this increase. Patients living in the South, and those insured by Medicare Advantage also have disproportionately higher prevalence of cirrhosis. Public health interventions are important to mitigate this concerning trajectory of strain to the health system.

Automated machine learning models for nonalcoholic fatty liver disease assessed by controlled attenuation parameter from the NHANES 2017-2020

Background

Nonalcoholic fatty liver disease (NAFLD) is recognized as one of the most common chronic liver diseases worldwide. This study aims to assess the efficacy of automated machine learning (AutoML) in the identification of NAFLD using a population-based cross-sectional database.

Methods

All data, including laboratory examinations, anthropometric measurements, and demographic variables, were obtained from the National Health and Nutrition Examination Survey (NHANES). NAFLD was defined by controlled attenuation parameter (CAP) in liver transient ultrasound elastography. The least absolute shrinkage and selection operator (LASSO) regression analysis was employed for feature selection. Six algorithms were utilized on the H2O-automated machine learning platform: Gradient Boosting Machine (GBM), Distributed Random Forest (DRF), Extremely Randomized Trees (XRT), Generalized Linear Model (GLM), eXtreme Gradient Boosting (XGBoost), and Deep Learning (DL). These algorithms were selected for their diverse strengths, including their ability to handle complex, non-linear relationships, provide high predictive accuracy, and ensure interpretability. The models were evaluated by area under receiver operating characteristic curves (AUC) and interpreted by the calibration curve, the decision curve analysis, variable importance plot, SHapley Additive exPlanation plot, partial dependence plots, and local interpretable model agnostic explanation plot.

Results

A total of 4177 participants (non-NAFLD 3167 vs NAFLD 1010) were included to develop and validate the AutoML models. The model developed by XGBoost performed better than other models in AutoML, achieving an AUC of 0.859, an accuracy of 0.795, a sensitivity of 0.773, and a specificity of 0.802 on the validation set.

Conclusions

We developed an XGBoost model to better evaluate the presence of NAFLD. Based on the XGBoost model, we created an R Shiny web-based application named Shiny NAFLD (http://39.101.122.171:3838/App2/). This application demonstrates the potential of AutoML in clinical research and practice, offering a promising tool for the real-world identification of NAFLD.

Application and impact of Lasso regression in gastroenterology: A systematic review

Least absolute shrinkage and selection operator (Lasso) regression is a statistical technique that can be used to study the effects of clinical variables in outcome prediction. In this study, we aimed at systematically reviewing the application of Lasso regression in gastroenterology for developing predictive models and providing a method of performing Lasso regression. A comprehensive search strategy was conducted in PubMed, Embase and Cochrane CENTRAL databases (Keywords: lasso regression; gastrointestinal tract/diseases) following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Studies were screened for eligibility based on pre-defined selection criteria and the data was extracted using a standardized form. Total 16 studies were included, comprising a diverse range of gastroenterological disease-related outcomes. Sample sizes ranged from 134 to 8861 subjects. Eleven studies reported liver disease-related prediction models, while five focused on non-hepatic etiology models. Lasso regression was applied for variable selection, risk prediction and model development, with various validation methods and performance metrics used. Model performance metrics included Area Under the Receiver Operating Characteristics (AUROC), C-index and calibration plots. In gastroenterology, Lasso regression has been used in various diseases such as inflammatory bowel disease, liver disease and esophageal cancer. It is valuable for complex scenarios with many predictors. However, its effectiveness depends on high-quality and complete data. While it identifies important variables, it doesn't provide causal interpretations. Therefore, cautious interpretation is necessary considering the study design and data quality.

Longitudinal Study of Comorbidities and Clinical Outcomes in Persons with Gallstone Disease Using Electronic Health Records

BACKGROUND

Gallstone disease (GSD) is common and leads to significant morbidity, mortality, and health care utilization in the USA. We examined comorbidities and clinical outcomes among persons with GSD using electronic health records (EHR).

METHODS

In this retrospective study of 1,381,004 adults, GSD was defined by ICD-9 code 574 or ICD-10 code K80 using Optum® longitudinal EHR from January 2007 to March 2021. We obtained diagnosis, procedure, prescription, and vital sign records and evaluated associations between demographics, comorbidities, and medications with cholecystectomy, digestive cancers, and mortality.

RESULTS

Among persons with GSD, 30% had a cholecystectomy and were more likely to be women, White, and younger, and less likely to have comorbidities, except for obesity, gastroesophageal reflux disease (GERD), abdominal pain, hyperlipidemia, and pancreatitis. Among persons with GSD, 2.2% had a non-colorectal digestive cancer diagnosis during follow-up and risk was 40% lower among persons with a cholecystectomy. Non-colorectal digestive cancer predictors included older age, male sex, non-White race-ethnicity, lower BMI, other cancers, diabetes, chronic liver disease, pancreatitis, GERD, and abdominal pain. Among persons with GSD, mortality was 15.1% compared with 9.7% for the whole EHR sample. Persons with a cholecystectomy had 40% lower mortality risk and mortality predictors included older age, male sex, Black race, lower BMI, and most comorbidities.

CONCLUSIONS

In this EHR analysis of persons with GSD, 30% had a cholecystectomy. Mortality was higher compared with the whole EHR sample. Persons with cholecystectomy were less likely to have non-colorectal digestive cancer or to die.

Association between hypoglycemic agent use and the risk of occurrence of nonalcoholic fatty liver disease in patients with type 2 diabetes mellitus

Nonalcoholic fatty liver disease (NAFLD) is a growing health concern with increasing prevalence and associated health impacts. Although no approved drugs are available for the NAFLD treatment, several hypoglycemic agents have been investigated as promising therapeutic agents. We aimed to compare the risk of occurrence of NAFLD with respect to the use of different hypoglycemic agents in patients with type 2 diabetes. This retrospective cohort study used data from the National Health Insurance Service-National Sample Cohort of South Korea. Participants newly diagnosed with type 2 diabetes (2003-2019) were included in this study. Two new user-active comparator cohorts were assembled: Cohort 1, new users of thiazolidinediones (TZD) and dipeptidyl peptidase-4 inhibitors (DPP-4i), and Cohort 2, new users of sodium-glucose cotransporter-2 inhibitors (SGLT-2i) and DPP-4i. The occurrence of NAFLD was defined based claims that include diagnostic codes. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazard models in 1:3 propensity score (PS)-matched cohorts. For 65,224 patients newly diagnosed with type 2 diabetes, the overall prevalence of NAFLD was 42.6%. The PS-matched Cohort 1 included 6,351 and 2,117 new users of DPP-4i and TZD, respectively. Compared to DPP-4i, TZD use was associated with the decreased risk of NAFLD (HR, 0.66; 95% CI: 0.55-0.78). Cohort 2 consisted of 6,783 and 2,261 new users of DPP-4i and SGLT-2i, respectively; SGLT-2i use was associated with a decreased risk of NAFLD (HR, 0.93; 95% CI: 0.80-1.08). This population-based cohort study supports the clinical implications of prioritizing TZD and SGLT-2i over DPP-4i in reducing the risk of occurrence of NAFLD in patients with type 2 diabetes. However, the findings lacked statistical significance, highlighting the need for further verification studies.

The association between non-alcoholic fatty liver disease and atopic dermatitis: a population-based cohort study

Background

In previous studies, it was reported that non-alcoholic fatty liver disease (NAFLD) incidence and prevalence increased in children with atopic dermatitis. Nevertheless, the actual association between the two diseases has not been fully proven in large-scale studies, and real-world evidence is missing. The objective of this nationwide, longitudinal cohort study was to evaluate the association between NAFLD and atopic dermatitis.

Methods

The National Health Insurance Research Database in Taiwan was utilized in this study. Patients with records of NAFLD diagnosis were recruited as the experimental group, and patients having less than three outpatient visits or one inpatient visiting record due to NAFLD were excluded from the study design. Non-NAFLD controls were matched based on a 1:4 propensity score matching. Potential confounders including age, gender, comorbidity, and medical utilization status were considered as covariates. The risk of future atopic dermatitis would be evaluated based on multivariate Cox proportional hazard regression.

Results

Compared with people without NAFLD, a decreased risk of atopic dermatitis in NALFD patients had been observed (aHR = 0.93, 95% CI 0.87-0.98). The trend was especially presented in young NAFLD patients. In patients younger than 40 years old, a 20% decreased risk of atopic dermatitis was reported (aHR = 0.80, 95% CI 0.70-0.92).

Conclusion

People with NAFLD were not associated with an increased risk of atopic dermatitis. Conversely, a 0.93-fold risk was noted in NAFLD patients, compared with NAFLD-free controls. Future studies are warranted to evaluate further the mechanism regarding the interplay between the inflammatory mechanisms of NAFLD and atopic dermatitis.

Large-scale identification of undiagnosed hepatic steatosis using natural language processing

Background

Nonalcoholic fatty liver disease (NAFLD) is a major cause of liver-related morbidity in people with and without diabetes, but it is underdiagnosed, posing challenges for research and clinical management. Here, we determine if natural language processing (NLP) of data in the electronic health record (EHR) could identify undiagnosed patients with hepatic steatosis based on pathology and radiology reports.

Methods

A rule-based NLP algorithm was built using a Linguamatics literature text mining tool to search 2.15 million pathology report and 2.7 million imaging reports in the Penn Medicine EHR from November 2014, through December 2020, for evidence of hepatic steatosis. For quality control, two independent physicians manually reviewed randomly chosen biopsy and imaging reports (n = 353, PPV 99.7%).

Findings

After exclusion of individuals with other causes of hepatic steatosis, 3007 patients with biopsy-proven NAFLD and 42,083 patients with imaging-proven NAFLD were identified. Interestingly, elevated ALT was not a sensitive predictor of the presence of steatosis, and only half of the biopsied patients with steatosis ever received an ICD diagnosis code for the presence of NAFLD/NASH. There was a robust association for PNPLA3 and TM6SF2 risk alleles and steatosis identified by NLP. We identified 234 disorders that were significantly over- or underrepresented in all subjects with steatosis and identified changes in serum markers (e.g., GGT) associated with presence of steatosis.

Interpretation

This study demonstrates clear feasibility of NLP-based approaches to identify patients whose steatosis was indicated in imaging and pathology reports within a large healthcare system and uncovers undercoding of NAFLD in the general population. Identification of patients at risk could link them to improved care and outcomes.

Funding

The study was funded by US and German funding sources that did provide financial support only and had no influence or control over the research process.

Administrative coding for non-alcoholic fatty liver disease is accurate in Swedish patients

Background and Aims: Epidemiological studies of non-alcoholic fatty liver disease (NAFLD) frequently use the International Classification of Disease (ICD) codes to identify patients. The validity of such ICD codes in a Swedish setting is unknown. Here, we aimed to validate the administrative code for NAFLD in Sweden.Methods: In total, 150 patients with an ICD-10 code for NAFLD (K76.0) from the Karolinska University Hospital between 1 January 2015 and 3 November 2021 were randomly selected. Patients were classified as true or false positives for NAFLD by medical chart review and the positive predictive value (PPV) for the ICD-10 code corresponding to NAFLD was calculated.Results: The PPV of the ICD-10 code for NAFLD was 0.82 (95% confidence interval [CI] = 0.76-0.89). After exclusion of patients with diagnostic coding for other liver diseases or alcohol abuse disorder (n = 14), the PPV was improved to 0.91 (95% CI 0.87-0.96). The PPV was higher in patients with coding for NAFLD in combination with obesity (0.95, 95%CI = 0.87-1.00) or type 2 diabetes (0.96, 95%CI = 0.89-1.00). However, in false-positive cases, a high alcohol consumption was common and such patients had somewhat higher Fibrosis-4 scores than true-positive patients (1.9 vs 1.3, p = 0.16)Conclusions: The ICD-10 code for NAFLD had a high PPV, that was further improved after exclusion of patients with coding for other liver diseases than NAFLD. This approach should be preferred when performing register-based studies to identify patients with NAFLD in Sweden. Still, residual alcohol-related liver disease might risk confound some findings seen in epidemiological studies which needs to be considered.

Leveraging electronic health record data for endometriosis research

Endometriosis is a chronic, complex disease for which there are vast disparities in diagnosis and treatment between sociodemographic groups. Clinical presentation of endometriosis can vary from asymptomatic disease-often identified during (in)fertility consultations-to dysmenorrhea and debilitating pelvic pain. Because of this complexity, delayed diagnosis (mean time to diagnosis is 1.7-3.6 years) and misdiagnosis is common. Early and accurate diagnosis of endometriosis remains a research priority for patient advocates and healthcare providers. Electronic health records (EHRs) have been widely adopted as a data source in biomedical research. However, they remain a largely untapped source of data for endometriosis research. EHRs capture diverse, real-world patient populations and care trajectories and can be used to learn patterns of underlying risk factors for endometriosis which, in turn, can be used to inform screening guidelines to help clinicians efficiently and effectively recognize and diagnose the disease in all patient populations reducing inequities in care. Here, we provide an overview of the advantages and limitations of using EHR data to study endometriosis. We describe the prevalence of endometriosis observed in diverse populations from multiple healthcare institutions, examples of variables that can be extracted from EHRs to enhance the accuracy of endometriosis prediction, and opportunities to leverage longitudinal EHR data to improve our understanding of long-term health consequences for all patients.

Screening for NAFLD-Current Knowledge and Challenges

Non-alcoholic fatty liver disease (NAFLD) is the most common cause of abnormal liver function tests worldwide, with an estimated prevalence ranging between 19-46% in the general population. Of note, NAFLD is also expected to become a leading cause of end-stage liver disease in the next decades. Given the high prevalence and severity of NAFLD, especially in high-risk populations (i.e., patients with type-2 diabetes mellitus and/or obesity), there is a major interest in early detection of the disease in primary care. Nevertheless, substantial uncertainties still surround the development of a screening policy for NAFLD, such as limitations in currently used non-invasive markers of fibrosis, cost-effectiveness and the absence of a licensed treatment. In this review, we summarise current knowledge and try to identify the limitations surrounding the screening policy for NAFLD in primary care.

Clinical decision support automates care gap detection among primary care patients with nonalcoholic fatty liver disease

BACKGROUND

Although guidelines recommend primary care-driven management of NAFLD, workflow constraints hinder feasibility. Leveraging electronic health records to risk stratify patients proposes a scalable, workflow-integrated strategy.

MATERIALS AND METHODS

We prospectively evaluated an electronic health record-embedded clinical decision support system's ability to risk stratify patients with NAFLD and detect gaps in care. Patients missing annual laboratory testing to calculate Fibrosis-4 Score (FIB-4) or those missing necessary linkage to further care were considered to have a gap in care. Linkage to care was defined as either referral for elastography-based testing or for consultation in hepatology clinic depending on clinical and biochemical characteristics.

RESULTS

Patients with NAFLD often lacked annual screening labs within primary care settings (1129/2154; 52%). Linkage to care was low in all categories, with <3% of patients with abnormal FIB-4 undergoing further evaluation.

DISCUSSION

Significant care gaps exist within primary care for screening and risk stratification of patients with NAFLD and can be efficiently addressed using electronic health record functionality.

Healthcare resource utilization and costs of care in the United States for patients with non-alcoholic steatohepatitis

AIMS

This retrospective, observational cohort study aimed to determine the burden of comorbidities, hospitalization, and healthcare costs among patients with non-alcoholic steatohepatitis (NASH) in the United States stratified by fibrosis-4 (FIB-4) or body mass index (BMI).

METHODS

Adults with NASH were identified in the Veradigm Health Insights Electronic Health Record Database and linked Komodo claims data. The index date was the earliest coded NASH diagnosis between 1 January 2016 and 31 December 2020 with valid FIB-4 and ≥6 months of database activity and continuous enrollment pre- and post-index. We excluded patients with viral hepatitis, alcohol-use disorder, or alcoholic liver disease. Patients were stratified by FIB-4: FIB-4 ≤ 0.95, 0.95 < FIB-4 ≤ 2.67, 2.67 < FIB-4 ≤ 4.12, FIB-4 > 4.12) or BMI (BMI <25, 25 ≤ BMI ≤30, BMI > 30). Multivariate analysis was used to assess the relationship of FIB-4 with costs and hospitalizations.

RESULTS

Among 6,743 qualifying patients, index FIB-4 was ≤0.95 for 2,345 patents, 0.95-2.67 for 3,289 patients, 2.67-4.12 for 571 patients, and >4.12 for 538 patients (mean age 55.8 years; 62.9% female). Mean age, comorbidity burden, cardiovascular disease risk, and healthcare utilization increased with increasing FIB-4. Mean ± SD annual costs increased from $16,744±$53,810 to $34,667±$67,691 between the lowest and highest FIB-4 cohorts and were higher among patients with BMI <25 ($24,568±$81,250) than BMI >30 ($21,542±$61,490). A one-unit increase in FIB-4 at index was associated with a 3.4% (95%CI: 1.7%-5.2%) increase in mean total annual cost and an 11.6% (95%CI: 8.0%-15.3%) increased likelihood of hospitalization.

CONCLUSIONS

A higher FIB-4 was associated with increased healthcare costs and risk of hospitalization in adults with NASH; however, even patients with FIB-4 ≤ 0.95 presented a significant burden.

Accuracy of Non-invasive Indices for Diagnosing Hepatic Steatosis Compared to Imaging in a Real-World Cohort

BACKGROUND & AIMS

Nonalcoholic fatty liver disease is common and under-diagnosed. This study evaluated the accuracy of several previously reported indices, including hepatic steatosis index, alanine aminotransferase (ALT) method, Framingham steatosis index, and Dallas steatosis index, to diagnose hepatic steatosis in a real-world cohort.

METHODS

This study included 701 randomly selected adult patients seen in our integrated healthcare system between 2015 and 2020 with appropriate abdominal imaging and routine outpatient laboratory studies. Information on demographics, comorbidities and existing liver disease, anthropometrics, laboratory studies, and abdominal imaging was collected. The sensitivity, specificity, and C-statistic of each method in detecting hepatic steatosis based on abdominal imaging were determined.

RESULTS

202/701 patients (28.8%) had hepatic steatosis on abdominal imaging. These patients were more likely to have metabolic syndrome components and higher body mass index. All indices performed similarly with moderate accuracy in detecting hepatic steatosis based on the C-statistic (95% confidence interval): Hepatic steatosis index 0.76 (0.72-0.79), Framingham steatosis index 0.78 (0.74-0.82), and Dallas steatosis index 0.80 (0.76-0.83). ALT method had sensitivity 44.7% (36.9-52.7%) and specificity 88.6% (85.0-91.7%). Several sensitivity analyses were performed, which did not significantly alter the performance of any index.

CONCLUSION

The findings support both the clinical utility of these indices in diagnosing hepatic steatosis in the absence of imaging in real-world settings and the research utility of these indices in generating reliable electronic medical record-based nonalcoholic fatty liver disease cohorts.

The Impact of NAFLD on Hospitalization Outcomes in Patients With Inflammatory Bowel Diseases: Nationwide Analysis

BACKGROUND

Nonalcoholic fatty liver disease (NAFLD) is highly prevalent in patients with inflammatory bowel diseases (IBD). Yet, the impact of NAFLD on outcomes, along with the contribution of nonmetabolic factors to NAFLD development, is unclear. To investigate these topics, we conducted a nationwide study examining the impact of NAFLD on hospitalization outcomes in IBD patients after adjusting for metabolic factors.

METHODS

Patients with IBD-related hospitalizations were identified using the Nationwide Readmissions Database from 2016 to 2018. Inflammatory bowel disease patients with and without NAFLD were matched based on IBD type, age, sex, metabolic syndrome, and diabetes mellitus. Primary outcomes were IBD-related readmission, IBD-related surgery, and death. Secondary outcomes were length of stay (LOS) and cost of care (COC). The primary multivariable model adjusted for obesity, dyslipidemia, Charlson-Deyo comorbidity index, hospital characteristics, payer, patient income, and elective status of admissions.

RESULTS

Nonalcoholic fatty liver disease was associated with a higher risk of IBD-related readmission (adjusted hazard ratio, 1.90; P < .01) and death (adjusted hazard ratio, 2.73; P < .01), 0.71-day longer LOS (P < .01), and $7312 higher COC (P < .01) in those with Crohn's disease. Nonalcoholic fatty liver disease was also associated with a higher risk of IBD-related readmission (adjusted hazard ratio, 1.65; P < .01), 0.64-day longer LOS (P < .01), and $9392 (P < .01) higher COC, but there was no difference in death in those with UC. No differences in risk of IBD-related surgery were observed.

CONCLUSIONS

Nonalcoholic fatty liver disease is associated with worse hospitalization outcomes in IBD patients after adjusting for metabolic factors. These data suggest nonmetabolic factors may be implicated in the pathogenesis of NAFLD in IBD patients and may contribute to worsened clinical outcomes.

Applications of artificial intelligence (AI) in researches on non-alcoholic fatty liver disease(NAFLD) : A systematic review

Non-alcoholic fatty liver disease (NAFLD) is one of the most important causes of chronic liver disease in the world, it has been found that cardiovascular and renal risks and diseases are also highly prevalent in adults with NAFLD. Diagnosis and treatment of NAFLD face many challenges, although the medical science has been very developed. Efficiency, accuracy and individualization are the main goals to be solved. Evaluation of the severity of NAFLD involves a variety of clinical parameters, how to optimize non-invasive evaluation methods is a necessary issue that needs to be discussed in this field. Artificial intelligence (AI) has become increasingly widespread in healthcare applications, and it has been also brought many new insights into better analyzing chronic liver disease, including NAFLD. This paper reviewed AI related researches in NAFLD field published recently, summarized diagnostic models based on electronic health record and lab test, ultrasound and radio imaging, and liver histopathological data, described the application of therapeutic models in personalized lifestyle guidance and the development of drugs for NAFLD. In addition, we also analyzed present AI models in distinguishing healthy VS NAFLD/NASH, and fibrosis VS non-fibrosis in the evaluation of NAFLD progression. We hope to provide alternative directions for the future research.

Patient-Reported Outcome Measures Modestly Enhance Prediction of Readmission in Patients with Cirrhosis

BACKGROUND & AIMS

Patients with cirrhosis have high rates of hospital readmission, but prediction models are suboptimal and have not included important patient-reported outcome measures (PROMs). In a large prospective cohort, we examined the impact of PROMs on prediction of 30-day readmissions.

METHODS

We performed a prospective cohort study of adults with cirrhosis admitted to a tertiary center between June 2014 and March 2020. We collected clinical information, socioeconomic status, and PROMs addressing functional status and quality of life. We used hierarchical competing risk time-to-event analysis to examine the impact of PROMs on readmission prediction.

RESULTS

A total of 654 patients were discharged alive, and 247 (38%) were readmitted within 30 days. Readmission was independently associated with cerebrovascular disease, ascites, prior hospital admission, admission via the emergency department, lower albumin, higher Model for End-Stage Liver Disease, discharge with public transportation, and impaired basic activities of daily living and quality-of-life activity domain. Reduced readmission was associated with cancer, admission for infection, children at home, and impaired emotional function. Compared with a model including only clinical variables, addition of functional status and quality-of-life variables improved the area under the receiver-operating characteristic curve from 0.72 to 0.73 and 0.75, with net reclassification indices of 0.22 and 0.18, respectively. Socioeconomic variables did not significantly improve prediction compared with clinical variables alone. Compared with a model using electronically available variables only, no models improved prediction when examined with integrated discrimination improvement.

CONCLUSIONS

PROMs may marginally add to the prediction of 30-day readmissions for patients with cirrhosis. Poor social support and disability are associated with readmissions and may be high-yield targets for future interventions.

Diversity in Machine Learning: A Systematic Review of Text-Based Diagnostic Applications

OBJECTIVE

 As the storage of clinical data has transitioned into electronic formats, medical informatics has become increasingly relevant in providing diagnostic aid. The purpose of this review is to evaluate machine learning models that use text data for diagnosis and to assess the diversity of the included study populations.

METHODS

 We conducted a systematic literature review on three public databases. Two authors reviewed every abstract for inclusion. Articles were included if they used or developed machine learning algorithms to aid in diagnosis. Articles focusing on imaging informatics were excluded.

RESULTS

 From 2,260 identified papers, we included 78. Of the machine learning models used, neural networks were relied upon most frequently (44.9%). Studies had a median population of 661.5 patients, and diseases and disorders of 10 different body systems were studied. Of the 35.9% (N = 28) of papers that included race data, 57.1% (N = 16) of study populations were majority White, 14.3% were majority Asian, and 7.1% were majority Black. In 75% (N = 21) of papers, White was the largest racial group represented. Of the papers included, 43.6% (N = 34) included the sex ratio of the patient population.

DISCUSSION

 With the power to build robust algorithms supported by massive quantities of clinical data, machine learning is shaping the future of diagnostics. Limitations of the underlying data create potential biases, especially if patient demographics are unknown or not included in the training.

CONCLUSION

 As the movement toward clinical reliance on machine learning accelerates, both recording demographic information and using diverse training sets should be emphasized. Extrapolating algorithms to demographics beyond the original study population leaves large gaps for potential biases.

Bidirectional Association Between Psoriasis and Nonalcoholic Fatty Liver Disease: Real-World Evidence From Two Longitudinal Cohort Studies

Background

Association between nonalcoholic fatty liver disease (NAFLD) and future psoriasis has not yet been confirmed, although the two diseases partially share a common pathogenesis pathway. Studies have revealed an association between psoriasis and subsequent NAFLD; however, these studies were limited to small sample sizes and a cross-sectional study design. Hence, the main objective of this population-based longitudinal cohort study was to evaluate the bidirectional association between psoriasis and NAFLD.

Methods

Data were retrieved from Taiwan’s National Health Insurance Research Database. Patients with new-onset NAFLD and psoriasis were respectively enrolled in two cohorts. For each comparison cohort, propensity-score-matched controls with no record of NAFLD or psoriasis were selected. An adjusted hazard ratio (aHR) was applied to evaluate subsequent risks.

Results

The risk of patients with new-onset NAFLD developing psoriasis was statistically significant, with an HR of 1.07 (95% CI, 1.01–1.14). For younger patients with NAFLD, the risk of developing psoriasis was 1.3-fold higher. The risk of patients with new-onset psoriasis developing NAFLD in the future was 1.28-fold higher than that of patients without psoriasis (95% CI, 1.21–1.35), and patients in younger psoriasis subgroups below the age of 40 years were at a higher risk than those in older subgroups, with an aHR of 1.55 (95% CI, 1.40–1.71).

Conclusion

Evidence supports a bidirectional association between NAFLD and psoriasis, especially in patients below the age of 40 years. The correlation between the two diseases and the subsequent risk of disease development should be considered when caring for patients.

The National Consensus statement on the management of adult patients with non-alcoholic fatty liver disease and main comorbidities

The National Consensus was prepared with the participation of the National Medical Association for the Study of the Multimorbidity, Russian Scientific Liver Society, Russian Association of Endocrinologists, Russian Association of Gerontologists and Geriatricians, National Society for Preventive Cardiology, Professional Foundation for the Promotion of Medicine Fund PROFMEDFORUM. The aim of the multidisciplinary consensus is a detailed analysis of the course of non-alcoholic fatty liver disease (NAFLD) and the main associated conditions. The definition of NAFLD is given, its prevalence is described, methods for diagnosing its components such as steatosis, inflammation and fibrosis are described. The association of NAFLD with a number of cardio-metabolic diseases (arterial hypertension, atherosclerosis, thrombotic complications, type 2 diabetes mellitus (T2DM), obesity, dyslipidemia, etc.), chronic kidney disease (CKD) and the risk of developing hepatocellular cancer (HCC) were analyzed. The review of non-drug methods of treatment of NAFLD and modern opportunities of pharmacotherapy are presented. The possibilities of new molecules in the treatment of NAFLD are considered: agonists of nuclear receptors, antagonists of pro-inflammatory molecules, etc. The positive properties and disadvantages of currently used drugs (vitamin E, thiazolidinediones, etc.) are described. Special attention is paid to the multi-target ursodeoxycholic acid (UDCA) molecule in the complex treatment of NAFLD as a multifactorial disease. Its anti-inflammatory, anti-oxidant and cytoprotective properties, the ability to reduce steatosis an independent risk factor for the development of cardiovascular pathology, reduce inflammation and hepatic fibrosis through the modulation of autophagy are considered. The ability of UDCA to influence glucose and lipid homeostasis and to have an anticarcinogenic effect has been demonstrated. The Consensus statement has advanced provisions for practitioners to optimize the diagnosis and treatment of NAFLD and related common pathogenetic links of cardio-metabolic diseases.

Administrative Coding in Electronic Health Care Record-Based Research of NAFLD: An Expert Panel Consensus Statement

BACKGROUND AND AIMS

Electronic health record (EHR)-based research allows the capture of large amounts of data, which is necessary in NAFLD, where the risk of clinical liver outcomes is generally low. The lack of consensus on which International Classification of Diseases (ICD) codes should be used as exposures and outcomes limits comparability and generalizability of results across studies. We aimed to establish consensus among a panel of experts on ICD codes that could become the reference standard and provide guidance around common methodological issues.

APPROACH AND RESULTS

Researchers with an interest in EHR-based NAFLD research were invited to collectively define which administrative codes are most appropriate for documenting exposures and outcomes. We used a modified Delphi approach to reach consensus on several commonly encountered methodological challenges in the field. After two rounds of revision, a high level of agreement (>67%) was reached on all items considered. Full consensus was achieved on a comprehensive list of administrative codes to be considered for inclusion and exclusion criteria in defining exposures and outcomes in EHR-based NAFLD research. We also provide suggestions on how to approach commonly encountered methodological issues and identify areas for future research.

CONCLUSIONS

This expert panel consensus statement can help harmonize and improve generalizability of EHR-based NAFLD research.

Non-Alcoholic Fatty Liver Disease: Implementing Complete Automated Diagnosis and Staging. A Systematic Review

Background: Non-alcoholic fatty liver disease (NAFLD) is a fast-growing pathology around the world, being considered the most common chronic liver disease. It is diagnosed based on the presence of steatosis in more than 5% of hepatocytes without significant alcohol consumption. This review aims to provide a comprehensive overview of current studies of artificial intelligence (AI) applications that may help physicians in implementing a complete automated NAFLD diagnosis and staging. Methods: PubMed, EMBASE, Cochrane Library, and WILEY databases were screened for relevant publications in relation to AI applications in NAFLD. The search terms included: (non-alcoholic fatty liver disease OR NAFLD) AND (artificial intelligence OR machine learning OR neural networks OR deep learning OR automated diagnosis OR computer-aided diagnosis OR digital pathology OR automated ultrasound OR automated computer tomography OR automated magnetic imaging OR electronic health records). Results: Our search identified 37 articles about automated NAFLD diagnosis, out of which 15 articles analyzed imagistic techniques, 15 articles analyzed digital pathology, and 7 articles analyzed electronic health records (EHC). All studies included in this review show an accurate capacity of automated diagnosis and staging in NAFLD using AI-based software. Conclusions: We found significant evidence demonstrating that implementing a complete automated system for NAFLD diagnosis, staging, and risk stratification is currently possible, considering the accuracy, sensibility, and specificity of available AI-based tools.

Application of Artificial Intelligence for the Diagnosis and Treatment of Liver Diseases

Modern medical care produces large volumes of multimodal patient data, which many clinicians struggle to process and synthesize into actionable knowledge. In recent years, artificial intelligence (AI) has emerged as an effective tool in this regard. The field of hepatology is no exception, with a growing number of studies published that apply AI techniques to the diagnosis and treatment of liver diseases. These have included machine-learning algorithms (such as regression models, Bayesian networks, and support vector machines) to predict disease progression, the presence of complications, and mortality; deep-learning algorithms to enable rapid, automated interpretation of radiologic and pathologic images; and natural-language processing to extract clinically meaningful concepts from vast quantities of unstructured data in electronic health records. This review article will provide a comprehensive overview of hepatology-focused AI research, discuss some of the barriers to clinical implementation and adoption, and suggest future directions for the field.

Non-alcoholic fatty liver and chronic kidney disease: Retrospect, introspect, and prospect

With the growing prevalence of obesity and diabetes in the United States and across the world, a rise in the overall incidence and prevalence of non-alcoholic fatty liver disease (NAFLD) is expected. The risk factors for NAFLD are also associated with the development of chronic kidney disease (CKD). We review the epidemiology, risk factors, genetics, implications of gut dysbiosis, and specific pathogenic mechanisms linking NAFLD to CKD. Mechanisms such as ectopic lipid accumulation, cellular signaling abnormalities, and the interplay between fructose consumption and uric acid accumulation have led to the emergence of potential therapeutic implications for this patient population. Transplant evaluation in the setting of both NAFLD and CKD is also reviewed. Potential strategies for surveillance and management include the monitoring of comorbidities, the use of non-invasive fibrosis scoring systems, and the measurement of laboratory markers. Lastly, we discuss the management of patients with NAFLD and CKD, from preventative measures to experimental interventions.

Development of an Algorithm to Identify Cases of Nonalcoholic Steatohepatitis Cirrhosis in the Electronic Health Record

BACKGROUND AND AIMS

Current genetic research of nonalcoholic steatohepatitis (NASH) cirrhosis is limited by our ability to accurately identify cases on a large scale. Our objective was to develop and validate an electronic health record (EHR) algorithm to accurately identify cases of NASH cirrhosis in the EHR.

METHODS

We used Clinical Query 2, a search tool at Beth Israel Deaconess Medical Center, to create a pool of potential NASH cirrhosis cases (n = 5415). We created a training set of 300 randomly selected patients for chart review to confirm cases of NASH cirrhosis. Test characteristics of different algorithms, consisting of diagnosis codes, laboratory values, anthropomorphic measurements, and medication records, were calculated. The algorithms with the highest positive predictive value (PPV) and the highest F score with a PPV ≥ 80% were selected for internal validation using a separate random set of 100 patients from the potential NASH cirrhosis pool. These were then externally validated in another random set of 100 individuals using the research patient data registry tool at Massachusetts General Hospital.

RESULTS

The algorithm with the highest PPV of 100% on internal validation and 92% on external validation consisted of ≥ 3 counts of "cirrhosis, no mention of alcohol" (571.5, K74.6) and ≥ 3 counts of "nonalcoholic fatty liver" (571.8-571.9, K75.81, K76.0) codes in the absence of any diagnosis codes for other common causes of chronic liver disease.

CONCLUSIONS

We developed and validated an EHR algorithm using diagnosis codes that accurately identifies patients with NASH cirrhosis.

Leveraging unstructured data to identify hereditary angioedema patients in electronic medical records

Background

The epidemiologic impact of hereditary angioedema (HAE) is difficult to quantify, due to misclassification in retrospective studies resulting from non-specific diagnostic coding. The aim of this study was to identify cohorts of patients with HAE-1/2 by evaluating structured and unstructured data in a US ambulatory electronic medical record (EMR) database.

Methods

A retrospective feasibility study was performed using the GE Centricity EMR Database (2006–2017). Patients with ≥ 1 diagnosis code for HAE-1/2 (International Classification of Diseases, Ninth Revision, Clinical Modification 277.6 or International Classification of Diseases, Tenth Revision, Clinical Modification D84.1) and/or ≥ 1 physician note regarding HAE-1/2 and ≥ 6 months’ data before and after the earliest code or note (index date) were included. Two mutually exclusive cohorts were created: probable HAE (≥ 2 codes or ≥ 2 notes on separate days) and suspected HAE (only 1 code or note). The impact of manually reviewing physician notes on cohort formation was assessed, and demographic and clinical characteristics of the 2 final cohorts were described.

Results

Initially, 1691 patients were identified: 190 and 1501 in the probable and suspected HAE cohorts, respectively. After physician note review, the confirmed HAE cohort comprised 254 patients and the suspected HAE cohort decreased to 1299 patients; 138 patients were determined not to have HAE and were excluded. The overall false-positive rate for the initial algorithms was 8.2%. Across final cohorts, the median age was 50 years and > 60% of patients were female. HAE-specific prescriptions were identified for 31% and 2% of the confirmed and suspected HAE cohorts, respectively.

Conclusions

Unstructured EMR data can provide valuable information for identifying patients with HAE-1/2. Further research is needed to develop algorithms for more representative HAE cohorts in retrospective studies.

Recent Epidemiology of Nonalcoholic Fatty Liver Disease

The ongoing obesity epidemic and the increasing recognition of metabolic syndrome have contributed to the growing prevalence of nonalcoholic fatty liver disease (NAFLD), the most common form of liver disease worldwide. It is imperative to understand the incidence and prevalence of NAFLD as it is associated with a profound economic burden of hospitalizations, including the shifting trends in liver transplantation. The long-term cumulative healthcare cost of NAFLD patients has been shown to be 80% higher than that of non-NAFLD patients. We explore diagnostic challenges in identifying those with NAFLD who have a higher predilection to progress to end-stage liver disease. We aim to assess all-cause and cause-specific mortality as it relates to NAFLD.

Electronic Medical Record-Based Case Phenotyping for the Charlson Conditions: Scoping Review

Background

Electronic medical records (EMRs) contain large amounts of rich clinical information. Developing EMR-based case definitions, also known as EMR phenotyping, is an active area of research that has implications for epidemiology, clinical care, and health services research.

Objective

This review aims to describe and assess the present landscape of EMR-based case phenotyping for the Charlson conditions.

Methods

A scoping review of EMR-based algorithms for defining the Charlson comorbidity index conditions was completed. This study covered articles published between January 2000 and April 2020, both inclusive. Embase (Excerpta Medica database) and MEDLINE (Medical Literature Analysis and Retrieval System Online) were searched using keywords developed in the following 3 domains: terms related to EMR, terms related to case finding, and disease-specific terms. The manuscript follows the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA) guidelines.

Results

A total of 274 articles representing 299 algorithms were assessed and summarized. Most studies were undertaken in the United States (181/299, 60.5%), followed by the United Kingdom (42/299, 14.0%) and Canada (15/299, 5.0%). These algorithms were mostly developed either in primary care (103/299, 34.4%) or inpatient (168/299, 56.2%) settings. Diabetes, congestive heart failure, myocardial infarction, and rheumatology had the highest number of developed algorithms. Data-driven and clinical rule–based approaches have been identified. EMR-based phenotype and algorithm development reflect the data access allowed by respective health systems, and algorithms vary in their performance.

Conclusions

Recognizing similarities and differences in health systems, data collection strategies, extraction, data release protocols, and existing clinical pathways is critical to algorithm development strategies. Several strategies to assist with phenotype-based case definitions have been proposed.

NAFLD and MAFLD as emerging causes of HCC: A populational study

Background & Aims

There are conflicting data regarding the epidemiology of hepatocellular carcinoma (HCC) arising in the context of non-alcoholic and metabolic-associated fatty liver disease (NAFLD and MAFLD). We aimed to examine the changing contribution of NAFLD and MAFLD, stratified by sex, in a well-defined geographical area and highly characterised HCC population between 1990 and 2014.

Methods

We identified all patients with HCC resident in the canton of Geneva, Switzerland, diagnosed between 1990 and 2014 from the prospective Geneva Cancer Registry and assessed aetiology-specific age-standardised incidence. NAFLD-HCC was diagnosed when other causes of liver disease were excluded in cases with type 2 diabetes, metabolic syndrome, or obesity. Criteria for MAFLD included one or more of the following criteria: overweight/obesity, presence of type 2 diabetes mellitus, or evidence of metabolic dysregulation.

Results

A total of 76/920 (8.3%) of patients were diagnosed with NAFLD-HCC in the canton of Geneva between 1990 and 2014. Between the time periods 1990–1994 and 2010–2014, there was a significant increase in HCC incidence in women (standardised incidence ratio [SIR] 1.83, 95% CI 1.08–3.13, p = 0.026) but not in men (SIR 1.10, 95% CI 0.85–1.43, p = 0.468). In the same timeframe, the proportion of NAFLD-HCC increased more in women (0–29%, p = 0.037) than in men (2–12%, p = 0.010) while the proportion of MAFLD increased from 21% to 68% in both sexes and from 7% to 67% in women (p <0.001). From 2000–2004 to 2010–2014, the SIR of NAFLD-HCC increased to 1.92 (95% CI 0.77–5.08) for men and 12.7 (95% CI 1.63–545) in women, whereas it decreased or remained stable for other major aetiologies of HCC.

Conclusions

In a populational cohort spanning 25 years, the burden of NAFLD and MAFLD associated HCCs increased significantly, driving an increase in HCC incidence, particularly in women.

Lay summary

Hepatocellular carcinoma (HCC) is the most common type of liver cancer, increasingly arising in patients with liver disease caused by metabolic syndrome, termed non-alcoholic fatty liver disease (NAFLD) or metabolic-associated fatty liver disease (MAFLD). We assessed all patients with HCC between 1990 and 2014 in the canton of Geneva (western Switzerland) and found an increase in all HCC cases in this timeframe, particularly in women. In addition, we found that HCC caused by NAFLD or MAFLD significantly increased over the years, particularly in women, possibly driving the increase in overall HCC cases.

•

The burden of HCC arising in the context of non-alcoholic and metabolic-associated fatty liver disease (NAFLD and MAFLD) remains unclear.

•

We assessed all HCC cases between 1990 and 2014 in an area of western Switzerland.

•

We found a significant increase in overall HCC incidence in women but not in men.

•

The proportion of NAFLD- and MAFLD-HCC increased in both sexes, particularly in women.

•

Liver function of MAFLD patients was intermediate between ‘pure’ NAFLD and non-MAFLD individuals.

Evolving Role and Future Directions of Natural Language Processing in Gastroenterology

In line with the current trajectory of healthcare reform, significant emphasis has been placed on improving the utilization of data collected during a clinical encounter. Although the structured fields of electronic health records have provided a convenient foundation on which to begin such efforts, it was well understood that a substantial portion of relevant information is confined in the free-text narratives documenting care. Unfortunately, extracting meaningful information from such narratives is a non-trivial task, traditionally requiring significant manual effort. Today, computational approaches from a field known as Natural Language Processing (NLP) are poised to make a transformational impact in the analysis and utilization of these documents across healthcare practice and research, particularly in procedure-heavy sub-disciplines such as gastroenterology (GI). As such, this manuscript provides a clinically focused review of NLP systems in GI practice. It begins with a detailed synopsis around the state of NLP techniques, presenting state-of-the-art methods and typical use cases in both clinical settings and across other domains. Next, it will present a robust literature review around current applications of NLP within four prominent areas of gastroenterology including endoscopy, inflammatory bowel disease, pancreaticobiliary, and liver diseases. Finally, it concludes with a discussion of open problems and future opportunities of this technology in the field of gastroenterology and health care as a whole.

The Use of Administrative Data to Investigate the Population Burden of Hepatic Encephalopathy

Hepatic encephalopathy (HE) is a devastating complication of cirrhosis with an increasing footprint in global public health. Although the condition is defined using a careful history and examination, we cannot accurately measure the true impact of HE relying on data collected exclusively from clinical studies. For this reason, administrative data sources are necessary to study the population burden of HE. Administrative data is generated with each health care encounter to account for health care resource utilization and is extracted into a dataset for the secondary purpose of research. In order to utilize such data for valid analysis, several pitfalls must be avoided—specifically, selecting the particular database capable of meeting the needs of the study’s aims, paying careful attention to the limits of each given database, and ensuring validity of case definition for HE specific to the dataset. In this review, we summarize the types of data available for and the results of administrative data studies of HE.

Validating a non-invasive, ALT-based non-alcoholic fatty liver phenotype in the million veteran program

Background & aims

Given ongoing challenges in non-invasive non-alcoholic liver disease (NAFLD) diagnosis, we sought to validate an ALT-based NAFLD phenotype using measures readily available in electronic health records (EHRs) and population-based studies by leveraging the clinical and genetic data in the Million Veteran Program (MVP), a multi-ethnic mega-biobank of US Veterans.

Methods

MVP participants with alanine aminotransferases (ALT) >40 units/L for men and >30 units/L for women without other causes of liver disease were compared to controls with normal ALT. Genetic variants spanning eight NAFLD risk or ALT-associated loci (LYPLAL1, GCKR, HSD17B13, TRIB1, PPP1R3B, ERLIN1, TM6SF2, PNPLA3) were tested for NAFLD associations with sensitivity analyses adjusting for metabolic risk factors and alcohol consumption. A manual EHR review assessed performance characteristics of the NAFLD phenotype with imaging and biopsy data as gold standards. Genetic associations with advanced fibrosis were explored using FIB4, NAFLD Fibrosis Score and platelet counts.

Results

Among 322,259 MVP participants, 19% met non-invasive criteria for NAFLD. Trans-ethnic meta-analysis replicated associations with previously reported genetic variants in all but LYPLAL1 and GCKR loci (P<6x10^-3), without attenuation when adjusted for metabolic risk factors and alcohol consumption. At the previously reported LYPLAL1 locus, the established genetic variant did not appear to be associated with NAFLD, however the regional association plot showed a significant association with NAFLD 279kb downstream. In the EHR validation, the ALT-based NAFLD phenotype yielded a positive predictive value 0.89 and 0.84 for liver biopsy and abdominal imaging, respectively (inter-rater reliability (Cohen’s kappa = 0.98)). HSD17B13 and PNPLA3 loci were associated with advanced fibrosis.

Conclusions

We validate a simple, non-invasive ALT-based NAFLD phenotype using EHR data by leveraging previously established NAFLD risk-associated genetic polymorphisms.

Nonalcoholic fatty liver disease increases risk of incident advanced chronic kidney disease: a propensity-matched cohort study

BACKGROUND

As the prevalence of nonalcoholic fatty liver disease (NAFLD) escalates, understanding its potential impact on the development of chronic kidney disease (CKD) is needed.

OBJECTIVE

To determine the longitudinal association of NAFLD with the development of advanced CKD in the United States.

METHODS

A retrospective cohort analysis of the Truven Health MarketScan Database (2006-2015) was conducted. We used Cox proportional hazards models to compare the risk of developing CKD stages 3-5 in patients with NAFLD versus non-NAFLD, identified by ICD-9 codes, after 1:3 propensity score (PS) matching.

RESULTS

In a cohort of 262 619 newly diagnosed patients with NAFLD and 769 878 PS (1:3)-matched non-NAFLD patients, we identified 5766 and 8655 new advanced (stage 3-5) CKD cases, respectively. The crude CKD incidence rate was 8.2 and 5.5 per 1000 person-years in NAFLD and non-NAFLD groups, respectively. In multivariable Cox model, patients with NAFLD had a 41% increased risk of developing advanced CKD compared with non-NAFLD patients [adjusted hazard ratio (aHR), 1.41; 95% confidence interval (CI), 1.36-1.46]. In the sensitivity analysis adjusting for time-varying covariates after NAFLD diagnosis, NAFLD persisted as a significant CKD risk factor (aHR, 1.58; 95% CI, 1.52-1.66) and the association remained significant when stratified by age, gender and pre-existing comorbidities. The risk of CKD increased in NAFLD with compensated cirrhosis (aHR, 1.47; 95% CI, 1.36-1.59) and decompensated cirrhosis (aHR, 2.28; 95% CI, 2.12-2.46).

CONCLUSION

Nonalcoholic fatty liver disease was independently associated with an increased risk of advanced CKD development suggesting renal function screening and regular monitoring are needed in this population.

Cohort Selection for Clinical Trials From Longitudinal Patient Records: Text Mining Approach

Background

Clinical trials are an important step in introducing new interventions into clinical practice by generating data on their safety and efficacy. Clinical trials need to ensure that participants are similar so that the findings can be attributed to the interventions studied and not to some other factors. Therefore, each clinical trial defines eligibility criteria, which describe characteristics that must be shared by the participants. Unfortunately, the complexities of eligibility criteria may not allow them to be translated directly into readily executable database queries. Instead, they may require careful analysis of the narrative sections of medical records. Manual screening of medical records is time consuming, thus negatively affecting the timeliness of the recruitment process.

Objective

Track 1 of the 2018 National Natural Language Processing Clinical Challenge focused on the task of cohort selection for clinical trials, aiming to answer the following question: Can natural language processing be applied to narrative medical records to identify patients who meet eligibility criteria for clinical trials? The task required the participating systems to analyze longitudinal patient records to determine if the corresponding patients met the given eligibility criteria. We aimed to describe a system developed to address this task.

Methods

Our system consisted of 13 classifiers, one for each eligibility criterion. All classifiers used a bag-of-words document representation model. To prevent the loss of relevant contextual information associated with such representation, a pattern-matching approach was used to extract context-sensitive features. They were embedded back into the text as lexically distinguishable tokens, which were consequently featured in the bag-of-words representation. Supervised machine learning was chosen wherever a sufficient number of both positive and negative instances was available to learn from. A rule-based approach focusing on a small set of relevant features was chosen for the remaining criteria.

Results

The system was evaluated using microaveraged F measure. Overall, 4 machine algorithms, including support vector machine, logistic regression, naïve Bayesian classifier, and gradient tree boosting (GTB), were evaluated on the training data using 10–fold cross-validation. Overall, GTB demonstrated the most consistent performance. Its performance peaked when oversampling was used to balance the training data. The final evaluation was performed on previously unseen test data. On average, the F measure of 89.04% was comparable to 3 of the top ranked performances in the shared task (91.11%, 90.28%, and 90.21%). With an F measure of 88.14%, we significantly outperformed these systems (81.03%, 78.50%, and 70.81%) in identifying patients with advanced coronary artery disease.

Conclusions

The holdout evaluation provides evidence that our system was able to identify eligible patients for the given clinical trial with high accuracy. Our approach demonstrates how rule-based knowledge infusion can improve the performance of machine learning algorithms even when trained on a relatively small dataset.

Augmented intelligence with natural language processing applied to electronic health records for identifying patients with non-alcoholic fatty liver disease at risk for disease progression

OBJECTIVE

Electronic health record (EHR) systems contain structured data (such as diagnostic codes) and unstructured data (clinical documentation). Clinical insights can be derived from analyzing both. The use of natural language processing (NLP) algorithms to effectively analyze unstructured data has been well demonstrated. Here we examine the utility of NLP for the identification of patients with non-alcoholic fatty liver disease, assess patterns of disease progression, and identify gaps in care related to breakdown in communication among providers.

MATERIALS AND METHODS

All clinical notes available on the 38,575 patients enrolled in the Mount Sinai BioMe cohort were loaded into the NLP system. We compared analysis of structured and unstructured EHR data using NLP, free-text search, and diagnostic codes with validation against expert adjudication. We then used the NLP findings to measure physician impression of progression from early-stage NAFLD to NASH or cirrhosis. Similarly, we used the same NLP findings to identify mentions of NAFLD in radiology reports that did not persist into clinical notes.

RESULTS

Out of 38,575 patients, we identified 2,281 patients with NAFLD. From the remainder, 10,653 patients with similar data density were selected as a control group. NLP outperformed ICD and text search in both sensitivity (NLP: 0.93, ICD: 0.28, text search: 0.81) and F2 score (NLP: 0.92, ICD: 0.34, text search: 0.81). Of 2281 NAFLD patients, 673 (29.5%) were believed to have progressed to NASH or cirrhosis. Among 176 where NAFLD was noted prior to NASH, the average progression time was 410 days. 619 (27.1%) NAFLD patients had it documented only in radiology notes and not acknowledged in other forms of clinical documentation. Of these, 170 (28.4%) were later identified as having likely developed NASH or cirrhosis after a median 1057.3 days.

DISCUSSION

NLP-based approaches were more accurate at identifying NAFLD within the EHR than ICD/text search-based approaches. Suspected NAFLD on imaging is often not acknowledged in subsequent clinical documentation. Many such patients are later found to have more advanced liver disease. Analysis of information flows demonstrated loss of key information that could have been used to help prevent the progression of early NAFLD (NAFL) to NASH or cirrhosis.

CONCLUSION

For identification of NAFLD, NLP performed better than alternative selection modalities. It then facilitated analysis of knowledge flow between physician and enabled the identification of breakdowns where key information was lost that could have slowed or prevented later disease progression.

Electronic health record year and country of birth testing and patient navigation to increase diagnosis of chronic viral hepatitis

The United States Preventive Services Task Force recommends hepatitis C testing people born from 1945 to 1965, "birth cohort" as well as hepatitis C and hepatitis B testing people from countries of birth with endemic infection risk. We automated the hospital electronic health record system to test birth cohort and those born in countries with endemic infection risk. A script is launched searching the laboratory database upon registration for any hepatitis C antibody, hepatitis C RNA and/or hepatitis B surface antigen result. If no positive result was found, a hepatitis C antibody/reflex RNA and/or hepatitis B surface antigen were ordered. A patient navigator received weekly results and assisted patients with positive serology to schedule an appointment with their primary care provider or treatment specialist. A total of 10 726 participants were hepatitis C antibody tested, with 6.9% antibody positive. Monthly hepatitis C testing from January to July 2016 compared to August 2016-August 2017 increased 342% as a result of "birth cohort" testing. Following country of birth testing, monthly hepatitis B and hepatitis C testing increased 91%, and 44%, respectively, during June-August 2017 compared to September 2017-March 2018. 67% of hepatitis C-positive patients were linked to care. If the navigator contacted the patient, 92% were linked to care, and 32% were treated. Of hepatitis B surface antigen-positive patients, 43% were linked to care, 5% were on treatment, and 15% started treatment. Automated electronic health record ordering of hepatitis C and/or hepatitis B testing is feasible and increases testing. In the population tested, much improvement is needed with linkage to care and treatment.

Evaluation of the Performance of Algorithms That Use Serial Hepatitis C RNA Tests to Predict Treatment Initiation and Sustained Virological Response Among Patients Infected With Hepatitis C Virus

The structure of electronic medical record data prevents easy population-level monitoring of hepatitis C virus (HCV) treatment uptake and cure. Using an HCV registry from a public hospital system in Atlanta, Georgia, we developed multiple algorithms that use serial HCV RNA test results as proxy measures for initiation of direct-acting antiviral (DAA) treatment and sustained virological response (SVR). We calculated sensitivity and positive predictive values (PPVs) by comparing the algorithms with the DAA initiation and SVR results from the registry. From December 2013 to August 2016, 1,807 persons actively infected with HCV were identified in the registry. Of those, 698 initiated DAA treatment on the basis of medical record abstraction; of 442 patients with treatment start and/or end dates, 314 had documented SVR. Treatment algorithm 2 (a detectable HCV RNA result followed by 2 sequential HCV RNA test results) and treatment algorithm 5 (a detectable HCV RNA result followed by 2 sequential HCV RNA test results >6 weeks apart) had the highest sensitivity (87% and 85%, respectively) and PPV (80% and 82%, respectively) combinations. Four SVR algorithms relied on fulfilling treatment algorithm definitions and having an undetectable HCV RNA test result ≥12 weeks after the last HCV RNA result; sensitivity for all 4 algorithms was 79%, and PPV was 92%-93%. Algorithms using serial quantitative HCV RNA results can serve as proxy measures for evaluating population-level DAA treatment and SVR outcomes.

High-Dimensional LASSO-Based Computational Regression Models: Regularization, Shrinkage, and Selection

Regression models are a form of supervised learning methods that are important for machine learning, statistics, and general data science. Despite the fact that classical ordinary least squares (OLS) regression models have been known for a long time, in recent years there are many new developments that extend this model significantly. Above all, the least absolute shrinkage and selection operator (LASSO) model gained considerable interest. In this paper, we review general regression models with a focus on the LASSO and extensions thereof, including the adaptive LASSO, elastic net, and group LASSO. We discuss the regularization terms responsible for inducing coefficient shrinkage and variable selection leading to improved performance metrics of these regression models. This makes these modern, computational regression models valuable tools for analyzing high-dimensional problems.

Application of Machine Learning Methods to Predict Non-Alcoholic Steatohepatitis (NASH) in Non-Alcoholic Fatty Liver (NAFL) Patients

Non-alcoholic fatty liver disease (NAFLD) is the leading cause of chronic liver disease worldwide. NAFLD patients have excessive liver fat (steatosis), without other liver diseases and without excessive alcohol consumption. NAFLD consists of a spectrum of conditions: benign steatosis or non-alcoholic fatty liver (NAFL), steatosis accompanied by inflammation and fibrosis or nonalcoholic steatohepatitis (NASH), and cirrhosis. Given a lack of clinical biomarkers and its asymptomatic nature, NASH is under-diagnosed. We use electronic health records from the Optum Analytics to (1) identify patients diagnosed with benign steatosis and NASH, and (2) train machine learning classifiers for NASH and healthy (non-NASH) populations to (3) predict NASH disease status on patients diagnosed with NAFL. Summarized temporal lab data for alanine aminotransferase, aspartate aminotransferase, and platelet counts, with basic demographic information and type 2 diabetes status were included in the models.

Transforming Diabetes Care Through Artificial Intelligence: The Future Is Here

An estimated 425 million people globally have diabetes, accounting for 12% of the world's health expenditures, and yet 1 in 2 persons remain undiagnosed and untreated. Applications of artificial intelligence (AI) and cognitive computing offer promise in diabetes care. The purpose of this article is to better understand what AI advances may be relevant today to persons with diabetes (PWDs), their clinicians, family, and caregivers. The authors conducted a predefined, online PubMed search of publicly available sources of information from 2009 onward using the search terms "diabetes" and "artificial intelligence." The study included clinically-relevant, high-impact articles, and excluded articles whose purpose was technical in nature. A total of 450 published diabetes and AI articles met the inclusion criteria. The studies represent a diverse and complex set of innovative approaches that aim to transform diabetes care in 4 main areas: automated retinal screening, clinical decision support, predictive population risk stratification, and patient self-management tools. Many of these new AI-powered retinal imaging systems, predictive modeling programs, glucose sensors, insulin pumps, smartphone applications, and other decision-support aids are on the market today with more on the way. AI applications have the potential to transform diabetes care and help millions of PWDs to achieve better blood glucose control, reduce hypoglycemic episodes, and reduce diabetes comorbidities and complications. AI applications offer greater accuracy, efficiency, ease of use, and satisfaction for PWDs, their clinicians, family, and caregivers.

Measurement error and timing of predictor values for multivariable risk prediction models are poorly reported

OBJECTIVE

Measurement error in predictor variables may threaten the validity of clinical prediction models. We sought to evaluate the possible extent of the problem. A secondary objective was to examine whether predictors are measured at the intended moment of model use.

METHODS

A systematic search of Medline was used to identify a sample of articles reporting the development of a clinical prediction model published in 2015. After screening according to a predefined inclusion criteria, information on predictors, strategies to control for measurement error, and intended moment of model use were extracted. Susceptibility to measurement error for each predictor was classified into low and high risks.

RESULTS

Thirty-three studies were reviewed, including 151 different predictors in the final prediction models. Fifty-one (33.7%) predictors were categorized as high risk of error; however, this was not accounted for in the model development. Only 8 (24.2%) studies explicitly stated the intended moment of model use and when the predictors were measured.

CONCLUSION

Reporting of measurement error and intended moment of model use is poor in prediction model studies. There is a need to identify circumstances where ignoring measurement error in prediction models is consequential and whether accounting for the error will improve the predictions.