Classifying Individuals With Rheumatic Conditions as Financially Insecure Using Electronic Health Record Data and Natural Language Processing: Algorithm Derivation and Validation.

OBJECTIVE: We aimed to examine the feasibility of applying natural language processing (NLP) to unstructured electronic health record (EHR) documents to detect the presence of financial insecurity among patients with rheumatologic disease enrolled in an integrated care management program (iCMP). METHODS: We incorporated supervised, rule-based NLP and statistical methods to identify financial insecurity among patients with rheumatic conditions enrolled in an iCMP (n = 20,395) in a multihospital EHR system. We constructed a lexicon for financial insecurity using data from available knowledge sources and then reviewed EHR notes from 538 randomly selected individuals (training cohort n = 366, validation cohort n = 172). We manually categorized records as having "definite," "possible," or "no" mention of financial insecurity. All available notes were processed using Narrative Information Linear Extraction, a rule-based version of NLP. Models were trained using the NLP features for financial insecurity using logistic, least absolute shrinkage operator (LASSO), and random forest performance characteristic and were compared with the reference standard. RESULTS: A total of 245,142 notes were processed from 538 individual patient records. Financial insecurity was present among 100 (27%) individuals in the training cohort and 63 (37%) in the validation cohort. The LASSO and random forest models performed identically and slightly better than logistic regression, with positive predictive values of 0.90, sensitivities of 0.29, and specificities of 0.98. CONCLUSION: The development of a context-driven lexicon used with rule-based NLP to extract data that identify financial insecurity is feasible for use and improved the capture for presence of financial insecurity with high accuracy. In the absence of a standard lexicon and construct definition for financial insecurity status, additional studies are needed to optimize the sensitivity of algorithms to categorize financial insecurity with construct validity.

Improving the accuracy of automated gout flare ascertainment using natural language processing of electronic health records and linked Medicare claims data.

BACKGROUND: We aimed to determine whether integrating concepts from the notes from the electronic health record (EHR) data using natural language processing (NLP) could improve the identification of gout flares. METHODS: Using Medicare claims linked with EHR, we selected gout patients who initiated the urate-lowering therapy (ULT). Patients' 12-month baseline period and on-treatment follow-up were segmented into 1-month units. We retrieved EHR notes for months with gout diagnosis codes and processed notes for NLP concepts. We selected a random sample of 500 patients and reviewed each of their notes for the presence of a physician-documented gout flare. Months containing at least 1 note mentioning gout flares were considered months with events. We used 60% of patients to train predictive models with LASSO. We evaluated the models by the area under the curve (AUC) in the validation data and examined positive/negative predictive values (P/NPV). RESULTS: We extracted and labeled 839 months of follow-up (280 with gout flares). The claims-only model selected 20 variables (AUC = 0.69). The NLP concept-only model selected 15 (AUC = 0.69). The combined model selected 32 claims variables and 13 NLP concepts (AUC = 0.73). The claims-only model had a PPV of 0.64 [0.50, 0.77] and an NPV of 0.71 [0.65, 0.76], whereas the combined model had a PPV of 0.76 [0.61, 0.88] and an NPV of 0.71 [0.65, 0.76]. CONCLUSION: Adding NLP concept variables to claims variables resulted in a small improvement in the identification of gout flares. Our data-driven claims-only model and our combined claims/NLP-concept model outperformed existing rule-based claims algorithms reliant on medication use, diagnosis, and procedure codes.

Social Determinants of Health Documentation Among Individuals With Rheumatic and Musculoskeletal Conditions in an Integrated Care Management Program.

OBJECTIVE: Social determinants of health (SDoH), such as poverty, are associated with increased burden and severity of rheumatic and musculoskeletal diseases. This study was undertaken to study the prevalence and documentation of SDoH-related needs in electronic health records (EHRs) of individuals with these conditions. METHODS: We randomly selected individuals with ≥1 International Classification of Diseases, Ninth/Tenth Revision (ICD-9/10) code for a rheumatic/musculoskeletal condition enrolled in a multihospital integrated care management program that coordinates care for medically and/or psychosocially complex individuals. We assessed SDoH documentation using terms for financial needs, food insecurity, housing instability, transportation, and medication access according to EHR note review and ICD-10 SDoH billing codes (Z codes). We used multivariable logistic regression to examine associations between demographic factors (age, gender, race, ethnicity, insurance) and ≥1 (versus 0) SDoH need as the odds ratio (OR) with 95% confidence interval (95% CI). RESULTS: Among 558 individuals with rheumatic/musculoskeletal conditions, 249 (45%) had ≥1 SDoH need documented in EHR notes by social workers, care coordinators, nurses, and physicians. A total of 171 individuals (31%) had financial insecurity, 105 (19%) had transportation needs, 94 (17%) had food insecurity; 5% had ≥1 related Z code. In the multivariable model, the odds of having ≥1 SDoH need was 2.45 times higher (95% CI 1.17-5.11) for Black versus White individuals and significantly higher for Medicaid or Medicare beneficiaries versus commercially insured individuals. CONCLUSION: Nearly half of this sample of complex care management patients with rheumatic/musculoskeletal conditions had SDoH documented within EHR notes; financial insecurity was the most prevalent. Only 5% of patients had representative billing codes suggesting that systematic strategies to extract SDoH from notes are needed.

Rheumatology Informatics System for Effectiveness: A National Informatics-Enabled Registry for Quality Improvement.

OBJECTIVE: The Rheumatology Informatics System for Effectiveness (RISE) is a national electronic health record (EHR)-enabled registry. RISE passively collects data from EHRs of participating practices, provides advanced quality measurement and data analytic capacities, and fulfills national quality reporting requirements. Here we report the registry's architecture and initial data, and we demonstrate how RISE is being used to improve the quality of care. METHODS: RISE is a certified Centers for Medicare and Medicaid Services Qualified Clinical Data Registry, allowing collection of data without individual patient informed consent. We analyzed data between October 1, 2014 and September 30, 2015 to characterize initial practices and patients captured in RISE. We also analyzed medication use among rheumatoid arthritis (RA) patients and performance on several quality measures. RESULTS: Across 55 sites, 312 clinicians contributed data to RISE; 72% were in group practice, 21% in solo practice, and 7% were part of a larger health system. Sites contributed data on 239,302 individuals. Among the subset with RA, 34.4% of patients were taking a biologic or targeted synthetic disease-modifying antirheumatic drug (DMARD) at their last encounter, and 66.7% were receiving a nonbiologic DMARD. Examples of quality measures include that 55.2% had a disease activity score recorded, 53.6% a functional status score, and 91.0% were taking a DMARD in the last year. CONCLUSION: RISE provides critical infrastructure for improving the quality of care in rheumatology and is a unique data source to generate new knowledge. Data validation and mapping are ongoing and RISE is available to the research and clinical communities to advance rheumatology.

Traditional Cardiovascular Disease Risk Factor Management in Rheumatoid Arthritis Compared to Matched Nonrheumatoid Arthritis in a US Managed Care Setting.

OBJECTIVE: To compare traditional cardiovascular (CV) risk factor management among patients with rheumatoid arthritis (RA) to that of matched non-RA controls within a large US managed care setting. METHODS: Adult patients with RA and age- and sex-matched general population (general controls) or osteoarthritis (OA) controls were identified between January 1, 2007 and December 31, 2011. We compared health care utilization, measurement, treatment, and treatment target achievement of traditional CV risk factors among subgroups of CV comorbidity during 1 year of followup between RA and controls. RESULTS: A total of 9,440 RA patients, 31,009 general controls, and 10,352 OA controls were included. The proportions with measurements (blood pressure [BP], low-density lipoprotein [LDL] cholesterol, or hemoglobin A1c ), treatment (antihypertensive, statin, or anti-diabetes mellitus medications), and treatment target achievement were slightly higher in patients with RA compared with general controls. Controlling for other factors, RA patients were more likely to have a measurement of BP (odds ratio [OR] 16.77 [95% confidence interval (95% CI) 10.01-28.08]) or LDL cholesterol (OR 1.25 [95% CI 1.13-1.39]), and to receive antihypertensive (OR 1.84 [95% CI 1.47-2.30]) or anti-diabetic medications (OR 1.26 [95% CI 1.01-1.56]) compared to general controls. RA was not associated with receiving a statin (OR 1.01 [95% CI 0.92-1.12]); however, a target LDL level was more likely to be achieved in RA compared to general controls (OR 1.27 [95% CI 1.17-1.37]) as well as target levels of BP and hemoglobin A1c . These results were consistent with results for OA controls except for a lower probability of receiving a statin in RA compared to OA. CONCLUSION: Traditional CV risk factors in patients with RA were not less aggressively managed compared to non-RA controls.

Mortality and extraintestinal cancers in patients with primary sclerosing cholangitis and inflammatory bowel disease.

INTRODUCTION: Primary sclerosing cholangitis (PSC) and inflammatory bowel disease (IBD) frequently co-occur. PSC is associated with increased risk for colorectal cancer (CRC). However, whether PSC is associated with increased risk of extraintestinal cancers or affects mortality in an IBD cohort has not been examined previously. METHODS: In a multi-institutional IBD cohort of IBD, we established a diagnosis of PSC using a novel algorithm incorporating narrative and codified data with high positive and negative predictive value. Our primary outcome was occurrence of extraintestinal and digestive tract cancers. Mortality was determined through monthly linkage to the social security master death index. RESULTS: In our cohort of 5506 patients with CD and 5522 patients with UC, a diagnosis of PSC was established in 224 patients (2%). Patients with IBD-PSC were younger and more likely to be male compared to IBD patients without PSC; three-quarters had UC. IBD-PSC patients had significantly increased overall risk of cancers compared to patients without PSC (OR 4.36, 95% CI 2.99-6.37). Analysis of specific cancer types revealed that a statistically significant excess risk for digestive tract cancer (OR 10.40, 95% CI 6.86-15.76), pancreatic cancer (OR 11.22, 95% CI 4.11-30.62), colorectal cancer (OR 5.00, 95% CI 2.80-8.95), and cholangiocarcinoma (OR 55.31, 95% CI 22.20-137.80) but not for other solid organ or hematologic malignancies. CONCLUSIONS: PSC is associated with increased risk of colorectal and pancreatobiliary cancer but not with excess risk of other solid organ cancers.