Early recognition of multiple sclerosis using natural language processing of the electronic health record

Abstract

BackgroundDiagnostic accuracy might be improved by algorithms that searched patients’ clinical notes in the electronic health record (EHR) for signs and symptoms of diseases such as multiple sclerosis (MS). The focus this study was to determine if patients with MS could be identified from their clinical notes prior to the initial recognition by their healthcare providers.MethodsAn MS-enriched cohort of patients with well-established MS (n = 165) and controls (n = 545), was generated from the adult outpatient clinic. A random sample cohort was generated from randomly selected patients (n = 2289) from the same adult outpatient clinic, some of whom had MS (n = 16). Patients’ notes were extracted from the data warehouse and signs and symptoms mapped to UMLS terms using MedLEE. Approximately 1000 MS-related terms occurred significantly more frequently in MS patients’ notes than controls’. Synonymous terms were manually clustered into 50 buckets and used as classification features. Patients were classified as MS or not using Naïve Bayes classification.ResultsClassification of patients known to have MS using notes of the MS-enriched cohort entered after the initial ICD9[MS] code yielded an ROC AUC, sensitivity, and specificity of 0.90 [0.87-0.93], 0.75[0.66-0.82], and 0.91 [0.87-0.93], respectively. Similar classification accuracy was achieved using the notes from the random sample cohort. Classification of patients not yet known to have MS using notes of the MS-enriched cohort entered before the initial ICD9[MS] documentation identified 40% [23–59%] as having MS. Manual review of the EHR of 45 patients of the random sample cohort classified as having MS but lacking an ICD9[MS] code identified four who might have unrecognized MS.ConclusionsDiagnostic accuracy might be improved by mining patients’ clinical notes for signs and symptoms of specific diseases using NLP. Using this approach, we identified patients with MS early in the course of their disease which could potentially shorten the time to diagnosis. This approach could also be applied to other diseases often missed by primary care providers such as cancer. Whether implementing computerized diagnostic support ultimately shortens the time from earliest symptoms to formal recognition of the disease remains to be seen.

Natural language processing of symptoms documented in free-text narratives of electronic health records: a systematic review

“…The main objectives of studies included in this review (Table 2) were to capture or detect symptoms (n ¼ 10) 20,23,27,30,31,35,[37][38][39]42 ; identify, classify, or characterize disease (n ¼ 8) 21,22,24,25,33,43,45,46 ; study adverse drug (n ¼ 5) 32,34,36,41,44 or vaccine (n ¼ 1) 29 events; and identify or detect readmission (n ¼ 1), 26 presence of a device (n ¼ 1), 28 or unplanned clinical encounters (n ¼ 1). 40 Approximately 52% (n ¼ 14) of studies presented symptom-related information as a primary outcome.…”

“…Free-text narratives were primarily from EHRs (n ¼ 13) 20,22,24,26,29,31,35,37,38,41,42,45,46 and data repositories (n ¼ 12). 21,23,25,27,28,32,33,36,39,40,43,44 Free-text narratives used in the 2 remaining studies were obtained from paper records converted into electronic free text 30 and Informatics for Integrating Biology & the Bedside Challenge datasets. 34 Narratives represented both inpatient (eg, admission documents, discharge summaries, emergency department documents, progress notes, nursing narratives) and outpatient (eg, primary care and specialty clinic documents, mental health encounters) settings and were written by various members of the clinical care team (eg, physicians, nurses).…”

Impact of a Web-Based Electronic Health Record on Behavioral Health Service Delivery for Children and Adolescents: Randomized Controlled Trial

“…Electronic health record (EHR) systems are a type of health information technology (HIT) that has been widely proposed as a mechanism for improving the quality and positive impact of health care services [ 1 - 4 ]. Research suggests that a well-implemented and fully-integrated EHR systems can promote complete record-keeping and more efficient access to documentation, facilitating information sharing and better coordination of care [ 1 , 5 , 6 ].…”

“…Electronic health record (EHR) systems are a type of health information technology (HIT) that has been widely proposed as a mechanism for improving the quality and positive impact of health care services [ 1 - 4 ]. Research suggests that a well-implemented and fully-integrated EHR systems can promote complete record-keeping and more efficient access to documentation, facilitating information sharing and better coordination of care [ 1 , 5 , 6 ]. Other proposed, but less well-validated, benefits of EHR systems include: (1) facilitating the use of standardized assessments that can promote progress monitoring, (2) better linkage to evidence-based interventions, (3) more effective communication between providers and supervisors, and (4) use of data to promote quality improvement and research [ 7 ].…”

Electronic medical records in multiple sclerosis research

“…A recent study showed that later initiation of disease‐modifying drug treatments is associated with a greater likelihood of relapse compared with earlier initiation . A study from Cleveland Clinic using EMR data of MS patients identified three subgroups of patients with different depression level trajectories, and another study applied a natural language processing (NLP)‐based classification on clinical notes to identify MS patients before the initial recognition by their healthcare providers …”