scholarly journals Use of commercially available natural language processing software to identify bleeding from the medical record

Author(s):  
Andrew L Walker ◽  
Cheri Watson ◽  
Ryan Butcher ◽  
Ryan Butcher ◽  
Mark Yandell ◽  
...  

Background: Real-world evidence derived from the electronic medical record (EMR) is increasingly prevalent. How best to ascertain cardiovascular outcomes from EMRs is unknown. We sought to validate a commercially available natural language processing (NLP) software to extract bleeding events. Methods: We included patients with atrial fibrillation and cancer seen at our cancer center from 1/1/2016 to 12/31/2019. A query set based on SNOMED CT expressions was created to represent bleeding from 11 different organ systems. We ran the query against the clinical notes and randomly selected a sample of notes for physician validation. The primary outcome was the positive predictive value (PPV) of the software to identify bleeding events stratified by organ system. Results: We included 1370 patients with mean age 72 years old (SD 1.5) and 35% female. We processed 66,130 notes; the NLP software identified 6522 notes including 654 unique patients with possible bleeding events. Among 1269 randomly selected notes, the PPV of the software ranged from 0.921 for neurologic bleeds to 0.571 for OB/GYN bleeds. Patterns related to false positive bleeding events identified by the software included historic bleeds, hypothetical bleeds, missed negatives, and word errors. Conclusions: NLP may provide an alternative for population-level screening for bleeding outcomes in cardiovascular studies. Human validation is still needed, but an NLP-driven screening approach may improve efficiency. 

2021 ◽  
Vol 27 ◽  
pp. 107602962110131
Author(s):  
Bela Woller ◽  
Austin Daw ◽  
Valerie Aston ◽  
Jim Lloyd ◽  
Greg Snow ◽  
...  

Real-time identification of venous thromboembolism (VTE), defined as deep vein thrombosis (DVT) and pulmonary embolism (PE), can inform a healthcare organization’s understanding of these events and be used to improve care. In a former publication, we reported the performance of an electronic medical record (EMR) interrogation tool that employs natural language processing (NLP) of imaging studies for the diagnosis of venous thromboembolism. Because we transitioned from the legacy electronic medical record to the Cerner product, iCentra, we now report the operating characteristics of the NLP EMR interrogation tool in the new EMR environment. Two hundred randomly selected patient encounters for which the imaging report assessed by NLP that revealed VTE was present were reviewed. These included one hundred imaging studies for which PE was identified. These included computed tomography pulmonary angiography—CTPA, ventilation perfusion—V/Q scan, and CT angiography of the chest/ abdomen/pelvis. One hundred randomly selected comprehensive ultrasound (CUS) that identified DVT were also obtained. For comparison, one hundred patient encounters in which PE was suspected and imaging was negative for PE (CTPA or V/Q) and 100 cases of suspected DVT with negative CUS as reported by NLP were also selected. Manual chart review of the 400 charts was performed and we report the sensitivity, specificity, positive and negative predictive values of NLP compared with manual chart review. NLP and manual review agreed on the presence of PE in 99 of 100 cases, the presence of DVT in 96 of 100 cases, the absence of PE in 99 of 100 cases and the absence of DVT in all 100 cases. When compared with manual chart review, NLP interrogation of CUS, CTPA, CT angiography of the chest, and V/Q scan yielded a sensitivity = 93.3%, specificity = 99.6%, positive predictive value = 97.1%, and negative predictive value = 99%.


Author(s):  
Rod Middleton ◽  
Ashley Akbari ◽  
Hazel Lockhart-Jones ◽  
Jemma Jones ◽  
Charlotte Owen ◽  
...  

ABSTRACT ObjectivesThe UK MS Register is a research project that aims to capture real world data about living with Multiple Sclerosis(MS) in the UK. Launched in 2011, identified data sources were: Directly from People with MS (PwMS) via the internet, from NHS treatment centers via ‘traditional’ database capture and by linkage to routine datasets from the SAIL databank. Data received from the NHS, though ‘gold standard’ in terms of diagnosis, is dependent on clinical staff finding both time and information to enter into a clinical system. System implementations across the NHS are variable, as is clinical time. Therefore, we looked to other complementary methodologies. ApproachThe Clix enrich natural language processing (NLP) software was chosen to see if it could capture a portion of the MS Register minimum clinical dataset, the software matches clinical phrases against SNOMED-CT.40 letters, from 2 NHS Trusts, from 28 patients were loaded. The letters were a mix of MS patients with differing disease subtypes and were dictated by Neurologists, Specialist General Practitioners and MS Specialist Nurses. 20 of the letters were in docx format and 20 as PDF. The letters were parsed by a domain expert for clinical content, scored by data item for sensitivity and specificity. Next the output from the software was scored by another researcher to see if the 12 relevant clinical concepts from the Register dataset had been elicited. Lastly a ruleset was created to look for particular clinical concepts and scored in the same way. ResultsOf the 40 letters one failed to load, the rest were analysed for the specific data items. Date related items were clearly challenging, with only 7% of appointment dates being matched and 22% for date of diagnosis. MS Type (93.3%) and EDSS score (93.75%) were well recognised, additionally symptoms of MS that would be poorly reported in traditional databases were recognised, with fatigue being well highlighted (78.5%) and gait and walking issues (68.7%) Of concern, were a number of false positive results in DMT’s with 15% patients being identified as being on a DMT when this was just being ‘considered’. ConclusionThe NLP pathway could be extremely useful for obtaining hard to capture clinical data for the Register. Further work is needed to reduce errors, even with the current minimal configuration, it's possible to ascertain MS Type, functional score of MS, current medication and potentially disabling symptomology within the condition.


2011 ◽  
Vol 32 (1) ◽  
pp. 188-197 ◽  
Author(s):  
Joshua C. Denny ◽  
Neesha N. Choma ◽  
Josh F. Peterson ◽  
Randolph A. Miller ◽  
Lisa Bastarache ◽  
...  

2007 ◽  
Vol 153 (4) ◽  
pp. 666-673 ◽  
Author(s):  
Serguei S.V. Pakhomov ◽  
Harry Hemingway ◽  
Susan A. Weston ◽  
Steven J. Jacobsen ◽  
Richard Rodeheffer ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document