Epidemiology of angina pectoris: role of natural language processing of the medical record

Fusion Modeling: Combining Clinical and Imaging Data to Advance Cardiac Care

In addition to the traditional clinical risk factors, an increasing amount of imaging biomarkers have shown value for cardiovascular risk prediction. Clinical and imaging data are captured from a variety of data sources during multiple patient encounters and are often analyzed independently. Initial studies showed that fusion of both clinical and imaging features results in superior prognostic performance compared with traditional scores. There are different approaches to fusion modeling, combining multiple data resources to optimize predictions, each with its own advantages and disadvantages. However, manual extraction of clinical and imaging data is time and labor intensive and often not feasible in clinical practice. An automated approach for clinical and imaging data extraction is highly desirable. Convolutional neural networks and natural language processing can be utilized for the extraction of electronic medical record data, imaging studies, and free-text data. This review outlines the current status of cardiovascular risk prediction and fusion modeling; and in addition gives an overview of different artificial intelligence approaches to automatically extract data from images and electronic medical records for this purpose.

Natural language processing with machine learning methods to analyze unstructured patient-reported outcomes derived from electronic health records: A systematic review

OBJECTIVE

Natural language processing (NLP) combined with machine learning (ML) techniques are increasingly used to process unstructured/free-text patient-reported outcome (PRO) data available in electronic health records (EHRs). This systematic review summarizes the literature reporting NLP/ML systems/toolkits for analyzing PROs in clinical narratives of EHRs and discusses the future directions for the application of this modality in clinical care.

METHODS

We searched PubMed, Scopus, and Web of Science for studies written in English between 1/1/2000 and 12/31/2020. Seventy-nine studies meeting the eligibility criteria were included. We abstracted and summarized information related to the study purpose, patient population, type/source/amount of unstructured PRO data, linguistic features, and NLP systems/toolkits for processing unstructured PROs in EHRs.

RESULTS

Most of the studies used NLP/ML techniques to extract PROs from clinical narratives (n = 74) and mapped the extracted PROs into specific PRO domains for phenotyping or clustering purposes (n = 26). Some studies used NLP/ML to process PROs for predicting disease progression or onset of adverse events (n = 22) or developing/validating NLP/ML pipelines for analyzing unstructured PROs (n = 19). Studies used different linguistic features, including lexical, syntactic, semantic, and contextual features, to process unstructured PROs. Among the 25 NLP systems/toolkits we identified, 15 used rule-based NLP, 6 used hybrid NLP, and 4 used non-neural ML algorithms embedded in NLP.

CONCLUSIONS

This study supports the potential utility of different NLP/ML techniques in processing unstructured PROs available in EHRs for clinical care. Though using annotation rules for NLP/ML to analyze unstructured PROs is dominant, deploying novel neural ML-based methods is warranted.

Development of a generalizable natural language processing pipeline to extract physician-reported pain from clinical reports: Generated using publicly-available datasets and tested on institutional clinical reports for cancer patients with bone metastases

OBJECTIVE

The majority of cancer patients suffer from severe pain at the advanced stage of their illness. In most cases, cancer pain is underestimated by clinical staff and is not properly managed until it reaches a critical stage. Therefore, detecting and addressing cancer pain early can potentially improve the quality of life of cancer patients. The objective of this research project was to develop a generalizable Natural Language Processing (NLP) pipeline to find and classify physician-reported pain in the radiation oncology consultation notes of cancer patients with bone metastases.

MATERIALS AND METHODS

The texts of 1249 publicly-available hospital discharge notes in the i2b2 database were used as a training and validation set. The MetaMap and NegEx algorithms were implemented for medical terms extraction. Sets of NLP rules were developed to score pain terms in each note. By averaging pain scores, each note was assigned to one of the three verbally-declared pain (VDP) labels, including no pain, pain, and no mention of pain. Without further training, the generalizability of our pipeline in scoring individual pain terms was tested independently using 30 hospital discharge notes from the MIMIC-III database and 30 consultation notes of cancer patients with bone metastasis from our institution's radiation oncology electronic health record. Finally, 150 notes from our institution were used to assess the pipeline's performance at assigning VDP.

RESULTS

Our NLP pipeline successfully detected and quantified pain in the i2b2 summary notes with 93% overall precision and 92% overall recall. Testing on the MIMIC-III database achieved precision and recall of 91% and 86% respectively. The pipeline successfully detected pain with 89% precision and 82% recall on our institutional radiation oncology corpus. Finally, our pipeline assigned a VDP to each note in our institutional corpus with 84% and 82% precision and recall, respectively.

CONCLUSION

Our NLP pipeline enables the detection and classification of physician-reported pain in our radiation oncology corpus. This portable and ready-to-use pipeline can be used to automatically extract and classify physician-reported pain from clinical notes where the pain is not otherwise documented through structured data entry.

Assessment of Agreement Between Human Ratings and Lexicon-Based Sentiment Ratings of Open-Ended Responses on a Behavioral Rating Scale

To date, there is a paucity of research conducting natural language processing (NLP) on the open-ended responses of behavior rating scales. Using three NLP lexicons for sentiment analysis of the open-ended responses of the Behavior Assessment System for Children-Third Edition, the researchers discovered a moderately positive correlation between the human composite rating and the sentiment score using each of the lexicons for strengths comments and a slightly positive correlation for the concerns comments made by guardians and teachers. In addition, the researchers found that as the word count increased for open-ended responses regarding the child's strengths, there was a greater positive sentiment rating. Conversely, as word count increased for open-ended responses regarding child concerns, the human raters scored comments more negatively. The authors offer a proof-of-concept to use NLP-based sentiment analysis of open-ended comments to complement other data for clinical decision making.

Application of optical character recognition with natural language processing for large-scale quality metric data extraction in colonoscopy reports

BACKGROUND AND AIMS

Colonoscopy is commonly performed for colorectal cancer screening in the United States. Reports are often generated in a non-standardized format and are not always integrated into electronic health records. Thus, this information is not readily available for streamlining quality management, participating in endoscopy registries, or reporting of patient- and center-specific risk factors predictive of outcomes. We aim to demonstrate the use of a new hybrid approach using natural language processing of charts that have been elucidated with optical character recognition processing (OCR/NLP hybrid) to obtain relevant clinical information from scanned colonoscopy and pathology reports, a technology co-developed by Cleveland Clinic and eHealth Technologies (West Henrietta, NY, USA).

METHODS

This was a retrospective study conducted at Cleveland Clinic, Cleveland, Ohio, and the University of Minnesota, Minneapolis, Minnesota. A randomly sampled list of outpatient screening colonoscopy procedures and pathology reports was selected. Desired variables were then collected. Two researchers first manually reviewed the reports for the desired variables. Then, the OCR/NLP algorithm was used to obtain the same variables from 3 electronic health records in use at our institution: Epic (Verona, Wisc, USA), ProVation (Minneapolis, Minn, USA) used for endoscopy reporting, and Sunquest PowerPath (Tucson, Ariz, USA) used for pathology reporting.

RESULTS

Compared with manual data extraction, the accuracy of the hybrid OCR/NLP approach to detect polyps was 95.8%, adenomas 98.5%, sessile serrated polyps 99.3%, advanced adenomas 98%, inadequate bowel preparation 98.4%, and failed cecal intubation 99%. Comparison of the dataset collected via NLP alone with that collected using the hybrid OCR/NLP approach showed that the accuracy for almost all variables was >99%.

CONCLUSIONS

Our study is the first to validate the use of a unique hybrid OCR/NLP technology to extract desired variables from scanned procedure and pathology reports contained in image format with an accuracy >95%.

Extracting Angina Symptoms from Clinical Notes Using Pre-Trained Transformer Architectures

Anginal symptoms can connote increased cardiac risk and a need for change in cardiovascular management. In this study, a pre-trained transformer architecture was used to automatically detect and characterize anginal symptoms from within the history of present illness sections of 459 primary care physician notes. Consecutive patients referred for cardiac testing were included. Notes were annotated for positive and negative mentions of chest pain and shortness of breath characterization. The results demonstrate high sensitivity and specificity for the detection of chest pain or discomfort, substernal chest pain, shortness of breath, and dyspnea on exertion. Model performance extracting factors related to provocation and palliation of chest pain were limited by small sample size. Overall, this study shows that pre-trained transformer architectures have promise in automating the extraction of anginal symptoms from clinical texts.

Extraction of radiographic findings from unstructured thoracoabdominal computed tomography reports using convolutional neural network based natural language processing

BACKGROUND

Heart failure (HF) is a major cause of morbidity and mortality. However, much of the clinical data is unstructured in the form of radiology reports, while the process of data collection and curation is arduous and time-consuming.

PURPOSE

We utilized a machine learning (ML)-based natural language processing (NLP) approach to extract clinical terms from unstructured radiology reports. Additionally, we investigate the prognostic value of the extracted data in predicting all-cause mortality (ACM) in HF patients.

MATERIALS AND METHODS

This observational cohort study utilized 122,025 thoracoabdominal computed tomography (CT) reports from 11,808 HF patients obtained between 2008 and 2018. 1,560 CT reports were manually annotated for the presence or absence of 14 radiographic findings, in addition to age and gender. Thereafter, a Convolutional Neural Network (CNN) was trained, validated and tested to determine the presence or absence of these features. Further, the ability of CNN to predict ACM was evaluated using Cox regression analysis on the extracted features.

RESULTS

11,808 CT reports were analyzed from 11,808 patients (mean age 72.8 ± 14.8 years; 52.7% (6,217/11,808) male) from whom 3,107 died during the 10.6-year follow-up. The CNN demonstrated excellent accuracy for retrieval of the 14 radiographic findings with area-under-the-curve (AUC) ranging between 0.83-1.00 (F1 score 0.84-0.97). Cox model showed the time-dependent AUC for predicting ACM was 0.747 (95% confidence interval [CI] of 0.704-0.790) at 30 days.

CONCLUSION

An ML-based NLP approach to unstructured CT reports demonstrates excellent accuracy for the extraction of predetermined radiographic findings, and provides prognostic value in HF patients.

Combining Inpatient and Outpatient Data for Diagnosis of Non-Valvular Atrial Fibrillation Using Electronic Health Records: A Validation Study

Purpose

Previous studies have demonstrated differences in atrial fibrillation (AF) detection based on data from hospital sources without data from outpatient sources. We investigated the detection of documented diagnoses of non-valvular AF in a large Israeli health-care organization using electronic health record data from multiple sources.

Patients and Methods

This was an open-chart validation study. Three distinct algorithms for identifying AF in electronic health records, differing in the source of their International Classification of Diseases, Ninth Revision code and use of the associated free text, were defined. Algorithm 1 incorporated inpatient data with outpatient data and the associated free text. Algorithm 2 incorporated inpatient and outpatient data regardless of the free text associated with AF diagnosis. Algorithm 3 used only inpatient data source. These algorithms were compared to a gold standard and their sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. To establish the gold standard (documentation of arrhythmia based on electrocardiography interpretation or a cardiologist’s written diagnosis), 200 patients at highest risk for having non-valvular AF were randomly selected for open-chart validation by two physicians.

Results

The algorithm that included hospital settings, outpatient settings, and incorporated associated free text in the outpatient records had the optimal balance between all validation measures, with a high level of sensitivity (85.4%), specificity (95.0%), PPV (81.4%), and NPV (96.2%). The alternative algorithm that combined inpatient and outpatient data without free text also performed better than the algorithm that included only hospital data (82.9%, 95.0%, 81.0%, and 95.6%, compared to 70.7%, 96.9%, 85.3%, and 92.8%, sensitivity, specificity, PPV, and NPV, respectively).

Conclusion

In this study, involving a comprehensive data collection from inpatient and outpatient sources, incorporating outpatient data with inpatient data improved the diagnosis of non-valvular AF compared to inpatient data alone.

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

OBJECTIVE

Natural language processing (NLP) of symptoms from electronic health records (EHRs) could contribute to the advancement of symptom science. We aim to synthesize the literature on the use of NLP to process or analyze symptom information documented in EHR free-text narratives.

MATERIALS AND METHODS

Our search of 1964 records from PubMed and EMBASE was narrowed to 27 eligible articles. Data related to the purpose, free-text corpus, patients, symptoms, NLP methodology, evaluation metrics, and quality indicators were extracted for each study.

RESULTS

Symptom-related information was presented as a primary outcome in 14 studies. EHR narratives represented various inpatient and outpatient clinical specialties, with general, cardiology, and mental health occurring most frequently. Studies encompassed a wide variety of symptoms, including shortness of breath, pain, nausea, dizziness, disturbed sleep, constipation, and depressed mood. NLP approaches included previously developed NLP tools, classification methods, and manually curated rule-based processing. Only one-third (n = 9) of studies reported patient demographic characteristics.

DISCUSSION

NLP is used to extract information from EHR free-text narratives written by a variety of healthcare providers on an expansive range of symptoms across diverse clinical specialties. The current focus of this field is on the development of methods to extract symptom information and the use of symptom information for disease classification tasks rather than the examination of symptoms themselves.

CONCLUSION

Future NLP studies should concentrate on the investigation of symptoms and symptom documentation in EHR free-text narratives. Efforts should be undertaken to examine patient characteristics and make symptom-related NLP algorithms or pipelines and vocabularies openly available.

Natural language processing for disease phenotyping in UK primary care records for research: a pilot study in myocardial infarction and death

BACKGROUND

Free text in electronic health records (EHR) may contain additional phenotypic information beyond structured (coded) information. For major health events - heart attack and death - there is a lack of studies evaluating the extent to which free text in the primary care record might add information. Our objectives were to describe the contribution of free text in primary care to the recording of information about myocardial infarction (MI), including subtype, left ventricular function, laboratory results and symptoms; and recording of cause of death. We used the CALIBER EHR research platform which contains primary care data from the Clinical Practice Research Datalink (CPRD) linked to hospital admission data, the MINAP registry of acute coronary syndromes and the death registry. In CALIBER we randomly selected 2000 patients with MI and 1800 deaths. We implemented a rule-based natural language engine, the Freetext Matching Algorithm, on site at CPRD to analyse free text in the primary care record without raw data being released to researchers. We analysed text recorded within 90 days before or 90 days after the MI, and on or after the date of death.

RESULTS

We extracted 10,927 diagnoses, 3658 test results, 3313 statements of negation, and 850 suspected diagnoses from the myocardial infarction patients. Inclusion of free text increased the recorded proportion of patients with chest pain in the week prior to MI from 19 to 27%, and differentiated between MI subtypes in a quarter more patients than structured data alone. Cause of death was incompletely recorded in primary care; in 36% the cause was in coded data and in 21% it was in free text. Only 47% of patients had exactly the same cause of death in primary care and the death registry, but this did not differ between coded and free text causes of death.

CONCLUSIONS

Among patients who suffer MI or die, unstructured free text in primary care records contains much information that is potentially useful for research such as symptoms, investigation results and specific diagnoses. Access to large scale unstructured data in electronic health records (millions of patients) might yield important insights.

The Current Research Landscape of the Application of Artificial Intelligence in Managing Cerebrovascular and Heart Diseases: A Bibliometric and Content Analysis

The applications of artificial intelligence (AI) in aiding clinical decision-making and management of stroke and heart diseases have become increasingly common in recent years, thanks in part to technological advancements and the heightened interest of the research and medical community. This study aims to provide a comprehensive picture of global trends and developments of AI applications relating to stroke and heart diseases, identifying research gaps and suggesting future directions for research and policy-making. A novel analysis approach that combined bibliometrics analysis with a more complex analysis of abstract content using exploratory factor analysis and Latent Dirichlet allocation, which uncovered emerging research domains and topics, was adopted. Data were extracted from the Web of Science database. Results showed topics with the most compelling growth to be AI for big data analysis, robotic prosthesis, robotics-assisted stroke rehabilitation, and minimally invasive surgery. The study also found an emerging landscape of research that was centered on population-specific and early detection of stroke and heart disease. Application of AI in health behavior tracking and improvement as well as the use of robotics in medical diagnostics and prognostication have also been found to attract significant research attention. In light of these findings, it is suggested that the currently under-researched issues of data management, AI model reliability, as well as validation of its clinical utility, need to be further explored in future research and policy decisions to maximize the benefits of AI applications in stroke and heart diseases.

Integrating terminologies into standard SQL: a new approach for research on routine data

Background

Most electronic medical records still contain large amounts of free-text data. Semantic evaluation of such data requires the data to be encoded with sufficient classifications or transformed into a knowledge-based database.

Methods

We present an approach that allows databases accessible via SQL (Structured Query Language) to be searched directly through semantic queries without the need for further transformations. Therefore, we developed I) an extension to SQL named Ontology-SQL (O-SQL) that allows to use semantic expressions, II) a framework that uses a standard terminology server to annotate free-text containing database tables and III) a parser that rewrites O-SQL to SQL, so that such queries can be passed to the database server.

Results

I) We compared several semantic queries published to date and were able to reproduce them in a reduced, highly condensed form. II) The quality of the annotation process was measured against manual annotation, and we found a sensitivity of 97.62% and a specificity of 100.00%. III) Different semantic queries were analyzed, and measured with F-scores between 0.91 and 0.98.

Conclusions

We showed that systematic analysis of free-text-containing medical records is possible with standard tools. The seamless connection of ontologies and standard technologies from the database field represents an important constituent of unstructured data analysis. The developed technology can be readily applied to relationally organized data and supports the increasingly important field of translational research.

Electronic supplementary material

The online version of this article (10.1186/s13326-019-0199-z) contains supplementary material, which is available to authorized users.

Uncovering the Genetic Architecture of Major Depression

There have been several recent studies addressing the genetic architecture of depression. This review serves to take stock of what is known now about the genetics of depression, how it has increased our knowledge and understanding of its mechanisms, and how the information and knowledge can be leveraged to improve the care of people affected. We identify four priorities for how the field of MD genetics research may move forward in future years, namely by increasing the sample sizes available for genome-wide association studies (GWASs), greater inclusion of diverse ancestries and low-income countries, the closer integration of psychiatric genetics with electronic medical records, and the development of the neuroscience toolkit for polygenic disorders.

Ivabradine for coronary artery disease and/or heart failure-a protocol for a systematic review of randomised clinical trials with meta-analysis and Trial Sequential Analysis

BACKGROUND

Coronary artery disease and heart failure are both highly prevalent diseases with a global prevalence of 93 million and 40 million. These patients are at increased risk of morbidity and mortality. The management of these patients involves medical therapy, and both diseases can be treated using the heart rate-lowering drug ivabradine. However, the evidence regarding the use of ivabradine in the treatment of coronary artery disease and/or heart failure is unclear. Our objective is to assess the beneficial and harmful effects of ivabradine in the treatment of coronary artery disease and/or heart failure.

METHODS

This protocol for a systematic review was undertaken using the recommendations of The Cochrane Collaboration, the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P), and the eight-step assessment procedure suggested by Jakobsen and colleagues. We plan to include all relevant randomised clinical trials assessing the use of ivabradine in the treatment of coronary artery disease and/or heart failure. We will search the Cochrane Central Register of Controlled Trials (CENTRAL), Medical Literature Analysis and Retrieval System Online (MEDLINE), Excerpta Medica database (EMBASE), Latin American and Caribbean Health Sciences Literature (LILACS), Science Citation Index Expanded on Web of Science, Chinese Biomedical Literature Database (CBM), China National Knowledge Infrastructure (CNKI), Chinese Science Journal Database (VIP), and BIOSIS in order to identify relevant trials. We will begin the searches in February 2019. All included trials will be assessed and classified at low risk of bias or at high risk of bias. Our primary conclusions will be based on the results from the primary outcomes at low risk of bias. Extracted data will be analysed using Review Manager 5.3 and Trial Sequential Analysis 0.9.5.10. We will create a 'Summary of Findings' table in which we will present our primary and secondary outcomes, and we will assess the quality of evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE).

DISCUSSION

The systematic review will have the potential to aid clinicians in decision-making regarding ivabradine and to benefit patients with coronary artery disease and/or heart failure.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42018112082.

Assessing Information Congruence of Documented Cardiovascular Disease between Electronic Dental and Medical Records

Dentists are more often treating patients with Cardiovascular Diseases (CVD) in their clinics; therefore, dentists may need to alter treatment plans in the presence of CVD. However, it's unclear to what extent patient-reported CVD information is accurately captured in Electronic Dental Records (EDRs). In this pilot study, we aimed to measure the reliability of patient-reported CVD conditions in EDRs. We assessed information congruence by comparing patients' self-reported dental histories to their original diagnosis assigned by their medical providers in the Electronic Medical Record (EMR). To enable this comparison, we encoded patients CVD information from the free-text data of EDRs into a structured format using natural language processing (NLP). Overall, our NLP approach achieved promising performance extracting patients' CVD-related information. We observed disagreement between self-reported EDR data and physician-diagnosed EMR data.

Big data from electronic health records for early and late translational cardiovascular research: challenges and potential

Aims

Cohorts of millions of people's health records, whole genome sequencing, imaging, sensor, societal and publicly available data present a rapidly expanding digital trace of health. We aimed to critically review, for the first time, the challenges and potential of big data across early and late stages of translational cardiovascular disease research.

Methods and results

We sought exemplars based on literature reviews and expertise across the BigData@Heart Consortium. We identified formidable challenges including: data quality, knowing what data exist, the legal and ethical framework for their use, data sharing, building and maintaining public trust, developing standards for defining disease, developing tools for scalable, replicable science and equipping the clinical and scientific work force with new inter-disciplinary skills. Opportunities claimed for big health record data include: richer profiles of health and disease from birth to death and from the molecular to the societal scale; accelerated understanding of disease causation and progression, discovery of new mechanisms and treatment-relevant disease sub-phenotypes, understanding health and diseases in whole populations and whole health systems and returning actionable feedback loops to improve (and potentially disrupt) existing models of research and care, with greater efficiency. In early translational research we identified exemplars including: discovery of fundamental biological processes e.g. linking exome sequences to lifelong electronic health records (EHR) (e.g. human knockout experiments); drug development: genomic approaches to drug target validation; precision medicine: e.g. DNA integrated into hospital EHR for pre-emptive pharmacogenomics. In late translational research we identified exemplars including: learning health systems with outcome trials integrated into clinical care; citizen driven health with 24/7 multi-parameter patient monitoring to improve outcomes and population-based linkages of multiple EHR sources for higher resolution clinical epidemiology and public health.

Conclusion

High volumes of inherently diverse ('big') EHR data are beginning to disrupt the nature of cardiovascular research and care. Such big data have the potential to improve our understanding of disease causation and classification relevant for early translation and to contribute actionable analytics to improve health and healthcare.

Incidence of and Risk Factors for Severe Adverse Events in Elderly Patients Taking Angiotensin-Converting Enzyme Inhibitors or Angiotensin II Receptor Blockers after an Acute Myocardial Infarction

STUDY OBJECTIVE

To assess the incidence of and risk factors associated with severe adverse events in elderly patients who used angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin II receptor blockers (ARBs) after an acute myocardial infarction (AMI).

DESIGN

Retrospective cohort study.

DATA SOURCES

Centers for Medicare & Medicaid Services Chronic Conditions Data Warehouse (Medicare service claims database), American Community Survey of the U.S. Census Bureau, and Multum Lexicon Drug database.

PATIENTS

A total of 101,588 eligible Medicare fee-for-service beneficiaries 66 years or older, who were hospitalized for AMI between January 1, 2008, and December 31, 2009, and used ACEIs or ARBs within 30 days after discharge.

MEASUREMENTS AND MAIN RESULTS

Primary outcomes were hospitalizations for acute renal failure (ARF) and hyperkalemia. The secondary outcome was discontinuation or suspension of ACEI/ARB therapy after a visit to a health care provider. The primary risk factors of interest were age, sex, race/ethnicity, and chronic kidney disease (CKD). Cumulative incidence curves and multivariable Fine-Gray proportional hazards models with 95% confidence intervals (CIs) were used with death as a competing risk in both intention-to-treat (ITT) and as-treated (AT) analyses. In the study cohort, 2.8% experienced ARF, 0.5% experienced hyperkalemia, and 63.7% discontinued ACEI/ARB therapy within 1 year after hospital discharge. Approximately half of the incidence of ARF and hyperkalemia occurred within 6 months after hospital discharge, but the cumulative incidence increased after 6 months. Patients older than 85 years had a higher rate of ARF (ITT hazard ratio [HR] 1.15, 95% CI 1.04-1.28) and hyperkalemia (ITT HR 1.33, 95% CI 1.05-1.68) compared with those aged 65-74 years. Patients with baseline CKD had higher rates of ARF (ITT HR 1.61, 95% CI 1.42-1.82), hyperkalemia (ITT HR 1.41, 95% CI 1.11-1.77), and ACEI/ARB therapy discontinuation or suspension (ITT HR 1.05, 95% CI 1.02-1.09).

CONCLUSION

We found a low incidence of ARF and hyperkalemia in elderly patients treated with ACEIs or ARBs after AMI hospitalization. However, a high rate of treatment discontinuation might prevent a higher rate of occurrence of these events. Long-term careful monitoring of severe adverse events and timely discontinuation of ACEIs or ARBs among elderly patients with advancing age and CKD after an AMI is warranted in clinical practice.

Clinical Data Reuse or Secondary Use: Current Status and Potential Future Progress

Objective: To perform a review of recent research in clinical data reuse or secondary use, and envision future advances in this field. Methods: The review is based on a large literature search in MEDLINE (through PubMed), conference proceedings, and the ACM Digital Library, focusing only on research published between 2005 and early 2016. Each selected publication was reviewed by the authors, and a structured analysis and summarization of its content was developed. Results: The initial search produced 359 publications, reduced after a manual examination of abstracts and full publications. The following aspects of clinical data reuse are discussed: motivations and challenges, privacy and ethical concerns, data integration and interoperability, data models and terminologies, unstructured data reuse, structured data mining, clinical practice and research integration, and examples of clinical data reuse (quality measurement and learning healthcare systems). Conclusion: Reuse of clinical data is a fast-growing field recognized as essential to realize the potentials for high quality healthcare, improved healthcare management, reduced healthcare costs, population health management, and effective clinical research.

Using nationwide ‘big data’ from linked electronic health records to help improve outcomes in cardiovascular diseases: 33 studies using methods from epidemiology, informatics, economics and social science in the ClinicAl disease research using LInked Bespoke studies and Electronic health Records (CALIBER) programme

Background

Electronic health records (EHRs), when linked across primary and secondary care and curated for research use, have the potential to improve our understanding of care quality and outcomes.

Objective

To evaluate new opportunities arising from linked EHRs for improving quality of care and outcomes for patients at risk of or with coronary disease across the patient journey.

Design

Epidemiological cohort, health informatics, health economics and ethnographic approaches were used.

Setting

230 NHS hospitals and 226 general practices in England and Wales.

Participants

Up to 2 million initially healthy adults, 100,000 people with stable coronary artery disease (SCAD) and up to 300,000 patients with acute coronary syndrome.

Main outcome measures

Quality of care, fatal and non-fatal cardiovascular disease (CVD) events.

Data platform and methods

We created a novel research platform [ClinicAl disease research using LInked Bespoke studies and Electronic health Records (CALIBER)] based on linkage of four major sources of EHR data in primary care and national registries. We carried out 33 complementary studies within the CALIBER framework. We developed a web-based clinical decision support system (CDSS) in hospital chest pain clinics. We established a novel consented prognostic clinical cohort of SCAD patients.

Results

CALIBER was successfully established as a valid research platform based on linked EHR data in nearly 2 million adults with > 600 EHR phenotypes implemented on the web portal (seehttps://caliberresearch.org/portal). Despite national guidance, key opportunities for investigation and treatment were missed across the patient journey, resulting in a worse prognosis for patients in the UK compared with patients in health systems in other countries. Our novel, contemporary, high-resolution studies showed heterogeneous associations for CVD risk factors across CVDs. The CDSS did not alter the decision-making behaviour of clinicians in chest pain clinics. Prognostic models using real-world data validly discriminated risk of death and events, and were used in cost-effectiveness decision models.

Conclusions

Emerging ‘big data’ opportunities arising from the linkage of records at different stages of a patient’s journey are vital to the generation of actionable insights into the diagnosis, risk stratification and cost-effective treatment of people at risk of, or with, CVD.

Future work

The vast majority of NHS data remain inaccessible to research and this hampers efforts to improve efficiency and quality of care and to drive innovation. We propose three priority directions for further research. First, there is an urgent need to ‘unlock’ more detailed data within hospitals for the scale of the UK’s 65 million population. Second, there is a need for scaled approaches to using EHRs to design and carry out trials, and interpret the implementation of trial results. Third, large-scale, disease agnostic genetic and biological collections linked to such EHRs are required in order to deliver precision medicine and to innovate discovery.

Study registration

CALIBER studies are registered as follows: study 2 – NCT01569139, study 4 – NCT02176174 and NCT01164371, study 5 – NCT01163513, studies 6 and 7 – NCT01804439, study 8 – NCT02285322, and studies 26–29 – NCT01162187. Optimising the Management of Angina is registered as Current Controlled Trials ISRCTN54381840.

Funding

The National Institute for Health Research (NIHR) Programme Grants for Applied Research programme (RP-PG-0407-10314) (all 33 studies) and additional funding from the Wellcome Trust (study 1), Medical Research Council Partnership grant (study 3), Servier (study 16), NIHR Research Methods Fellowship funding (study 19) and NIHR Research for Patient Benefit (study 33).

A Natural Language Processing Tool for Large-Scale Data Extraction from Echocardiography Reports

Large volumes of data are continuously generated from clinical notes and diagnostic studies catalogued in electronic health records (EHRs). Echocardiography is one of the most commonly ordered diagnostic tests in cardiology. This study sought to explore the feasibility and reliability of using natural language processing (NLP) for large-scale and targeted extraction of multiple data elements from echocardiography reports. An NLP tool, EchoInfer, was developed to automatically extract data pertaining to cardiovascular structure and function from heterogeneously formatted echocardiographic data sources. EchoInfer was applied to echocardiography reports (2004 to 2013) available from 3 different on-going clinical research projects. EchoInfer analyzed 15,116 echocardiography reports from 1684 patients, and extracted 59 quantitative and 21 qualitative data elements per report. EchoInfer achieved a precision of 94.06%, a recall of 92.21%, and an F1-score of 93.12% across all 80 data elements in 50 reports. Physician review of 400 reports demonstrated that EchoInfer achieved a recall of 92–99.9% and a precision of >97% in four data elements, including three quantitative and one qualitative data element. Failure of EchoInfer to correctly identify or reject reported parameters was primarily related to non-standardized reporting of echocardiography data. EchoInfer provides a powerful and reliable NLP-based approach for the large-scale, targeted extraction of information from heterogeneous data sources. The use of EchoInfer may have implications for the clinical management and research analysis of patients undergoing echocardiographic evaluation.

Defining disease phenotypes using national linked electronic health records: a case study of atrial fibrillation

BACKGROUND

National electronic health records (EHR) are increasingly used for research but identifying disease cases is challenging due to differences in information captured between sources (e.g. primary and secondary care). Our objective was to provide a transparent, reproducible model for integrating these data using atrial fibrillation (AF), a chronic condition diagnosed and managed in multiple ways in different healthcare settings, as a case study.

METHODS

Potentially relevant codes for AF screening, diagnosis, and management were identified in four coding systems: Read (primary care diagnoses and procedures), British National Formulary (BNF; primary care prescriptions), ICD-10 (secondary care diagnoses) and OPCS-4 (secondary care procedures). From these we developed a phenotype algorithm via expert review and analysis of linked EHR data from 1998 to 2010 for a cohort of 2.14 million UK patients aged ≥ 30 years. The cohort was also used to evaluate the phenotype by examining associations between incident AF and known risk factors.

RESULTS

The phenotype algorithm incorporated 286 codes: 201 Read, 63 BNF, 18 ICD-10, and four OPCS-4. Incident AF diagnoses were recorded for 72,793 patients, but only 39.6% (N = 28,795) were recorded in primary care and secondary care. An additional 7,468 potential cases were inferred from data on treatment and pre-existing conditions. The proportion of cases identified from each source differed by diagnosis age; inferred diagnoses contributed a greater proportion of younger cases (≤ 60 years), while older patients (≥ 80 years) were mainly diagnosed in SC. Associations of risk factors (hypertension, myocardial infarction, heart failure) with incident AF defined using different EHR sources were comparable in magnitude to those from traditional consented cohorts.

CONCLUSIONS

A single EHR source is not sufficient to identify all patients, nor will it provide a representative sample. Combining multiple data sources and integrating information on treatment and comorbid conditions can substantially improve case identification.

Optimising the use of electronic health records to estimate the incidence of rheumatoid arthritis in primary care: what information is hidden in free text?

Background

Primary care databases are a major source of data for epidemiological and health services research. However, most studies are based on coded information, ignoring information stored in free text. Using the early presentation of rheumatoid arthritis (RA) as an exemplar, our objective was to estimate the extent of data hidden within free text, using a keyword search.

Methods

We examined the electronic health records (EHRs) of 6,387 patients from the UK, aged 30 years and older, with a first coded diagnosis of RA between 2005 and 2008. We listed indicators for RA which were present in coded format and ran keyword searches for similar information held in free text. The frequency of indicator code groups and keywords from one year before to 14 days after RA diagnosis were compared, and temporal relationships examined.

Results

One or more keyword for RA was found in the free text in 29% of patients prior to the RA diagnostic code. Keywords for inflammatory arthritis diagnoses were present for 14% of patients whereas only 11% had a diagnostic code. Codes for synovitis were found in 3% of patients, but keywords were identified in an additional 17%. In 13% of patients there was evidence of a positive rheumatoid factor test in text only, uncoded. No gender differences were found. Keywords generally occurred close in time to the coded diagnosis of rheumatoid arthritis. They were often found under codes indicating letters and communications.

Conclusions

Potential cases may be missed or wrongly dated when coded data alone are used to identify patients with RA, as diagnostic suspicions are frequently confined to text. The use of EHRs to create disease registers or assess quality of care will be misleading if free text information is not taken into account. Methods to facilitate the automated processing of text need to be developed and implemented.

Optimising use of electronic health records to describe the presentation of rheumatoid arthritis in primary care: a strategy for developing code lists

Background

Research using electronic health records (EHRs) relies heavily on coded clinical data. Due to variation in coding practices, it can be difficult to aggregate the codes for a condition in order to define cases. This paper describes a methodology to develop ‘indicator markers’ found in patients with early rheumatoid arthritis (RA); these are a broader range of codes which may allow a probabilistic case definition to use in cases where no diagnostic code is yet recorded.

Methods

We examined EHRs of 5,843 patients in the General Practice Research Database, aged ≥30y, with a first coded diagnosis of RA between 2005 and 2008. Lists of indicator markers for RA were developed initially by panels of clinicians drawing up code-lists and then modified based on scrutiny of available data. The prevalence of indicator markers, and their temporal relationship to RA codes, was examined in patients from 3y before to 14d after recorded RA diagnosis.

Findings

Indicator markers were common throughout EHRs of RA patients, with 83.5% having 2 or more markers. 34% of patients received a disease-specific prescription before RA was coded; 42% had a referral to rheumatology, and 63% had a test for rheumatoid factor. 65% had at least one joint symptom or sign recorded and in 44% this was at least 6-months before recorded RA diagnosis.

Conclusion

Indicator markers of RA may be valuable for case definition in cases which do not yet have a diagnostic code. The clinical diagnosis of RA is likely to occur some months before it is coded, shown by markers frequently occurring ≥6 months before recorded diagnosis. It is difficult to differentiate delay in diagnosis from delay in recording. Information concealed in free text may be required for the accurate identification of patients and to assess the quality of care in general practice.

Longitudinal analysis of pain in patients with metastatic prostate cancer using natural language processing of medical record text

Objectives

To test the feasibility of using text mining to depict meaningfully the experience of pain in patients with metastatic prostate cancer, to identify novel pain phenotypes, and to propose methods for longitudinal visualization of pain status.

Materials and methods

Text from 4409 clinical encounters for 33 men enrolled in a 15-year longitudinal clinical/molecular autopsy study of metastatic prostate cancer (Project to ELIminate lethal CANcer) was subjected to natural language processing (NLP) using Unified Medical Language System-based terms. A four-tiered pain scale was developed, and logistic regression analysis identified factors that correlated with experience of severe pain during each month.

Results

NLP identified 6387 pain and 13 827 drug mentions in the text. Graphical displays revealed the pain ‘landscape’ described in the textual records and confirmed dramatically increasing levels of pain in the last years of life in all but two patients, all of whom died from metastatic cancer. Severe pain was associated with receipt of opioids (OR=6.6, p<0.0001) and palliative radiation (OR=3.4, p=0.0002). Surprisingly, no severe or controlled pain was detected in two of 33 subjects’ clinical records. Additionally, the NLP algorithm proved generalizable in an evaluation using a separate data source (889 Informatics for Integrating Biology and the Bedside (i2b2) discharge summaries).

Discussion

Patterns in the pain experience, undetectable without the use of NLP to mine the longitudinal clinical record, were consistent with clinical expectations, suggesting that meaningful NLP-based pain status monitoring is feasible. Findings in this initial cohort suggest that ‘outlier’ pain phenotypes useful for probing the molecular basis of cancer pain may exist.

Limitations

The results are limited by a small cohort size and use of proprietary NLP software.

Conclusions

We have established the feasibility of tracking longitudinal patterns of pain by text mining of free text clinical records. These methods may be useful for monitoring pain management and identifying novel cancer phenotypes.

The freetext matching algorithm: a computer program to extract diagnoses and causes of death from unstructured text in electronic health records

Background

Electronic health records are invaluable for medical research, but much information is stored as free text rather than in a coded form. For example, in the UK General Practice Research Database (GPRD), causes of death and test results are sometimes recorded only in free text. Free text can be difficult to use for research if it requires time-consuming manual review. Our aim was to develop an automated method for extracting coded information from free text in electronic patient records.

Methods

We reviewed the electronic patient records in GPRD of a random sample of 3310 patients who died in 2001, to identify the cause of death. We developed a computer program called the Freetext Matching Algorithm (FMA) to map diagnoses in text to the Read Clinical Terminology. The program uses lookup tables of synonyms and phrase patterns to identify diagnoses, dates and selected test results. We tested it on two random samples of free text from GPRD (1000 texts associated with death in 2001, and 1000 general texts from cases and controls in a coronary artery disease study), comparing the output to the U.S. National Library of Medicine’s MetaMap program and the gold standard of manual review.

Results

Among 3310 patients registered in the GPRD who died in 2001, the cause of death was recorded in coded form in 38.1% of patients, and in the free text alone in 19.4%. On the 1000 texts associated with death, FMA coded 683 of the 735 positive diagnoses, with precision (positive predictive value) 98.4% (95% confidence interval (CI) 97.2, 99.2) and recall (sensitivity) 92.9% (95% CI 90.8, 94.7). On the general sample, FMA detected 346 of the 447 positive diagnoses, with precision 91.5% (95% CI 88.3, 94.1) and recall 77.4% (95% CI 73.2, 81.2), which was similar to MetaMap.

Conclusions

We have developed an algorithm to extract coded information from free text in GP records with good precision. It may facilitate research using free text in electronic patient records, particularly for extracting the cause of death.

Extracting diagnoses and investigation results from unstructured text in electronic health records by semi-supervised machine learning

BACKGROUND

Electronic health records are invaluable for medical research, but much of the information is recorded as unstructured free text which is time-consuming to review manually.

AIM

To develop an algorithm to identify relevant free texts automatically based on labelled examples.

METHODS

We developed a novel machine learning algorithm, the 'Semi-supervised Set Covering Machine' (S3CM), and tested its ability to detect the presence of coronary angiogram results and ovarian cancer diagnoses in free text in the General Practice Research Database. For training the algorithm, we used texts classified as positive and negative according to their associated Read diagnostic codes, rather than by manual annotation. We evaluated the precision (positive predictive value) and recall (sensitivity) of S3CM in classifying unlabelled texts against the gold standard of manual review. We compared the performance of S3CM with the Transductive Vector Support Machine (TVSM), the original fully-supervised Set Covering Machine (SCM) and our 'Freetext Matching Algorithm' natural language processor.

RESULTS

Only 60% of texts with Read codes for angiogram actually contained angiogram results. However, the S3CM algorithm achieved 87% recall with 64% precision on detecting coronary angiogram results, outperforming the fully-supervised SCM (recall 78%, precision 60%) and TSVM (recall 2%, precision 3%). For ovarian cancer diagnoses, S3CM had higher recall than the other algorithms tested (86%). The Freetext Matching Algorithm had better precision than S3CM (85% versus 74%) but lower recall (62%).

CONCLUSIONS

Our novel S3CM machine learning algorithm effectively detected free texts in primary care records associated with angiogram results and ovarian cancer diagnoses, after training on pre-classified test sets. It should be easy to adapt to other disease areas as it does not rely on linguistic rules, but needs further testing in other electronic health record datasets.

The role of the electronic medical record in the assessment of health related quality of life

We applied a hybrid Natural Language Processing (NLP) and machine learning (ML) approach (NLP-ML) to assessment of health related quality of life (HRQOL). The approach uses text patterns extracted from HRQOL inventories and electronic medical records (EMR) as predictive features for training ML classifiers. On a cohort of 200 patients, our approach agreed with patient self-report (EQ5D) and manual audit of the EMR 65-74% of the time. In an independent cohort of 285 patients, we found no association of HRQOL (by EQ5D or NLP-ML) with quality measures of metabolic control (HbA1c, Blood Pressure, Lipids). In addition; while there was no association between patient self-report of HRQOL and cost of care, abnormalities in Usual Activities and Anxiety/Depression assessed by NLP-ML were 40-70% more likely to be associated with greater health care costs. Our method represents an efficient and scalable surrogate measure of HRQOL to predict healthcare spending in ambulatory diabetes patients.

Under-documentation of chronic kidney disease in the electronic health record in outpatients

OBJECTIVE

To ascertain if outpatients with moderate chronic kidney disease (CKD) had their condition documented in their notes in the electronic health record (EHR).

DESIGN

Outpatients with CKD were selected based on a reduced estimated glomerular filtration rate and their notes extracted from the Columbia University data warehouse. Two lexical-based classification tools (classifier and word-counter) were developed to identify documentation of CKD in electronic notes.

MEASUREMENTS

The tools categorized patients' individual notes on the basis of the presence of CKD-related terms. Patients were categorized as appropriately documented if their notes contained reference to CKD when CKD was present.

RESULTS

The sensitivities of the classifier and word-count methods were 95.4% and 99.8%, respectively. The specificity of both was 99.8%. Categorization of individual patients as appropriately documented was 96.9% accurate. Of 107 patients with manually verified moderate CKD, 32 (22%) lacked appropriate documentation. Patients whose CKD had not been appropriately documented were significantly less likely to be on renin-angiotensin system inhibitors or have urine protein quantified, and had the illness for half as long (15.1 vs 30.7 months; p<0.01) compared to patients with documentation.

CONCLUSION

Our studies show that lexical-based classification tools can accurately ascertain if appropriate documentation of CKD is present in a EHR. Using this method, we demonstrated under-documentation of patients with moderate CKD. Under-documented patients were less likely to receive CKD guideline recommended care. A tool that prompts providers to document CKD might shorten the time to implementing guideline-based recommendations.

A high throughput semantic concept frequency based approach for patient identification: a case study using type 2 diabetes mellitus clinical notes

Current research on high throughput identification of patients with a specific phenotype is in its infancy. There is an urgent need to develop a general automatic approach for patient identification.

OBJECTIVE

We took advantage of Mayo Clinic electronic clinical notes and proposed a novel method of combining NLP, machine learning, and ontology for automatic patient identification. We also investigated the benefits of involving existing SNOMED semantic knowledge in a patient identification task.

METHODS

the SVM algorithm was applied on SNOMED concept units extracted from T2DM case/control clinical notes. Precision, recall, and F-score were calculated to evaluate the performance.

RESULTS

This approach achieved an F-score of above 0.950 for both groups when using all identified concept units as features. Concept units from semantic type-Disease or Syndrome contain the most important information for patient identification. Our results also implied that the coarse level concepts contain enough information to classify T2DM cases/controls.

Automated evaluation of electronic discharge notes to assess quality of care for cardiovascular diseases using Medical Language Extraction and Encoding System (MedLEE)

The objective of this study was to develop and validate an automated acquisition system to assess quality of care (QC) measures for cardiovascular diseases. This system combining searching and retrieval algorithms was designed to extract QC measures from electronic discharge notes and to estimate the attainment rates to the current standards of care. It was developed on the patients with ST-segment elevation myocardial infarction and tested on the patients with unstable angina/non-ST-segment elevation myocardial infarction, both diseases sharing almost the same QC measures. The system was able to reach a reasonable agreement (kappa value) with medical experts from 0.65 (early reperfusion rate) to 0.97 (beta-blockers and lipid-lowering agents before discharge) for different QC measures in the test set, and then applied to evaluate QC in the patients who underwent coronary artery bypass grafting surgery. The result has validated a new tool to reliably extract QC measures for cardiovascular diseases.

Automated detection of follow-up appointments using text mining of discharge records

OBJECTIVE

To determine whether text mining can accurately detect specific follow-up appointment criteria in free-text hospital discharge records.

DESIGN

Cross-sectional study.

SETTING

Mayo Clinic Rochester hospitals.

PARTICIPANTS

Inpatients discharged from general medicine services in 2006 (n = 6481).

INTERVENTIONS

Textual hospital dismissal summaries were manually reviewed to determine whether the records contained specific follow-up appointment arrangement elements: date, time and either physician or location for an appointment. The data set was evaluated for the same criteria using SAS Text Miner software. The two assessments were compared to determine the accuracy of text mining for detecting records containing follow-up appointment arrangements.

MAIN OUTCOME MEASURES

Agreement of text-mined appointment findings with gold standard (manual abstraction) including sensitivity, specificity, positive predictive and negative predictive values (PPV and NPV).

RESULTS

About 55.2% (3576) of discharge records contained all criteria for follow-up appointment arrangements according to the manual review, 3.2% (113) of which were missed through text mining. Text mining incorrectly identified 3.7% (107) follow-up appointments that were not considered valid through manual review. Therefore, the text mining analysis concurred with the manual review in 96.6% of the appointment findings. Overall sensitivity and specificity were 96.8 and 96.3%, respectively; and PPV and NPV were 97.0 and 96.1%, respectively.

ANALYSIS

of individual appointment criteria resulted in accuracy rates of 93.5% for date, 97.4% for time, 97.5% for physician and 82.9% for location.

CONCLUSION

Text mining of unstructured hospital dismissal summaries can accurately detect documentation of follow-up appointment arrangement elements, thus saving considerable resources for performance assessment and quality-related research.

Description of a rule-based system for the i2b2 challenge in natural language processing for clinical data

The Obesity Challenge, sponsored by Informatics for Integrating Biology and the Bedside (i2b2), a National Center for Biomedical Computing, asked participants to build software systems that could "read" a patient's clinical discharge summary and replicate the judgments of physicians in evaluating presence or absence of obesity and 15 comorbidities. The authors describe their methodology and discuss the results of applying Lockheed Martin's rule-based natural language processing (NLP) capability, ClinREAD. We tailored ClinREAD with medical domain expertise to create assigned default judgments based on the most probable results as defined in the ground truth. It then used rules to collect evidence similar to the evidence that the human judges likely relied upon, and applied a logic module to weigh the strength of all evidence collected to arrive at final judgments. The Challenge results suggest that rule-based systems guided by human medical expertise are capable of solving complex problems in machine processing of medical text.

Presentation of stable angina pectoris among women and South Asian people

BACKGROUND

There is speculation that women and South Asian people are more likely than men and white people to report atypical angina and that they are less likely to undergo invasive management of angina. We sought to determine whether atypical symptoms of angina pectoris in women and South Asians impacted clinically important outcomes and clinical management.

METHODS

We prospectively identified 2189 South Asian people and 5605 white people with recent-onset chest pain at 6 chest-pain clinics in the United Kingdom. We documented hospital admissions for acute coronary syndromes, coronary deaths as well as coronary angiography and revascularization procedures.

RESULTS

Atypical chest pain was reported by more women than men (56.5% vs 54.5%, p < 0.054) and by more South Asian patients than white patients (59.9% vs 52.5%, p < 0.001). Typical symptoms were associated with coronary death or acute coronary syndromes among women (hazard ratio [HR] 2.30, 95% CI 1.70-3.11, p < 0.001) but not among men (HR 1.23, 95% CI 0.96-1.57, p = 0.10). Typical symptoms were associated with coronary outcomes in both South Asian and white patients. Among those with typical symptoms, women (HR 0.76, 95% CI 0.63-0.92, p = 0.004) and South Asian patients (HR 0.52, 95% CI 0.41-0.67, p < 0.001) were less likely than men and white patients to receive angiography.

INTERPRETATION

Compared to those with atypical chest pain, women and South Asian patients with typical pain had worse clinical outcomes. However, sex and ethnic background did not explain differences in the use of invasive procedures.

Agreement between patient-reported symptoms and their documentation in the medical record

OBJECTIVES

To determine the agreement between patient-reported symptoms of chest pain, dyspnea, and cough and the documentation of these symptoms by physicians in the electronic medical record.

METHODS

Symptoms reported on patient-provided information forms between January 1, 2006, and June 30, 2006, were compared with those identified by natural language processing of the text of clinical notes from care providers. Terms that represent the 3 symptoms were used to search clinical notes electronically with subsequent manual identification of the context (eg, affirmative, negated, family history) in which they occurred. Results were reported using positive and negative agreement, and kappa statistics.

RESULTS

Symptoms reported by 1119 patients age 18 years or older were compared with the nonnegated terms identified in their clinical notes. Positive agreement was 74, 70, and 63 for chest pain, dyspnea, and cough, while negative agreement was 78, 76, and 75, respectively. Kappa statistics were 0.52 (95% confidence interval [CI] = 0.44, 0.60) for chest pain, 0.46 (95% CI = 0.37, 0.54) for dyspnea, and 0.38 (95% CI = 0.28, 0.48) for cough. Positive agreement was higher for older men (P >.05), and negative agreement was higher for younger women (P >.05).

CONCLUSIONS

We found discordance between patient self-report and documentation of symptoms in the medical record. This discordance has important implications for research studies that rely on symptom information for patient identification and may have clinical implications that must be evaluated for potential impact on quality of care, patient safety, and outcomes.

Automatic classification of foot examination findings using clinical notes and machine learning

We examine the feasibility of a machine learning approach to identification of foot examination (FE) findings from the unstructured text of clinical reports. A Support Vector Machine (SVM) based system was constructed to process the text of physical examination sections of in- and out-patient clinical notes to identify if the findings of structural, neurological, and vascular components of a FE revealed normal or abnormal findings or were not assessed. The system was tested on 145 randomly selected patients for each FE component using 10-fold cross validation. The accuracy was 80%, 87% and 88% for structural, neurological, and vascular component classifiers, respectively. Our results indicate that using machine learning to identify FE findings from clinical reports is a viable alternative to manual review and warrants further investigation. This application may improve quality and safety by providing inexpensive and scalable methodology for quality and risk factor assessments at the point of care.

Performing stable angina pectoris: an ethnographic study

Symptoms play a crucial part in the formulation of medical diagnoses, yet the construction and interpretation of symptom narratives is not well understood. The diagnosis of angina is largely based on symptoms, but a substantial minority of patients diagnosed with "non-cardiac" chest pain go on to have a heart attack. In this ethnographic study our aims were to understand: (1) how the patients' accounts are performed or enacted in consultations with doctors; (2) the ways in which ambiguity in the symptom narrative is managed by doctors; and (3) how doctors reach or do not reach a diagnostic decision. We observed 59 consultations of patients in a UK teaching hospital with new onset chest pain who had been referred for a specialist opinion in ambulatory care. We found that patients rarely gave a history that, without further interrogation, satisfied the doctors, who actively restructured the complex narrative until it fitted a diagnostic canon, detaching it from the patient's interpretation and explanation. A minority of doctors asked about chest pain symptoms outside the canon. Re-structuring into the canonical classification was sometimes resisted by patients who contested key concepts, like exertion. Symptom narratives were sometimes unstable, with central features changing on interrogation and re-telling. When translation was required for South Asian patients, doctors considered the history less relevant to the diagnosis. Diagnosis and effective treatment could be enhanced by research on the diagnostic and prognostic value of the terms patients use to describe their symptoms.

Prognosis of stable angina pectoris: why we need larger population studies with higher endpoint resolution

The prognosis of angina was described as "unhappy" by the Framingham investigators and as little different from that of 1-year survivors of acute myocardial infarction. Yet recent clinical trials now report that angina has a good prognosis with adverse outcomes reduced to "normal levels". These disparate prognostic assessments may not be incompatible, applying as they do to population cohorts (Framingham) and selected participants in clinical trials. Comparisons between studies are further complicated by the absence of agreed case definitions for stable angina (contrast this with acute coronary syndromes). Our recent data show that for patients with recent onset symptoms attending chest pain clinics, angina remains a high-risk diagnosis and although many patients receive symptomatic benefit from revascularisation, prognosis is usually unaffected. This leaves little room for complacency and, with angina the commonest initial manifestation of coronary artery disease, there is the opportunity for early detection, risk stratification and treatment to modify outcomes. Meanwhile, larger population-based studies are needed to define the patient journey from earliest presentation through the various syndrome transitions to coronary or noncardiac death in order to increase understanding of the aetiological and prognostic differences between the different coronary disease phenotypes.