A-Life Medical: Scientific Papers

Scientific Papers

        The technology of Natural Language Processing is not easy to understand, and as with many scientific endeavors, there are a variety of approaches each with a devoted following. In this regard, A-Life Medical’s innovative and robust LifeCode® NLP system takes a creative and eclectic approach. The What is NLP? white paper, below, gives a brief introduction to the topic. See our NLP page, for more details about what makes LifeCode® the leading application for medical NLP. The publications that are listed below provide in depth treatment of scientific details and controlled studies that evaluate LifeCode’s® performance.

        So that you can explore the field of NLP even further, we have assembled an annotated directory of web links that lead to a wealth of scientific and technical information on NLP. We have particularly listed sites and publications that will be helpful for understanding aspects of NLP that are useful for medical text processing.

        Many of the papers on this page are in PDF format. If you do not have the Adobe Acrobat Reader, you can get it here:

WHITE PAPERS

•	What is NLP? An introduction to the field of Natural Language Processing with particular attention to the applications and issues pertinent to the medical industry.

•	NLP Comparison Comparing LifeCode® to tool-box or statistical systems.

PUBLISHED PAPERS

•	i2b2 NLP Smoking Challenge – The A-Life Medical System Architecture and Methodology (Journal of the American Medical Informatics Association - Jan/Feb 2008) We describe the architecture of LifeCode® (A-Life Medical, Inc.), a natural language processing system for free-text clinical information extraction, our methodology in applying LifeCode® to the I2B2 NLP smoking challenge, and statistical measures for performance evaluation that include a coefficient of variation modeling the observed or expected accuracy of the test standard. Informatics for Integrating Biology & the Bedside (i2b2) is an NIH-funded National Center for Biomedical Computing based at Partners HealthCare System in Boston.

•	Computer Assisted Coding At Its Limits – An Analysis of More Complex Coding Scenarios (Computer Assisted Coding Software Standards Workshop – October 2007) Healthcare organizations that have adopted computer assisted coding applications have benefited from enhanced coder productivity, elimination of manual processes, flexibility of remote coding, improved coding accuracy and consistency, and an overall more manageable and auditable coding process. In this paper, we define three dimensions that characterize complexity which go beyond categorization by medical specialty or setting. Also, we present an analysis of CAC performance for two more complex coding areas: E/M for emergency medicine and interventional radiology.

•	The Impact of Physician Quality Measures on the Coding Process (Computer Assisted Coding Software Standards Workshop – October 2007) Physician coding and billing is undergoing a major change that has expanded the responsibility of coders. The change is taking the form of CMS quality measures that are part of the new Physician Quality Reporting Initiative (PQRI). This paper presents an analysis of the impact of PQRI on three aspects of the coding process: (1) the frequency of occurrence of PQRI reportable cases, (2) the performance of a CAC system to assist the quality measure coding process, and (3) measurement of the coding effort required per case.

•	A-Life Medical I2B2 NLP Smoking Challenge System Architecture & Methodology (Workshop on Challenges in Natural Language Processing for Clinical Data – November 2006) Informatics for Integrating Biology & the Bedside (I2B2) is an NIH-funded National Center for Biomedical Computing based at Partners HealthCare System in Boston, MA. I2B2 sponsored the First Shared Task for Challenges in Natural Language Processing for Clinical Data. In this paper we describe our methodology in applying LifeCode® to the I2B2 NLP Smoking Challenge

•	Computer Assisted Auditing for High Volume Medical Coding (Computer Assisted Coding Software Standards Workshop – September 2006) This paper presents techniques and statistical methods that are developed and implemented for the purpose of auditing the medical coding process and meeting the objectives of producing scores that accurately reflect the quality of the coding work, are comparable across time and between coders and auditors, and that employ statistical methods for production control. The techniques and methods here described are patent pending and are implemented in the A-Life Medical, Inc. CoAudit™ system that is commercially available for auditing both computerized and human coding.

•	Software Engineering of NLP-based Computer Assisted Coding Application (Computer Assisted Coding Software Standards Workshop – September 2006) The development of production quality Natural Language Processing (NLP)-based computer assisted coding applications requires a process-driven approach to software development and quality assurance. This paper describes the aspects of software development and approaches to testing that yield consistent and high quality results for NLP-based Computer Assisted Coding (CAC) applications.

•	Automated Interpretation of Clinical Encounters with Cultural Cues and Electronic Health Records Generation (HLT/NAACL: First International Workshop on Medical Speech Translation - June 2006) This paper describes Accultran, A-Life Medical's prototype system that combines automated speech recognition (ASR) and natural language processing (NLP) to produce Electronic Health Records (EHRs) that are represented using SNOMED-CT terminology and the HL7 Clinical Document Architecture, Release 2 (CDA2). Interpretation (translation) of spoken clinical content is accomplished using the multi-lingual implementations of SNOMED-CT and the unambiguous nature of CDA2.

•	CM-Extractor: An Application for Automating Medical Quality Measures Abstraction in a Hospital Setting (Innovative Applications of Artificial Intelligence Conference - July 2006) This paper discusses the implementation and evaluation of CM-Extractor, a computerized system that automates the medical data abstraction process using natural language processing and a rule-based approach. The results showed that the NLP performed with high accuracy across multiple types of medical documents, and users were able to significantly improve productivity.

•	A Cognitive Grammar Motivated Ontology For Processing With Term Cycles and Non-Classical Negation (FLAIRS - AAAI Conference - May 2005) A technical discussion of an ontology (knowledge representation method) that enables, among other capabilities, correct representation and interpretation of circular and counterfactual statements that are common in natural language.

•	Factors in Deploying Automated Tools for Clinical Abstraction and Coding (IT in Health Care International Conference - Sep 2004) In this paper, we explore human and technical factors involved in justifying, selecting and successfully deploying tools that automate clinical abstraction and coding.

•	Automated Extraction and Normalization of Findings from Cancer-Related Free-text Radiology Reports (AMIA Annual Symposium - Nov 2003) Analysis by Indiana University School of Medicine of LifeCode® for highly detailed information extraction from radiology reports as part of the National Cancer Institute (NCI) Shared Pathology Information Network (SPIN) grant.

•	Text Mining Medical Records with LifeCode (AMIA Annual Symposium - Nov 2001) Analysis of a large corpora text-mining project examining medical methods and outcomes for the treatment of asthma, acute myocardial infarction and gallbladder surgery.

•	LifeCode: A Deployed Application for Automated Medical Coding (AI Magazine, Summer 2001) An overview of the algorithms and system architecture of A-Life Medical’s NLP system. Enhanced version of a paper that garnered A-Life Medical a “Best Innovative Applications of Artificial Intelligence” award from the American Association for Artificial Intelligence (AAAI).

•	LifeCode E/M Coding Performance (AMIA Annual Symposium - Nov 2000) A formal 7-way blind study with both Agreement and Kappa-statistic comparisons of LifeCode® to experienced human coders, particularly as related to Evaluation and Management (E/M) coding.

PRESENTATIONS

•	Extracting Clinical Information from Transcribed Physician Notes (INPUT 2003: Session 18) Power Point presentation on the basics of NLP and how LifeCode® works.