Improving Outpatient Medication Lists Using Temporal Reasoning and Clinical Texts (Massachusetts)

Project Final Report (PDF, 178.13 KB) Disclaimer

Project Details - Ended

Project Categories

Summary:

An accurate and complete medication list in a patient’s electronic health record (EHR) is critical to prevent medication prescribing and administration errors. Most prior research uses aggregate structured medication data from the EHR to generate and maintain a reconciled list. However, certain critical information for medication reconciliation and decision support exists in free-text clinical notes that may be unavailable in structured data. Structured data in a standard, predictable form can be easily processed by a computer, but narrative data are not codified and thus pose challenges. Natural language processing (NLP) is any system that manipulates free-form text or speech. NLP applications have been developed to identify and extract medical information from non-structured sources; however, few projects have examined the use of NLP as a method for improving the accuracy of medication lists and facilitating medication reconciliation. This study investigated the feasibility of extracting medication information from non-structured electronic clinical sources within the Longitudinal Medical Record (LMR) system, the ambulatory-care EHR at Brigham and Women’s Hospital in Boston. The extracted information can be subsequently used by clinicians at the point of care, thereby reducing prescription and administrative errors.

Specifically, this project: 1) designed and developed a NLP application called Medical Text Extraction, Reasoning and Mapping System (MTERMS), which identifies medication names and drug signatures and other information from free-text clinical notes, encodes medication names using RxNorm and local terminology in the LMR, conducts terminology mapping simultaneously, and structures the extracted information; 2) evaluated the tool by verifying the NLP output against manual review; and 3) identified requirements for a user interface to efficiently use NLP output for medication reconciliation. MTERMs was noted to function with high accuracy, with an F-measure of 90.6 for free-text notes, and 94.0 for structured notes. The F-measure is a measure of a test’s accuracy in which the highest score is 100. For free-text notes, RxNorm covered 98 percent of the terms, while the local drug dictionary only had 83 percent coverage. When mapping between terminologies, only a 62 percent exact match was achieved.

The study showed that the use of NLP to mine free-text medications in order to assemble a list for medication reconciliation is feasible. In addition, such a system could be used to detect potential medication errors. Application to the clinical setting has challenges however, particularly around the updating and mapping of terminologies.

Improving Outpatient Medication Lists Using Temporal Reasoning and Clinical Texts - 2011

Summary Highlights

  • Principal Investigator: 
  • Funding Mechanism: 
    PAR: HS08-268: Small Research Grant to Improve Health Care Quality Through Health Information Technology (R03)
  • Grant Number: 
    R03 HS 018288
  • Project Period: 
    October 2009 - September 2011
  • AHRQ Funding Amount: 
    $99,949
  • PDF Version: 
    (PDF, 199.2 KB)

Summary: An accurate and complete medication list in a patient's electronic health record (EHR) is critical to prevent medication prescribing and administration errors. Most software systems aggregate structured medication data from the EHR to generate and maintain a reconciled list. However, certain critical information for medication reconciliation and decision support exists in free-text clinical notes that may be unavailable in structured data. Structured data in a standard, predictable form can be processed easily by a computer, but narrative data are not codified and thus pose challenges. Natural language processing (NLP) is any system that manipulates free-form text or speech. NLP applications have been developed to identify and extract medical information from non-structured sources, but few projects have examined the use of NLP as a method for improving the accuracy of medication lists and facilitating medication reconciliation.

One challenge for medication reconciliation is that the drug names from various EHR applications and NLP systems are usually coded using different terminologies (e.g., a local terminology for a specific organization or a commercial terminology) and therefore not interoperable. This study investigated the feasibility of extracting medication information from non-structured electronic clinical sources within the Longitudinal Medical Record (LMR) system, the ambulatory care EHR at Partners HealthCare System. The extracted information can be used by clinicians at the point of care to reduce prescription and administrative errors. This project: 1) designed and developed an NLP application that identifies medication names and drug signatures (e.g., dose amount) and other contextual information (e.g., status) from free-text clinical notes; 2) encoded medication names using RxNorm and local terminology in the LMR; 3) conducted terminology mapping simultaneously; 4) structured the extracted information; 5) evaluated the tool by verifying the NLP output against manual review; and 6) identified requirements for a user interface to use NLP output for efficient medication reconciliation.

Specific Aims:

  • Extract and encode medication information from clinical texts available in an ambulatory electronic medical record system. (Achieved)
  • Apply temporal information (a controlled terminology, domain knowledge, and linguistic knowledge) to develop a mechanism to represent the timing of medication use, detect the changes, and then to organize medications in a chronological order and classify them into appropriate groups. (Achieved)
  • Measure the feasibility and efficiency of the proposed methods and tools for improving the process of medication reconciliation. (Achieved)

2011 Activities: Dr. Zhou and her project team completed the development of the NLP system at the beginning of the year. The system, called the Medical Text Extraction, Reasoning and Mapping System (MTERMS), applies a modular, pipeline approach flowing from a preprocessor to a semantic tagger, a terminology mapper, and a context analyzer to a parser. It extracts free-text medication information (e.g., drug name, dose, and frequency), encodes drug names using different terminologies, and establishes dynamic mappings between them to improve data interoperability.

Thereafter, the project team evaluated the performance of MTERMS in processing medication information from clinical free-text documents. They focused on free-text outpatient clinical notes created mainly by patients’ primary care physicians and medical specialists. Evaluators manually reviewed and compared 30 free-text and 10 structured outpatient notes with MTERMS output. The mapping between RxNorm and a local medication terminology in the LMR was also assessed, and requirements for integrating NLP output to the medication reconciliation process were studied.

Dr. Zhou and her team disseminated the results of the project in Using Medical Text Extraction, Reasoning and Mapping System (MTERMS) to Process Medication Information in Outpatient Clinical Notes, an article in the Proceedings of 2011 Annual Symposium of the American Medical Informatics Association, and in Mapping Partners Master Drug Dictionary to RxNorm using an NLP-based Approach, which was published in the Journal of Biomedical Informatics.

As last self-reported in the AHRQ Research Reporting System, project progress and activities were on track and project budget spending was on target. The project was completed in September 2011.

Preliminary Impact and Findings: Dr. Zhou and her team found that real-time clinical use of NLP in assembling the medication reconciliation list is feasible. However, a real-life application will require change management. For example, a terminology management process to review how updates to terminologies will affect the mappings and to track retired concepts is needed. A common occurrence in electronic order entry systems is free-text medication entries, which represent something of a 'black box' to the systems that process them. NLP could be used to extract coded medications from these entries and allow duplication alerts or a drug interaction system to catch potential medication errors.

Target Population: General

Strategic Goal: Develop and disseminate health IT evidence and evidence-based tools to improve the quality and safety of medication management via the integration and utilization of medication management systems and technologies.

Business Goal: Knowledge Creation

Improving Outpatient Medication Lists Using Temporal Reasoning and Clinical Texts - 2010

Summary Highlights

  • Principal Investigator: 
  • Funding Mechanism: 
    PAR: HS08-268: Small Research Grant to Improve Health Care Quality Through Health Information Technology (R03)
  • Grant Number: 
    R03 HS 018288
  • Project Period: 
    October 2009 – September 2011
  • AHRQ Funding Amount: 
    $99,949
  • PDF Version: 
    (PDF, 308.13 KB)


Target Population: General

Summary: Accurate and complete medication information at the point of care is crucial for delivery of high-quality care and prevention of adverse events. Medication reconciliation has been mandated by the Joint Commission on Accreditation of Healthcare Organizations. Most reconciliation is done by verbally asking the patient what they are taking and comparing it against a medication list. With electronic systems able to do medication reconciliation, structured data from electronic medical records (EMRs) are aggregated with computerized physician order entry systems’ data into a single reconciled medication list. However, critical information such as a change in medication regimen is often in non-structured narrative sources, such as clinical notes. This information must also be reconciled to document the patient’s complete and accurate medication record.

Structured data in a standard, predictable form can be easily processed by a computer, but narrative data are not codified and thus pose challenges. “Natural language processing” (NLP) is any system that manipulates free-form text or speech. NLP applications have been developed to identify and extract medical information from non-structured sources; however, few projects have examined the use of NLP as a method for improving medication reconciliation.

This study is investigating the feasibility of extracting medication information from non-structured electronic clinical sources within the Longitudinal Medical Record system, the Certification Commission for Health Information Technology-certified ambulatory-care EMR at Partners HealthCare System. The extracted information can be subsequently used by clinicians at the point of care, thereby reducing prescription and administrative errors. The project is piloting and testing the use of NLP and temporal-reasoning applications, which identify the timing of medication use, to automatically extract and encode medication and associated temporal information from clinical texts, and to chronologically order and classify medications. The study will measure the feasibility and efficiency of these methods and identify tools for improving medication reconciliation.

Specific Aims:
  • Extract and encode medication information from clinical texts available in an ambulatory electronic medical record system. (Achieved)
  • Apply temporal information (a controlled terminology, domain knowledge, and linguistic knowledge) to develop a mechanism to represent the timing of medication use, detect the changes, and then to organize medications in a chronological order and classify them into appropriate groups. (Ongoing)
  • Measure the feasibility and efficiency of the proposed methods and tools for improving the process of medication reconciliation. (Ongoing)

2010 Activities: The project team identified and sampled patients with chronic diseases in the EMR system and had at least one clinic note per year in the two-year study timeframe. The chronic diseases considered in this study include diabetes, hypertension, congestive heart failure, chronic obstructive pulmonary disease, and coronary artery disease. Two types of data were extracted for these patients: clinical notes and patients’ medication information from their structured medication list (SML). The project team manually compared the differences between medications listed on the SML and those recorded in clinical notes to identify the challenges in extracting and encoding medication information included in clinical text.

The research team developed and is currently refining an NLP tool to extract medication names and signatures from free text clinical notes using standard (RxNorm) and local terminologies as a lexicon. The NegEx and ConText algorithms are used to capture contextual information and to tag identified concepts. The TimeText system, a temporal reasoning application, is used to capture durations and other temporal information found within a clinical note.

Grantee's Most Recent Self-Reported Quarterly Status (as of December 2010): Project progress is completely on track, meeting all milestones on time and project spending is roughly on target.

Preliminary Impact and Findings: Based on the manual comparison of the medications listed in the SML and the clinical notes, the project team identified several unique characteristics in clinical texts that present challenges for using NLP for clinical texts. These challenges include: notes containing detailed dosing regimen adjustment and status changes; some medications discussed in notes, but not ordered; negation, e.g., “except for the Lasix”; abbreviations; misspellings; and coreference. Based on the preliminary results, they also found that the NLP tool can be used to accurately extract and encode medication names and signatures from clinical notes.

Strategic Goal: Develop and disseminate health IT evidence and evidence-based tools to improve the quality and safety of medication management via the integration and utilization of medication management systems and technologies.

Business Goal: Knowledge Creation

Improving Outpatient Medication Lists Using Temporal Reasoning and Clinical Texts - Final Report

Citation:
Zhou L. Improving Outpatient Medication Lists Using Temporal Reasoning and Clinical Texts - Final Report. (Prepared by Brigham and Women's Hospital under Grant No. R03 HS018288). Rockville, MD: Agency for Healthcare Research and Quality, 2011. (PDF, 178.13 KB)

The findings and conclusions in this document are those of the author(s), who are responsible for its content, and do not necessarily represent the views of AHRQ. No statement in this report should be construed as an official position of AHRQ or of the U.S. Department of Health and Human Services.
Principal Investigator: 
Document Type: 
This project does not have any related resource.
This project does not have any related survey.
This project does not have any related project spotlight.
This project does not have any related survey.
This project does not have any related story.
This project does not have any related emerging lesson.