학술논문
Rule-Based Identification of Individuals with Mild Cognitive Impairment or Alzheimer's Disease Using Clinical Notes from the United States Veterans Affairs Healthcare System
Original Research
Original Research
Document Type
Report
Author
Source
Neurology and Therapy. December 2023, Vol. 12 Issue 6, p2067, 12 p.
Subject
Language
English
Abstract
Author(s): Byron J. Aguilar [sup.1] , Donald Miller [sup.2] , Guneet Jasuja [sup.3] , Xuyang Li [sup.1] , Ekaterina Shishova [sup.2] , Maureen K. O'Connor [sup.1] [sup.4] , Andrew Nguyen [...]
Background Early identification of individuals with mild cognitive impairment (MCI) and Alzheimer's disease (AD) is a clinical and research imperative. Use of diagnostic codes for MCI and AD identification has limitations. We used clinical notes to supplement diagnostic codes in the Veterans Affairs Healthcare System (VAHS) electronic health records (EHR) to identify and establish cohorts of Veterans recorded with MCI or AD. Methods Targeted keyword searches for MCI ("Mild cognitive impairment;" "MCI") and AD ("Alz*") were used to extract clinical notes from the VAHS EHR from fiscal year (FY) 2010 through FY 2019. Iterative steps of inclusion and exclusion were applied until searches achieved a positive predictive value [greater than or equal to] 80%. MCI and AD cohorts were identified via clinical notes and/or diagnostic codes (i.e., including Veterans recorded by "Notes Only," "Notes + Code," or "Codes Only"). Results A total of 2,134,661 clinical notes from 339,007 Veterans met the iterative search criteria for MCI due to any cause and 4,231,933 notes from 572,063 Veterans met the iterative search criteria for AD. Over the 10-year study period, the number of clinical notes recording AD was generally stable, whereas the number for MCI more than doubled. More Veterans were identified for the MCI or AD cohorts via clinical notes than by diagnostic codes, particularly in the AD cohort. Among Veterans identified by having "Notes + Code" for MCI, the number first recorded by a code was lower than the number first recorded by a note until FY 2015 and then gradually became comparable after FY 2015. Among Veterans identified by having "Notes + Code" for AD, the number first recorded by a note was more than double the number first recorded by a code AD in each of the FYs. Conclusions Clinical note-based identification captured more Veterans recorded with MCI and AD than diagnostic code-based identification.
Background Early identification of individuals with mild cognitive impairment (MCI) and Alzheimer's disease (AD) is a clinical and research imperative. Use of diagnostic codes for MCI and AD identification has limitations. We used clinical notes to supplement diagnostic codes in the Veterans Affairs Healthcare System (VAHS) electronic health records (EHR) to identify and establish cohorts of Veterans recorded with MCI or AD. Methods Targeted keyword searches for MCI ("Mild cognitive impairment;" "MCI") and AD ("Alz*") were used to extract clinical notes from the VAHS EHR from fiscal year (FY) 2010 through FY 2019. Iterative steps of inclusion and exclusion were applied until searches achieved a positive predictive value [greater than or equal to] 80%. MCI and AD cohorts were identified via clinical notes and/or diagnostic codes (i.e., including Veterans recorded by "Notes Only," "Notes + Code," or "Codes Only"). Results A total of 2,134,661 clinical notes from 339,007 Veterans met the iterative search criteria for MCI due to any cause and 4,231,933 notes from 572,063 Veterans met the iterative search criteria for AD. Over the 10-year study period, the number of clinical notes recording AD was generally stable, whereas the number for MCI more than doubled. More Veterans were identified for the MCI or AD cohorts via clinical notes than by diagnostic codes, particularly in the AD cohort. Among Veterans identified by having "Notes + Code" for MCI, the number first recorded by a code was lower than the number first recorded by a note until FY 2015 and then gradually became comparable after FY 2015. Among Veterans identified by having "Notes + Code" for AD, the number first recorded by a note was more than double the number first recorded by a code AD in each of the FYs. Conclusions Clinical note-based identification captured more Veterans recorded with MCI and AD than diagnostic code-based identification.