Using Unsupervised Learning to Identify Clinical Subtypes of Alzheimer's Disease in Electronic Health Records

Nonie Alexander; Daniel C Alexander; Frederik Barkhof; Spiros Denaxas

doi:https://doi.org/10.3233/SHTI200210

Using Unsupervised Learning to Identify Clinical Subtypes of Alzheimer's Disease in Electronic Health Records

Nonie Alexander, Daniel C Alexander, Frederik Barkhof, Spiros Denaxas

Research output: Contribution to journal › Article › Academic › peer-review

8 Citations (Scopus)

Abstract

Identifying subtypes of Alzheimer's Disease (AD) can lead towards the creation of personalized interventions and potentially improve outcomes. In this study, we use UK primary care electronic health records (EHR) from the CALIBER resource to identify and characterize clinically-meaningful clusters patients using unsupervised learning approaches of MCA and K-means. We discovered and characterized five clusters with different profiles (mental health, non-typical AD, typical AD, CVD and men with cancer). The mental health cluster had faster rate of progression than all the other clusters making it a target for future research and intervention. Our results demonstrate that unsupervised learning approaches can be utilized on EHR to identify subtypes of heterogeneous conditions.

Original language	English
Pages (from-to)	499-503
Number of pages	5
Journal	Studies in health technology and informatics
Volume	270
DOIs	https://doi.org/10.3233/SHTI200210
Publication status	Published - 16 Jun 2020

Keywords

Alzheimer's disease
Electronic health records
Machine learning
Phenotyping

Access to Document

https://doi.org/10.3233/SHTI200210

Cite this

@article{4eb3aaae14de4d33a9251eaef0fdec22,

title = "Using Unsupervised Learning to Identify Clinical Subtypes of Alzheimer's Disease in Electronic Health Records",

abstract = "Identifying subtypes of Alzheimer's Disease (AD) can lead towards the creation of personalized interventions and potentially improve outcomes. In this study, we use UK primary care electronic health records (EHR) from the CALIBER resource to identify and characterize clinically-meaningful clusters patients using unsupervised learning approaches of MCA and K-means. We discovered and characterized five clusters with different profiles (mental health, non-typical AD, typical AD, CVD and men with cancer). The mental health cluster had faster rate of progression than all the other clusters making it a target for future research and intervention. Our results demonstrate that unsupervised learning approaches can be utilized on EHR to identify subtypes of heterogeneous conditions.",

keywords = "Alzheimer's disease, Electronic health records, Machine learning, Phenotyping",

author = "Nonie Alexander and Alexander, {Daniel C} and Frederik Barkhof and Spiros Denaxas",

year = "2020",

month = jun,

day = "16",

doi = "https://doi.org/10.3233/SHTI200210",

language = "English",

volume = "270",

pages = "499--503",

journal = "Studies in health technology and informatics",

issn = "0926-9630",

publisher = "IOS Press",

}

TY - JOUR

T1 - Using Unsupervised Learning to Identify Clinical Subtypes of Alzheimer's Disease in Electronic Health Records

AU - Alexander, Nonie

AU - Alexander, Daniel C

AU - Barkhof, Frederik

AU - Denaxas, Spiros

PY - 2020/6/16

Y1 - 2020/6/16

N2 - Identifying subtypes of Alzheimer's Disease (AD) can lead towards the creation of personalized interventions and potentially improve outcomes. In this study, we use UK primary care electronic health records (EHR) from the CALIBER resource to identify and characterize clinically-meaningful clusters patients using unsupervised learning approaches of MCA and K-means. We discovered and characterized five clusters with different profiles (mental health, non-typical AD, typical AD, CVD and men with cancer). The mental health cluster had faster rate of progression than all the other clusters making it a target for future research and intervention. Our results demonstrate that unsupervised learning approaches can be utilized on EHR to identify subtypes of heterogeneous conditions.

AB - Identifying subtypes of Alzheimer's Disease (AD) can lead towards the creation of personalized interventions and potentially improve outcomes. In this study, we use UK primary care electronic health records (EHR) from the CALIBER resource to identify and characterize clinically-meaningful clusters patients using unsupervised learning approaches of MCA and K-means. We discovered and characterized five clusters with different profiles (mental health, non-typical AD, typical AD, CVD and men with cancer). The mental health cluster had faster rate of progression than all the other clusters making it a target for future research and intervention. Our results demonstrate that unsupervised learning approaches can be utilized on EHR to identify subtypes of heterogeneous conditions.

KW - Alzheimer's disease

KW - Electronic health records

KW - Machine learning

KW - Phenotyping

UR - http://www.scopus.com/inward/record.url?scp=85086909384&partnerID=8YFLogxK

U2 - https://doi.org/10.3233/SHTI200210

DO - https://doi.org/10.3233/SHTI200210

M3 - Article

C2 - 32570434

SN - 0926-9630

VL - 270

SP - 499

EP - 503

JO - Studies in health technology and informatics

JF - Studies in health technology and informatics

ER -

Using Unsupervised Learning to Identify Clinical Subtypes of Alzheimer's Disease in Electronic Health Records

Abstract

Keywords

Access to Document

Other files and links

Cite this