TY - GEN
T1 - The use of clustering to understand disease progression in Rheumatoid Arthritis
AU - De La Iglesia, Beatriz
AU - Nawongs, Kathapet
AU - Dainty, Jack R.
AU - Macgregor, Alexander
N1 - Funding Information: We acknowledge support from Grant Number ES/L011859/1, from The Business and Local Government Data Research Centre, funded by the Economic and Social Research Council to provide economic, scientific and social researchers and business analysts with secure data services.
PY - 2022/10/28
Y1 - 2022/10/28
N2 - In this paper we examine data representing patients with Rheumatoid Arthritis (RA). This is an important medical conditions that affects a proportion of the adult population and is very disabling. The data contains some demographics as well as follow up for up to 20 years where objective measures of 'joint involvement', e.g. counts of how many tender or swollen joints are present in a given follow up year, are recorded.To date the patterns of disease progression and joint involvement have not been investigated in detail for RA. We propose a clustering approach to extract patterns of joint involvement in disease progression for groups of patients. For this, we investigate how to measure distance for the type of data we analyse which consists of multiple attributes each corresponding to years of follow up measuring a particular objective measure. We settle for an aggregate Dynamic Time Warping measure of distance between patients and use it in combination with K-means clustering to cluster our patient trajectories. Our preliminary results, with some interpretation, show that it is possible to cluster such complex data to extract some meaningful patterns of joint involvement in disease progression.
AB - In this paper we examine data representing patients with Rheumatoid Arthritis (RA). This is an important medical conditions that affects a proportion of the adult population and is very disabling. The data contains some demographics as well as follow up for up to 20 years where objective measures of 'joint involvement', e.g. counts of how many tender or swollen joints are present in a given follow up year, are recorded.To date the patterns of disease progression and joint involvement have not been investigated in detail for RA. We propose a clustering approach to extract patterns of joint involvement in disease progression for groups of patients. For this, we investigate how to measure distance for the type of data we analyse which consists of multiple attributes each corresponding to years of follow up measuring a particular objective measure. We settle for an aggregate Dynamic Time Warping measure of distance between patients and use it in combination with K-means clustering to cluster our patient trajectories. Our preliminary results, with some interpretation, show that it is possible to cluster such complex data to extract some meaningful patterns of joint involvement in disease progression.
KW - Clustering
KW - rheumatoid arthritis
KW - sequence distance metrics
KW - unsupervised learning
UR - http://www.scopus.com/inward/record.url?scp=85144625389&partnerID=8YFLogxK
U2 - 10.1109/MetroXRAINE54828.2022.9967609
DO - 10.1109/MetroXRAINE54828.2022.9967609
M3 - Conference contribution
T3 - 2022 IEEE International Workshop on Metrology for Extended Reality, Artificial Intelligence and Neural Engineering, MetroXRAINE 2022 - Proceedings
SP - 444
EP - 448
BT - IEEE MetroXRAINE 2022
PB - The Institute of Electrical and Electronics Engineers (IEEE)
CY - Rome
ER -