Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences

Nikoloas Papachristou, Payam Barnaghi, Bruce A. Cooper, Xiao Hu, Roma Maguire, Kathi Apostolidis, Jo Armes, Yvette P. Conley, Marilyn Hammer, Stylianos Katsaragakis, Kord M. Kober, Jon D. Levine, Lisa McCann, Elisabeth Patiraki, Steven M. Paul, Emma Ream, Fay Wright, Christine Miaskowski

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

CONTEXT:
Risk profiling of oncology patients based on their symptom experience assists clinicians to provide more personalized symptom management interventions. Recent findings suggest that oncology patients with distinct symptom profiles can be identified using a variety of analytic methods.
OBJECTIVES:
The objective of this study was to evaluate the concordance between the number and types of subgroups of patients with distinct symptom profiles using latent class analysis and K-modes analysis.
METHODS:
Using data on the occurrence of 25 symptoms from the Memorial Symptom Assessment Scale, that 1329 patients completed prior to their next dose of chemotherapy (CTX), Cohen's kappa coefficient was used to evaluate for concordance between the two analytic methods. For both latent class analysis and K-modes, differences among the subgroups in demographic, clinical, and symptom characteristics, as well as quality of life outcomes were determined using parametric and nonparametric statistics.
RESULTS:
Using both analytic methods, four subgroups of patients with distinct symptom profiles were identified (i.e., all low, moderate physical and lower psychological, moderate physical and higher Psychological, and all high). The percent agreement between the two methods was 75.32%, which suggests a moderate level of agreement. In both analyses, patients in the all high group were significantly younger and had a higher comorbidity profile, worse Memorial Symptom Assessment Scale subscale scores, and poorer QOL outcomes.
CONCLUSION:
Both analytic methods can be used to identify subgroups of oncology patients with distinct symptom profiles. Additional research is needed to determine which analytic methods and which dimension of the symptom experience provide the most sensitive and specific risk profiles
LanguageEnglish
JournalJournal of Pain and Symptom Management
Early online date28 Aug 2017
DOIs
Publication statusE-pub ahead of print - 28 Aug 2017

Fingerprint

Symptom Assessment
Psychology
Nonparametric Statistics
Comorbidity
Quality of Life
Demography
Drug Therapy
Research

Keywords

  • symptom clusters
  • cancer
  • latent class analysis
  • machine learning
  • clustering
  • chemotherapy
  • k-modes analysis

Cite this

Papachristou, Nikoloas ; Barnaghi, Payam ; Cooper, Bruce A. ; Hu, Xiao ; Maguire, Roma ; Apostolidis, Kathi ; Armes, Jo ; Conley, Yvette P. ; Hammer, Marilyn ; Katsaragakis, Stylianos ; Kober, Kord M. ; Levine, Jon D. ; McCann, Lisa ; Patiraki, Elisabeth ; Paul, Steven M. ; Ream, Emma ; Wright, Fay ; Miaskowski, Christine. / Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences. In: Journal of Pain and Symptom Management. 2017.
@article{7032b3b08b4d4406b15366ce39b743e9,
title = "Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences",
abstract = "CONTEXT:Risk profiling of oncology patients based on their symptom experience assists clinicians to provide more personalized symptom management interventions. Recent findings suggest that oncology patients with distinct symptom profiles can be identified using a variety of analytic methods.OBJECTIVES:The objective of this study was to evaluate the concordance between the number and types of subgroups of patients with distinct symptom profiles using latent class analysis and K-modes analysis.METHODS:Using data on the occurrence of 25 symptoms from the Memorial Symptom Assessment Scale, that 1329 patients completed prior to their next dose of chemotherapy (CTX), Cohen's kappa coefficient was used to evaluate for concordance between the two analytic methods. For both latent class analysis and K-modes, differences among the subgroups in demographic, clinical, and symptom characteristics, as well as quality of life outcomes were determined using parametric and nonparametric statistics.RESULTS:Using both analytic methods, four subgroups of patients with distinct symptom profiles were identified (i.e., all low, moderate physical and lower psychological, moderate physical and higher Psychological, and all high). The percent agreement between the two methods was 75.32{\%}, which suggests a moderate level of agreement. In both analyses, patients in the all high group were significantly younger and had a higher comorbidity profile, worse Memorial Symptom Assessment Scale subscale scores, and poorer QOL outcomes.CONCLUSION:Both analytic methods can be used to identify subgroups of oncology patients with distinct symptom profiles. Additional research is needed to determine which analytic methods and which dimension of the symptom experience provide the most sensitive and specific risk profiles",
keywords = "symptom clusters, cancer, latent class analysis, machine learning, clustering, chemotherapy, k-modes analysis",
author = "Nikoloas Papachristou and Payam Barnaghi and Cooper, {Bruce A.} and Xiao Hu and Roma Maguire and Kathi Apostolidis and Jo Armes and Conley, {Yvette P.} and Marilyn Hammer and Stylianos Katsaragakis and Kober, {Kord M.} and Levine, {Jon D.} and Lisa McCann and Elisabeth Patiraki and Paul, {Steven M.} and Emma Ream and Fay Wright and Christine Miaskowski",
year = "2017",
month = "8",
day = "28",
doi = "10.1016/j.jpainsymman.2017.08.020",
language = "English",
journal = "Journal of Pain and Symptom Management",
issn = "0885-3924",

}

Papachristou, N, Barnaghi, P, Cooper, BA, Hu, X, Maguire, R, Apostolidis, K, Armes, J, Conley, YP, Hammer, M, Katsaragakis, S, Kober, KM, Levine, JD, McCann, L, Patiraki, E, Paul, SM, Ream, E, Wright, F & Miaskowski, C 2017, 'Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences' Journal of Pain and Symptom Management. https://doi.org/10.1016/j.jpainsymman.2017.08.020

Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences. / Papachristou, Nikoloas ; Barnaghi, Payam; Cooper, Bruce A.; Hu, Xiao; Maguire, Roma; Apostolidis, Kathi; Armes, Jo; Conley, Yvette P.; Hammer, Marilyn ; Katsaragakis, Stylianos ; Kober, Kord M.; Levine, Jon D.; McCann, Lisa; Patiraki, Elisabeth ; Paul, Steven M.; Ream, Emma; Wright, Fay; Miaskowski, Christine.

In: Journal of Pain and Symptom Management, 28.08.2017.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Congruence between latent class and k-modes analyses in the identification of oncology patients with distinct symptom experiences

AU - Papachristou, Nikoloas

AU - Barnaghi, Payam

AU - Cooper, Bruce A.

AU - Hu, Xiao

AU - Maguire, Roma

AU - Apostolidis, Kathi

AU - Armes, Jo

AU - Conley, Yvette P.

AU - Hammer, Marilyn

AU - Katsaragakis, Stylianos

AU - Kober, Kord M.

AU - Levine, Jon D.

AU - McCann, Lisa

AU - Patiraki, Elisabeth

AU - Paul, Steven M.

AU - Ream, Emma

AU - Wright, Fay

AU - Miaskowski, Christine

PY - 2017/8/28

Y1 - 2017/8/28

N2 - CONTEXT:Risk profiling of oncology patients based on their symptom experience assists clinicians to provide more personalized symptom management interventions. Recent findings suggest that oncology patients with distinct symptom profiles can be identified using a variety of analytic methods.OBJECTIVES:The objective of this study was to evaluate the concordance between the number and types of subgroups of patients with distinct symptom profiles using latent class analysis and K-modes analysis.METHODS:Using data on the occurrence of 25 symptoms from the Memorial Symptom Assessment Scale, that 1329 patients completed prior to their next dose of chemotherapy (CTX), Cohen's kappa coefficient was used to evaluate for concordance between the two analytic methods. For both latent class analysis and K-modes, differences among the subgroups in demographic, clinical, and symptom characteristics, as well as quality of life outcomes were determined using parametric and nonparametric statistics.RESULTS:Using both analytic methods, four subgroups of patients with distinct symptom profiles were identified (i.e., all low, moderate physical and lower psychological, moderate physical and higher Psychological, and all high). The percent agreement between the two methods was 75.32%, which suggests a moderate level of agreement. In both analyses, patients in the all high group were significantly younger and had a higher comorbidity profile, worse Memorial Symptom Assessment Scale subscale scores, and poorer QOL outcomes.CONCLUSION:Both analytic methods can be used to identify subgroups of oncology patients with distinct symptom profiles. Additional research is needed to determine which analytic methods and which dimension of the symptom experience provide the most sensitive and specific risk profiles

AB - CONTEXT:Risk profiling of oncology patients based on their symptom experience assists clinicians to provide more personalized symptom management interventions. Recent findings suggest that oncology patients with distinct symptom profiles can be identified using a variety of analytic methods.OBJECTIVES:The objective of this study was to evaluate the concordance between the number and types of subgroups of patients with distinct symptom profiles using latent class analysis and K-modes analysis.METHODS:Using data on the occurrence of 25 symptoms from the Memorial Symptom Assessment Scale, that 1329 patients completed prior to their next dose of chemotherapy (CTX), Cohen's kappa coefficient was used to evaluate for concordance between the two analytic methods. For both latent class analysis and K-modes, differences among the subgroups in demographic, clinical, and symptom characteristics, as well as quality of life outcomes were determined using parametric and nonparametric statistics.RESULTS:Using both analytic methods, four subgroups of patients with distinct symptom profiles were identified (i.e., all low, moderate physical and lower psychological, moderate physical and higher Psychological, and all high). The percent agreement between the two methods was 75.32%, which suggests a moderate level of agreement. In both analyses, patients in the all high group were significantly younger and had a higher comorbidity profile, worse Memorial Symptom Assessment Scale subscale scores, and poorer QOL outcomes.CONCLUSION:Both analytic methods can be used to identify subgroups of oncology patients with distinct symptom profiles. Additional research is needed to determine which analytic methods and which dimension of the symptom experience provide the most sensitive and specific risk profiles

KW - symptom clusters

KW - cancer

KW - latent class analysis

KW - machine learning

KW - clustering

KW - chemotherapy

KW - k-modes analysis

U2 - 10.1016/j.jpainsymman.2017.08.020

DO - 10.1016/j.jpainsymman.2017.08.020

M3 - Article

JO - Journal of Pain and Symptom Management

T2 - Journal of Pain and Symptom Management

JF - Journal of Pain and Symptom Management

SN - 0885-3924

ER -