Cargando…

Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters

BACKGROUND: The majority of patients with advanced cancer experience symptom pairs or clusters among pain, fatigue, and insomnia. Improved methods are needed to detect and interpret interactions among symptoms or diesease markers to reveal influential pairs or clusters. In prior work, I developed an...

Descripción completa

Detalles Bibliográficos
Autor principal: Francoeur, Richard B
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Dove Medical Press 2014
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4278795/
https://www.ncbi.nlm.nih.gov/pubmed/25565865
http://dx.doi.org/10.2147/OTT.S66465
_version_ 1782350574747910144
author Francoeur, Richard B
author_facet Francoeur, Richard B
author_sort Francoeur, Richard B
collection PubMed
description BACKGROUND: The majority of patients with advanced cancer experience symptom pairs or clusters among pain, fatigue, and insomnia. Improved methods are needed to detect and interpret interactions among symptoms or diesease markers to reveal influential pairs or clusters. In prior work, I developed and validated sequential residual centering (SRC), a method that improves the sensitivity of multiple regression to detect interactions among predictors, by conditioning for multicollinearity (shared variation) among interactions and component predictors. MATERIALS AND METHODS: Using a hypothetical three-way interaction among pain, fatigue, and sleep to predict depressive affect, I derive and explain SRC multiple regression. Subsequently, I estimate raw and SRC multiple regressions using real data for these symptoms from 268 palliative radiation outpatients. RESULTS: Unlike raw regression, SRC reveals that the three-way interaction (pain × fatigue/weakness × sleep problems) is statistically significant. In follow-up analyses, the relationship between pain and depressive affect is aggravated (magnified) within two partial ranges: 1) complete-to-some control over fatigue/weakness when there is complete control over sleep problems (ie, a subset of the pain–fatigue/weakness symptom pair), and 2) no control over fatigue/weakness when there is some-to-no control over sleep problems (ie, a subset of the pain–fatigue/weakness–sleep problems symptom cluster). Otherwise, the relationship weakens (buffering) as control over fatigue/weakness or sleep problems diminishes. CONCLUSION: By reducing the standard error, SRC unmasks a three-way interaction comprising a symptom pair and cluster. Low-to-moderate levels of the moderator variable for fatigue/weakness magnify the relationship between pain and depressive affect. However, when the comoderator variable for sleep problems accompanies fatigue/weakness, only frequent or unrelenting levels of both symptoms magnify the relationship. These findings suggest that a countervailing mechanism involving depressive affect could account for the effectiveness of a cognitive behavioral intervention to reduce the severity of a pain, fatigue, and sleep disturbance cluster in a previous randomized trial.
format Online
Article
Text
id pubmed-4278795
institution National Center for Biotechnology Information
language English
publishDate 2014
publisher Dove Medical Press
record_format MEDLINE/PubMed
spelling pubmed-42787952015-01-06 Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters Francoeur, Richard B Onco Targets Ther Methodology BACKGROUND: The majority of patients with advanced cancer experience symptom pairs or clusters among pain, fatigue, and insomnia. Improved methods are needed to detect and interpret interactions among symptoms or diesease markers to reveal influential pairs or clusters. In prior work, I developed and validated sequential residual centering (SRC), a method that improves the sensitivity of multiple regression to detect interactions among predictors, by conditioning for multicollinearity (shared variation) among interactions and component predictors. MATERIALS AND METHODS: Using a hypothetical three-way interaction among pain, fatigue, and sleep to predict depressive affect, I derive and explain SRC multiple regression. Subsequently, I estimate raw and SRC multiple regressions using real data for these symptoms from 268 palliative radiation outpatients. RESULTS: Unlike raw regression, SRC reveals that the three-way interaction (pain × fatigue/weakness × sleep problems) is statistically significant. In follow-up analyses, the relationship between pain and depressive affect is aggravated (magnified) within two partial ranges: 1) complete-to-some control over fatigue/weakness when there is complete control over sleep problems (ie, a subset of the pain–fatigue/weakness symptom pair), and 2) no control over fatigue/weakness when there is some-to-no control over sleep problems (ie, a subset of the pain–fatigue/weakness–sleep problems symptom cluster). Otherwise, the relationship weakens (buffering) as control over fatigue/weakness or sleep problems diminishes. CONCLUSION: By reducing the standard error, SRC unmasks a three-way interaction comprising a symptom pair and cluster. Low-to-moderate levels of the moderator variable for fatigue/weakness magnify the relationship between pain and depressive affect. However, when the comoderator variable for sleep problems accompanies fatigue/weakness, only frequent or unrelenting levels of both symptoms magnify the relationship. These findings suggest that a countervailing mechanism involving depressive affect could account for the effectiveness of a cognitive behavioral intervention to reduce the severity of a pain, fatigue, and sleep disturbance cluster in a previous randomized trial. Dove Medical Press 2014-12-22 /pmc/articles/PMC4278795/ /pubmed/25565865 http://dx.doi.org/10.2147/OTT.S66465 Text en © 2015 Francoeur. This work is published by Dove Medical Press Limited, and licensed under Creative Commons Attribution – Non Commercial (unported, v3.0) License The full terms of the License are available at http://creativecommons.org/licenses/by-nc/3.0/. Non-commercial uses of the work are permitted without any further permission from Dove Medical Press Limited, provided the work is properly attributed.
spellingShingle Methodology
Francoeur, Richard B
Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title_full Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title_fullStr Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title_full_unstemmed Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title_short Using an innovative multiple regression procedure in a cancer population (Part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
title_sort using an innovative multiple regression procedure in a cancer population (part 1): detecting and probing relationships of common interacting symptoms (pain, fatigue/weakness, sleep problems) as a strategy to discover influential symptom pairs and clusters
topic Methodology
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4278795/
https://www.ncbi.nlm.nih.gov/pubmed/25565865
http://dx.doi.org/10.2147/OTT.S66465
work_keys_str_mv AT francoeurrichardb usinganinnovativemultipleregressionprocedureinacancerpopulationpart1detectingandprobingrelationshipsofcommoninteractingsymptomspainfatigueweaknesssleepproblemsasastrategytodiscoverinfluentialsymptompairsandclusters