Denise Soltow Hershey1, Steven J Pierce2. 1. College of Nursing, Michigan State University, USA. Electronic address: denise.hershey@hc.msu.edu. 2. Center for Statistical Training and Consulting, Michigan State University, Rm 178 Giltner Hall, 293 Farm Ln, East Lansing, MI 48824, USA. Electronic address: pierces1@msu.edu.
Abstract
PURPOSE: Diabetes and cancer are chronic conditions that produce symptoms, some unique to each condition and others common to both. Previous studies have reported on subgroups of patients who experience distinct combinations of symptoms in cross-sectional samples and the univariate longitudinal trajectories of individual symptoms. The literature currently lacks examples of research that take a multivariate longitudinal perspective to understanding patients' symptom experiences. The purpose of this study was to identify subgroups of patients who share distinct multivariate longitudinal profiles with respect to how symptom severity changes over time for a set of five symptoms (pain, fatigue, change in appetite, nausea, and numbness and tingling). METHODS: This exploratory study included 43 participants with pre-existing diabetes from eight community-based cancer centers who were receiving chemotherapy for a solid tumor. Using baseline and 8-week data, a model-based cluster analysis with Bayesian regularization was used to identify subgroups. RESULTS: Two groups were identified. Group 1 experienced mild symptoms that changed very little at 8 weeks; group 2 experienced mild to moderate symptom severity, with small increases in fatigue, nausea, and numbness and tingling. Effect size confidence intervals suggest that level of depression, length of time with diabetes, and severity of diabetes at baseline may differ between groups. CONCLUSIONS: More research in this area is needed to further test this model, address limitations associated with analyzing a small sample, and explore factors that may be associated with changes in the overall symptom experience for patients with diabetes and cancer.
PURPOSE:Diabetes and cancer are chronic conditions that produce symptoms, some unique to each condition and others common to both. Previous studies have reported on subgroups of patients who experience distinct combinations of symptoms in cross-sectional samples and the univariate longitudinal trajectories of individual symptoms. The literature currently lacks examples of research that take a multivariate longitudinal perspective to understanding patients' symptom experiences. The purpose of this study was to identify subgroups of patients who share distinct multivariate longitudinal profiles with respect to how symptom severity changes over time for a set of five symptoms (pain, fatigue, change in appetite, nausea, and numbness and tingling). METHODS: This exploratory study included 43 participants with pre-existing diabetes from eight community-based cancer centers who were receiving chemotherapy for a solid tumor. Using baseline and 8-week data, a model-based cluster analysis with Bayesian regularization was used to identify subgroups. RESULTS: Two groups were identified. Group 1 experienced mild symptoms that changed very little at 8 weeks; group 2 experienced mild to moderate symptom severity, with small increases in fatigue, nausea, and numbness and tingling. Effect size confidence intervals suggest that level of depression, length of time with diabetes, and severity of diabetes at baseline may differ between groups. CONCLUSIONS: More research in this area is needed to further test this model, address limitations associated with analyzing a small sample, and explore factors that may be associated with changes in the overall symptom experience for patients with diabetes and cancer.