Symptom clusters in patients with bone metastases--a reanalysis comparing different statistical methods.

PURPOSE: The aims of this study were to determine whether symptom clusters in patients with bone metastases varied when derived using three different statistical methods and to compare the presentation of symptom clusters over time in responders and nonresponders to palliative radiation treatment (RT). METHODS: Secondary analysis of a previously reported data set compiled using the brief pain inventory from 348 patients with bone metastases. Hierarchical cluster analysis (HCA) and exploratory factor analysis (EFA) were performed to identify symptom clusters at baseline, 1, 2, and 3 months following radiation treatment. Clusters derived were compared with the findings obtained using principal component analysis (PCA) in our previous study. The total patient sample was further separated into two subgroups: responders and nonresponders to RT. PCA, HCA, and EFA identified symptom clusters experienced by each subgroup at the same time points as before. RESULTS: Little correlation was observed in the symptom cluster findings of PCA, EFA, and HCA in the total patient sample. Absolute consensus among all three statistical methods was never reached at any assessment time point in the present study. Varying patterns of symptom cluster presentation over time were observed in the responders versus nonresponders subgroups regardless of the analytical method employed. A core cluster of symptoms composed of worst pain, general activity, walking ability, normal work, and enjoyment of life frequently presented in the same cluster. CONCLUSION: The presence and composition of symptom clusters derived varied depending on which statistical analysis method was employed. A key step in attaining consistency in symptom cluster research necessitates the utilization of a common method.

Symptom Clusters in Patients With Bone Metastases: A Sub-Analysis of Patients Reporting Exclusively Non-Zero BPI Scores.

BACKGROUND: The use of different statistical methods and inclusion criteria when deriving symptom clusters in cancer patients are contributing factors in cluster inconsistencies across studies. Primary objective was to extract symptom clusters in a subgroup of patients reporting non-zero Brief Pain Inventory (BPI) scores at baseline, and to compare clusters with those identified in the total patient sample. METHODS: Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA) and Exploratory Factor Analysis (EFA) were performed on the non-zero subgroup and total patient sample to identify symptom clusters at baseline and 1, 2 and 3 months following radiotherapy. RESULTS: At baseline, different symptom clusters were derived from the non-zero subgroup and the total patient population. Only PCA identified identical clusters. Over time, clusters extracted using the three statistical methods varied, with a few exceptions where the same clusters were extracted using two different methods at a specific time point. A complete consensus between all three methods was not noted at any time. The BPI, which is a short assessment tool, may lead to the extraction of oversimplified clusters. In addition, since this study analyzed results in the non-zero subgroup, clusters derived may be reflective of patients with poorer prognosis as these patients experienced all symptoms. CONCLUSION: Analyzing data compiled from all eligible consenting patients may not provide clinically relevant clustering among all symptoms in the assessment tool. The composition of symptom clusters varied with the inclusion of patients with zero symptom severity scores and with the statistical method employed.

Symptom Clusters Using the Edmonton Symptom Assessment System in Patients With Bone Metastases: A Reanalysis Comparing Different Statistical Methods.

BACKGROUND: To determine whether symptom clusters in patients with bone metastases vary when extracted using three different statistical methods. To compare the temporal composition of symptom clusters in responders versus non-responders to palliative radiation treatment. METHODS: A previous dataset of 518 bone metastases patients who completed the Edmonton Symptom Assessment System (ESAS) was used in this study. Clusters derived using Principal Component Analysis (PCA) in our previous study were compared to symptom clusters extracted using Hierarchical Cluster Analysis (HCA) and Exploratory Factor Analysis (EFA). Clusters were derived at baseline, and 1, 2, 4, 8 and 12 weeks after radiation treatment. The patient sample was further divided into responders versus non-responders to radiotherapy. The three statistical methods were performed to identify clusters in the subgroups at each time point. RESULTS: A complete consensus between HCA, EFA and PCA for the number and composition of symptom clusters was not reached at any time point. Furthermore, little correlation in clusters was found between the three statistical methods despite the use of an identical data set. As expected, different symptom clusters were observed in the responders and non-responders with all three statistical methods. In addition, clusters varied at each time point within each subgroup. Depression and anxiety were consistently found in the same cluster. CONCLUSION: The quantity, composition, and occurrence of symptom clusters varied based on which statistical method was employed. The use of a common analytical method is necessary for consistency and comparison purposes in future symptom cluster research.

Symptom Clusters in Cancer Patients With Bone Metastases: Subanalysis of Patients Reporting Exclusively Non-zero ESAS Scores.

BACKGROUND: To identify symptom clusters in a subgroup of patients reporting exclusively non-zero Edmonton Symptom Assessment System (ESAS) scores at baseline, and to compare clusters with those identified in the total patient population. Secondary objective was to determine whether symptom clusters in patients with bone metastases vary when extracted using different statistical methods. METHODS: An existing dataset compiled from bone metastases patients was used to identify a "non-zero" subgroup of patients reporting severity scores > 0 for all nine ESAS symptoms at baseline. Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA) and Exploratory Factor Analysis (EFA) were performed on the non-zero subgroup to derive symptom clusters at baseline and 1, 2, 4, 8 and 12 weeks following radiation treatment. Symptom clusters in the total patient sample at baseline were also derived using the three statistical methods. RESULTS: At baseline, different symptom clusters were identified in the non-zero subgroup compared with the total patient population regardless of the statistical method utilized. When comparing clusters derived using different statistical methods, symptom cluster results varied depending on the method employed, with a few exceptions where analogous clusters were derived using two different statistical methods at a specific time point. A complete consensus between all three methods was never observed. Only two ESAS symptoms, anxiety and depression, consistently occurred in the same cluster across different methods and over time. CONCLUSION: Compiling data from all eligible consenting patients may not provide an accurate overview of clustering among all the symptoms included in the assessment tool. The quantity and composition of symptom clusters identified varied based on whether patients with zero symptom severity scores were included at baseline and which statistical method was utilized.