The BEST study, a multicenter randomized trial of group psychosocial support in metastatic breast cancer, had several unusual features that may have influenced recruitment, notably the group nature of the intervention and the need for close collaboration between medical and psychosocial investigators. The recruitment process was examined in light of these features. Establishment of study centers was facilitated by involvement of experienced medical investigators who had successfully collaborated in previous research projects. Systematic evaluation of potential subjects or direct recruitment by psychosocial investigators optimized recruitment; however, the group nature of the intervention prolonged recruitment. Overall, 652 women were approached and 237 (43.3% of those medically eligible) randomized. Using population-based estimates, 24.3% of women with metastatic breast cancer were assessed for the study and 8.7% randomized. A randomization ratio of 2:1 was required to form and maintain groups. Competing clinical trials were the greatest barrier to recruitment. Five lessons were learned during recruitment for this trial: (1) multicenter randomized trials of psychosocial interventions are feasible, even in very ill patients, (2) the use of a group intervention effectively increased the required sample size by 50%, (3) similarity of randomization rates suggests that generalizability of study results will probably be comparable to that of other randomized cancer trials, (4) multidisciplinary collaborations and involvement of experienced researchers facilitated enrollment, and (5) most challenges encountered in recruitment were similar to those seen in all clinical trials.
BACKGROUND: Supportive-expressive group therapy has been reported to prolong survival among women with metastatic breast cancer. However, in recent studies, various psychosocial interventions have not prolonged survival.
METHODS: In a multicenter trial, we randomly assigned 235 women with metastatic breast cancer who were expected to survive at least three months in a 2:1 ratio to an intervention group that participated in weekly supportive-expressive group therapy (158 women) or to a control group that received no such intervention (77 women). All the women received educational materials and any medical or psychosocial care that was deemed necessary. The primary outcome was survival; psychosocial function was assessed by self-reported questionnaires.
RESULTS: Women assigned to supportive-expressive therapy had greater improvement in psychological symptoms and reported less pain (P=0.04) than women in the control group. A significant interaction of treatment-group assignment with base-line psychological score was found (P</=0.003 for the comparison of mood variables; P=0.04 for the comparison of pain); women who were more distressed benefited, whereas those who were less distressed did not. The psychological intervention did not prolong survival (median survival, 17.9 months in the intervention group and 17.6 months in the control group; hazard ratio for death according to the univariate analysis, 1.06 [95 percent confidence interval, 0.78 to 1.45]; hazard ratio according to the multivariate analysis, 1.23 [95 percent confidence interval, 0.88 to 1.72]).
CONCLUSIONS: Supportive-expressive group therapy does not prolong survival in women with metastatic breast cancer. It improves mood and the perception of pain, particularly in women who are initially more distressed.
PURPOSE: To evaluate the effect of a standardized group psychosocial intervention on health-related quality of life (HrQOL) in women with metastatic breast cancer and to explore the effect of missing data in HrQOL analyses.
PATIENTS AND METHODS: Between 1993 and 1998, seven Canadian centers randomly assigned 235 eligible women to participate in a weekly, 90-minute, therapist-led support group that adhered to principles of supportive-expressive (SE) therapy or to a control arm (no SE). All women received educational material and any type of medical or psychosocial care deemed necessary. HrQOL data were prospectively collected using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C30 (EORTC QLQ-C30) at baseline, 4, 8, and 12 months. The primary HrQOL analyses compared scores in the two study arms. Analyses were limited to women with appropriate baseline HrQOL information (n = 215).
RESULTS: Baseline EORTC QLQ-C30 scores were not different between the two study arms (all P >.05). Primary analysis of all subscales failed to show a significant influence of the intervention on HrQOL (all P >.05). There was a significant deterioration over time in several functional scales of the EORTC QLQ-C30: global (P =.03), physical (P =.0002), role (P =.01), and cognitive functioning (P =.04); and in symptom scales: dyspnea (P =.007), appetite loss (P =.04), and fatigue (P =.003); these changes were independent of randomization allocation. Results were similar in additional analyses of overall HrQOL using a variety of approaches to handling missing data.
CONCLUSION: Supportive-expressive group therapy in patients with metastatic breast cancer does not appear to influence HrQOL, as measured by the EORTC QLQ-C30.
