B ased on a very small subsample of 25 estrogen receptor (ER)- negative women with breast cancer, 13 of whom received group therapy, Spiegel et al. 1 conducted an exploratory analysis that found increased survival for these women. They have argued that these women were not protected by hormonal therapies.
In our much larger Australian study of group therapy with women with advanced breast cancer, there were 70 ER-negative women. 2 Kaplan-Meier survival analysis showed no difference between the 43 hormone receptor-negative (estrogen) patients in the group intervention condition and the 27 hormone receptor-negative patients in the control condition. The median survival durations were at 15.5 months (95% confidence interval [CI], 9.5-25.8) for patients in the intervention condition and 17.2 months (95% CI, 8.1-26.5) for patients in the control condition, respectively. The logrank test for equality of survivor functions was not statistically significant (chi-square (2) ΒΌ 0.55, P ΒΌ .46).
We also tried to replicate their analytic method exactly. In a Cox proportional hazards model we tested the effect of a treatment by positive hormone receptor status interaction, after controlling for intervention site, intervention site by treatment interaction, a treatment main effect, and a positive hormone receptor status main effect. We also centered the predictors as recommended in Kraemer and Blasey, 3 with coding of 1/2 and Γ1/2 in dichotomous variables (for both the randomization and hormone receptor status variables) and 1-1/m and Γ1/m for the site dummy variables (where m ΒΌ 5 in our study). Again, the result was consistent with the Kaplan-Meier survival analysis. There was no hormone receptor status by treatment interaction (Cox proportional hazards B ΒΌ 1.14, Wald statistic ΒΌ 0.38, P ΒΌ .71).
Given our inability to replicate their latest exploratory finding, we conclude that it is a chance finding. Women should not be swayed into false hope, but appreciate that strong evidence has emerged that group therapy does not prolong survival. Group therapy does prevent depression and improve quality of life, thus representing a wonderful adjunctive psychosocial treatment for those with cancer.