Letting Go of the Hope That Psychotherapy Prolongs Cancer Survival

David W. Kissane  Journal of Clinical Oncology, Vol 25, No 36 (December 20), 2007: pp. 5689-5690

Memorial Sloan-Kettering Cancer Center, Department of Psychiatry and Behavioral Sciences, New York, NY

The contribution of psychological health to the length of cancer survival has been a controversial and highly debated issue throughout the last decade. Does living better result in living longer? Two influential US randomized controlled trials (RCTs) reported by Spiegel in 1989 and Fawzy in 1993 created support for the idea that psychotherapy promotes survival. Yet both had methodological weaknesses, with small sample sizes and the lack of survival as an a priori end point at the heart of critiques.

Fox compared the Stanford patients with metastatic breast cancer with the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) data, which suggested that 32% of women with metastatic breast cancer during that time would have been expected to be alive beyond 5 years, whereas only 2.8% of the controls of Spiegel et al controls and 24% of their group intervention were alive. This highlighted the inadvertent sampling bias that can emerge in small trials, despite efforts through random assignment to balance extraneous sources of group differences.

Similarly, Coyne and colleagues pointed out, with respect to the Fawzy et al study, that the reclassification of a single patient would have made their survival findings nonsignificant. Analyses were not based on intention-to-treat principles. Again, for this era, the 5-year survival rate for stage I melanoma was 92%, whereas the control arm patients retained for analysis had a 5-year survival rate of only 72%. Despite these criticisms,the general public opinion about the mind being able to exert power over the body made adequately powered replication of these studies (with survival as a primary end point) mandatory.

In this issue, Boesen and colleagues from the Danish Cancer Society report no survival benefit from their replication study of the psychoeducational intervention done by Fawzy et al for patients with cutaneous malignant melanoma, from which Fawzy's group had found a survival gain persisting at 10-year follow-up.This intervention is a brief six-session, manual-based, structured program teaching health-promoting behaviors, stress management, problem-solving, and mutual support. The original University of California, Los Angeles, RCT involved a small cohort of 68, whereas the Danish replication study was better powered with 259 patients with melanoma. Only Breslow depth and lymph node status were significant prognostic factors for time to recurrence and length of survival in the Danish study.

This report comes hot on the trail of other RCTs conducted as replication studies of Supportive-Expressive Group Therapy (SEGT) to examine its ability to prolong survival in women with metastatic breast cancer. Again, larger cohort sizes achieved adequate power in these replications—the Canadian multisite SEGT study, 235 women; Australia's SEGT,227 women; and Spiegel's group,125 women. Another Australian study tested the survival benefit of group therapy in a cohort of 303 women with early-stage breast cancer. All of these prospective studies failed to find survival benefits, despite reducing distress and preventing onset of new cases of depression.

Might there be cross-cultural differences in the efficacy of different interventions? One always needs to allow for the adaptation of a therapy to the health beliefs and coping styles of an ethnic group. The current evidence comes from Western societies where psychological mindedness, rather than a somatic expression of distress, has developed across the past century. There is a caveat that every intervention needs to be culturally appropriate for the group to which it is applied.

Nonetheless, taken together, these replication studies have been of the utmost importance to psycho-oncology in adopting the necessary scientific rigor, care with treatment integrity, intention-to-treat analyses, application of the CONSORT guidelines, adequate power, and a priori specification of survival as the primary end point. Other studies of psychosocial interventions suggesting a survival benefit have substantial methodological flaws and cannot compete with the rigor of these five recent replication studies. They provide incontrovertible evidence that such psychotherapy interventions are not able to extend survival.

Where then does this leave the field? Two points are worth emphasis here: first, the impact of socioeconomic status and ethnicity, and second, the quality-of-life benefit of psychological and psychiatric treatments. Boesen and colleagues were able to compare nationally available recurrence and survival data on 137 patients who had refused to participate in their study with the 259 patients who did. To their surprise, they found a two-fold increase in morbidity in this former group, who also had lower socioeconomic status. Social inequality has long been associated with poorer access to and perseverance with anticancer treatments; Boesen et al review evidence about this in England, Wales, and Sweden.In the United States, the 5-year relative survival rate for all cancers combined is lower for African Americans (57%) than it is for whites (68%), while the death rate for African American men is 38% higher than whites and 17% higher than white women. Hispanics have higher rates of cancers associated with infection, such as cervix, liver, and stomach. Asians have high rates of liver and stomach cancers, but Vietnamese women, for instance, have a four times higher incidence of cervical cancer than all other Asians, a finding associated with their recent immigrant status and lower social status.

Psycho-oncology needs to promote behaviors leading to health promotion (smoking cessation, obesity reduction, and exercise), preventive screening and early detection, and culturally sensitive interventions to encourage adherence to anticancer treatments. Issues of medical mistrust and fatalism leading to treatment drop-outs could be addressed through these interventions. Social advocacy for access to health care is crucial. Some 24% of African Americans and 23% of Hispanics/Latinos live below the poverty line.Health care partnerships and outreach programs are paramount to disseminate psychosocial interventions into these communities. SEGT has been shown to promote adherence to hormonal and chemotherapeutic treatments for breast cancer.8 Future research could test this application in socially disadvantaged minorities.

Behavioral scientists have nevertheless debated the effect size and quality-of-life benefits from psychotherapeutic interventions in cancer care. Some meta-analyses and systematic reviews have concluded that the effectiveness of psychotherapeutic interventions is tentative and modest; others have accepted that the preponderance of evidence casts its weight in support of the relief of distress, with the strongest evidence for anxiety-related outcomes and when participants are prescreened for distress.One welcome aspect of this recent batch of RCTs addressing survival has been improved rigor in the conduct of these trials. Replication of interventions by independent research teams, sharing of treatment manuals, and training of therapists with due attention to treatment integrity remain vital. Effectiveness will be further improved through highly targeted interventions such as those addressing the fear of recurrence, a search for meaning, promotion of intimacy, or conservation of dignity.

Coyne recently noted that no intervention had been shown to prevent the onset of depression as cancer progresses, a conclusion that was overturned by the recently reported Australian SEGT replication. Here, using the Diagnostic and Statistical Manual of Mental Disorders diagnoses for psychiatric disorders, clinical depression was both actively ameliorated and new cases prevented across the course of the group therapy compared with controls. This provides evidence for the prophylactic benefit of a psychotherapeutic intervention being provided to all interested subjects rather than only those who are actively distressed.

For the patients observing this debate across recent years, one message is that there should be no guilt about not joining a group if that is a person's preference. Another is the importance of expressing concern and seeking help with any fears or existential threats that arise. Interventions are being continually refined and do assuage many of the sources of distress that patients experience. The survival debate has assisted the maturation of psycho-oncology as a discipline. There is still much to do in responding to the many challenges for patients coping with cancer and its related treatment and optimizing their quality of life. We remain a healing profession.