Evaluation of oncology registry follow-up methods

Making effective use of tumor registries for cancer survivorship research

BACKGROUND

The growing number of cancer survivors has created an increased need for survivorship research; however, the identification and recruitment of cancer survivors present some challenges. This report describes how two hospital cancer registries were used to recruit a large sample of breast cancer survivors (BCS) for a study examining the late reproductive effects of breast cancer treatments. Limitations and opportunities associated with this type of recruitment strategy are described, and the overall success of recruitment using this approach is presented.

METHODS

Tumor registries from a comprehensive cancer center and a community hospital were used to identify BCS who met the study screening criteria. Invitations and response forms were mailed to all potentially eligible women, and those who did not respond by mail also were contacted by telephone. Women who indicated interest and met the study requirements were asked to give written consent, were enrolled in the study, and were sent a self-report questionnaire.

RESULTS

Seventy percent of the eligible women (n = 733 women) responded to the mailing. Seventy-seven percent of the 512 respondents indicated a willingness to participate and were sent a questionnaire. Of these, 78% (n = 368 women) completed questionnaires. BCS recruited from the cancer center registry were more likely than those from the community hospital registry to respond to the invitation form (P = 0.033) and were more likely to return a completed questionnaire (P = 0.001). However, the community hospital provided access to a more ethnically diverse sample of survivors.

CONCLUSIONS

The two participating cancer registries were an excellent source for identifying a large sample of long-term BCS, and the different types of registries provided greater sample size and diversity. Although there are some limitations to this approach, including nonresponse of a significant number of BCS, tumor registries represent an important resource for the rapid identification of cancer survivors for research studies. Findings from this study suggest several enhancements for future studies that may increase the yield from registry recruitment.

Survival differences in breast cancer among racial/ethnic groups: a population-based study

In women, breast cancer is the most frequent solid tumor and the second leading cause of cancer death. Differences in survival of breast cancer have been noted among racial/ethnic groups, but the reasons are unclear. This report presents the characteristics and the survival experience of four racial/ethnic groups and evaluates the effects of stage, age, histology, and treatment on survival time. The distributions of prognostic factors and treatment among racial/ethnic groups are compared using female breast cancer patients from two population-based registries in Southern California. The main end points are observed survival time and survival by cause of death. The Cox model is used to estimate the relative risk of death in three minority groups compared with non-Hispanic whites, while controlling for several covariates. Breast cancer cases included in this study were 10,937 non-Hispanic whites, 185 blacks, 875 Hispanics, and 412 Asians. The median follow-up period was 76 months (range: 48-132). The median age at diagnosis was 64 years among non-Hispanic whites, 55 years among Hispanics (p = 0.001), 52 years among blacks (p = 0.001), and 50 years among Asians (p = 0. 001). There was more localized disease among non-Hispanic whites (61. 4%) than among blacks (50.8%) and Hispanics (52.2%), but not compared to Asians (59.7%). After controlling for stage, age, histology, treatment, and registry, overall survival significantly differed between non-Hispanic whites and blacks [relative risk (RR) = 2.27, 95% confidence interval (95% CI) 1.82-2.84) and between non-Hispanic whites and Hispanics (RR = 1.18, 95% CI 1.04-1.34). The same results were found for breast cancer death in blacks (RR = 2.32, 95% CI 1.76-3.07) and Hispanics (RR = 1.28, 95% CI 1.10-1.50). We found no difference between Asians and non-Hispanic whites in overall and cancer-related survival. These results show that stage of disease, age at diagnosis, histologic features and treatment for breast cancer differed among racial/ethnic groups. Moreover, black women, in particular, and Hispanic women with breast cancer had a higher risk of death compared to non-Hispanic white women, even after controlling for prognostic factors. These findings underline the necessity of improved screening and access to appropriate treatment among minority women for breast cancer.