Breast Cancer Studies: Design & Limitations
The complex nature of lifestyle modification and the potential benefit to breast cancer survivors is difficult to assess in human populations, which often leads to contradictory results among studies. The primary study designs outlined in this review include observational studies (prospective cohort studies) and randomized controlled trials. The observational cohort studies followed groups (cohorts) of participants over time (up to ~25 years) to identify associations between lifestyle factors and breast cancer recurrence and mortality.[119] Conversely, randomized controlled trials are considered the gold standard for measuring causality and the efficacy of treatments on breast cancer prognosis.[120] These trials randomly assigned patients with similar characteristics to receive the intervention or to serve as controls who received standard clinical care. At the conclusion of the trials, if a significant difference between the intervention and control groups in recurrence or survival could be detected, this difference was attributed to the intervention.
In general, cohort studies have the advantages of being able to assess multiple aspects of lifestyle and better represent the typical breast cancer patient population; however, as discussed below, potential bias in patient selection and limitations of recall and self-reporting of dietary and physical activity data may have confounded the overall results. Although randomized trials may provide the best evidence of treatment efficacy, these studies are expensive to conduct and thus may not follow patients as long as prospective cohort studies, which may bias estimates of long-term survival and recurrence. In addition, targeted accrual and exclusion of patients with advanced disease, patients undergoing chemotherapy, and patients with evidence of other forms of cancer may restrict the applicability of randomized trials to the general population.
Most of the cohort studies and randomized trials reported here were ethnically homogeneous, containing a large proportion of Caucasian participants, which makes generalizing results to patients with different demographics or genetic backgrounds difficult. Cultural differences in accepting and maintaining lifestyle changes may necessitate alternative intervention strategies for certain groups. Furthermore, a number of published cohort studies as well as randomized trials attempted to evaluate the effects of lifestyle factors in relatively small numbers of women, and thus lacked adequate statistical power to effectively assess potential benefit. The methods of recruitment used to enroll participants in both cohort studies and clinical trials may have confounded the observed relationships between lifestyle factors and outcomes because participants contacted through different media tend to exhibit different disease and motivational characteristics.[121] In the prospective cohort studies utilizing questionnaires to investigate past behaviors, inherent inaccuracies in self-reported data, such as under-reporting caloric intake and overestimating fat reduction[122] due to memory and recall bias is a potentially significant limitation. Randomized trials assessing long-term outcomes such as relapse-free survival are susceptible to contamination bias, where controls begin to engage in lifestyle changes on their own[123] and a gradual erosion of adherence over time occurs,[124] which may impact their ability to quantify the effects of lifestyle changes. As the purpose of this review was to critically weigh the available evidence regarding lifestyle behaviors and breast cancer outcomes, we utilized evidence from both observational studies and clinical trials, which was often complementary, as well as meta-analyses, to derive maximum information on the effects of lifestyle factors on breast cancer prognosis and recurrence.
Expert Rev Pharmacoeconomics Outcomes Res. 2012;12(4):451-464. © 2012 Expert Reviews Ltd.
