What do cancer survivors believe caused their cancer? A secondary analysis of cross-sectional survey data.

PURPOSE: Given that risk reduction and healthy lifestyles can prevent 4 in 10 cancers, it is important to understand what survivors believe caused their cancer to inform educational initiatives. METHODS: In this secondary analysis, we analyzed cancer survivor responses on the Causes Subscale of the Revised Illness Perception Questionnaire, which lists 18 possible causes of illness and a free text question. We used descriptive statistics to determine cancer survivors' agreement with the listed causes and conducted separate partial proportional odds models for the top three causes to examine their associations with sociodemographic and clinical characteristics. Content analysis was used to examine free text responses. RESULTS: Of the 1,001 participants, most identified as Caucasian (n = 764, 77%), female (n = 845, 85%), and were diagnosed with breast cancer (n = 656, 66%). The most commonly believed causes of cancer were: stress or worry (n = 498, 51%), pollution in the environment (n = 471, 48%), and chance or bad luck (n = 412, 42%). The associations of sociodemographic and clinical variables varied across the models. Free text responses indicated that hereditary and genetic causes (n = 223, 22.3%) followed by trauma and stress (n = 218, 21.8%) and bad luck or chance (n = 79, 7.9%) were the most important causes of cancer. CONCLUSIONS: Study results illuminate cancer survivors' beliefs about varying causes of their cancer diagnosis and identify characteristics of survivors who are more likely to believe certain factors caused their cancer. Results can be used to plan cancer education and risk-reduction campaigns and highlight for whom such initiatives would be most suitable.

ESTIMATION OF A MONOTONE DENSITY IN S-SAMPLE BIASED SAMPLING MODELS.

We study the nonparametric estimation of a decreasing density function g 0 in a general s-sample biased sampling model with weight (or bias) functions wi for i = 1, …, s. The determination of the monotone maximum likelihood estimator gn and its asymptotic distribution, except for the case when s = 1, has been long missing in the literature due to certain non-standard structures of the likelihood function, such as non-separability and a lack of strictly positive second order derivatives of the negative of the log-likelihood function. The existence, uniqueness, self-characterization, consistency of gn and its asymptotic distribution at a fixed point are established in this article. To overcome the barriers caused by non-standard likelihood structures, for instance, we show the tightness of gn via a purely analytic argument instead of an intrinsic geometric one and propose an indirect approach to attain the n -rate of convergence of the linear functional ∫ wi gn.