Introduction

In Japan, it is common for physicians to withhold information about cancer diagnoses from patients [18, 27, 32], but that is not the case with breast cancer because it generally has a relatively good prognosis. The information has the potential to assist patients who are coping with the illness and has a positive impact on their quality of life. Furthermore, with the information, patients may have a better clinical outcome [1, 24, 29, 36]. Breast cancer patients who are informed of their diagnosis may begin to think about the cause of their cancer. With the progress of research in breast cancer prevention, information on breast cancer risk factors is well-documented and patients may try to correlate the risk factors with their own circumstances.

Other studies, on the other hand, have provided consistent reports from patients that God’s will, stress, the environment, chance, and injury or physical abuse to the breast are the causes of breast cancer; however, these causes are not generally accepted as the established breast cancer risk factors [2, 19, 26, 31, 33, 35]. Several studies have indicated that forming causal attributions or self-blame decreases psychological adjustment to the illness among breast and cervical cancer patients [6, 7, 9, 17, 33]. Other studies have reported that none of the attributions formed among breast cancer patients was associated with depression or overall adjustment to the illness [19, 35]

We hypothesized that causal attributions to each risk factor have a different influence on an individual’s psychological adjustment to the illness. The major motivation for studying risk factors has been disease prevention; however, the study of risk factors relative to the formation of causal attributions may lead to other findings. In the current study, we investigated the formation of causal attributions to risk factors and other explanations and prospectively evaluated the relationship between causal attributions and psychological adjustment to the illness among Japanese breast cancer patients.

Materials and methods

Study participants

This study was conducted in the Department of Breast Surgery at a general hospital in Gifu, Japan. The subjects of the current study were a part of our previously reported case–control study, and the participation rate was estimated at about 70% [23]. The participants’ ages ranged from 25 to 77 years. Newly diagnosed breast cancer patients were histopathologically confirmed, and those who were admitted to the hospital for surgery were invited to participate in the study. Eligible subjects were given information about the study, and only those who agreed to participate and completed the informed consent form were registered for the study. This study was approved by the institutional review board of Gifu University Graduate School of Medicine.

Study procedure

During the hospital admission of study participants, one nurse, hired and trained for the current study, conducted a personal interview with each participant regarding causal attributions. The interviews were conducted in the hospital after surgery, and the time between the diagnosis and the interview ranged from 8 to 88 days with 83% of the participants having the interview within 30 days. At the same time as the interview, a self-administered questionnaire was used to collect information on each participant’s sociodemographics and other characteristics such as marital status, number of children, smoking status, alcohol consumption, employment status, religion, medical history, family history of mammary gland disease, age at onset of menstruation, and pre/postmenopausal status. Dispositional optimism as a personality trait was also measured by using a questionnaire developed by Scheier and Carver [28]. For each participant, cancer stage at diagnosis was identified from a review of medical records. About 2 months after the initial interview, the psychological state of each participant was measured by the use of a follow-up questionnaire, which was mailed to each participant. Inclusion in the current study was based on the availability of complete data for all of the measures noted above. As a result, 63 women were included in the study.

Measures of causal attributions

The interviewer first asked each participant if she had ever thought about the reason that she had breast cancer. Next, the interviewer went through the list including risk factors and other conventional or common explanations for breast cancer and asked the participant whether she thought any of the items were the cause of her illness. The interviewer prompted the participant to answer based on a four-point scale ranging from 1 (“do not think so”) to 4 (“strongly think so”). The selected breast cancer risk factors were as follows: “heredity,” “never having children,” “alcohol consumption,” “body size,” and “issues regarding past child delivery or menstruation” [3, 4, 8, 15, 20, 25]. Also, we included the following items as risk factors, even though their reported association with breast cancer development is relatively weak or the findings across studies are inconsistent: “never breast-feeding,” “dietary habit,” “tobacco,” and “exposure to chemical substances” [3, 5, 10, 13, 34]. The selected conventional or common explanations for the breast cancer were “stress,” “personal characteristics,” “chance,” “occupation,” “physical fragility,” “fate,” “medication,” “God’s will,” “air pollution,” and “blow to the breast or breast injury.”

Psychological adjustment

To measure psychological adjustment to the illness at the follow-up, we employed two sets of questionnaires: the General Health Questionnaire (GHQ-28) [22] and the Center for Epidemiologic Studies Depression Scale (CES-D) [11]. The GHQ-28 is a self-administered questionnaire originally designed to detect current diagnosable psychiatric disorders. With four separate scores based on subscales ranging from 0 to 21, it measures the identifiable elements of distress: somatic symptoms, anxiety and insomnia, social dysfunction, and severe depression. For each element, the higher the score, the greater the severity of that element. The CES-D is a self-reporting scale designed to measure depressive symptomatology during the past week. It consists of 20 questions with scores ranging from 0 to 60. On the CES-D, higher scores indicate the existence of more symptoms of depression. Among the 63 participants who filled in the questionnaires at the follow-up, some left several items blank, which made it impossible to obtain the score of each measure for these participants; therefore, participants with missing scores were excluded from each analysis. The numbers of participants included in the analysis of the scores for the GHQ-28 was as follows: 59 for somatic symptoms, 59 for anxiety and insomnia, 63 for social dysfunction, and 60 for severe depression. For the analysis of the CES-D score, 53 participants were included. At the follow-up, we also measured the perceived degree of physical discomfort from nausea, pain, mobility, and bowel pattern, which might act as a confounder, by using a symptom distress scale developed by McCorkle and Young [21].

Statistical analysis

The mean and standard deviation (SD) for each causal attribution were calculated. We also determined the prevalence of each attribution from the percentage of participants who gave it a high score [3–4]. The relationship between each causal attribution and measure of psychological state was assessed by using the Spearman rank correlation. To consider the possible confounding effect of age, marital status, cancer stage at diagnosis, time between the day of diagnosis and follow-up day, and physical symptoms at the follow-up were partially adjusted in the correlation analysis.

For certain risk factors, we restricted the analysis to the participants who were eligible to be discussed. The attribution to “never having children” was analyzed only among nulliparous participants, and that to “occupation” was analyzed only among participants who had a full-time or part-time job. The attribution to “heredity” was analyzed only among participants who reported a family history of mammary gland disease. The attribution to “alcohol consumption” was analyzed only among participants who reported drinking at least once a week. All the statistical analyses were performed with SAS software (SAS Institute, Cary, NC, USA).

Results

The baseline characteristics of the study participants are summarized in Table 1. Of the participants, 60% answered yes to the question: “Have you ever thought about why you have this illness?” Table 2 presents the mean score of each causal attribution for breast cancer in descending order with the percentage of participants who attributed their illness to each factor. Of the attributions to breast cancer risk factors, “never having children” obtained the highest score. The attributions to “diet” and “heredity” received the second- and third-highest scores, respectively. The attribution to “tobacco” had the lowest mean score, but, when we conducted a subsidiary analysis including only the participants who were current or past smokers, the score increased (mean = 2.44, SD = 0.88), and 44% of them attributed their illness to this factor.

Table 1 Baseline characteristics of Japanese women diagnosed with breast cancer
Full size table
Table 2 Causal attributions to risk and non-risk factors among Japanese breast cancer patients, in ranking order
Full size table

The relationship between causal attributions and psychological state at the follow-up is shown in Table 3. Of the attributions to breast cancer risk factors, “body size” was significantly inversely correlated with the measure of anxiety and insomnia and with the measure of social dysfunction on the GHQ-28. Although there was a lack of statistical significance, “body size” was also inversely associated with the measure on the CES-D. Because obesity was reported as a risk factor for breast cancer in postmenopausal women [23], we additionally conducted a correlation analysis of the attribution to “body size” and psychological measures among the participants who were postmenopausal and whose body mass index (BMI) was 23 kg/m2 or over, a reported optimal cut-off point for obesity among the Japanese population (n = 18) [14]. The result did not alter that obtained from the original analysis (data not shown). The attribution to “tobacco” was significantly positively correlated with the measure of anxiety and insomnia (r = 0.34, p = 0.01) and the measure of severe depression (r = 0.31, p = 0.02) on the GHQ-28. The attribution to “never having children” showed a nonsignificant but relatively consistent inverse association across the psychological measures. Of the conventional or common explanations for breast cancer, the attribution to “stress,” “occupation,” “physical fragility,” and “personal characteristics” were significantly positively correlated with some of the measures on the GHQ-28 and CES-D. Only the attribution to “fate” showed a significant inverse correlation with the psychological measure on the GHQ-28 of the attribution to the conventional or common explanations. We found that the level of education obtained by the participants varied according to employment status, but an additional adjustment for education in years and a stratified analysis with employment status did not alter the original result (data not shown). We also conducted an additional analysis adjusting for dispositional optimism, but the result did not alter our original findings either (data not shown).

Table 3 Spearman rank correlation between causal attributions and later psychological adjustment, partially adjusted for confounding variables among Japanese breast cancer patients
Full size table

Discussion

The attribution to a breast cancer risk factor, “body size,” significantly decreased the level of distress measures, indicating better psychological adjustment to the illness. The attribution to “never having children” also seemed to result in consistently better adjustment. An insignificant but relatively consistent inverse correlation pattern was observed between other breast cancer risk factors and the score on the CES-D as well; the attributions to five out of nine evaluated breast cancer risk factors showed an inverse correlation, indicating better psychological adjustment to the illness. The finding suggests that causal attributions to some breast cancer risk factors are likely to help patients adjust psychologically to the illness, although it has frequently been indicated in previous studies that forming causal attributions, mostly to non-risk factors, was associated with worsening psychological adjustment to the illness [6, 7, 9, 33]. Similar to the results from a current study, those from a recent study about health behavior among breast cancer patients indicated that patients who attributed their illness to certain breast cancer risk factors later modified their behaviors appropriately in response to the attributions, indicating better behavioral adjustment to the illness [26].

The attribution to “tobacco” was associated with anxiety and insomnia on the GHQ-28, suggesting that the attribution worsened the adjustment. It is widely known that tobacco has various adverse health effects, and the participants may have linked the attribution to self-blame; eight of ten participants who indicated the attribution were current or former smokers. The effort of tobacco control greatly contributes to public health by preventing various detrimental health problems. Preventing smoking may also reduce the psychological burden and enhance adjustment to the illness for people who develop breast cancer.

Previous studies reported a link between a nonwhite or minority background and the formation of nonscientific or rather religious explanations for the cause of the illness [2, 12, 16]. The attribution to God’s will, however, was not common among Japanese participants in the current study in which about 40% of the participants considered themselves Buddhists and close to half of them reported having no specific religious belief. Further investigation of our data showed that religious background and psychological state were not associated with any causal attribution (data not shown).

There are several limitations to the current study. The sample size may not have been large enough for statistical analysis, especially for an analysis restricted to certain participants. Ten participants had to be excluded from the analysis involving the CES-D score because their questionnaires were incomplete. Our follow-up time may not have been long enough to measure the psychological adjustment of the participants. It was reported that, even a year or longer after mastectomy, the patients were still suffering from distress, depression, or anxiety caused by the surgery [30]. Two previous studies of causal attributions among breast cancer patients assessed the relationship between the time of the diagnosis and the time of adjustment to the illness and obtained inconsistent results; one study reported a higher level of psychological coping with a longer period after the diagnosis [31], and the other study showed no correlation between time and each measure of coping [33]. Moreover, it is possible that each participant’s innate personality acted as a determinant of causal attributions on the one hand and influenced the measures of later psychological adjustment on the other hand. However, our additional adjustment for optimistic personality did not alter the results; still, other personality types may need to be considered.

Despite the limitations, the findings of the current study present valuable information on causal attributions of Japanese breast cancer patients. The formation of several causal attributions increased distress and depression, but the causal attributions to certain breast cancer risk factors showed potential beneficial effects on later psychological state, suggesting a better adjustment to the illness. The findings suggest that further studies would be valuable to confirm the potential benefits of forming causal attributions to risk factors.