The present study reports on a psychometric evaluation of the BCIA, including identification of factors and creation and validation of factor-based scales, in a sample of 783 breast cancer survivors on average 40 months post diagnosis. Factor analysis revealed four meaningful scales that measure the impact of cancer on caregiving abilities and finances, exercise and diet behaviors, social and emotional functioning, and religiosity. Results indicated that each of the four scales had good internal consistency reliability, the items within each scale were sufficiently highly related but not redundant, and each scale represented a homogeneous construct.
The BCIA scales identified here only minimally correspond to the scales defined by the factor analysis of the newly-created Impact of Cancer Scale [14
]. Change in religious faith or spirituality is measured both by the Positive Outlook scale of the IOC and the Religiosity scale of the BCIA. However, while both the IOC and the BCIA scales assess the impact of cancer on social life, the BCIA social/emotional scale includes domains such as family plans, love life, and living arrangements, whereas the IOC focuses on life interferences and changes in the value of relationships. While the IOC scale does include three items assessing potential change in employment, these items were not subjected to the factor analysis and so it is not known whether an employment factor would emerge, corresponding to the BCIA caregiving/finances scale. Further, the IOC scale does not measure impact on exercise or diet behaviors. The psychometric analyses of the BCIA and the IOC, suggest that, apart from commonalities in religiosity/spirituality, the two instruments measure different aspects of the impact of cancer among survivors. Future research is needed to elucidate which domains of well-being are most strongly impacted by cancer.
We computed scale scores for each of the four cancer impact scales. The correlations among these scale scores were generally low, indicating that the scale scores are measuring distinct but related dimensions of the impact of cancer and can be used as separate outcomes in future research. We correlated the four cancer impact scales with psychosocial and health-related QOL measures to establish convergent and discriminant validity. For the most part, these correlations were low. Though the pattern and direction of these correlations generally supported the validity of the cancer impact scales, the low correlations indicate that the four cancer impact scales measure a construct distinct from optimism, fear of recurrence, perceived stress, post-traumatic growth, or health-related QOL.
The pattern of results reported here supported the validity of the Caregiving/Finances scale. Consistent with our hypothesis, Caregiving/Finances scores were negative and much greater among those participants who likely have greater financial concerns: those who were unemployed on leave, or looking for work. The pattern for participants with lower household income to report more negative Caregiving/Finances scores is also consistent. Further support for the validity of the scale can be drawn from the pattern of correlations of the Caregiving/Finances scale scores with the psychosocial and health-related QOL scale scores. Participants who scored low on optimism (i.e., pessimists), reported higher fear of recurrence and perceived stress, perceived fewer positive growth experiences after cancer, or reported poor physical and emotional functioning were more likely to report a greater negative impact on their caregiving and financial responsibilities. It makes sense that participants with poor emotional functioning would have a harder time caring for others and would have a tendency to perceive the impact of cancer on the current financial situation more negatively.
Support for the validity of the Exercise/Diet
scale was demonstrated in several ways. Younger women reported a greater positive impact of cancer on Exercise/Diet compared to those over age 70. Since physical activity has been shown to decline with age [26
], it is likely that older participants exercised less before their cancer diagnosis and would be less likely to improve their exercise behavior after cancer. Consistent with our hypothesis, married participants reported higher positive Exercise/Diet scale scores, and there was a pattern for women with higher incomes to report greater positive changes in exercise and diet behaviors as well. Married and higher-income women may have had better psychosocial adjustment or simply more social support and available resources to support improvement in their diet and exercise behaviors after cancer. Contrary to our hypothesis, women who were not working outside the home reported a less positive impact of cancer on diet and exercise. This may be due in part to the inclusion of retired and disabled women in this employment group. Women who took an early retirement or became disabled as a result of their cancer experience may not have the resources to improve their diet or exercise behaviors. Further, if their retirement or disability reflects a more demanding or negative cancer experience, the positive impact of cancer on diet and exercise may be limited.
Correlations with other health scale scores further supported the validity of the Exercise/Diet scale. People who reported greater optimism, physical and emotional functioning, and positive growth after cancer, and lower fear of recurrence and perceived stress were more likely to report a greater positive impact on their exercise and diet behaviors. Optimists, who by definition have positive future expectancies, should respond to difficulty (i.e., a cancer diagnosis) with continued effort by changing their diet and exercise behaviors, while pessimists (who have negative expectancies) would be more likely to give up [27
], and not exercise or change their diet. Better physical and emotional functioning (i.e., low distress) should be associated with greater positive impact on diet and exercise. Previous research has shown that higher levels of distress have been associated with poor diet [28
] and depressed mood is frequently associated with lack of physical activity [30
The pattern of results supported the convergent and discriminant validity of the Social/Emotional scale
in several ways. Social/Emotional scores were more negative for participants who were on leave or looking for work, possibly due to the loss of social support at work. As hypothesized, the Social/Emotional scale showed good discriminant validity, correlating more strongly with the SF-36 mental component summary scale than with the physical component. Consistent with the concept of better social/emotional adjustment, Social/Emotional scores were positively correlated with optimism and post-traumatic growth, and negatively correlated with fear of recurrence and perceived stress. Of note, the impact of cancer on social/emotional functioning was much more negative in women treated with surgery and chemotherapy compared to women who had other combinations of treatments. Post-hoc analyses revealed that these women tended to be younger and were more likely to have had a total mastectomy (data not shown). Previous research has shown that women treated with mastectomy may have poorer emotional adjustment, due for example, to negative body image [31
]. This finding further supports the validity of the Social/Emotional impact scale.
The convergent and discriminant validity of the Religiosity scale score
was demonstrated in several ways. As hypothesized, a positive impact of cancer on religiosity was correlated with positive post-traumatic growth on all five PTGI scales, and the correlation with the Spiritual Change scale was highest. Further, the Religiosity scale was not correlated with fear of recurrence, perceived stress, or physical health scales on the SF-36. The impact of cancer on religiosity was lower among participants 70 years of age or older and among those with lower household incomes. This could be because of a ceiling effect, as the elderly [32
] and lower SES individuals [33
] tend to report higher levels of religiosity and would have less room for their religiosity to improve through the cancer experience.
The findings with the BCIA in the present multiethnic sample are similar to those with the original scale in a predominately White sample in the study by Ganz et al., 2002. Participants reported greatest positive impact on exercise, diet, and religiosity items, and the greatest negative impact on love life and financial situation. Further, greater impact was associated with younger age. Also consistent with Ganz and colleagues, the means on many impact scale items corresponded to “no impact.” Future studies evaluating the impact of cancer with the BCIA should investigate whether there are identifiable subgroups of survivors who report greater positive or negative impact of cancer. For example, studies have shown that although many of the QOL difficulties experienced by survivors in the short term resolve over time, there is a subgroup of approximately 20–25% of women who still report decrements in several aspects of QOL up to four years after diagnosis [34
]. It may be that these survivors would report a greater impact of cancer on the BCIA scales.
It is interesting that there was no unique effect of time since diagnosis (2–5 years) on any of the BCIA scale scores. On one hand, this is not surprising given that these participants were at least 2 years post diagnosis and most research finds that the many adverse aspects of cancer diagnosis and treatment are experienced over the course of the first year.(e.g., [35
]). On the other hand, it is interesting that the impact of cancer diagnosis and treatment is still detectable 5 years after diagnosis, and that the positive impact on Religiosity and Exercise/Diet is so durable.
The strengths of this study include its large ethnically and socio-economically diverse sample of breast cancer survivors and the opportunity to relate BCIA scales to relevant psychosocial and health-related QOL instruments. The main limitation of this study is the limited number of instruments to use for construct validation. For example, the impact of cancer on diet and exercise behaviors should be validated with actual changes in measures that assess diet and exercise behaviors. More information is also needed on test-retest reliability, face validity, how the impact of cancer changes in effect over time (e.g. from diagnosis to five or more years post-diagnosis), and how the factor structure presented here relates to these changes. This study relied on retrospective reports of the perceived impact of cancer, a less than ideal approach. Future studies should further evaluate the validity of the BCIA by comparing these retrospective perceived impact scores with change in scores given on corresponding instruments completed both before and after cancer treatment.
In sum, the evidence presented here suggests that the four scales from the BCIA are reasonably reliable and valid and can be used to assess the impact of cancer among breast cancer survivors who are 2 or more years post-diagnosis. Including the BCIA along with measures of QOL, symptoms, and functioning will allow researchers to gain a more comprehensive assessment of the biopsychosocial impact of cancer in survivors. While the BCIA does not measure all the domains of impact captured by the IOC instrument [14
], this brief assessment may be useful to future studies in which a longer scale is not feasible, or where researchers are specifically interested in the impact of cancer on caregiving and finances, exercise and diet behaviors, social and emotional functioning, or religiosity. Future studies of the health and well-being of cancer survivors would benefit from reliable and valid assessment of the potential negative and positive impacts of cancer on these domains.