This methodological study was conducted in 2018. Study participants were 400 women with BC who fulfilled the following criteria: definite diagnosis of BC at least three months before the study or receiving BC treatments (i.e. surgery, radiation therapy, or hormone therapy), stable health status, no cognitive disorder, and no history of metastasis, recurrent cancer, or psychiatric disorders. Because of the dramatic effects of chemotherapy on body image, women who were receiving chemotherapy were not included.
Participants were conveniently recruited from Shahid Rajaei leading cancer care center, Babolsar, Iran.
Data collection
Data were collected using a demographic questionnaire and BIBCQ. The demographic questionnaire included items on age, length of affliction by BC, marital status, type of surgery (mastectomy or breast-conserving surgery),as well as satisfaction with financial status, health status, sexual function, and quality of life. Satisfaction-related items were scored on a three-point scale as “Good”, “Moderate”, or “Low”. BIBCQ contains 53 items in the six main subscales of vulnerability, body stigma, transparency, body limitations, body concerns, and arm concerns. Items are scored on a five-point Likert scale from “Completely agree” to “Completely disagree” for items 1–23 and from “Never’ to “Always” for items 24–53. Items 1–23 and 29–50 are the same for patients with mastectomy or breast-conserving surgery, while items 24 and 51 are specific for patients with mastectomy and items 25–28, 52, and 53 are specific for those with breast-conserving surgery (20).
Forward- back ward translation
Initially, we contacted the developer of BIBCQ and obtained her permission for using the questionnaire. Then, the questionnaire was translated using the World Health Organization standard protocol for instrument translation (22). Accordingly, the questionnaire was translated into Persian by two English-Persian translators, one of whom was familiar with medical concepts and terminology. Their translations were combined to produce a single Persian BIBCQ. Then, the Persian BIBCQ was back-translated into English by two other translators and their translations were combined to produce a single English BIBCQ. The final English BIBCQ was sent to the developer of the original BIBCQ. She approved the soundness of the translation.
Psychometric evaluation
Content validity evaluation
The content validity of BIBCQ was evaluated both qualitatively and quantitatively. In qualitative content validity evaluation, twelve experts in BC and body image (including two oncologists, two gynecologists, two psychologists, four reproductive health specialists with PhD degree, and two nurses with PhD degree) were asked to evaluate the grammar, wording, and allocation of the BIBCQ items. BIBCQ was revised based on their comments (23). In quantitative content validity evaluation, content validity ratio (CVR) and index (CVI) were calculated to respectively determine the essentiality and the relevance of the items. Accordingly, BIBCQ was provided to the same twelve experts to evaluate the essentiality of each item on a three-point scale with the following three points: 1: “Not essential”; 2: “Useful but not essential”; and 3: “Essential”. Then, CVR of each item was calculated using the following formula: CVR = (Ne – N/2) / (N / 2), where Ne was the number of experts who considered the intended item essential and N was the total number of experts (24). Items with a CVR of more than 0.56 were considered appropriate (25). The same experts were also asked to rate the relevance of each item on a four-point scale as the following: 1: “Irrelevant”; 2: “Somewhat relevant”; 3: “Relatively relevant”; and 4: “Completely relevant”. Then, the CVI of each item (i.e. I-CVI) was calculated through dividing the number of experts who rated that item 3 or 4 by the total number of experts. Then, the average scale-level CVI (S-CVI/Ave) was calculated through dividing the sum of I-CVIs by the total number of the items. S-CVI/Ave value more than 0.90 was considered appropriate (24).
Construct validity evaluation
The construct validity of BIBCQ was evaluated through factor analysis. Sample size for factor analysis was estimated using the rule of thumb in methodological studies which considers 100–200 participants adequate (26). Thus, 200 participants were recruited for exploratory factor analysis and 200 for confirmatory factor analysis.
Exploratory factor analysis was conducted using the maximum-likelihood method and with Promax rotation. Sample adequacy was tested through the Kaiser-Meyer-Olkin (KMO) and the Bartlett’s tests. KMO values of 0.7–0.8 and 0.8–0.9 were interpreted as good and excellent, respectively (27). The presence of an item in a factor was determined based on a factor loading of almost 0.3, which was calculated using the following formula: CV = 5.152/ √(n – 2), where CV was the number of extractable factors and n was the sample size (28). The number of latent factors was estimated using parallel analysis (29). Items with communalities less than 0.2 were excluded from exploratory factor analysis (30). After factor extraction, the Pearson correlation analysis was employed to test the correlations of the extracted factors with items 24 and 51 (which were specific for patients with mastectomy) and items 25–28, 52, and 53 (which were specific for those with breast-conserving surgery). Each of these items was loaded on the factor with which it had the strongest significant correlation.
After exploratory factor analysis, first-order confirmatory factor analysis was conducted using the maximum-likelihood method and the most common goodness of fit indices in structural equation modeling. Then, second-order confirmatory factor analysis was conducted with the presumption that the latent factors extracted in the first-order analysis reflected another level of the intended concept and could indicate a broader concept at a higher level (31).
Convergent and discriminant validity evaluation
The convergent and discriminant validity of BIBCQ were evaluated using Fornell and Larcker’s approach and through the Average Variance Extracted (AVE), the Maximum Shared Squared Variance (MSV), and the Average Shared Squared Variance (ASV) (32). An AVE of more than 0.5 reflects acceptable convergent validity and an AVE greater than MSV and ASV confirm discriminant validity (33, 34).
Reliability evaluation
The internal consistency of BIBCQ was evaluated via calculating average inter-item correlation (AIC), Cronbach’s alpha, and McDonald’s omega (35). Cronbach’s alpha value of more than 0.6 (36), McDonald’s omega value of more than 0.7 (25), and AIC value of 0.2–0.4 (37) were considered acceptable. Moreover, composite reliability was calculated in confirmatory factor analysis (29, 38). Composite reliability is a substitute for Cronbach’s alpha in structural equation modeling. Composite reliability of more than 0.7 is acceptable (39).
Normal distribution of the data, outliers, and missing data
Univariate distribution of the data was evaluated using the skewness coefficient (±3) and the kurtosis coefficient (±3), while multivariate distribution of the data was evaluated using the Mardia coefficient (more than 8). Moreover, the existence of multivariate outliers was evaluated through Mahalanobis distance (P < 0.001). The rate of missing values was evaluated through multiple imputations and then, missing values were replaced with the mean of participants’ responses (25). All data analyses were performed using the SPSS Amos (v. 25.0) and the JASP (v. 0.9.2.0).