**2. Methods**

28 Topics in Cancer Survivorship

showed that chemotherapy (Fehlauer *et al.*, 2005, Joly *et al.*, 2000), hormone therapy (Ganz *et al.*, 1998), and radiotherapy (Fehlauer *et al.*, 2005, Hopwood *et al.*) did not negatively affect body image, Schover et al. (Schover *et al.*, 1995) concluded that chemotherapy did have a

Women's perception about their bodies may be influenced by the length of time since treatment. In general, most studies found that body image improved over time (Hopwood *et al.*), for MTC and BCT (Ganz *et al.*, 1992, King *et al.*, 2000).However, in a more recent study, it was found that most body image scores were quite stable, especially for MTC patients. Only BCT patients felt more attractive and feminine after two years (Engel *et al.*, 2004). Information about the effect of surgery across time is lacking and few studies measured body image before diagnosis making it impossible to know the effect of treatment on patients' body image. Therefore, the first aim of this prospective follow-up study was to examine changes in body image across one year, starting before diagnosis and comparing women with benign breast problems (BBP group) with women with breast cancer (MTC and

The impact of disease and treatment on general QOL seems to vary with age, marital status, and educational level, with younger women and women with lower levels of education reporting lower QOL scores when patients received chemotherapy (Janz *et al.*, 2005, King *et al.*, 2000). Two studies reported a strong relationship between age and body image (Al-Ghazal *et al.*, 1999, King *et al.*, 2000, Yeo *et al.*, 2004). The largest negative impact of MTC on body image was found amongst young, married women (King *et al.*, 2000, Yeo *et al.*, 2004). In contrast, other studies did not find differences in scores on body image between younger and older women (Engel *et al.*, 2004, Hartl *et al.*, , Kenny *et al.*, 2000, Zimmermann *et al.*,

Body image is not only influenced by life events such as having breast cancer, but also by culture, socio-economic status, and personality (Diener *et al.*, 2003). The personality traits extraversion (the disposition towards cheerfulness, sociability, and high activity) and neuroticism (the tendency to experience distressing emotions, such as fear, guilt, and frustration) may have an effect on QOL (Diener *et al.*, 2003). Only one study examined these characteristics and found that neuroticism was acting as a vulnerability factor for anxiety and/or depressive symptoms one year after breast cancer surgery (Millar *et al.*, 2005). Besides, Costa et al. (Costa *et al.*, 1992) found that neuroticism was correlated with a negative body image and extraversion was correlated with a positive body image. However, there is a lack of prospective data on possible relationships between psychological, clinical, and demographic factors and body image (Hartl *et al.*, 2003, Zimmermann *et al.*, 2009). Therefore, the second aim of this prospective study was to examine the effects of personality, sociodemographic factors, and type of surgery on body image in breast cancer patients at different time points after treatment over a one-year period. In contrast with previous studies on satisfaction with body image in breast cancer patients, this study examined which factors from a combination of factors (age, marital status, educational level, work status, disease stage, type of surgery, chemotherapy, radiotherapy, hormone therapy, neuroticism, extraversion, agreeableness, openness to experiences, conscientiousness) predicted body image. It is hypothesized that body image problems arise from surgery and are most commonly experienced following MTC. In addition, body image problems are a

negative impact on body image, while hormonal and radiation therapy did not.

BCT).

2009).

function of personality (i.e., neuroticism).

#### **2.1 Participants**

Women with a palpable lump in the breast or an abnormality on a screening mammography were referred by their general practitioner to the outpatient clinic of the St. Elisabeth Hospital (Tilburg), the Maasland Hospital (Sittard; since August 2004), or the Jeroen Bosch Hospital (Den Bosch; since January 2006) in the Netherlands between September 2002 and September 2007. Women were included if they had an abnormality in the breast, were able to read, speak and write Dutch, and were 18 years or older. Women who had a history of abnormalities in the breast, benign or malignant, or had a breast tumor that was too large (>5 centimeter) for BCT, were excluded from the study. After written informed consent and before the first appointment with the surgeon, the participating women completed a set of questionnaires. Thus, women completed the first set of questionnaires when the diagnosis was still unknown. After this baseline measurement (Time-1), a set of questionnaires was also completed one (Time-2), three (Time-3), six (Time-4), and 12 months (Time-5) after diagnosis (BBP) and/or surgical treatment (BC). The breast cancer group consisted of 219 patients; the women with benign breast problems (BBP group) consisted of 381 patients (See Figure 1). Non-participants (57.8 ± 10.1 yrs) were older than participants (55.0 ± 10.4 yrs; p=.001) in the study. They did not differ on other sociodemographic (i.e., living with a partner, having children, educational level) or clinical characteristics (i.e., disease stage, type of surgery, adjuvant therapy). The length of the questionnaires and the amount of stress the women experienced during the diagnostic period were the reasons for not participating in the study. This study was approved by the local ethics committee.

#### **2.2 Questionnaires**

All women completed questionnaires on personality factors (only at baseline) and the WHOQOL-100 Body Image and Appearance facet (all time points). The BC group also completed the EORTC-QLQ-BR23 Body image subscale from Time-2 onwards. Both instruments were chosen since both subscales complement each other, i.e., the Body and Appearance facet covers more general concerns (satisfaction with they way the body looks, acceptance of bodily appearance, and inhibition by own looks), while the Body image subscale covers feelings of low attractiveness and femininity as a result of cancer or treatment.

The Neuroticism-Extraversion-Openness Five Factor Inventory (NEO-FFI) (Costa and McCrae, 1992, Hoekstra *et al.*, 1996) was developed to study an individual's personality by testing the five domains of the Five Factor or Big Five Model: neuroticism (i.e., the tendency to experience distressing emotions, such as fear, guilt, and frustration), extraversion (i.e., the disposition towards cheerfulness, sociability, and high activity), openness (i.e., the tendency to have a receptive orientation towards varied experiences and ideas), agreeableness (i.e., the inclination towards interpersonal trust and consideration of others) and conscientiousness (i.e., the tendency towards persistence, sense of duty, organizing, planning, and self-discipline). This self-report questionnaire consists of 60 statements. Each statement is rated on a 5-point scale ranging from 1 (*strongly disagree*) to 5 (*strongly agree*), resulting in dimension scores ranging from 12 to 60. The psychometric properties are acceptable to good.(Costa and McCrae, 1992)

The World Health Organization Quality of Life assessment instrument-100 (WHOQOL-100) (De Vries and Van Heck, 1997, WHOQOL Group, 1998) is a cross-culturally developed generic multi-dimensional quality of life measure. This questionnaire consists of 100 items

Changes in Body Image in Women with Early Stage Breast Cancer 31

The EORTC QLQ-BR23 is a 23-item disease-specific questionnaire measuring health status in breast cancer patients. The BR-23 is a supplementary module of the EORTC QLQ C30, which covers the physical, personal, cognitive, emotional, and social domains.(Montazeri *et al.*, 2000, Sprangers *et al.*, 1996) The EORTC QLQ BR-23 incorporates two functional scales (Body Image and Sexual Functioning) and three symptom scales (Arm Symptoms, Breast Symptoms, and Systematic Therapy Side Effects). The remaining items assess sexual enjoyment and being upset by hair loss. The reliability and validity are adequate(Montazeri *et al.*, 2000, Sprangers *et al.*, 1996, Yun *et al.*, 2004). In this study, only the scale Body Image was used. The Body Image scale consists of four items, for instance 'Did you feel less feminine as a consequence of your illness or treatment?' This scale was linearly transformed (score range 0-100). A higher score represent higher levels of functioning. The reliability and validity of this scale is adequate.(Sprangers *et al.*, 1996, Yun *et al.*, 2004) The Cronbach's alpha coefficients of the Body Image scale in this study were .87 (BCT group) and .89 (MTC

Patients were asked to respond to a number of questions concerning age, marital status, having children, and years of education. Marital status was dichotomized in two categories,

Data concerning diagnosis, type of surgical treatment (BCT or MTC), disease stage, and type of adjuvant treatment (chemotherapy, hormone therapy, and radiotherapy) were obtained

Student t-tests and chi-square tests were used to examine differences between participants and non-participants and women who had undergone BCT and MTC. General linear model for repeated measures was used to examine if scores on Body Image changed over time, if scores on Body Image were different for (women with a benign diagnosis,) women who undergone BCT or MTC, and if the pattern of Body Image scores over time was different for BCT and MTC. Subsequently, multivariate analysis of covariance (MANCOVA) with repeated measures was performed to adjust for the effect of potential confounders on the relationship between body image and group (benign breast problems, MTC, BCT). Radiotherapy and disease stage were selected as covariates based on statistical differences between treatment groups on baseline characteristics. Post-hoc paired samples t-tests were conducted to determine differences in body image for group separately. Linear regression analyses were performed to examine which, and to what extent, sociodemographic, clinical, and personality variables predicted the scores on body image (WHOQOL-100, EORTC QLQ-BR23). For the time points, one, three, six, and 12 months after surgery, a hierarchical multiple regression analysis (method: enter) was conducted. As a first step, aiming at minimizing the number of independent variables in the final regression analysis, separate preliminary regression analyses were performed with sociodemographic (age, marital status, educational level, work status), clinical (disease stage, type of surgery, chemotherapy, radiotherapy, and hormone therapy), psychological (body image at baseline), and personality factors (neuroticism, extraversion, agreeableness, openness to experiences, conscientiousness) as independent variables. Subsequently, significant predictors (p<.05) were entered in the final regression analysis. Chemotherapy, radiotherapy, and hormone therapy were not entered at Time-2 as a factor, since women with breast cancer did not yet received adjuvant therapy at that time. Mean and standard

i.e. being involved in a relationship or not being involved in a relationship.

deviations are provided as (M ± SD). SPSS 17.0 was used for all calculations.

from the medical records of the included patients.

**2.3 Statistical procedures** 

group).

that are divided in 24 facets covering four domains (Physical health, Psychological health, Social Relationships, and Environment) and an Overall Quality of Life and General Health facet. Each facet is measured with four items using 5-point Likert scales. In the present study, only the facet Body Image and Appearance was used. The facet body image consists of four items, for instance 'Are you able to accept your bodily appearance?' A high facet score indicates good body image (score range: 4 - 20). The reliability and validity are adequate and sensitivity is high. (De Vries and Van Heck, 1997, Den Oudsten *et al.*, 2009b, O'Carroll *et al.*, 2000) The Cronbach's alpha coefficients of the facet Body Image and Appearance in this study were .85 (BCT group), .87 (BBP group), and .88 (MTC group).

Fig. 1. Flow chart

The EORTC QLQ-BR23 is a 23-item disease-specific questionnaire measuring health status in breast cancer patients. The BR-23 is a supplementary module of the EORTC QLQ C30, which covers the physical, personal, cognitive, emotional, and social domains.(Montazeri *et al.*, 2000, Sprangers *et al.*, 1996) The EORTC QLQ BR-23 incorporates two functional scales (Body Image and Sexual Functioning) and three symptom scales (Arm Symptoms, Breast Symptoms, and Systematic Therapy Side Effects). The remaining items assess sexual enjoyment and being upset by hair loss. The reliability and validity are adequate(Montazeri *et al.*, 2000, Sprangers *et al.*, 1996, Yun *et al.*, 2004). In this study, only the scale Body Image was used. The Body Image scale consists of four items, for instance 'Did you feel less feminine as a consequence of your illness or treatment?' This scale was linearly transformed (score range 0-100). A higher score represent higher levels of functioning. The reliability and validity of this scale is adequate.(Sprangers *et al.*, 1996, Yun *et al.*, 2004) The Cronbach's alpha coefficients of the Body Image scale in this study were .87 (BCT group) and .89 (MTC group).

Patients were asked to respond to a number of questions concerning age, marital status, having children, and years of education. Marital status was dichotomized in two categories, i.e. being involved in a relationship or not being involved in a relationship.

Data concerning diagnosis, type of surgical treatment (BCT or MTC), disease stage, and type of adjuvant treatment (chemotherapy, hormone therapy, and radiotherapy) were obtained from the medical records of the included patients.

#### **2.3 Statistical procedures**

30 Topics in Cancer Survivorship

that are divided in 24 facets covering four domains (Physical health, Psychological health, Social Relationships, and Environment) and an Overall Quality of Life and General Health facet. Each facet is measured with four items using 5-point Likert scales. In the present study, only the facet Body Image and Appearance was used. The facet body image consists of four items, for instance 'Are you able to accept your bodily appearance?' A high facet score indicates good body image (score range: 4 - 20). The reliability and validity are adequate and sensitivity is high. (De Vries and Van Heck, 1997, Den Oudsten *et al.*, 2009b, O'Carroll *et al.*, 2000) The Cronbach's alpha coefficients of the facet Body Image and Appearance in this study were .85 (BCT group), .87 (BBP group), and .88 (MTC group).

Fig. 1. Flow chart

Student t-tests and chi-square tests were used to examine differences between participants and non-participants and women who had undergone BCT and MTC. General linear model for repeated measures was used to examine if scores on Body Image changed over time, if scores on Body Image were different for (women with a benign diagnosis,) women who undergone BCT or MTC, and if the pattern of Body Image scores over time was different for BCT and MTC. Subsequently, multivariate analysis of covariance (MANCOVA) with repeated measures was performed to adjust for the effect of potential confounders on the relationship between body image and group (benign breast problems, MTC, BCT). Radiotherapy and disease stage were selected as covariates based on statistical differences between treatment groups on baseline characteristics. Post-hoc paired samples t-tests were conducted to determine differences in body image for group separately. Linear regression analyses were performed to examine which, and to what extent, sociodemographic, clinical, and personality variables predicted the scores on body image (WHOQOL-100, EORTC QLQ-BR23). For the time points, one, three, six, and 12 months after surgery, a hierarchical multiple regression analysis (method: enter) was conducted. As a first step, aiming at minimizing the number of independent variables in the final regression analysis, separate preliminary regression analyses were performed with sociodemographic (age, marital status, educational level, work status), clinical (disease stage, type of surgery, chemotherapy, radiotherapy, and hormone therapy), psychological (body image at baseline), and personality factors (neuroticism, extraversion, agreeableness, openness to experiences, conscientiousness) as independent variables. Subsequently, significant predictors (p<.05) were entered in the final regression analysis. Chemotherapy, radiotherapy, and hormone therapy were not entered at Time-2 as a factor, since women with breast cancer did not yet received adjuvant therapy at that time. Mean and standard deviations are provided as (M ± SD). SPSS 17.0 was used for all calculations.

Changes in Body Image in Women with Early Stage Breast Cancer 33

Patients who underwent MTC were more often treated with chemotherapy and hormone therapy, compared to patients who received BCT (p *<*.05). As expected, based on standard treatment, women with BCT were more often treated with radiotherapy (p<.0001) and differed regarding disease stage (p<.0001). No other differences between the surgical groups were found with regard to other sociodemographic, clinical, and psychological variables. Figure 2 shows the change in scores on WHOQOL-facet Body Image. Body Image changed significantly over time [F(4,239) = 3.0; p =.020], after correcting for potential confounders. Furthermore, an interaction effect was found for time by surgical treatment, indicating that the pattern of change over time in Body Image is different for women with MTC, women with BCT, and women with BBP [F(8,480) = 2.8; p =.004]. From Time-1 to Time-2, women with MTC reported a significant deterioration in their Body Image (p=.035), while women with BCT and BBP were stable. Although their Body Image improved in time, they had significantly lower scores at Time-5 when compared to Time-1 (p = .004). Radiotherapy and disease stage did not interact with Body Image (p >.05). Overall, women with BBP and women with BCT and MTC did not score differently on Body Image, except at Time-2

Fig. 2. Mean scores on Body Image and Appearance (WHOQOL-100) across time for women who undergonebreast-conserving therapy (BCT) and mastectomy (MTC) and women with a

(p<.036).

benign diagnosis
