Qual Life Res DOI 10.1007/s11136-014-0907-1
BRIEF COMMUNICATION
Reliability and validity of the Dutch-translated Body Image Scale V. M. T. van Verschuer • W. W. Vrijland I. Mares-Engelberts • T. M. A. L. Klem
•
Accepted: 22 December 2014 Ó Springer International Publishing Switzerland 2015
Abstract Purpose Lacking a comprehensible and widely applicable Dutch test to assess body image changes in cancer patients, we validated Hopwood’s Body Image Scale (BIS) for the Dutch language. Methods The BIS consists of 10 items scored 0–3. Total scores range from 0 (minimum body image-related distress) to 30 (maximum distress). After forward and backward translation of the BIS, we evaluated its psychometric characteristics in breast cancer patients. We assessed feasibility by missing answer rates and positive response prevalence (score C1) per item (criterion C30 %), test– retest reliability with a 2-week interval, internal consistence using Cronbach’s a and discriminant ability by comparing body image after breast-conserving therapy (BCT) versus mastectomy. Results Psychometric evaluation of 108 BCT and 101 mastectomy patients showed high feasibility (0.2 % missing answers), high positive response prevalence of C30 % in 9/10 items and high internal consistency (a [ 0.90). Test–retest reliability and correlation were high with 5.78 (test) versus 5.75 (retest; P = 0.86) and Spearman’s q = 0.92 (P \ 0.01). Discriminant ability was good with BIS scores of 4.56 after BCT versus 7.19 after mastectomy
V. M. T. van Verschuer Department of Surgical Oncology, Erasmus MC Cancer Institute, Rotterdam, The Netherlands V. M. T. van Verschuer W. W. Vrijland I. Mares-Engelberts T. M. A. L. Klem (&) Sint Franciscus Vlietland Group, Department of Surgery, Sint Franciscus Gasthuis, Kleiweg 500, 3045 PM Rotterdam, The Netherlands e-mail: [email protected]
(P \ 0.01). All results were comparable to the results of the original BIS. Conclusion The Dutch-translated BIS showed excellent psychometric results very similar to the original BIS. Its concise and simple design further supports wide application in clinical practice. Keywords Body image Questionnaires Translating Psychometrics Mastectomy Breast-conserving therapy
Introduction Healthcare quality indicators are increasingly determined by patient-reported outcome measures (PROMs). Commonly used PROMs are quality of life (QoL) assessments of which body image is one of the presumed determinants [1, 2]. Several types of cancer and cancer treatments may cause body image changes. The majority of studies assessing body image changes in cancer patients have focused on breast cancer patients [3–6], but other types of cancer are likely to induce body image distress as well [1, 7–9]. Lacking a comprehensible and widely applicable Dutch test to assess body image changes in cancer patients, we validated Hopwood’s Body Image Scale (BIS) [10] for the Dutch language.
Patients and methods Hopwood’s Body Image Scale (BIS) The BIS assesses body image and body image changes after cancer treatment (Table 1). Respondents are asked to answer questions with reference to the past week. The scale
123
Qual Life Res Table 1 Hopwood’s Body Image Scale (BIS) BIS items
Not at all
A little
Quite a bit
Very much
1. Have you been feeling self-conscious about your appearance?
–
–
–
– –
2. Have you felt less physically attractive as a result of your disease or treatment?
–
–
–
3. Have you been dissatisfied with your appearance when dressed?
–
–
–
–
4. Have you been feeling less feminine/masculine as a result of your disease or treatment?
–
–
–
–
5. Did you find it difficult to look at yourself naked?
–
–
–
–
6. Have you been feeling less sexually attractive as a result of your disease or treatment?
–
–
–
–
7. Did you avoid people because of the way you felt about your appearance? 8. Have you been feeling the treatment has left your body less whole?
– –
– –
– –
– –
9. Have you felt dissatisfied with your body?
–
–
–
–
10. Have you been dissatisfied with the appearance of your scar?
–
–
–
–
N/A
N/A not applicable
consists of ten items including affective items (e.g., feeling ‘‘self-conscious,’’ ‘‘less feminine/masculine,’’ ‘‘less physically attractive’’), cognitive items (e.g., dissatisfied ‘‘with appearance,’’ ‘‘with scar’’) and behavioral items (e.g., ‘‘avoid people,’’ ‘‘difficult to look at yourself naked’’). Response options range from ‘‘not at all’’ (score 0), ‘‘a little’’ (score 1), ‘‘quite a bit’’ (score 2) to ‘‘very much’’ (score 3). Question 10 (‘‘dissatisfied with scar’’) has an additional response option ‘‘not applicable.’’ Summing up the scores, a total score ranging from 0 to 30 per patient is obtained with 0 representing no distress or symptoms, whereas increasing scores represent increasing distress and symptoms. Translation and adaptation of the BIS The adaptation process was previously described by Bullinger et al. [11]. Three Dutch native speakers with extensive knowledge of the English language provided a forward translation into Dutch. Emphasis was lying on conceptual equivalence using simple language, rather than achieving literal translation. Translators discussed difficulties with the principal investigator until consensus was reached on one optimal Dutch formulation. Two native English speakers who were fluent in Dutch provided a backward translation to English. Both backward translations were compared with the original BIS, and any differences were analyzed. Finally, necessary adaptations of the Dutch version were made. The resulting version (Table 2) was given to three patients who had been treated for breast cancer. They were asked to comment on readability and comprehension of the scale. After this last test, no adaptations had to be made and the scale was administered to the study population for psychometric data collection.
123
Study population and data collection We chose to conduct psychometric evaluation in breast cancer patients, since this subgroup of patients is relatively large, homogeneous and likely to have body image-related distress due the type of surgery. The fact that the original version was (partly) validated in breast cancer patients easier the comparison [10]. Female breast cancer patients C18 years who visited our outpatient clinic for follow-up were eligible. We aimed to include 100 breast-conserving therapy (BCT) and 100 mastectomy patients. Exclusion criteria were breast reconstruction after mastectomy and insufficient knowledge of the Dutch language. In order to determine test–retest reliability, participants were asked to complete the Dutch BIS twice with a two-week interval. Patients who had to undergo breast cancer treatment between the test and the retest or who did not return the second copy by mail were excluded. Medical ethical board approval was obtained. All participants gave written informed consent. Validation The SPSS computer package (version 20.0) was used for statistical analyses. Psychometric results are compared to the results of the psychometric evaluation of the final version of the original BIS. Feasibility Missing or non-unique responses (0 or [1 box ticked, respectively) were considered invalid. Questionnaires with \9/10 valid responses to all items were excluded from analyses. When one item was not answered, the maximum score possible at that item was subtracted from the maximum achievable score of the scale (30 points). A valid
Qual Life Res Table 2 Dutch translation of the Body Image Scale Body Image Scale vragen
Helemaal niet
Een beetje
Nogal
Heel erg
1. Heeft u zich onzeker gevoeld over uw uiterlijk?
–
–
–
–
2. Heeft u zich lichamelijk minder aantrekkelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
3. Bent u ontevreden geweest over uw uiterlijk als u aangekleed was?
–
–
–
–
4. Heeft u zich minder vrouwelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
5. Heeft u moeite gehad om uzelf naakt te zien?
–
–
–
–
6. Heeft u zichzelf seksueel minder aantrekkelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
7. Heeft u andere mensen vermeden vanwege hoe u zich voelde over uw uiterlijk? 8. Heeft u het gevoel dat de behandeling uw lichaam minder compleet heeft gemaakt?
– –
– –
– –
– –
9. Bent u ontevreden geweest over uw lichaam?
–
–
–
–
10. Bent u ontevreden geweest over hoe uw litteken eruit ziet?
–
–
–
–
n.v.t.
n.v.t. niet van toepassing
score was calculated by dividing the achieved total score by the new maximum score and multiplying this by 30. Feasibility of the scale was evaluated by response rates and missing answers. Response prevalence was defined as the frequency of positive ratings (score of C1) for each item, indicating a change in some aspect of body image. Per item a criterion of C30 % of positive ratings of the total sample was used. Reliability To assess whether items evaluate the same concept (e.g., body image), internal consistency of scale items was measured using Cronbach’s a, which should exceed 0.70 [12]. Test–retest reliability was tested using Spearman’s correlation coefficient rho (q), paired Student’s t tests and effect size (Cohen’s d = 0.2: small effect, d = 0.5: medium effect and d [ 0.8: large effect size) [13, 14]. Clinical validity Discriminant ability between lumpectomy and the mastectomy subgroups was assessed using Student’s t tests and effect size [13, 14].
Results Response rates and feasibility Both questionnaires were returned by 108/150 BCT patients (72 %) and by 101/150 mastectomy patients (67 %; Table 3). In both questionnaires, a total of 0.2 % of all answers was missing. There were no non-unique answers. None of the questionnaires had [1 invalid responses. All items reached the 30 % response prevalence criterion, except for item 7 (‘‘avoid people because of the
way you felt about your appearance?’’) which had a response prevalence of 12 %. In the first version of the original BIS, three items including item 7 had a response prevalence of B30 %, which decreased to zero items in the final version [10]. Test–retest reliability and internal consistency Test–retest reliability was high. Mean BIS scores were 5.78 (95 % CI 4.97–6.59) in Questionnaire 1 versus 5.75 (95 % CI 4.93–6.57) in Questionnaire 2 (P = 0.86). This is in line with the outcomes of the original BIS with test–retest scores of 8.1 and 9.0, respectively, in the breast subgroup [10]. Correlation between the questionnaires was high for total BIS score (q = 0.92; P \ 0.01) as well as for all items (Table 4), comparable to the correlation coefficient of the original BIS (q = 0.70; P \ 0.01) [10]. Test–retest effect size was low (Cohen’s d = 0.005). Both questionnaires showed high internal consistency with Cronbach’s a of 0.91 and 0.92, respectively, similar to the 0.93 of the original BIS [10]. Clinical validity: discriminant ability between breastconserving therapy and mastectomy Mean total BIS scores of both questionnaires were 4.56 (95 % CI 3.49–5.41) in the BCT group and 7.19 (95 % CI 5.91–8.48) in the mastectomy group (P \ 0.01; Table 5) representing good discriminant ability. Cohen’s d was 0.47, representing medium effect size. In the original BIS, this difference was even more pronounced with median total scores of 2.5 after BCT and 12.0 after mastectomy (P \ 0.01) [10].
Discussion The original BIS was developed in 2000 in the UK for use in clinical trials to assess body image changes in cancer
123
Qual Life Res Table 3 Patient characteristics
BCT group
Mastectomy group
Total patients
108
101
Age, years (median, range)
59.0 (32.0–82.0)
60.0 (33.0–92.0)
Follow-up since surgery, years (median, range)
4.0 (1.0–13.0)
4.0 (1.0–20.0)
Chemotherapy
52 (49 %)
65 (65 %)
Radiotherapy
97 (91 %)
39 (39 %)
Hormonal therapy
49 (46 %)
60 (61 %)
History of
BCT Breast-conserving therapy
Table 4 Scores of Body Image Scale (BIS) items, total BIS scores and test–retest reliability
Questionnaire 1—test
Questionnaire 2—retest
Correlation
Mean score (95 % CI)
Mean score (95 % CI)
q
P valuea
1
0.80 (0.69–0.92)
0.72 (0.61–0.83)
0.80
\0.01
2
0.80 (0.68–0.92)
0.79 (0.68–0.91)
0.80
\0.01
3
0.40 (0.30–0.49)
0.39 (0.30–0.47)
0.68
\0.01
4
0.59 (0.47–0.70)
0.51 (0.40–0.61)
0.74
\0.01
5
0.59 (0.47–0.71)
0.65 (0.52–0.77)
0.80
\0.01
6
0.70 (0.58–0.82)
0.75 (0.62–0.87)
0.80
\0.01
7
0.17 (0.10–0.24)
0.17 (0.11–0.24)
0.63
\0.01
8
0.65 (0.54–0.76)
0.63 (0.52–0.75)
0.73
\0.01
9
0.63 (0.53–0.74)
0.67 (0.56–0.78)
0.72
\0.01
10
0.45 (0.34–0.56)
0.47 (0.36–0.58)
0.79
\0.01
P value of Spearman’s correlation coefficient
Total BIS
5.78 (4.97–6.59)
5.75 (4.93–6.57)
0.92
\0.01
Table 5 Discriminative ability of the Body Image Scale (BIS)
Total BIS score of
N
BCT group N = 108 Mean score (95 % CI)
Mastectomy N = 101 Mean score (95 % CI)
95 % CI 95 % confidence interval, q Spearman’s correlation coefficient q a
BCT Breast-conserving therapy
BIS item
All questionnaires
418
4.56 (3.85–5.26)
7.05 (6.16–7.94)
\0.01
Questionnaire 1—test
209
4.45 (3.49–5.41)
7.19 (5.91–8.48)
\0.01
Questionnaire 2—retest
209
4.66 (3.61–5.71)
6.90 (5.65–8.16)
\0.01
patients, resulting from changes in a patient’s appearance due to cancer treatment [10]. The scale assesses affective, behavioral and cognitive changes corresponding with a multidimensional approach of body image [15–17]. We validated the Dutch-translated version of the BIS. Psychometric evaluation of the Dutch BIS showed results comparable to the original version. Internal consistency was high with Cronbach’s a of 0.92 and 0.91, similar to the original BIS [14]. Clinical validity based on response prevalence indicating a change in an aspect of body image also was comparable to the original tool [10]. Only one item, ‘‘avoidance of other people because of the way a patient feels about his or her appearance,’’ failed to reach the response prevalence criterion of C30 % compared with none of the items in the original version [10].
123
P value
This item indeed addresses an issue of body image likely not frequently impacted in breast cancer patients because the breast surgical area is hidden in most social situations. Discriminant ability was moderate to good between breast cancer patients treated by mastectomy and BCT, representing two groups with expected differences in body image changes. BIS scores differed significantly between the groups, but the effect size was only moderate. The discriminant ability between BCT and mastectomy was comparable to the original BIS, although the difference between the two groups was larger in the original tool [10]. A possible explanation for the moderate effect size is that body image distress may not simply be related to the amount of breast tissue visibly removed during surgery. For example, body image distress is also found to be highly
Qual Life Res
impacted after mastectomy followed by direct breast reconstruction [3]. Both the original and the Dutch BIS showed high test–retest reliability [10]. A limitation of this study was the fact that many women did not return the second questionnaire and were excluded, most likely due to unawareness that it was important to complete the same questionnaire twice. Comprehensibility of the scale did not seem to be a problem since rates of missing or non-unique answers were very low. After starting to emphasize on the importance of the second questionnaire at inclusion, response rates improved. Potential future applications of the BIS include its use as a healthcare quality indicator, for example in combination with assessment of QoL or satisfaction with care. Healthcare quality indicators, however, are not yet well defined [18, 19]. To our knowledge, the BIS is not yet being implemented for the assessment of quality of care. Potential applications of interest include screening for body image-related issues after cancer surgery and evaluation of treatment effects after psychological therapy or reconstructive surgery. Before the BIS can be used for evaluation of treatment effects, however, further specific validation is warranted. To conclude, psychometric evaluation of the Dutch BIS showed excellent results that were comparable those of the original version. The concise and simple design makes the Dutch BIS suitable for assessment of body image issues in routine clinical practice.
4.
5.
6.
7.
8.
9.
10.
11.
12. Acknowledgments We thank all participating patients and acknowledge Dr. P. Hopwood and colleagues for the development of the original Body Image Scale.
13. 14.
Conflict of interest None of the authors declare financial or personal relationships that could inappropriately influence their work. 15.
References 16. 1. Taylor-Ford, M., Meyerowitz, B. E., D’Orazio, L. M., Christie, K. M., Gross, M. E., & Agus, D. B. (2013). Body image predicts quality of life in men with prostate cancer. Psychooncology, 22(4), 756–761. 2. Zebrack, B., & Isaacson, S. (2012). Psychosocial care of adolescent and young adult patients with cancer and survivors. Journal of Clinical Oncology, 30(11), 1221–1226. 3. den Heijer, M., Seynaeve, C., Timman, R., Duivenvoorden, H. J., Vanheusden, K., Tilanus-Linthorst, M., et al. (2012). Body image and psychological distress after prophylactic mastectomy and breast reconstruction in genetically predisposed women: A
17.
18. 19.
prospective long-term follow-up study. European Journal of Cancer, 48(9), 1263–1268. Pusic, A. L., Klassen, A. F., Snell, L., Cano, S. J., McCarthy, C., Scott, A., et al. (2012). Measuring and managing patient expectations for breast reconstruction: Impact on quality of life and patient satisfaction. Expert Rev Pharmacoecon Outcomes Res, 12(2), 149–158. Denewer, A., Farouk, O., Kotb, S., Setit, A., Abd El-Khalek, S., & Shetiwy, M. (2012). Quality of life among Egyptian women with breast cancer after sparing mastectomy and immediate autologous breast reconstruction: A comparative study. Breast Cancer Research and Treatment, 133(2), 537–544. Bellino, S., Fenocchio, M., Zizza, M., Rocca, G., Bogetti, P., & Bogetto, F. (2011). Quality of life of patients who undergo breast reconstruction after mastectomy: Effects of personality characteristics. Plastic and Reconstructive Surgery, 127(1), 10–17. Gurevich, M., Bishop, S., Bower, J., Malka, M., & Nyhof-Young, J. (2004). (Dis)embodying gender and sexuality in testicular cancer. Social Science and Medicine, 58(9), 1597–1607. Park, S. Y., Bae, D. S., Nam, J. H., Park, C. T., Cho, C. H., Lee, J. M., et al. (2007). Quality of life and sexual problems in diseasefree survivors of cervical cancer compared with the general population. Cancer, 110(12), 2716–2725. Pinar, G., Okdem, S., Dogan, N., Buyukgonenc, L., & Ayhan, A. (2012). The effects of hysterectomy on body image, self-esteem, and marital adjustment in Turkish women with gynecologic cancer. Clinical Journal of Oncology Nursing, 16(3), E99–E104. Hopwood, P., Fletcher, I., Lee, A., & Al Ghazal, S. (2001). A Body Image Scale for use with cancer patients. European Journal of Cancer, 37(2), 189–197. Bullinger, M., Alonso, J., Apolone, G., Leplege, A., Sullivan, M., Wood-Dauphinee, S., et al. (1998). Translating health status questionnaires and evaluating their quality: The IQOLA project approach. international quality of life assessment. Journal of Clinical Epidemiology, 51(11), 913–923. Bland, J. M., & Altman, D. G. (1997). Cronbach’s alpha. BMJ, 314(7080), 572. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New Jersey: Lawrence Erlbaum Associates. Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods and data analysis (2nd ed.). New York: McGraw Hill. Jakatdar, T. A., Cash, T. F., & Engle, E. K. (2006). Body-image thought processes: The development and initial validation of the assessment of body-image cognitive distortions. Body Image, 3(4), 325–333. Cash, T. F., & Pruzinsky, T. E. (1990). Body images: Development, deviance, and change. New York: Guilford Press. Cash, T. F., & Pruzinsky, T. E. (2002). Body image: A handbook of theory, research and clinical practice. New York: The Guilford Press. Donabedian, A. (1988). The quality of care. How can it be assessed? JAMA, 260(12), 1743–1748. Damman, O. C., Hendriks, M., & Sixma, H. J. (2009). Towards more patient centred healthcare: A new Consumer quality index instrument to assess patients’ experiences with breast care. European Journal of Cancer, 45(9), 1569–1577.
123
BRIEF COMMUNICATION
Reliability and validity of the Dutch-translated Body Image Scale V. M. T. van Verschuer • W. W. Vrijland I. Mares-Engelberts • T. M. A. L. Klem
•
Accepted: 22 December 2014 Ó Springer International Publishing Switzerland 2015
Abstract Purpose Lacking a comprehensible and widely applicable Dutch test to assess body image changes in cancer patients, we validated Hopwood’s Body Image Scale (BIS) for the Dutch language. Methods The BIS consists of 10 items scored 0–3. Total scores range from 0 (minimum body image-related distress) to 30 (maximum distress). After forward and backward translation of the BIS, we evaluated its psychometric characteristics in breast cancer patients. We assessed feasibility by missing answer rates and positive response prevalence (score C1) per item (criterion C30 %), test– retest reliability with a 2-week interval, internal consistence using Cronbach’s a and discriminant ability by comparing body image after breast-conserving therapy (BCT) versus mastectomy. Results Psychometric evaluation of 108 BCT and 101 mastectomy patients showed high feasibility (0.2 % missing answers), high positive response prevalence of C30 % in 9/10 items and high internal consistency (a [ 0.90). Test–retest reliability and correlation were high with 5.78 (test) versus 5.75 (retest; P = 0.86) and Spearman’s q = 0.92 (P \ 0.01). Discriminant ability was good with BIS scores of 4.56 after BCT versus 7.19 after mastectomy
V. M. T. van Verschuer Department of Surgical Oncology, Erasmus MC Cancer Institute, Rotterdam, The Netherlands V. M. T. van Verschuer W. W. Vrijland I. Mares-Engelberts T. M. A. L. Klem (&) Sint Franciscus Vlietland Group, Department of Surgery, Sint Franciscus Gasthuis, Kleiweg 500, 3045 PM Rotterdam, The Netherlands e-mail: [email protected]
(P \ 0.01). All results were comparable to the results of the original BIS. Conclusion The Dutch-translated BIS showed excellent psychometric results very similar to the original BIS. Its concise and simple design further supports wide application in clinical practice. Keywords Body image Questionnaires Translating Psychometrics Mastectomy Breast-conserving therapy
Introduction Healthcare quality indicators are increasingly determined by patient-reported outcome measures (PROMs). Commonly used PROMs are quality of life (QoL) assessments of which body image is one of the presumed determinants [1, 2]. Several types of cancer and cancer treatments may cause body image changes. The majority of studies assessing body image changes in cancer patients have focused on breast cancer patients [3–6], but other types of cancer are likely to induce body image distress as well [1, 7–9]. Lacking a comprehensible and widely applicable Dutch test to assess body image changes in cancer patients, we validated Hopwood’s Body Image Scale (BIS) [10] for the Dutch language.
Patients and methods Hopwood’s Body Image Scale (BIS) The BIS assesses body image and body image changes after cancer treatment (Table 1). Respondents are asked to answer questions with reference to the past week. The scale
123
Qual Life Res Table 1 Hopwood’s Body Image Scale (BIS) BIS items
Not at all
A little
Quite a bit
Very much
1. Have you been feeling self-conscious about your appearance?
–
–
–
– –
2. Have you felt less physically attractive as a result of your disease or treatment?
–
–
–
3. Have you been dissatisfied with your appearance when dressed?
–
–
–
–
4. Have you been feeling less feminine/masculine as a result of your disease or treatment?
–
–
–
–
5. Did you find it difficult to look at yourself naked?
–
–
–
–
6. Have you been feeling less sexually attractive as a result of your disease or treatment?
–
–
–
–
7. Did you avoid people because of the way you felt about your appearance? 8. Have you been feeling the treatment has left your body less whole?
– –
– –
– –
– –
9. Have you felt dissatisfied with your body?
–
–
–
–
10. Have you been dissatisfied with the appearance of your scar?
–
–
–
–
N/A
N/A not applicable
consists of ten items including affective items (e.g., feeling ‘‘self-conscious,’’ ‘‘less feminine/masculine,’’ ‘‘less physically attractive’’), cognitive items (e.g., dissatisfied ‘‘with appearance,’’ ‘‘with scar’’) and behavioral items (e.g., ‘‘avoid people,’’ ‘‘difficult to look at yourself naked’’). Response options range from ‘‘not at all’’ (score 0), ‘‘a little’’ (score 1), ‘‘quite a bit’’ (score 2) to ‘‘very much’’ (score 3). Question 10 (‘‘dissatisfied with scar’’) has an additional response option ‘‘not applicable.’’ Summing up the scores, a total score ranging from 0 to 30 per patient is obtained with 0 representing no distress or symptoms, whereas increasing scores represent increasing distress and symptoms. Translation and adaptation of the BIS The adaptation process was previously described by Bullinger et al. [11]. Three Dutch native speakers with extensive knowledge of the English language provided a forward translation into Dutch. Emphasis was lying on conceptual equivalence using simple language, rather than achieving literal translation. Translators discussed difficulties with the principal investigator until consensus was reached on one optimal Dutch formulation. Two native English speakers who were fluent in Dutch provided a backward translation to English. Both backward translations were compared with the original BIS, and any differences were analyzed. Finally, necessary adaptations of the Dutch version were made. The resulting version (Table 2) was given to three patients who had been treated for breast cancer. They were asked to comment on readability and comprehension of the scale. After this last test, no adaptations had to be made and the scale was administered to the study population for psychometric data collection.
123
Study population and data collection We chose to conduct psychometric evaluation in breast cancer patients, since this subgroup of patients is relatively large, homogeneous and likely to have body image-related distress due the type of surgery. The fact that the original version was (partly) validated in breast cancer patients easier the comparison [10]. Female breast cancer patients C18 years who visited our outpatient clinic for follow-up were eligible. We aimed to include 100 breast-conserving therapy (BCT) and 100 mastectomy patients. Exclusion criteria were breast reconstruction after mastectomy and insufficient knowledge of the Dutch language. In order to determine test–retest reliability, participants were asked to complete the Dutch BIS twice with a two-week interval. Patients who had to undergo breast cancer treatment between the test and the retest or who did not return the second copy by mail were excluded. Medical ethical board approval was obtained. All participants gave written informed consent. Validation The SPSS computer package (version 20.0) was used for statistical analyses. Psychometric results are compared to the results of the psychometric evaluation of the final version of the original BIS. Feasibility Missing or non-unique responses (0 or [1 box ticked, respectively) were considered invalid. Questionnaires with \9/10 valid responses to all items were excluded from analyses. When one item was not answered, the maximum score possible at that item was subtracted from the maximum achievable score of the scale (30 points). A valid
Qual Life Res Table 2 Dutch translation of the Body Image Scale Body Image Scale vragen
Helemaal niet
Een beetje
Nogal
Heel erg
1. Heeft u zich onzeker gevoeld over uw uiterlijk?
–
–
–
–
2. Heeft u zich lichamelijk minder aantrekkelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
3. Bent u ontevreden geweest over uw uiterlijk als u aangekleed was?
–
–
–
–
4. Heeft u zich minder vrouwelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
5. Heeft u moeite gehad om uzelf naakt te zien?
–
–
–
–
6. Heeft u zichzelf seksueel minder aantrekkelijk gevoeld door uw ziekte of behandeling?
–
–
–
–
7. Heeft u andere mensen vermeden vanwege hoe u zich voelde over uw uiterlijk? 8. Heeft u het gevoel dat de behandeling uw lichaam minder compleet heeft gemaakt?
– –
– –
– –
– –
9. Bent u ontevreden geweest over uw lichaam?
–
–
–
–
10. Bent u ontevreden geweest over hoe uw litteken eruit ziet?
–
–
–
–
n.v.t.
n.v.t. niet van toepassing
score was calculated by dividing the achieved total score by the new maximum score and multiplying this by 30. Feasibility of the scale was evaluated by response rates and missing answers. Response prevalence was defined as the frequency of positive ratings (score of C1) for each item, indicating a change in some aspect of body image. Per item a criterion of C30 % of positive ratings of the total sample was used. Reliability To assess whether items evaluate the same concept (e.g., body image), internal consistency of scale items was measured using Cronbach’s a, which should exceed 0.70 [12]. Test–retest reliability was tested using Spearman’s correlation coefficient rho (q), paired Student’s t tests and effect size (Cohen’s d = 0.2: small effect, d = 0.5: medium effect and d [ 0.8: large effect size) [13, 14]. Clinical validity Discriminant ability between lumpectomy and the mastectomy subgroups was assessed using Student’s t tests and effect size [13, 14].
Results Response rates and feasibility Both questionnaires were returned by 108/150 BCT patients (72 %) and by 101/150 mastectomy patients (67 %; Table 3). In both questionnaires, a total of 0.2 % of all answers was missing. There were no non-unique answers. None of the questionnaires had [1 invalid responses. All items reached the 30 % response prevalence criterion, except for item 7 (‘‘avoid people because of the
way you felt about your appearance?’’) which had a response prevalence of 12 %. In the first version of the original BIS, three items including item 7 had a response prevalence of B30 %, which decreased to zero items in the final version [10]. Test–retest reliability and internal consistency Test–retest reliability was high. Mean BIS scores were 5.78 (95 % CI 4.97–6.59) in Questionnaire 1 versus 5.75 (95 % CI 4.93–6.57) in Questionnaire 2 (P = 0.86). This is in line with the outcomes of the original BIS with test–retest scores of 8.1 and 9.0, respectively, in the breast subgroup [10]. Correlation between the questionnaires was high for total BIS score (q = 0.92; P \ 0.01) as well as for all items (Table 4), comparable to the correlation coefficient of the original BIS (q = 0.70; P \ 0.01) [10]. Test–retest effect size was low (Cohen’s d = 0.005). Both questionnaires showed high internal consistency with Cronbach’s a of 0.91 and 0.92, respectively, similar to the 0.93 of the original BIS [10]. Clinical validity: discriminant ability between breastconserving therapy and mastectomy Mean total BIS scores of both questionnaires were 4.56 (95 % CI 3.49–5.41) in the BCT group and 7.19 (95 % CI 5.91–8.48) in the mastectomy group (P \ 0.01; Table 5) representing good discriminant ability. Cohen’s d was 0.47, representing medium effect size. In the original BIS, this difference was even more pronounced with median total scores of 2.5 after BCT and 12.0 after mastectomy (P \ 0.01) [10].
Discussion The original BIS was developed in 2000 in the UK for use in clinical trials to assess body image changes in cancer
123
Qual Life Res Table 3 Patient characteristics
BCT group
Mastectomy group
Total patients
108
101
Age, years (median, range)
59.0 (32.0–82.0)
60.0 (33.0–92.0)
Follow-up since surgery, years (median, range)
4.0 (1.0–13.0)
4.0 (1.0–20.0)
Chemotherapy
52 (49 %)
65 (65 %)
Radiotherapy
97 (91 %)
39 (39 %)
Hormonal therapy
49 (46 %)
60 (61 %)
History of
BCT Breast-conserving therapy
Table 4 Scores of Body Image Scale (BIS) items, total BIS scores and test–retest reliability
Questionnaire 1—test
Questionnaire 2—retest
Correlation
Mean score (95 % CI)
Mean score (95 % CI)
q
P valuea
1
0.80 (0.69–0.92)
0.72 (0.61–0.83)
0.80
\0.01
2
0.80 (0.68–0.92)
0.79 (0.68–0.91)
0.80
\0.01
3
0.40 (0.30–0.49)
0.39 (0.30–0.47)
0.68
\0.01
4
0.59 (0.47–0.70)
0.51 (0.40–0.61)
0.74
\0.01
5
0.59 (0.47–0.71)
0.65 (0.52–0.77)
0.80
\0.01
6
0.70 (0.58–0.82)
0.75 (0.62–0.87)
0.80
\0.01
7
0.17 (0.10–0.24)
0.17 (0.11–0.24)
0.63
\0.01
8
0.65 (0.54–0.76)
0.63 (0.52–0.75)
0.73
\0.01
9
0.63 (0.53–0.74)
0.67 (0.56–0.78)
0.72
\0.01
10
0.45 (0.34–0.56)
0.47 (0.36–0.58)
0.79
\0.01
P value of Spearman’s correlation coefficient
Total BIS
5.78 (4.97–6.59)
5.75 (4.93–6.57)
0.92
\0.01
Table 5 Discriminative ability of the Body Image Scale (BIS)
Total BIS score of
N
BCT group N = 108 Mean score (95 % CI)
Mastectomy N = 101 Mean score (95 % CI)
95 % CI 95 % confidence interval, q Spearman’s correlation coefficient q a
BCT Breast-conserving therapy
BIS item
All questionnaires
418
4.56 (3.85–5.26)
7.05 (6.16–7.94)
\0.01
Questionnaire 1—test
209
4.45 (3.49–5.41)
7.19 (5.91–8.48)
\0.01
Questionnaire 2—retest
209
4.66 (3.61–5.71)
6.90 (5.65–8.16)
\0.01
patients, resulting from changes in a patient’s appearance due to cancer treatment [10]. The scale assesses affective, behavioral and cognitive changes corresponding with a multidimensional approach of body image [15–17]. We validated the Dutch-translated version of the BIS. Psychometric evaluation of the Dutch BIS showed results comparable to the original version. Internal consistency was high with Cronbach’s a of 0.92 and 0.91, similar to the original BIS [14]. Clinical validity based on response prevalence indicating a change in an aspect of body image also was comparable to the original tool [10]. Only one item, ‘‘avoidance of other people because of the way a patient feels about his or her appearance,’’ failed to reach the response prevalence criterion of C30 % compared with none of the items in the original version [10].
123
P value
This item indeed addresses an issue of body image likely not frequently impacted in breast cancer patients because the breast surgical area is hidden in most social situations. Discriminant ability was moderate to good between breast cancer patients treated by mastectomy and BCT, representing two groups with expected differences in body image changes. BIS scores differed significantly between the groups, but the effect size was only moderate. The discriminant ability between BCT and mastectomy was comparable to the original BIS, although the difference between the two groups was larger in the original tool [10]. A possible explanation for the moderate effect size is that body image distress may not simply be related to the amount of breast tissue visibly removed during surgery. For example, body image distress is also found to be highly
Qual Life Res
impacted after mastectomy followed by direct breast reconstruction [3]. Both the original and the Dutch BIS showed high test–retest reliability [10]. A limitation of this study was the fact that many women did not return the second questionnaire and were excluded, most likely due to unawareness that it was important to complete the same questionnaire twice. Comprehensibility of the scale did not seem to be a problem since rates of missing or non-unique answers were very low. After starting to emphasize on the importance of the second questionnaire at inclusion, response rates improved. Potential future applications of the BIS include its use as a healthcare quality indicator, for example in combination with assessment of QoL or satisfaction with care. Healthcare quality indicators, however, are not yet well defined [18, 19]. To our knowledge, the BIS is not yet being implemented for the assessment of quality of care. Potential applications of interest include screening for body image-related issues after cancer surgery and evaluation of treatment effects after psychological therapy or reconstructive surgery. Before the BIS can be used for evaluation of treatment effects, however, further specific validation is warranted. To conclude, psychometric evaluation of the Dutch BIS showed excellent results that were comparable those of the original version. The concise and simple design makes the Dutch BIS suitable for assessment of body image issues in routine clinical practice.
4.
5.
6.
7.
8.
9.
10.
11.
12. Acknowledgments We thank all participating patients and acknowledge Dr. P. Hopwood and colleagues for the development of the original Body Image Scale.
13. 14.
Conflict of interest None of the authors declare financial or personal relationships that could inappropriately influence their work. 15.
References 16. 1. Taylor-Ford, M., Meyerowitz, B. E., D’Orazio, L. M., Christie, K. M., Gross, M. E., & Agus, D. B. (2013). Body image predicts quality of life in men with prostate cancer. Psychooncology, 22(4), 756–761. 2. Zebrack, B., & Isaacson, S. (2012). Psychosocial care of adolescent and young adult patients with cancer and survivors. Journal of Clinical Oncology, 30(11), 1221–1226. 3. den Heijer, M., Seynaeve, C., Timman, R., Duivenvoorden, H. J., Vanheusden, K., Tilanus-Linthorst, M., et al. (2012). Body image and psychological distress after prophylactic mastectomy and breast reconstruction in genetically predisposed women: A
17.
18. 19.
prospective long-term follow-up study. European Journal of Cancer, 48(9), 1263–1268. Pusic, A. L., Klassen, A. F., Snell, L., Cano, S. J., McCarthy, C., Scott, A., et al. (2012). Measuring and managing patient expectations for breast reconstruction: Impact on quality of life and patient satisfaction. Expert Rev Pharmacoecon Outcomes Res, 12(2), 149–158. Denewer, A., Farouk, O., Kotb, S., Setit, A., Abd El-Khalek, S., & Shetiwy, M. (2012). Quality of life among Egyptian women with breast cancer after sparing mastectomy and immediate autologous breast reconstruction: A comparative study. Breast Cancer Research and Treatment, 133(2), 537–544. Bellino, S., Fenocchio, M., Zizza, M., Rocca, G., Bogetti, P., & Bogetto, F. (2011). Quality of life of patients who undergo breast reconstruction after mastectomy: Effects of personality characteristics. Plastic and Reconstructive Surgery, 127(1), 10–17. Gurevich, M., Bishop, S., Bower, J., Malka, M., & Nyhof-Young, J. (2004). (Dis)embodying gender and sexuality in testicular cancer. Social Science and Medicine, 58(9), 1597–1607. Park, S. Y., Bae, D. S., Nam, J. H., Park, C. T., Cho, C. H., Lee, J. M., et al. (2007). Quality of life and sexual problems in diseasefree survivors of cervical cancer compared with the general population. Cancer, 110(12), 2716–2725. Pinar, G., Okdem, S., Dogan, N., Buyukgonenc, L., & Ayhan, A. (2012). The effects of hysterectomy on body image, self-esteem, and marital adjustment in Turkish women with gynecologic cancer. Clinical Journal of Oncology Nursing, 16(3), E99–E104. Hopwood, P., Fletcher, I., Lee, A., & Al Ghazal, S. (2001). A Body Image Scale for use with cancer patients. European Journal of Cancer, 37(2), 189–197. Bullinger, M., Alonso, J., Apolone, G., Leplege, A., Sullivan, M., Wood-Dauphinee, S., et al. (1998). Translating health status questionnaires and evaluating their quality: The IQOLA project approach. international quality of life assessment. Journal of Clinical Epidemiology, 51(11), 913–923. Bland, J. M., & Altman, D. G. (1997). Cronbach’s alpha. BMJ, 314(7080), 572. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New Jersey: Lawrence Erlbaum Associates. Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods and data analysis (2nd ed.). New York: McGraw Hill. Jakatdar, T. A., Cash, T. F., & Engle, E. K. (2006). Body-image thought processes: The development and initial validation of the assessment of body-image cognitive distortions. Body Image, 3(4), 325–333. Cash, T. F., & Pruzinsky, T. E. (1990). Body images: Development, deviance, and change. New York: Guilford Press. Cash, T. F., & Pruzinsky, T. E. (2002). Body image: A handbook of theory, research and clinical practice. New York: The Guilford Press. Donabedian, A. (1988). The quality of care. How can it be assessed? JAMA, 260(12), 1743–1748. Damman, O. C., Hendriks, M., & Sixma, H. J. (2009). Towards more patient centred healthcare: A new Consumer quality index instrument to assess patients’ experiences with breast care. European Journal of Cancer, 45(9), 1569–1577.
123
