- Email: [email protected]
Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors
European Journal of Oncology Nursing xxx (2015) 1e6
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors Jun-E Liu a, *, Hui-Ying Wang a, Lin Hua b, Jing Chen a, Mu-Lan Wang a, Yi-Ying Li c a
School of Nursing, Capital Medical University, Beijing, 100069, China School of Biomedical Engineering, Capital Medical University, Beijing, 100069, China c Beijing Neurosurgical Institute, Beijing Tiantan Hospital affiliated to Capital Medical University, Beijing, 100050, China b
a b s t r a c t Keywords: Breast cancer Instrument development Posttraumatic growth Psychometric evaluation Construct validity
Purpose: Due to the rapid increase in the number of breast cancer survivors in China, it is important to have a valid instrument to assess their posttraumatic growth. We carried out a study to validate the psychometric testing of the Simplified Chinese Version of the Posttraumatic Growth Inventory (PTGI-SC) in breast cancer survivors. Methods and sample: A convenience sampling method was used to collect data from 1227 breast cancer survivors at eight tertiary hospitals and some anticancer groups in Beijing between April 2010 and April 2012. We tested the item discrimination, content validity, construct validity, and internal consistency of the PTGI-SC. Key results: The difficulties of the items ranged from 0.432 to 0.737, and their discrimination correlation coefficients ranged from 0.324 to 0.721. The content validity index of the inventory was 0.98. Five factors were extracted using exploratory factor analysis, and their cumulative contribution was determined to be 68.3%. The results of the confirmatory factor analysis include c2 =df ¼ 3:912, SRMR ¼ 0.046, RMSEA ¼ 0.055, IFI ¼ 0.932, CFI ¼ 0.932, and Cronbach's a ¼ 0.90. Conclusion: The validity and reliability of the PTGI-SC support its use for evaluating Chinese breast cancer survivors. This reliable and valid inventory can be used in practice to measure PTG in breast cancer survivors and provide information about their psychological adjustment. It can also facilitate further psychological research among Chinese breast cancer survivors. © 2015 Elsevier Ltd. All rights reserved.
Introduction Breast cancer is one of the most commonly diagnosed cancers worldwide. Its incidence has increased globally over the last few decades (Anderson and Jakesz, 2008; Porter, 2008), especially in Asian countries (Green and Raina, 2008). Although China is considered a low-incidence country, its average annual growth rate of breast cancer diagnosis is 3% (Xu, 2010). In developed areas of China, breast cancer has become the most common malignant tumor in women. Breast cancer diagnosis and treatment have various physical and psychological impacts on women, including fatigue, sexual
* Corresponding author. School of Nursing, Capital Medical University, You An Men, Beijing, 100069, China. Tel.: þ86 10 83911771 (office); fax: þ86 10 83911641. E-mail addresses: [email protected] (J.-E. Liu), [email protected] (H.-Y. Wang), [email protected] (L. Hua), [email protected] (J. Chen), [email protected] (M.-L. Wang), [email protected] (Y.-Y. Li).
disorders, anxiety, depression, potential feelings of social isolation, and fear of cancer recurrence (Cordova and Andrykowski, 2003; Harrington et al., 2010; Wang, 2011). Tedeschi and Calhoun (1996) reported positive changes subsequent to stressful events or crises and coined the most widely used term e “posttraumatic growth (PTG) ” e which is considered to be the “positive psychological change experienced as a result of the struggle with highly challenging life circumstances” (Calhoun and Tedeschi, 1999). To quantify PTG, Tedeschi and Calhoun (1996) developed the posttraumatic growth inventory (PTGI), which consists of five factors that are widely accepted to be personal strength, new possibilities, relating to others, appreciation of life, and spiritual change. It was reported that PTG was common among breast cancer survivors, whose PTGI scores have been reported to range from 47 to 73 (Manne et al., 2004; Sears et al., 2003; Weiss, 2004). As one of the most popular quantitative measurements of PTG, the PTGI has been examined with regard to its factor structure in many different people (Ho et al., 2004; Jaarsma et al., 2006; Joseph
http://dx.doi.org/10.1016/j.ejon.2015.01.002 1462-3889/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
2
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
et al., 2004; Levine et al., 2008; Powell et al., 2003; Sears et al., 2003; Taku et al., 2008). Levine et al. (2008) recommended a two-factor model based on their investigation of 4054 Israeli adolescents who had been exposed to terror. Powell et al. (2003) employed a three-factor model to assess former refugees and displaced individuals, which consisted of “changes in perception of self,” “changes in interpersonal relationships,” and “changes in philosophy of life.” Over the last few years, some authors (Linley et al., 2007; Taku et al., 2008) have reported results that support the use of a five-factor model in people who have experienced a range of adverse life events and consider a PTGI to be a meaningful way to understand PTG. In 2004, one study (Ho et al., 2004) translated the English version of the PTGI into traditional Chinese (PTGI-C) and studied a four-factor PTGI model in adult cancer survivors in Hong Kong. We believe that there are only two articles relating to a psychometric evaluation of the PTGI in breast cancer survivors. One of these studies was conducted in the United States by Sears et al. (2003), who used the PTGI in early-stage breast cancer survivors and explored a single factor for the scale; the other was conducted in Canada (Brunet et al., 2010) and assessed five factors. To date, most of the studies conducted on PTGI factor structure have assessed samples gathered from survivors following a variety of traumatic events (Ho et al., 2004; Jaarsma et al., 2006; Levine et al., 2008; Linley et al., 2007; Taku et al., 2008). This limits the interpretation of these findings to particular groups of people. Despite the prevalence of breast cancer worldwide, few studies in the international literature (Brunet et al., 2010; Sears et al., 2003) have examined PTGI factor structure when assessing breast cancer survivors, and there are no existing studies on PTGI factor structure and breast cancer survivors in China. This study attempted to determine the psychometric characteristics, especially the factor structure, of Chinese breast cancer survivors using the simplified Chinese version of the PTGI (PTGI-SC) to provide a valid instrument to assess PTG in breast cancer survivors on the Chinese mainland.
items), new possibilities (five items), personal strength (four items), appreciation of life (three items), and spiritual change (two items). The answers are rated from 0 to 5 (where 0 indicates “I did not experience this change as a result of my crisis” and 5 indicates “I experienced this change to a very great degree as a result of my crisis”). In the study by Tedeschi and Calhoun (1996), the PTGI demonstrated both good internal reliability (a ¼ 0.90) and testeretest reliability (0.71) over a 2-month period. A simplified Chinese version of the posttraumatic growth inventory (PTGI-SC) was translated, modified, and validated. This was based on the original English version developed in 1996 by Tedeschi and Calhoun and the Hong Kong Chinese version (PTGI-C) translated in 2004 by Ho. Two bilingual nursing experts independently translated and back-translated the original English PTGI to develop the original Chinese mainland version, which was then compared with the Hong Kong (HK) version; items that were consistent with the HK version were retained, while the inconsistent items were modified. Next, the PTGI-SC was preliminarily evaluated by seven breast cancer survivors with differing educational levels and three nursing experts in a tertiary Beijing hospital. They identified questions that were not clearly expressed or that did not fit with Chinese idiomatic expressions. Finally, 21 items from the original English and HK versions of the PTGI-SC were left unchanged from the original English and HK versions, while 4 items were modified (items 1, 3, 5, and 21) from the PTGI-C (Liu et al., 2014). Five experts (including one scale development expert and four expert cancer care nurses) were invited to assess the content validity of the PTGI-SC. They were asked to rate how adequately the items matched the PTG domain using the following four-point scale: (1) irrelevant, (2) somewhat relevant, (3) very relevant (relevant but needs minor alteration), or (4) very relevant and succinct. We used the content validity index (CVI) (Lynn, 1986) to determine the content validity of the PTGI-SC. The CVI was computed by summing the percentage of agreement between all items that were given a rating of 3 or 4 by the experts. Our calculations indicated that the CVI of this scale was 0.980.
Methods Statistical analyses Sample We used a descriptive research design with convenience sampling to collect the data. The inclusive criteria for breast cancer survivors were: (1) 18 years of age, (2) no prior psychiatric history, (3) confirmed histopathological diagnosis of breast cancer, and (4) had undergone breast cancer surgery. Procedure The study was conducted at the breast cancer departments of eight tertiary hospitals and some anticancer groups in Beijing, China, between April 2010 and April 2012. All investigators received communication skills training and were instructed how to administer the scale. In total, 1253 breast cancer survivors were enrolled. After providing informed consent, all participants answered the written questionnaire and returned it to the hospital, whereupon researchers checked the completeness of the questionnaires. Twenty-six incomplete datasets were excluded from the analysis. Subsequently, 1227 patients were included, to give an overall response rate of 97.9%. Instrument The version of the PTGI, developed by Tedeschi and Calhoun, consists of 21 items and five factors: relating to others (seven
Data were input using EpiData 3.0 software. After systematic logic error detection, the database was imported into an SPSS 16.0 software system. First, item analysis was used to confirm discrimination between each item. Generally, we consider that the difficulty of each item should have a value close to 0.5, which indicates that it is a more reliable and distinct item. Lord reported that additional choices lower the true difficulty values for each item (Lord, 1952). The PTGI-SC, which has six choices for each item, should have a true difficulty value of 0.5e0.7. Next, the factor structure of the PTGI-SC was determined using exploratory factor analysis. We used analysis of moment structures (AMOS, version 17.0) to conduct the confirmatory factor analysis (CFA, i.e., maximum likelihood) to verify the fitness of the hypothesized model and the data of the 1227 breast cancer survivors. The model's goodness-of-fit was assessed using c2/degrees of freedom (df), standardized root mean squared residual (SRMR), root mean square error of approximation (RMSEA), goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), comparative fit index (CFI), TuckereLewis index (TLI), normed fit index (NFI), and incremental fit index (IFI). The ideal value of c2/df is < 2 (Kit-Tai et al., 2004); however, this value is sensitive to the sample size (Marsh and Balla, 1988). Therefore, an c2/df value between 2 and 5 is generally €reskog and So €rbom, 2006). SRMR values 0.08 and acceptable (Jo RMSEA values 0.06 (Hu and Bentler, 1999; Thompson, 2004) generally indicate reasonable model fit. GFI, AGFI, NFI, IFI, TLI, and
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
CFI values are acceptable if they are >0.90 (Byrne, 1994; Kit-Tai et al., 2004). Finally, the Cronbach a coefficient of each PTGI factor was also obtained. Ethical considerations The University of Institutional Review Board approved this study (code number 2010 SY24), and verbal informed consent was obtained from each participant.
3
Table 2 PTGI-SC inter-factor correlation and factoretotal correlation (n ¼ 840).
1 2 3 4 5
PTGI-SC total score
1
2
3
4
5
0.803** 0.763** 0.736** 0.647** 0.652**
e 0.376** 0.495** 0.319** 0.351**
e 0.464** 0.429** 0.544**
e 0.459** 0.460**
e 0.398**
e
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, ** p < 0.01 (2-tailed); 1 ¼ Relating to others, 2 ¼ New possibilities, 3 ¼ Personal strength, 4 ¼ Appreciation of life, 5 ¼ Spiritual change.
Results two groups at a 2:1 ratio, using SPSS software (840 and 387 participants in samples 1 and 2, respectively).
Demographic data The participants were aged between 26 and 79 years, with an average age of 53.30 ± 8.05 years. The time since the diagnosis of breast cancer ranged from 4 months to 32 years, with a median of 3.5 years. The majority of patients reported no religious beliefs (90.4%). Among 1227 survivors, 324 (26.4%), 163 (13.3%), and 632 (51.5%) were diagnosed as TNM stage I, II, and III, respectively. With regard to surgery, 82.6% underwent mastectomy and the remaining 17.4% underwent breast-conserving surgery. Item analyses The item analyses consisted of determining the difficulty and discrimination of each item. The formula for determining difficulty is p ¼ x=xmax (p ¼ difficulty; x ¼ average score of each item; xmax ¼ maximum score of each item). The difficulty values ranged between 0.432 and 0.737. Discrimination was calculated using item-total Pearson correlation coefficients, which ranged between 0.324 and 0.721, all with p < 0.01 (Table 1).
Exploratory factor analysis The Kaiser-Meyer-Olkin (KMO) test for measuring sampling adequacy and Bartlett's test of sphericity were used to assess the adequacies of the correlation matrices used in the factor analysis. The KMO result for sample 1 was 0.899, which is considered “good” (Hutcheson and Sofroniou, 1999) and indicates that reliable and distinct factors can be extracted. Furthermore, Bartlett's test of sphericity determined that c2 ¼ 4348.826 and p < 0.01, which indicated that the correlation matrices were not identity matrices; hence, the factor model was appropriate (Snedecor and Cochran, 1989). The initial principal-axis factor analysis of sample 1 provided five factors with eigenvalues >1, which collectively accounted for Table 3 PTGI-SC exploratory factor analysis (n ¼ 840). Item
Construct validity The data for exploratory factor analysis and confirmatory factor analysis should be two samples from the same population (Hurley et al., 1997). Therefore, the samples were randomly divided into
Table 1 Difficulty and discrimination of each PTGI-SC item (n ¼ 1227). Item
p
r
1. My priorities about what is important in life. 2. I'm more likely to try to change things that need changing. 3. An appreciation for the value of my own life. 4. A feeling of self-reliance. 5. A better understanding of spiritual matters. 6. Knowing that I can count on people in times of trouble. 7. A sense of closeness with others. 8. Knowing I can handle difficulties. 9. A willingness to express my emotions. 10. Being able to accept the way things work out. 11. Appreciating each day. 12. Having compassion for others. 13. I'm able to do better things with my life. 14. New opportunities are available that wouldn't have been otherwise. 15. Putting effort into my relationship. 16. I have a stronger religious faith. 17. I discovered that I'm stronger than I thought I was. 18. I learned a great deal about how wonderful people are. 19. I developed new interests. 20. I accept needing others. 21. I established a new path for my life.
0.632 0.686
0.453** 0.468**
0.700 0.692 0.693 0.637 0.687 0.729 0.685 0.699 0.722 0.737 0.732 0.588
0.425** 0.511** 0.611** 0.617** 0.721** 0.601** 0.713** 0.527** 0.605** 0.678** 0.616** 0.575**
0.703 0.432 0.736 0.708 0.596 0.661 0.631
0.668** 0.324** 0.584** 0.712** 0.565** 0.645** 0.608**
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, ** p < 0.01.
Relating to others 6 7 9 12 15 18 20 New possibilities 2 13 14 19 21 Personal strength 4 8 10 17 Appreciation of life 1 3 11 Spiritual change 5 16 Eigenvalues Percentage of variance explained Cumulative % of variance explained
Factor loading Factor 1
Factor 2
Factor 3
Factor 4
Factor 5
0.827 0.856 0.828 0.799 0.842 0.815 0.792
0.029 0.141 0.221 0.190 0.072 0.238 0.106
0.107 0.114 0.164 0.242 0.204 0.149 0.124
0.114 0.148 0.040 0.031 0.035 0.028 0.083
0.004 0.001 0.034 0.088 0.032 0.002 0.184
0.125 0.188 0.150 0.094 0.178
0.683 0.786 0.739 0.716 0.792
0.038 0.273 0.117 0.155 0.022
0.221 0.089 0.161 0.017 0.145
0.147 0.107 0.088 0.326 0.188
0.227 0.219 0.197 0.154
0.010 0.369 0.110 0.391
0.821 0.570 0.791 0.573
0.219 0.136 0.083 0.085
0.047 0.114 0.106 0.278
0.010 0.110 0.152
0.196 0.099 0.237
0.037 0.291 0.313
0.830 0.570 0.729
0.274 0.203 0.039
0.279 0.011 5.128 36.821
0.259 0.160 3.519 13.456
0.290 0.058 2.633 7.968
0.019 0.075 1.866 5.145
0.574 0.832 1.203 4.941
36.821
50.277
58.245
63.389
68.331
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory. Extraction method: principal-axis analysis. Rotation method: varimax with Kaiser normalization. The bold indicates the factor loadings are more than 0.5.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
4
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
Table 4 Goodness-of-fit indices for PTGI-SC (n ¼ 387).
Table 5 Internal consistency of PTGI-SC (n ¼ 1227).
Model
c2/df
SRMR
RMSEA
GFI
AGFI
NFI
IFI
TLI
CFI
Subscale
Number of items
Cronbach's a
5-factor
3.912
0.046
0.055
0.967
0.913
0.942
0.932
0.945
0.932
Relating to others New possibilities Personal strength Appreciation of life Spiritual change Total PTGI-SC
7 5 4 3 2 21
0.93 0.85 0.73 0.65 0.60 0.90
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, SRMR: standardized root mean squared residual, RMSEA: root mean square error of approximation, GFI: goodness-of-fit index, AGFI: adjusted goodness-of-fit index, NFI: normed fit index, IFI: incremental fit index, TLI: TuckereLewis index, CFI: comparative fit index.
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory.
68.3% of the variance. After appraising the scree plot (Cattell, 1978), we decided to retain all five factors. The correlations of the five factors with the total PTGI-SC score ranged from 0.647 to 0.803, and the inter-factor correlations ranged from 0.319 to 0.544 (Table 2). This indicated that the five factors estimate common aspects, but that each factor also has a unique contribution. We found that varimax-rotated factor matrix of the five-factor PTGI-SC explained 68.3% of the variance (Table 3). The factors were identical to the original English version, but were inconsistent with the results of the four-factor traditional Chinese version.
Reliability The internal consistency of Cronbach's a coefficients for the total scale was 0.90, which is considered high according to DeVellis (0.80 is considered high) (DeVellis, 2002). The subscales demonstrated high Cronbach's a coefficients for “relating to others” and “new possibilities,” good values (0.70e0.80) for “personal strength,” and acceptable values (0.65e0.70) for “appreciation of life” (Table 5). However, “spiritual change” demonstrated a low internal consistency of 0.602, which means that this subscale should be modified in some way (DeVellis, 2002).
Confirmatory factor analysis Discussion The preliminary analysis of the five-factor PTGI-SC using sample 2 demonstrated no negative error variance, and all of the standardized regression weights were <1 (range 0.29e0.83). This signified that there is no improper solution to the model (Hair et al., 1998), and we can test the model's overall fitness (Table 4). The loading values were 0.65e0.83, 0.52e0.80, 0.51e0.69, 0.42e0.76, and 0.38e0.63 for “relating to others,” “new possibilities,” “personal strength,” “appreciation of life,” and “spiritual change,” respectively. Correlations between the five subscales ranged from 0.29 to 0.69 (Fig. 1).
This study reported the psychometric properties of Chinese breast cancer survivors according to the PTGI-SC results. With regard to the item analysis, some inconsistencies were found between difficulty and discrimination. The results showed that all of the items demonstrated a true difficulty except “I have a stronger religious faith.” We believe that this is likely due to the lack of religious populations in China compared with other countries (Ruan et al., 2011). In this study, 90.4% of patients reported no religious beliefs. That might have influenced the lower score for this
Fig. 1. Structural equation model of the PTGI-SC.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
item, which would also reduce the difficulty. However, the itemtotal correlation coefficients were 0.324e0.721, which meets the criterion (0.3e0.8) set by Tucker (1963). Therefore, we consider that the item “I have a stronger religious faith” still has acceptable and differentiated ability. We also tested the content and construct validity of the PTGI-SC. Content validity is the degree to which the items of an instrument adequately represent the targeted construct for a particular assessment purpose. In the translation and back-translation periods, most of the items in the simplified Chinese version were appropriately translated, and the conceptual meaning of the items was deemed by expert review (Bracken and Barona, 1991) to be similar to that of the original version. Next, the CVI was applied to the PTGI-SC. Different evaluators may hold different opinions about whether the items reflect the scale domains. To reduce the subjective influence of the evaluator, we asked five experts to score each item. A CVI value of 0.980 indicated that the item was correctly loaded in the proper domain. Construct validity indicates whether a measure relates to other observed variables in a way that is consistent with theoretically derived predictions (Bollen, 1989). The factor structure of the PTGI-SC according to the exploratory factor analysis was comparable to the original English version, and CFA was used to verify this result. All of the GFI, AGFI, NFI, IFI, TLI, and CFI values were >0.90. The SRMR and RMSEA values indicated a sound model fit, while the correlation matrix indicated moderate correlation among items. Therefore, the structural equation model fits the data of the 1227 breast cancer survivors moderately well, signifying that the PTGI-SC structure is reasonable. The total PTGI-SC demonstrated good internal consistency as reflected by the Cronbach's a values. However, the factors “appreciation of life” and “spiritual change” were responsible for the relatively low a coefficients. It is known that the strength of a depends on the length of the subscale. It is best to use Cronbach's a when evaluating subscales with 3 items (McDowell and Newell, 2006). Thus, the number of items in the “appreciation of life” and “spiritual change” subscales may explain the low a coefficients. A limitation of this study is the convenience sampling method, which may have reduced the representativeness of the sample. However, the large population may have reduced the selection bias of the study, and validation is a continuous process. Future research should focus on the relationship between the PTGI-SC and alternative measures. Conclusions The present findings regarding validity and reliability support its use for evaluating breast cancer survivors in China. The PTGI-SC is a reliable and valid inventory that can be used in practice to measure PTG levels in breast cancer survivors to provide information about their psychological adjustment. These findings will facilitate further psychological research in Chinese breast cancer survivors. Conflict of interest None declared. Acknowledgments We are grateful to nurses in the 8 first-class hospitals for their assistance of data collection and to all breast cancer survivors who took part in the study. This study was funded by National Natural Science Foundation of China (grant no. 30870770), and the Importation and Development of High-Caliber Talents Project of
5
Beijing Municipal Institutions (grant no. CIT&TCD20140328). The authors declare no conflicts of interest. References Anderson, B.O., Jakesz, R., 2008. Breast cancer issues in developing countries: an overview of the Breast Health Global Initiative. World Journal of Surgery 32 (12), 2578e2585. http://dx.doi.org/10.1007/s00268-007-9454-z. Bollen, K., 1989. Structural Equations with Latent Variables. John Wiley and Sons, New York. Bracken, B.A., Barona, A., 1991. State of the art procedures for translating, validating, and using psychoeducational tests in cross-cultural assessment. School Psychology International 12, 119e132. Brunet, J., McDonough, M.H., Hadd, V., Crocker, P.R., Sabiston, C.M., 2010. The posttraumatic growth inventory: an examination of the factor structure and invariance among breast cancer survivors. Psycho-Oncology 19 (8), 830e838. http://dx.doi.org/10.1002/pon.1640. Byrne, B.M., 1994. Structural Equation Modeling with EQS and EQS/Windows. Sage Publications, Thousand Oaks, CA. Calhoun, L.G., Tedeschi, R.G., 1999. Facilitating Posttraumatic Growth: a Clinician's Guide. Lawrence Erlbaum, Mahwah. Cattell, R.B., 1978. The Scientific Use of Factor Analysis in Behavioral and Life Sciences. Plenum Press, New York. Cordova, M.J., Andrykowski, M.A., 2003. Responses to cancer diagnosis and treatment: posttraumatic stress and posttraumatic growth. Seminars in Clinical Neuropsychiatry 8 (4), 286e296. DeVellis, R.F., 2002. Scale Development Theory and Applications, second ed. SAGE, London. Green, M., Raina, V., 2008. Epidemiology, screening and diagnosis of breast cancer in the Asia-Pacific region: current perspectives and important considerations. Asia-Pacific Journal of Clinical Oncology 4, S5eS13. http://dx.doi.org/10.1111/ j.1743-7563.2008.00191.x. Hair, J.F., Anderson, R.E., Tatham, R.L., Black, W.C., 1998. Multivariate Data Analysis with Readings. Maxwell Macmillan International, New York. Harrington, C.B., Hansen, J.A., Moskowitz, M., Todd, B.L., Feuerstein, M., 2010. It's not over when it's over: long-term symptoms in cancer survivorsda systematic review. International Journal of Psychiatry in Medicine 40 (2), 163e181. Ho, S.M., Chan, C.L., Ho, R.T., 2004. Posttraumatic growth in Chinese cancer survivors. Psycho-Oncology 13 (6), 377e389. http://dx.doi.org/10.1002/pon.758. Hu, L.T., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling 6, 1e55. Hurley, A.E., Scandura, T.A., Schriesheim, C.A., Brannick, M.T., Seers, A., Vandenberg, R.J., et al., 1997. Exploratory and confirmatory factor analysis: guidelines, issues, and alternatives. Journal of Organizational Behavior 18, 667e683. Hutcheson, G., Sofroniou, N., 1999. The Multivariate Social Scientist: Introductory Statistics Using Generalized Linear Models. Sage Publications, Thousand Oaks, CA. €reskog, K.G., So € rbom, D., 2006. LISREL 8.80. Science Software International, Chicago. Jo Jaarsma, T.A., Pool, G., Sanderman, R., Ranchor, A.V., 2006. Psychometric properties of the Dutch version of the posttraumatic growth inventory among cancer patients. Psycho-Oncology 15 (10), 911e920. http://dx.doi.org/10.1002/ pon.1026. Joseph, S., Linley, P.A., Harris, G.J., 2004. Understanding positive change following trauma and adversity: structural clarification. Journal of Loss and Trauma 10 (1), 83e96. http://dx.doi.org/10.1080/15325020490890741. Kit-Tai, H., Wen, Z.L., Cheng, Z.J., 2004. Structural Equation Model and its Applications. Educational Science Publishing House, Beijing. Levine, S.Z., Laufer, A., Hamama-Raz, Y., Stein, E., Solomon, Z., 2008. Posttraumatic growth in adolescence: examining its components and relationship with PTSD. Journal of Traumatic Stress 21 (5), 492e496. http://dx.doi.org/10.1002/jts.20361. Linley, P.A., Andrews, L., Joseph, S., 2007. Confirmatory factor analysis of the posttraumatic growth inventory. Journal of Loss and Trauma 12 (4), 321e332. http:// dx.doi.org/10.1080/15325020601162823. Liu, J.E., Wang, H.Y., Wang, M.L., Su, Y.L., Wang, P.L., 2014. Posttraumatic growth and psychological distress in Chinese early-stage breast cancer survivors: a longitudinal study. Psycho-Oncology 23 (4), 437e443. http://dx.doi.org/10.1002/ pon.3436. Lord, F.M., 1952. The relationship of the reliability of multiple-choice tests to the distribution of item difficulties. Psychometrika 18, 181e194. Lynn, M.R., 1986. Determination and quantification of content validity. Nursing Research 35 (6), 382e385. Manne, S., Ostroff, J., Winkel, G., Goldstein, L., Fox, K., Grana, G., 2004. Posttraumatic growth after breast cancer: patient, partner, and couple perspectives. Psychosomatic Medicine 66 (3), 442e454. Marsh, H.W., Balla, J.R., 1988. Goodness-of-fit indexes in confirmatory factor analysis: the effect of sample size. Psychological Bulletin 103 (3), 391e410. McDowell, I., Newell, C., 2006. Measuring Health: a Guide to Rating Scales and Questionnaires, third ed. Oxford University Press, Oxford. Porter, P., 2008. “Westernizing” women's risks? Breast cancer in lower-income countries. New England Journal of Medicine 358 (3), 213e216. http:// dx.doi.org/10.1056/NEJMp0708307.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
6
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
Powell, S., Rosner, R., Butollo, W., Tedeschi, R.G., Calhoun, L.G., 2003. Posttraumatic growth after war: a study with former refugees and displaced people in Sarajevo. Journal of Clinical Psychology 59 (1), 71e83. http://dx.doi.org/10.1002/jclp.10117. Ruan, R.P., Zheng, F.T., Liu, L., 2011. Religious faith, religious participation and subjective benefits. China Rural Observation 2, 74e86. Sears, S.R., Stanton, A.L., Danoff-Burg, S., 2003. The yellow brick road and the emerald city: benefit finding, positive reappraisal coping and posttraumatic growth in women with early-stage breast cancer. Health Psychology 22 (5), 487e497. http://dx.doi.org/10.1037/0278-6133.22.5.487. Snedecor, G.W., Cochran, W.G., 1989. Statistical Methods, eighth ed. Iowa State University Press, Ames, IA. Taku, K., Cann, A., Calhoun, L.G., Tedeschi, R.G., 2008. The factor structure of the posttraumatic growth inventory: a comparison of five models using confirmatory factor analysis. Journal of Traumatic Stress 21 (2), 158e164. http:// dx.doi.org/10.1002/jts.20305.
Tedeschi, R.G., Calhoun, L.G., 1996. The posttraumatic growth inventory: measuring the positive legacy of trauma. Journal of Traumatic Stress 9 (3), 455e471. Thompson, B., 2004. Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications. American Psychological Association, Washington DC. Tucker, L.R., 1963. Formal Models for a Central Prediction System (Psychometric Monograph). William Byrd Press, Richmond, VA. Wang, H.Y., 2011. Investigation and Analysis on the Level and Associated Factors of Posttraumatic Growth in Breast Cancer Patients. Thesis. Capital Medical University, Beijing. Weiss, T., 2004. Correlates of posttraumatic growth in husbands of breast cancer survivors. Psycho-Oncology 13 (4), 260e268. Xu, G., 2010. The screening of breast cancer. Chinese Medicine Herald 7 (2), 4e6.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors Jun-E Liu a, *, Hui-Ying Wang a, Lin Hua b, Jing Chen a, Mu-Lan Wang a, Yi-Ying Li c a
School of Nursing, Capital Medical University, Beijing, 100069, China School of Biomedical Engineering, Capital Medical University, Beijing, 100069, China c Beijing Neurosurgical Institute, Beijing Tiantan Hospital affiliated to Capital Medical University, Beijing, 100050, China b
a b s t r a c t Keywords: Breast cancer Instrument development Posttraumatic growth Psychometric evaluation Construct validity
Purpose: Due to the rapid increase in the number of breast cancer survivors in China, it is important to have a valid instrument to assess their posttraumatic growth. We carried out a study to validate the psychometric testing of the Simplified Chinese Version of the Posttraumatic Growth Inventory (PTGI-SC) in breast cancer survivors. Methods and sample: A convenience sampling method was used to collect data from 1227 breast cancer survivors at eight tertiary hospitals and some anticancer groups in Beijing between April 2010 and April 2012. We tested the item discrimination, content validity, construct validity, and internal consistency of the PTGI-SC. Key results: The difficulties of the items ranged from 0.432 to 0.737, and their discrimination correlation coefficients ranged from 0.324 to 0.721. The content validity index of the inventory was 0.98. Five factors were extracted using exploratory factor analysis, and their cumulative contribution was determined to be 68.3%. The results of the confirmatory factor analysis include c2 =df ¼ 3:912, SRMR ¼ 0.046, RMSEA ¼ 0.055, IFI ¼ 0.932, CFI ¼ 0.932, and Cronbach's a ¼ 0.90. Conclusion: The validity and reliability of the PTGI-SC support its use for evaluating Chinese breast cancer survivors. This reliable and valid inventory can be used in practice to measure PTG in breast cancer survivors and provide information about their psychological adjustment. It can also facilitate further psychological research among Chinese breast cancer survivors. © 2015 Elsevier Ltd. All rights reserved.
Introduction Breast cancer is one of the most commonly diagnosed cancers worldwide. Its incidence has increased globally over the last few decades (Anderson and Jakesz, 2008; Porter, 2008), especially in Asian countries (Green and Raina, 2008). Although China is considered a low-incidence country, its average annual growth rate of breast cancer diagnosis is 3% (Xu, 2010). In developed areas of China, breast cancer has become the most common malignant tumor in women. Breast cancer diagnosis and treatment have various physical and psychological impacts on women, including fatigue, sexual
* Corresponding author. School of Nursing, Capital Medical University, You An Men, Beijing, 100069, China. Tel.: þ86 10 83911771 (office); fax: þ86 10 83911641. E-mail addresses: [email protected] (J.-E. Liu), [email protected] (H.-Y. Wang), [email protected] (L. Hua), [email protected] (J. Chen), [email protected] (M.-L. Wang), [email protected] (Y.-Y. Li).
disorders, anxiety, depression, potential feelings of social isolation, and fear of cancer recurrence (Cordova and Andrykowski, 2003; Harrington et al., 2010; Wang, 2011). Tedeschi and Calhoun (1996) reported positive changes subsequent to stressful events or crises and coined the most widely used term e “posttraumatic growth (PTG) ” e which is considered to be the “positive psychological change experienced as a result of the struggle with highly challenging life circumstances” (Calhoun and Tedeschi, 1999). To quantify PTG, Tedeschi and Calhoun (1996) developed the posttraumatic growth inventory (PTGI), which consists of five factors that are widely accepted to be personal strength, new possibilities, relating to others, appreciation of life, and spiritual change. It was reported that PTG was common among breast cancer survivors, whose PTGI scores have been reported to range from 47 to 73 (Manne et al., 2004; Sears et al., 2003; Weiss, 2004). As one of the most popular quantitative measurements of PTG, the PTGI has been examined with regard to its factor structure in many different people (Ho et al., 2004; Jaarsma et al., 2006; Joseph
http://dx.doi.org/10.1016/j.ejon.2015.01.002 1462-3889/© 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
2
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
et al., 2004; Levine et al., 2008; Powell et al., 2003; Sears et al., 2003; Taku et al., 2008). Levine et al. (2008) recommended a two-factor model based on their investigation of 4054 Israeli adolescents who had been exposed to terror. Powell et al. (2003) employed a three-factor model to assess former refugees and displaced individuals, which consisted of “changes in perception of self,” “changes in interpersonal relationships,” and “changes in philosophy of life.” Over the last few years, some authors (Linley et al., 2007; Taku et al., 2008) have reported results that support the use of a five-factor model in people who have experienced a range of adverse life events and consider a PTGI to be a meaningful way to understand PTG. In 2004, one study (Ho et al., 2004) translated the English version of the PTGI into traditional Chinese (PTGI-C) and studied a four-factor PTGI model in adult cancer survivors in Hong Kong. We believe that there are only two articles relating to a psychometric evaluation of the PTGI in breast cancer survivors. One of these studies was conducted in the United States by Sears et al. (2003), who used the PTGI in early-stage breast cancer survivors and explored a single factor for the scale; the other was conducted in Canada (Brunet et al., 2010) and assessed five factors. To date, most of the studies conducted on PTGI factor structure have assessed samples gathered from survivors following a variety of traumatic events (Ho et al., 2004; Jaarsma et al., 2006; Levine et al., 2008; Linley et al., 2007; Taku et al., 2008). This limits the interpretation of these findings to particular groups of people. Despite the prevalence of breast cancer worldwide, few studies in the international literature (Brunet et al., 2010; Sears et al., 2003) have examined PTGI factor structure when assessing breast cancer survivors, and there are no existing studies on PTGI factor structure and breast cancer survivors in China. This study attempted to determine the psychometric characteristics, especially the factor structure, of Chinese breast cancer survivors using the simplified Chinese version of the PTGI (PTGI-SC) to provide a valid instrument to assess PTG in breast cancer survivors on the Chinese mainland.
items), new possibilities (five items), personal strength (four items), appreciation of life (three items), and spiritual change (two items). The answers are rated from 0 to 5 (where 0 indicates “I did not experience this change as a result of my crisis” and 5 indicates “I experienced this change to a very great degree as a result of my crisis”). In the study by Tedeschi and Calhoun (1996), the PTGI demonstrated both good internal reliability (a ¼ 0.90) and testeretest reliability (0.71) over a 2-month period. A simplified Chinese version of the posttraumatic growth inventory (PTGI-SC) was translated, modified, and validated. This was based on the original English version developed in 1996 by Tedeschi and Calhoun and the Hong Kong Chinese version (PTGI-C) translated in 2004 by Ho. Two bilingual nursing experts independently translated and back-translated the original English PTGI to develop the original Chinese mainland version, which was then compared with the Hong Kong (HK) version; items that were consistent with the HK version were retained, while the inconsistent items were modified. Next, the PTGI-SC was preliminarily evaluated by seven breast cancer survivors with differing educational levels and three nursing experts in a tertiary Beijing hospital. They identified questions that were not clearly expressed or that did not fit with Chinese idiomatic expressions. Finally, 21 items from the original English and HK versions of the PTGI-SC were left unchanged from the original English and HK versions, while 4 items were modified (items 1, 3, 5, and 21) from the PTGI-C (Liu et al., 2014). Five experts (including one scale development expert and four expert cancer care nurses) were invited to assess the content validity of the PTGI-SC. They were asked to rate how adequately the items matched the PTG domain using the following four-point scale: (1) irrelevant, (2) somewhat relevant, (3) very relevant (relevant but needs minor alteration), or (4) very relevant and succinct. We used the content validity index (CVI) (Lynn, 1986) to determine the content validity of the PTGI-SC. The CVI was computed by summing the percentage of agreement between all items that were given a rating of 3 or 4 by the experts. Our calculations indicated that the CVI of this scale was 0.980.
Methods Statistical analyses Sample We used a descriptive research design with convenience sampling to collect the data. The inclusive criteria for breast cancer survivors were: (1) 18 years of age, (2) no prior psychiatric history, (3) confirmed histopathological diagnosis of breast cancer, and (4) had undergone breast cancer surgery. Procedure The study was conducted at the breast cancer departments of eight tertiary hospitals and some anticancer groups in Beijing, China, between April 2010 and April 2012. All investigators received communication skills training and were instructed how to administer the scale. In total, 1253 breast cancer survivors were enrolled. After providing informed consent, all participants answered the written questionnaire and returned it to the hospital, whereupon researchers checked the completeness of the questionnaires. Twenty-six incomplete datasets were excluded from the analysis. Subsequently, 1227 patients were included, to give an overall response rate of 97.9%. Instrument The version of the PTGI, developed by Tedeschi and Calhoun, consists of 21 items and five factors: relating to others (seven
Data were input using EpiData 3.0 software. After systematic logic error detection, the database was imported into an SPSS 16.0 software system. First, item analysis was used to confirm discrimination between each item. Generally, we consider that the difficulty of each item should have a value close to 0.5, which indicates that it is a more reliable and distinct item. Lord reported that additional choices lower the true difficulty values for each item (Lord, 1952). The PTGI-SC, which has six choices for each item, should have a true difficulty value of 0.5e0.7. Next, the factor structure of the PTGI-SC was determined using exploratory factor analysis. We used analysis of moment structures (AMOS, version 17.0) to conduct the confirmatory factor analysis (CFA, i.e., maximum likelihood) to verify the fitness of the hypothesized model and the data of the 1227 breast cancer survivors. The model's goodness-of-fit was assessed using c2/degrees of freedom (df), standardized root mean squared residual (SRMR), root mean square error of approximation (RMSEA), goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), comparative fit index (CFI), TuckereLewis index (TLI), normed fit index (NFI), and incremental fit index (IFI). The ideal value of c2/df is < 2 (Kit-Tai et al., 2004); however, this value is sensitive to the sample size (Marsh and Balla, 1988). Therefore, an c2/df value between 2 and 5 is generally €reskog and So €rbom, 2006). SRMR values 0.08 and acceptable (Jo RMSEA values 0.06 (Hu and Bentler, 1999; Thompson, 2004) generally indicate reasonable model fit. GFI, AGFI, NFI, IFI, TLI, and
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
CFI values are acceptable if they are >0.90 (Byrne, 1994; Kit-Tai et al., 2004). Finally, the Cronbach a coefficient of each PTGI factor was also obtained. Ethical considerations The University of Institutional Review Board approved this study (code number 2010 SY24), and verbal informed consent was obtained from each participant.
3
Table 2 PTGI-SC inter-factor correlation and factoretotal correlation (n ¼ 840).
1 2 3 4 5
PTGI-SC total score
1
2
3
4
5
0.803** 0.763** 0.736** 0.647** 0.652**
e 0.376** 0.495** 0.319** 0.351**
e 0.464** 0.429** 0.544**
e 0.459** 0.460**
e 0.398**
e
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, ** p < 0.01 (2-tailed); 1 ¼ Relating to others, 2 ¼ New possibilities, 3 ¼ Personal strength, 4 ¼ Appreciation of life, 5 ¼ Spiritual change.
Results two groups at a 2:1 ratio, using SPSS software (840 and 387 participants in samples 1 and 2, respectively).
Demographic data The participants were aged between 26 and 79 years, with an average age of 53.30 ± 8.05 years. The time since the diagnosis of breast cancer ranged from 4 months to 32 years, with a median of 3.5 years. The majority of patients reported no religious beliefs (90.4%). Among 1227 survivors, 324 (26.4%), 163 (13.3%), and 632 (51.5%) were diagnosed as TNM stage I, II, and III, respectively. With regard to surgery, 82.6% underwent mastectomy and the remaining 17.4% underwent breast-conserving surgery. Item analyses The item analyses consisted of determining the difficulty and discrimination of each item. The formula for determining difficulty is p ¼ x=xmax (p ¼ difficulty; x ¼ average score of each item; xmax ¼ maximum score of each item). The difficulty values ranged between 0.432 and 0.737. Discrimination was calculated using item-total Pearson correlation coefficients, which ranged between 0.324 and 0.721, all with p < 0.01 (Table 1).
Exploratory factor analysis The Kaiser-Meyer-Olkin (KMO) test for measuring sampling adequacy and Bartlett's test of sphericity were used to assess the adequacies of the correlation matrices used in the factor analysis. The KMO result for sample 1 was 0.899, which is considered “good” (Hutcheson and Sofroniou, 1999) and indicates that reliable and distinct factors can be extracted. Furthermore, Bartlett's test of sphericity determined that c2 ¼ 4348.826 and p < 0.01, which indicated that the correlation matrices were not identity matrices; hence, the factor model was appropriate (Snedecor and Cochran, 1989). The initial principal-axis factor analysis of sample 1 provided five factors with eigenvalues >1, which collectively accounted for Table 3 PTGI-SC exploratory factor analysis (n ¼ 840). Item
Construct validity The data for exploratory factor analysis and confirmatory factor analysis should be two samples from the same population (Hurley et al., 1997). Therefore, the samples were randomly divided into
Table 1 Difficulty and discrimination of each PTGI-SC item (n ¼ 1227). Item
p
r
1. My priorities about what is important in life. 2. I'm more likely to try to change things that need changing. 3. An appreciation for the value of my own life. 4. A feeling of self-reliance. 5. A better understanding of spiritual matters. 6. Knowing that I can count on people in times of trouble. 7. A sense of closeness with others. 8. Knowing I can handle difficulties. 9. A willingness to express my emotions. 10. Being able to accept the way things work out. 11. Appreciating each day. 12. Having compassion for others. 13. I'm able to do better things with my life. 14. New opportunities are available that wouldn't have been otherwise. 15. Putting effort into my relationship. 16. I have a stronger religious faith. 17. I discovered that I'm stronger than I thought I was. 18. I learned a great deal about how wonderful people are. 19. I developed new interests. 20. I accept needing others. 21. I established a new path for my life.
0.632 0.686
0.453** 0.468**
0.700 0.692 0.693 0.637 0.687 0.729 0.685 0.699 0.722 0.737 0.732 0.588
0.425** 0.511** 0.611** 0.617** 0.721** 0.601** 0.713** 0.527** 0.605** 0.678** 0.616** 0.575**
0.703 0.432 0.736 0.708 0.596 0.661 0.631
0.668** 0.324** 0.584** 0.712** 0.565** 0.645** 0.608**
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, ** p < 0.01.
Relating to others 6 7 9 12 15 18 20 New possibilities 2 13 14 19 21 Personal strength 4 8 10 17 Appreciation of life 1 3 11 Spiritual change 5 16 Eigenvalues Percentage of variance explained Cumulative % of variance explained
Factor loading Factor 1
Factor 2
Factor 3
Factor 4
Factor 5
0.827 0.856 0.828 0.799 0.842 0.815 0.792
0.029 0.141 0.221 0.190 0.072 0.238 0.106
0.107 0.114 0.164 0.242 0.204 0.149 0.124
0.114 0.148 0.040 0.031 0.035 0.028 0.083
0.004 0.001 0.034 0.088 0.032 0.002 0.184
0.125 0.188 0.150 0.094 0.178
0.683 0.786 0.739 0.716 0.792
0.038 0.273 0.117 0.155 0.022
0.221 0.089 0.161 0.017 0.145
0.147 0.107 0.088 0.326 0.188
0.227 0.219 0.197 0.154
0.010 0.369 0.110 0.391
0.821 0.570 0.791 0.573
0.219 0.136 0.083 0.085
0.047 0.114 0.106 0.278
0.010 0.110 0.152
0.196 0.099 0.237
0.037 0.291 0.313
0.830 0.570 0.729
0.274 0.203 0.039
0.279 0.011 5.128 36.821
0.259 0.160 3.519 13.456
0.290 0.058 2.633 7.968
0.019 0.075 1.866 5.145
0.574 0.832 1.203 4.941
36.821
50.277
58.245
63.389
68.331
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory. Extraction method: principal-axis analysis. Rotation method: varimax with Kaiser normalization. The bold indicates the factor loadings are more than 0.5.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
4
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
Table 4 Goodness-of-fit indices for PTGI-SC (n ¼ 387).
Table 5 Internal consistency of PTGI-SC (n ¼ 1227).
Model
c2/df
SRMR
RMSEA
GFI
AGFI
NFI
IFI
TLI
CFI
Subscale
Number of items
Cronbach's a
5-factor
3.912
0.046
0.055
0.967
0.913
0.942
0.932
0.945
0.932
Relating to others New possibilities Personal strength Appreciation of life Spiritual change Total PTGI-SC
7 5 4 3 2 21
0.93 0.85 0.73 0.65 0.60 0.90
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory, SRMR: standardized root mean squared residual, RMSEA: root mean square error of approximation, GFI: goodness-of-fit index, AGFI: adjusted goodness-of-fit index, NFI: normed fit index, IFI: incremental fit index, TLI: TuckereLewis index, CFI: comparative fit index.
PTGI-SC: Simplified Chinese Version of the Posttraumatic Growth Inventory.
68.3% of the variance. After appraising the scree plot (Cattell, 1978), we decided to retain all five factors. The correlations of the five factors with the total PTGI-SC score ranged from 0.647 to 0.803, and the inter-factor correlations ranged from 0.319 to 0.544 (Table 2). This indicated that the five factors estimate common aspects, but that each factor also has a unique contribution. We found that varimax-rotated factor matrix of the five-factor PTGI-SC explained 68.3% of the variance (Table 3). The factors were identical to the original English version, but were inconsistent with the results of the four-factor traditional Chinese version.
Reliability The internal consistency of Cronbach's a coefficients for the total scale was 0.90, which is considered high according to DeVellis (0.80 is considered high) (DeVellis, 2002). The subscales demonstrated high Cronbach's a coefficients for “relating to others” and “new possibilities,” good values (0.70e0.80) for “personal strength,” and acceptable values (0.65e0.70) for “appreciation of life” (Table 5). However, “spiritual change” demonstrated a low internal consistency of 0.602, which means that this subscale should be modified in some way (DeVellis, 2002).
Confirmatory factor analysis Discussion The preliminary analysis of the five-factor PTGI-SC using sample 2 demonstrated no negative error variance, and all of the standardized regression weights were <1 (range 0.29e0.83). This signified that there is no improper solution to the model (Hair et al., 1998), and we can test the model's overall fitness (Table 4). The loading values were 0.65e0.83, 0.52e0.80, 0.51e0.69, 0.42e0.76, and 0.38e0.63 for “relating to others,” “new possibilities,” “personal strength,” “appreciation of life,” and “spiritual change,” respectively. Correlations between the five subscales ranged from 0.29 to 0.69 (Fig. 1).
This study reported the psychometric properties of Chinese breast cancer survivors according to the PTGI-SC results. With regard to the item analysis, some inconsistencies were found between difficulty and discrimination. The results showed that all of the items demonstrated a true difficulty except “I have a stronger religious faith.” We believe that this is likely due to the lack of religious populations in China compared with other countries (Ruan et al., 2011). In this study, 90.4% of patients reported no religious beliefs. That might have influenced the lower score for this
Fig. 1. Structural equation model of the PTGI-SC.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
item, which would also reduce the difficulty. However, the itemtotal correlation coefficients were 0.324e0.721, which meets the criterion (0.3e0.8) set by Tucker (1963). Therefore, we consider that the item “I have a stronger religious faith” still has acceptable and differentiated ability. We also tested the content and construct validity of the PTGI-SC. Content validity is the degree to which the items of an instrument adequately represent the targeted construct for a particular assessment purpose. In the translation and back-translation periods, most of the items in the simplified Chinese version were appropriately translated, and the conceptual meaning of the items was deemed by expert review (Bracken and Barona, 1991) to be similar to that of the original version. Next, the CVI was applied to the PTGI-SC. Different evaluators may hold different opinions about whether the items reflect the scale domains. To reduce the subjective influence of the evaluator, we asked five experts to score each item. A CVI value of 0.980 indicated that the item was correctly loaded in the proper domain. Construct validity indicates whether a measure relates to other observed variables in a way that is consistent with theoretically derived predictions (Bollen, 1989). The factor structure of the PTGI-SC according to the exploratory factor analysis was comparable to the original English version, and CFA was used to verify this result. All of the GFI, AGFI, NFI, IFI, TLI, and CFI values were >0.90. The SRMR and RMSEA values indicated a sound model fit, while the correlation matrix indicated moderate correlation among items. Therefore, the structural equation model fits the data of the 1227 breast cancer survivors moderately well, signifying that the PTGI-SC structure is reasonable. The total PTGI-SC demonstrated good internal consistency as reflected by the Cronbach's a values. However, the factors “appreciation of life” and “spiritual change” were responsible for the relatively low a coefficients. It is known that the strength of a depends on the length of the subscale. It is best to use Cronbach's a when evaluating subscales with 3 items (McDowell and Newell, 2006). Thus, the number of items in the “appreciation of life” and “spiritual change” subscales may explain the low a coefficients. A limitation of this study is the convenience sampling method, which may have reduced the representativeness of the sample. However, the large population may have reduced the selection bias of the study, and validation is a continuous process. Future research should focus on the relationship between the PTGI-SC and alternative measures. Conclusions The present findings regarding validity and reliability support its use for evaluating breast cancer survivors in China. The PTGI-SC is a reliable and valid inventory that can be used in practice to measure PTG levels in breast cancer survivors to provide information about their psychological adjustment. These findings will facilitate further psychological research in Chinese breast cancer survivors. Conflict of interest None declared. Acknowledgments We are grateful to nurses in the 8 first-class hospitals for their assistance of data collection and to all breast cancer survivors who took part in the study. This study was funded by National Natural Science Foundation of China (grant no. 30870770), and the Importation and Development of High-Caliber Talents Project of
5
Beijing Municipal Institutions (grant no. CIT&TCD20140328). The authors declare no conflicts of interest. References Anderson, B.O., Jakesz, R., 2008. Breast cancer issues in developing countries: an overview of the Breast Health Global Initiative. World Journal of Surgery 32 (12), 2578e2585. http://dx.doi.org/10.1007/s00268-007-9454-z. Bollen, K., 1989. Structural Equations with Latent Variables. John Wiley and Sons, New York. Bracken, B.A., Barona, A., 1991. State of the art procedures for translating, validating, and using psychoeducational tests in cross-cultural assessment. School Psychology International 12, 119e132. Brunet, J., McDonough, M.H., Hadd, V., Crocker, P.R., Sabiston, C.M., 2010. The posttraumatic growth inventory: an examination of the factor structure and invariance among breast cancer survivors. Psycho-Oncology 19 (8), 830e838. http://dx.doi.org/10.1002/pon.1640. Byrne, B.M., 1994. Structural Equation Modeling with EQS and EQS/Windows. Sage Publications, Thousand Oaks, CA. Calhoun, L.G., Tedeschi, R.G., 1999. Facilitating Posttraumatic Growth: a Clinician's Guide. Lawrence Erlbaum, Mahwah. Cattell, R.B., 1978. The Scientific Use of Factor Analysis in Behavioral and Life Sciences. Plenum Press, New York. Cordova, M.J., Andrykowski, M.A., 2003. Responses to cancer diagnosis and treatment: posttraumatic stress and posttraumatic growth. Seminars in Clinical Neuropsychiatry 8 (4), 286e296. DeVellis, R.F., 2002. Scale Development Theory and Applications, second ed. SAGE, London. Green, M., Raina, V., 2008. Epidemiology, screening and diagnosis of breast cancer in the Asia-Pacific region: current perspectives and important considerations. Asia-Pacific Journal of Clinical Oncology 4, S5eS13. http://dx.doi.org/10.1111/ j.1743-7563.2008.00191.x. Hair, J.F., Anderson, R.E., Tatham, R.L., Black, W.C., 1998. Multivariate Data Analysis with Readings. Maxwell Macmillan International, New York. Harrington, C.B., Hansen, J.A., Moskowitz, M., Todd, B.L., Feuerstein, M., 2010. It's not over when it's over: long-term symptoms in cancer survivorsda systematic review. International Journal of Psychiatry in Medicine 40 (2), 163e181. Ho, S.M., Chan, C.L., Ho, R.T., 2004. Posttraumatic growth in Chinese cancer survivors. Psycho-Oncology 13 (6), 377e389. http://dx.doi.org/10.1002/pon.758. Hu, L.T., Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling 6, 1e55. Hurley, A.E., Scandura, T.A., Schriesheim, C.A., Brannick, M.T., Seers, A., Vandenberg, R.J., et al., 1997. Exploratory and confirmatory factor analysis: guidelines, issues, and alternatives. Journal of Organizational Behavior 18, 667e683. Hutcheson, G., Sofroniou, N., 1999. The Multivariate Social Scientist: Introductory Statistics Using Generalized Linear Models. Sage Publications, Thousand Oaks, CA. €reskog, K.G., So € rbom, D., 2006. LISREL 8.80. Science Software International, Chicago. Jo Jaarsma, T.A., Pool, G., Sanderman, R., Ranchor, A.V., 2006. Psychometric properties of the Dutch version of the posttraumatic growth inventory among cancer patients. Psycho-Oncology 15 (10), 911e920. http://dx.doi.org/10.1002/ pon.1026. Joseph, S., Linley, P.A., Harris, G.J., 2004. Understanding positive change following trauma and adversity: structural clarification. Journal of Loss and Trauma 10 (1), 83e96. http://dx.doi.org/10.1080/15325020490890741. Kit-Tai, H., Wen, Z.L., Cheng, Z.J., 2004. Structural Equation Model and its Applications. Educational Science Publishing House, Beijing. Levine, S.Z., Laufer, A., Hamama-Raz, Y., Stein, E., Solomon, Z., 2008. Posttraumatic growth in adolescence: examining its components and relationship with PTSD. Journal of Traumatic Stress 21 (5), 492e496. http://dx.doi.org/10.1002/jts.20361. Linley, P.A., Andrews, L., Joseph, S., 2007. Confirmatory factor analysis of the posttraumatic growth inventory. Journal of Loss and Trauma 12 (4), 321e332. http:// dx.doi.org/10.1080/15325020601162823. Liu, J.E., Wang, H.Y., Wang, M.L., Su, Y.L., Wang, P.L., 2014. Posttraumatic growth and psychological distress in Chinese early-stage breast cancer survivors: a longitudinal study. Psycho-Oncology 23 (4), 437e443. http://dx.doi.org/10.1002/ pon.3436. Lord, F.M., 1952. The relationship of the reliability of multiple-choice tests to the distribution of item difficulties. Psychometrika 18, 181e194. Lynn, M.R., 1986. Determination and quantification of content validity. Nursing Research 35 (6), 382e385. Manne, S., Ostroff, J., Winkel, G., Goldstein, L., Fox, K., Grana, G., 2004. Posttraumatic growth after breast cancer: patient, partner, and couple perspectives. Psychosomatic Medicine 66 (3), 442e454. Marsh, H.W., Balla, J.R., 1988. Goodness-of-fit indexes in confirmatory factor analysis: the effect of sample size. Psychological Bulletin 103 (3), 391e410. McDowell, I., Newell, C., 2006. Measuring Health: a Guide to Rating Scales and Questionnaires, third ed. Oxford University Press, Oxford. Porter, P., 2008. “Westernizing” women's risks? Breast cancer in lower-income countries. New England Journal of Medicine 358 (3), 213e216. http:// dx.doi.org/10.1056/NEJMp0708307.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002
6
J.-E. Liu et al. / European Journal of Oncology Nursing xxx (2015) 1e6
Powell, S., Rosner, R., Butollo, W., Tedeschi, R.G., Calhoun, L.G., 2003. Posttraumatic growth after war: a study with former refugees and displaced people in Sarajevo. Journal of Clinical Psychology 59 (1), 71e83. http://dx.doi.org/10.1002/jclp.10117. Ruan, R.P., Zheng, F.T., Liu, L., 2011. Religious faith, religious participation and subjective benefits. China Rural Observation 2, 74e86. Sears, S.R., Stanton, A.L., Danoff-Burg, S., 2003. The yellow brick road and the emerald city: benefit finding, positive reappraisal coping and posttraumatic growth in women with early-stage breast cancer. Health Psychology 22 (5), 487e497. http://dx.doi.org/10.1037/0278-6133.22.5.487. Snedecor, G.W., Cochran, W.G., 1989. Statistical Methods, eighth ed. Iowa State University Press, Ames, IA. Taku, K., Cann, A., Calhoun, L.G., Tedeschi, R.G., 2008. The factor structure of the posttraumatic growth inventory: a comparison of five models using confirmatory factor analysis. Journal of Traumatic Stress 21 (2), 158e164. http:// dx.doi.org/10.1002/jts.20305.
Tedeschi, R.G., Calhoun, L.G., 1996. The posttraumatic growth inventory: measuring the positive legacy of trauma. Journal of Traumatic Stress 9 (3), 455e471. Thompson, B., 2004. Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications. American Psychological Association, Washington DC. Tucker, L.R., 1963. Formal Models for a Central Prediction System (Psychometric Monograph). William Byrd Press, Richmond, VA. Wang, H.Y., 2011. Investigation and Analysis on the Level and Associated Factors of Posttraumatic Growth in Breast Cancer Patients. Thesis. Capital Medical University, Beijing. Weiss, T., 2004. Correlates of posttraumatic growth in husbands of breast cancer survivors. Psycho-Oncology 13 (4), 260e268. Xu, G., 2010. The screening of breast cancer. Chinese Medicine Herald 7 (2), 4e6.
Please cite this article in press as: Liu, J.-E., et al., Psychometric evaluation of the Simplified Chinese Version of the Posttraumatic Growth Inventory for assessing breast cancer survivors, European Journal of Oncology Nursing (2015), http://dx.doi.org/10.1016/j.ejon.2015.01.002