- Email: [email protected]
The Chinese version of hospital anxiety and depression scale: Psychometric properties in Chinese cancer patients and their family caregivers
European Journal of Oncology Nursing 25 (2016) 16e23
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
The Chinese version of hospital anxiety and depression scale: Psychometric properties in Chinese cancer patients and their family caregivers Qiuping Li a, Yi Lin a, Caiping Hu b, Yinghua Xu c, *, Huiya Zhou c, Liping Yang d, Yongyong Xu e, ** a
Wuxi Medical School, Jiangnan University, Wuxi, Jiangsu Province, China Shanxi Provincial Tomor Hospital, Taiyuan, Shanxi Province, China Wuxi People's Hospital, Wuxi, Jiangsu Province, China d Xijing Hospital, Xi'an, Shaanxi Province, China e Department of Health Statistics, Fourth Military Medical University, Xi'an, Shaanxi Province, China b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 January 2016 Received in revised form 29 August 2016 Accepted 9 September 2016
Purpose: The Hospital Anxiety and Depression Scale (HADS) acts as one of the most frequently used selfreported measures in cancer practice. The evidence for construct validity of HADS, however, remains inconclusive. The objective of this study is to evaluate the psychometric properties of the Chinese version HADS (C-HADS) in terms of construct validity, internal consistency reliability, and concurrent validity in dyads of Chinese cancer patients and their family caregivers. Methods: This was a cross-sectional study, conducted in multiple centers: one hospital in each of the seven different administrative regions in China from October 2014 to May 2015. A total of 641 dyads, consisting of cancer patients and family caregivers, completed a survey assessing their demographic and background information, anxiety and depression using C-HADS, and quality of life (QOL) using Chinese version SF-12. Data analysis methods included descriptive statistics, confirmatory factor analysis (CFA), and Pearson correlations. Results: Both the two-factor and one-factor models offered the best and adequate fit to the data in cancer patients and family caregivers respectively. The comparison of the two-factor and single-factor models supports the basic assumption of two-factor construct of C-HADS. The overall and two subscales of CHADS in both cancer patients and family caregivers had good internal consistency and acceptable concurrent validity. Conclusions: The Chinese version of the HADS may be a reliable and valid screening tool, as indicated by its original two-factor structure. The finding supports the basic assumption of two-factor construct of HADS. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Anxiety Depression Hospital anxiety and depression scale Psychometric properties Cancer Family caregivers Caregiver-patient dyads Chinese
1. Introduction As the leading cause of morbidity and mortality worldwide, cancers reached up to 14.1 million new cases, with 8.2 million cancer-related deaths in 2012. Of these, 21% (3 million) of new cancer cases, and 27% (2.2 million) of cancer deaths occurred in China (WHO, 2015). It is unfortunate that in developing countries,
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (Y. Xu), [email protected] (Y. Xu). http://dx.doi.org/10.1016/j.ejon.2016.09.004 1462-3889/© 2016 Elsevier Ltd. All rights reserved.
where most new cancer cases are frequently diagnosed at an advanced stage, treatment options are both limited and expensive pez-Go mez et al., 2013). With the increase in cancer cases, there (Lo is a need for a similar number of family caregivers, who are expected to provide care or support to cancer patients. Both cancer patients and their family caregivers need to cope together and adjust to the challenge of the profound emotional and social adversity imposed by a cancer diagnosis and its treatment (Kayser et al., 2007). Accumulating evidence has shown that psychological distress represents a significant adversity in cancer populations (Saboonchi
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
et al., 2013; Satin et al., 2009,Carlson et al., 2004) and their family caregivers (Li and Loke, 2013,Haley, 2003). Anxiety and depression - the most common presenting symptoms of psychological distress (Rodgers et al., 2005) - are reported to be prevalent in approximately one-third, or as many as 45% of cancer patients (Rodgers et al., 2005,Schreier and Williams, 2004,Grassi et al., 1996) and ttir et al., 2011). their family caregivers (Janda et al., 2007,Fridriksdo It has been reported that the psychological distress of family caregivers could be as high or even higher, than that of cancer patients themselves (Hagedoorn et al., 2008; Northouse et al., 2000). Anxiety and depression also exert a significant impact on quality of life (QOL) in both cancer patients (Grassi et al., 1996,Li et al., 2014; Saevarsdottir et al., 2010), and their family caregivers ttir et al., 2011; Kim et al., 2008; Northouse et al., 2000). (Fridriksdo Evidence from a meta-analysis concluded that depression, in particular, constitutes a predictor of mortality in cancer patients (Satin et al., 2009). Studies have also shown there is a mutual impact between dyads of cancer patients and family caregivers in terms of QOL and psychological distress (Kim et al., 2008; Northouse et al., 2000). Evidence also indicates that cancer affects caregiver-patient dyads as a unit, rather than as isolated individuals (Hagedoorn et al., 2008), leading to the primary focus of cancer care research to shift from the individual experiences of cancer patients or family caregivers, to the dyadic level of caregiver-patient dyads (Fletcher et al., 2012). For a better understanding of the related experiences of caregiver-patient dyads from the dyadic level, the participants in the present study included both cancer patients and their family caregivers. The growing recognition of the common prevalence of anxiety and depression, and the significant impact on the lives of cancer patients and their family caregivers, highlights the need for valid assessment and screening methods for anxiety and depression in cancer practice (Saboonchi et al., 2013). Self-report questionnaires appear to be specifically appropriate, and a practical tool in this context (Mitchell, 2010). The Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, 1983) stands out as one of the most frequently used self-reported measures, which is considered an effective screening measure for both anxiety and depression, and has been widely used across a variety of cancer populations and family caregivers (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Gough and Hudson, 2009; Mitchell et al., 2010). In terms of the instrument's psychometric properties, although internal consistency reliability with Cronbach's alpha for HADS anxiety varied from 0.68 to 0.93 (mean 0.83), and for HADS depression from 0.67 to 0.90 (mean 0.82), this suggests the instrument has good reliability and shows a capacity to consistently capture reliable data (Bjelland et al., 2002); however, the evidence for construct validity, which is based on analysis for instrument dimensionality, remains inconclusive. According to a 10-year systematic review of the latent structure of the HADS, the largest degree of heterogeneity of construct validity occurs in studies of cancer populations (Cosco et al., 2012). The heterogeneity of the factorial structure of HADS in cancer populations consists of singlefactor (Smith et al., 2006; Razavi et al., 1990), two-factor (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006), three-factor (Rodgers et al., 2005; Brandberg et al., 1992), and four-factor structure (Lloyd-Williams et al., 2001). Given that the original English version of HADS has been translated into and validated in many different languages (Herrmann, 1997), including Chinese (Leung et al., 1993; Wang et al., 2009), the measurement properties of the HADS, such as the inconclusiveness of the construct validity, could be affected by cross-lingual and cross-cultural adaptation processes (Guillemin
17
et al., 1993). Although several studies have been conducted to validate the factorial structure of the Chinese version of HADS (CHADS), including in a sample of medical students (Leung et al., 1993) and in coronary heart disease (Wang et al., 2009), no studies have been conducted in mainland China, to our knowledge, to validate its psychometric properties in cancer patients and their family caregivers. Consequently, this study's aim was to evaluate the psychometric properties of the C-HADS from a dyadic perspective. To be specific, to evaluate the psychometric properties of the C-HADS in terms of construct validity, internal consistency reliability, and concurrent validity in dyads of Chinese cancer patients and their family caregivers. This study provides the psychometric properties of the CHADS when applied to a sample of Chinese cancer patients and their family caregivers. 2. Methods 2.1. Study design and participants This was a cross-sectional study, conducted in multiple centers: one hospital in each of the seven different administrative regions in China. The seven administrative regions cover different areas of China, and one of the high-ranking hospitals within each of the administrative regions was recruited by convenience sampling. The seven regions were: (i) East China; (ii) Southern China; (iii) North China; (iv) Central China; (v) Northwest China; (vi) Southwest China; and (vii) Northeast China. Participants consisted of 641 cancer patient and their family caregiver dyads, and were recruited by convenience sampling from October 2014 to May 2015. The study criteria inclusions were as follows: (i) dyads of Chinese adult cancer patients and family caregivers (age >18 years old); (ii) a medical diagnosis of any type of cancer in patients, who had no other diseases, such as dementia, which could lead to unconsciousness; (iii) a primary family caregiver who provides informal care to cancer patients; (iv) both patients and their family caregivers could communicate in Mandarin-Chinese, and consent to take part in the study. Sample size calculation: the sample size in the HADS evaluation was calculated by n¼ (uas/d)2 and n¼ (uas/d)2z4 s2, given a ¼ 0.05 and d ¼ ±1, the error in estimation of population means for HADS -Total, HADS-Anxiety and HADS- Depression. s2 was estimated separately by deff s2 Total, deff s2 Anxiety and deff s2 Depression (Statistics. health, 2015). From our pilot study in a region, the estimated value of s2 Total was 9.5, and the value of s2 Anxiety and s2 Depression was 5. The design effect (deff) for sampling from different regions was given as 1.5. The sample size for estimating HADS -Total was 4 1.5 9.52 ¼ 541, and for estimating HADSAnxiety and HADS- Depression it was 4 1.5 52 ¼ 150. One hundred more cases were added to the actual survey in case of no response and missing data; the final sample size was 641. The sample sizes for each subgroup of HADS-Anxiety and HADSDepression should be no fewer than 150 cases. To simplify, the sample size of the present study is large enough to ensure the statistical power for doing confirmatory factor analysis (CFA). 2.2. Instruments Three groups of variables were collected: socio-demographic characteristics and clinical data, the Chinese version of HADS (CHADS) for anxiety and depression (Zigmond and Snaith, 1983), and the Chinese version of Medical Outcomes Study 12-item Short Form (C-SF-12) (version 2) for QOL. The purpose of using the C-SF-12 was to examine the concurrent validity of the C-HADS. Information on socio-demographic characteristics and clinical
18
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
data from both patients and their family caregivers was solicited through a demographic and background information sheet. The solicited socio-demographic and clinical data included age, gender, marital status, caregiver's relationship with the patient, education level, working status, cancer type, cancer stage, time since cancer diagnosis, and financial burden on the family due to medical treatment costs. The HADS contains 14 items making up two 7-item subscales, one measuring anxiety and the other depression (Zigmond and Snaith, 1983). A four-point Likert-type scale ranging from 0 (not a problem) to 3 (high level of problems) is used for scoring each item. The score for each subscale is obtained by simple summation of individual items with scores in each subscale (anxiety and depression) ranging from 0 to 21. A score of seven or less indicates a non-anxiety/depression case, 8e10 a borderline case of anxiety/ depression, and 11 or above a definite case of anxiety/depression (Arving et al., 2008). The SF-12 consists of 12 items making up eight scales that measure the following eight QOL domains (Ware et al., 1996): Physical Functioning (PF), Role-Physical (RP), Bodily Pain (BP), General Health (GH), Vitality (VT), Role-Emotional (RE), Social Functioning (SF), and Mental Health (MH). The eight SF-12 domains hypothetically form two dimensions of the Physical Component Summary (PCS) and the Mental Component Summary (MCS) (Ware et al., 1996). The eight scales and two dimensions were transformed and calculated according to the SF-12 (version 2) score manual to a scale with a theoretical range of 0e100 (Ware et al., 2002). A higher score indicates a higher level of QOL. The two dimensions (PCS and MCS) of the SF-12 version 2 achieved R squares of 0.905 with PCS and 0.938 with MCS when the SF-36 was used in a cross-validated Medical Outcomes Study. Test-retest (2-week) correlations of 0.89 and 0.76 were observed for the SF-12 PCS and MCS respectively, in the general U.S. population (n ¼ 232) (Ware et al., 1996). The internal consistency of the present study for the total SF-12 was Cronbach's a ¼ 0.821 (n ¼ 596) and 0.812 (n ¼ 593) for cancer patients and family caregivers respectively. 2.3. Data collection procedure Before study commencement, ethical approval was granted by the research ethics committee of Jiangnan University (HSEARS20140701001), and access approval was obtained from the related hospitals. Nurses working in the oncology unit in the related hospitals were provided with explanations of both the study and the instruments, prior to the study commencing. The research was conducted with utmost respect for participants, and the confidentiality of the collected data was held in accordance with the Helsinki Declaration of 1975, as revised in 2000. The hospital oncologists identified the cancer dyads in accordance with the eligibility criteria. Cancer patient-family caregiver dyads who met the criteria for inclusion were approached in the oncology wards when they were admitted for chemotherapy treatment during the recovery stage. After their written informed consent was obtained, the dyad participants were invited to complete the questionnaire separately, with the help of a nurse if required. According to their preference, they filled out the questionnaire in a private in-patient room or in a nurse's office, so their privacy could be protected and they would be removed from any possible disturbances. The questionnaire took approximately 10e15 min to complete. None of the study participants reported discomfort or distress. 2.4. Data analysis Data were analyzed using SPSS version 21.0. Descriptive
statistics such as frequencies, percentages, means, and standard deviations were used to describe subject characteristics. The C-HADS factor structure was determined using confirmatory factor analysis (CFA) based on Maximum Likelihood, using Amos version 21.0. Six models developed from HADS validity and psychometric studies in the sample of cancer population were tested. The six models were: (i) single-factor model: the same factor structure of single factor model from Razavi et al. (1990) and Smith et al. (2006); (ii) the original two-factor model (Saboonchi et al., 2013; Mystakidou et al., 2004); (iii) the modified two-factor model (Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Gough and Hudson, 2009); (iv) three-factor model 1 (Brandberg et al., 1992); (v) three-factor model 2 (Rodgers et al., 2005); (vi) and four-factor model (Lloyd-Williams et al., 2001). Four indices were used to evaluate the overall fit of CFA models, namely: Chi-Square c2, a confirmatory fit index (CFI), a root mean square error of approximation (RMSEA), and the standardized root mean squared residual (SRMR). Generally, an insignificant p value (p > 0.05) in the Chi-Square c2, CFI greater than 0.95, both a RMSEA and an SRMR of less than 0.08 are considered indicative of good model fit (Hooper et al., 2008). The competing models were compared by means of likelihood ratio (Dc2) test. The internal reliability was evaluated by Cronbach's coefficient alpha. Cut-offs for a good reliability for Cronbach's alpha was set at 0.75 (Portney and Watkins, 2009). The concurrent validity of the C-HADS was examined by comparing it with the C-SF-12 to determine whether it was psychometrically sound for measuring its respective attributes. A significant negative correlation between the measurements indicates good concurrent validity (Portney and Watkins, 2009). 3. Results 3.1. Participant characteristics Out of a total 750 approached eligible cancer patient and family caregiver dyads, 42 dyads declined to participate in the study. The remaining 708 dyads became our participants and completed the questionnaire (response rate ¼ 94.4%). Of those, 641 dyads provided valid data for study variables (valid rate ¼ 90.5%). Table 1 summarizes the dyad characteristics. In this study, about half of the patients (50.1%) and family caregivers (48.5%) were male. Approximately 90% of patients (91.4%) and family caregivers (89.1%) were married, and more than half of family caregivers (53.5%) were spouses of cancer patients. All patients were at an advanced stage of cancer, namely, stage III (n ¼ 364, 56.8%) or stage IV (n ¼ 277, 43.2%). Nearly half of the dyads (46.3%) were coping with digestive system cancer (e.g., esophageal, gastric, liver, or colorectal cancer). All patients were receiving chemotherapy, and 102 (15.9%) patients had had surgery. Approximately two-thirds of the families (64.6%) experienced serious financial burden due to the cost of treating the cancer. 3.2. Construct validity Of the six models included in the analysis, the solution for the two three-factor models is not admissible; no related c2 value was calculated; the four-factor model cannot be converged in either sample of cancer patients or family caregivers. The results of the CFA fit statistics of the remaining three models of single-factor and two-factor models are presented in Table 2. For cancer patients (the upper part of Table 2), both the single-factor and two-factor models offered the best fit for the data. In the three best fit models, although the chi-square (c2) was significant, other fit indices had acceptable values.
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
19
Table 1 Characteristics of patients and family caregivers (n ¼ 641). Characteristics
Patients [ n (%)]*
FC [ n (%)]*
Age (mean ± SD), years Gender Male Female Marital status Married Not married** FCs' relationship with patients Spouses Offspring Parents Sibling Others Education levels Primary school or less High school University or above Working status Working Not working Type of cancer*** Breast cancer Ovarian and cervical cancer Esophageal and gastric cancer Colorectal cancer Liver cancer Lung cancer Others Type of treatment Chemotherapy Chemotherapy þ surgery Chemotherapy þ radiotherapy Chemotherapy þ others Average time since diagnosis/duration in their role as a FC
54.6 ± 12.9(ranging from 18 to 88)
46.6 ± 13.2(ranging from 18 to 79)
321(50.1) 318(49.6)
311(48.5) 329(51.3)
586(91.4) 55(8.6)
571(89.1) 69(10.8)
357(55.7) 102(15.9) 114(17.8) 61(9.5) 13.3 ± 24.0 months (ranging from 1 to 228 months)
Financial burden on the family due to cancer treatment
Serious: 414 (64.6); Normal: 191 (29.8); Mild or None: 33 (5.1)
343(53.5) 215(33.5) 20(3.1) 41(6.4) 21(3.3) 356(55.5) 198(30.9) 85(13.3)
256(40.0) 222(34.6) 159(24.8)
373(58.2) 263(41.0)
411(64.1) 227(35.4)
70(10.9) 93(14.5) 152(23.7) 86(13.4) 59(9.2) 88(13.7) 76(11.9)
<6 months: 348(54.3) 6 monthse2 years: 192(30.0) >2 yearse5 years: 55 (8.6) >5 years: 39(6.1)
Note: FC¼ Family caregivers; SD ¼ standard deviation; * The total n does not equal 641 because of missing values; ** Not married include: Divorced, Widowed, and Never married; *** All of the patients had advanced cancer, were in stage III (n ¼ 364, 56.8%) and stage IV (n ¼ 277, 43.2%).
Table 2 Comparisons of goodness-ofefit indicators of nested single-factor and two-factor models of C-HADS. Model Patients Single-factor Original two-factor D Dc2a Modified two-factor D Dc2b Family caregivers Single-factor Original two-factor Dc2a Modified two-factor D Dc2b
FLI 1
FLI 2
All items 1,3,5,7,9,11,13
2,4,6,8,10,12,14
1,3,5,9,11,13
2,4,6,7,8,10,12,14
All items 1,3,5,7,9,11,13
2,4,6,8,10,12,14
1,3,5,9,11,13
2,4,6,7,8,10,12,14
c2
df
CFI
RMSEA
SRMR
283.11* 265.50* 17.61* 248.35* 34.76*
77 76 1 76 1
0.954 0.958
0.066 0.064
0.035 0.034
0.962
0.061
0.032
358.80* 358.00* 0.80 338.59* 20.21*
77 76 1 76 1
0.930 0.932
0.077 0.077
0.041 0.041
0.935
0.075
0.040
Note: Dc2a ¼ c2 (Original two-factor) -c2 (Single-factor), Dc2b ¼ c2 (Modified two-factor) -c2 (Single-factor). A significant change in chi-square (c2) indicates better model fit in the two-factor models than the single-factor model does. *P < 0.001. D A significant improvement in the model fit compared to the single-factor model.
Chi-square (c2) test statistics provide a measure of discrepancy between the measurement model and data, however, they are also overly sensitive to large sample sizes, sometimes even resulting in the rejection of models (Hoyle, 2011) with a good fit. CFI is a fit index that can be used in combination with chi-square (c2) test statistics to minimize this risk (Hoyle, 2011). Furthermore, a combination of the RMSEA and SRMR was used to minimize the risk of
rejection of well-fitting models. Therefore, although the chi-square (c2) was significant in this context, both the single-factor and twofactor models were considered the best fit to the data. The comparison of the two-factor and single-factor models revealed the two-factor model fits the data significantly better than the singlefactor model does, both in the original two-factor model (Dc2 17.61, df ¼ 1, P < 0.001) (Saboonchi et al., 2013; Zigmond and Snaith,
20
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
1983; Mystakidou et al., 2004) and the modified two-factor model (Dc2 34.76, df ¼ 1, P < 0.001) (Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006). Fig. 1 shows the factor structure of the CHADS for patients. The correlation between the C-HADS anxiety and C-HADS depression subscales indicated that the C-HADS subscales were positively correlated (r ¼ 0.83, and 0.82 for the original two-factor model and the modified two-factor model respectively). For family caregivers (the lower part of Table 2), the same profile of model fit was displayed in family caregivers as in cancer patients, except that the model fit in both the single-factor and two-factor models indicated adequate fit to the data. CFI values ranging from 0.93 to 0.94 were greater than the cutoff values of 0.90 for adequacy (Hoyle, 2011). The modified two-factor model also displayed significant improvement in fit (Dc2 20.21, df ¼ 1, P < 0.001) compared to the single-factor model. Fig. 2 shows the factor structure of the CHADS for family caregivers. The correlation between the C-HADS anxiety and C-HADS depression subscales also indicated that the CHADS subscales were positively correlated (r ¼ 0.84, and 0.81 for the original two-factor model and the modified two-factor model respectively). 3.3. Internal consistency reliability Cronbach's alpha coefficients for each subscale are displayed in Table 3. As is shown, the internal consistency of the C-HADS total, C-HADS anxiety and depression in both cancer patients and family caregivers displayed overall acceptable levels with all of the Cronbach's alpha 0.840. 3.4. Concurrent validity The concurrent validity of the C-HADS was established by the significant negative correlation with the relevant subscales of the SF-12 (Table 4). The overall significant correlations were found among the C-HADS, including C-HADS total, C-HADS anxiety and depression, and all the C-SF-12 subscales both in patients (the upper part of Table 4) and family caregivers (the lower part of Table 4) (all ps < 0.001). There were significantly moderate
correlations between the C-HADS total, C-HADS anxiety and depression, and the C-SF-12 subscales pertaining to mental health, such as role emotional (RE), mental health (MH), and mental component summary (MCS) both in cancer patients (r ¼ 0.40e0.55) and in family caregivers (r ¼ 0.41e0.53). The same correlation profile was displayed in the original two-factor model as in the modified two-factor model. The negative correlation coefficients indicated that higher levels of psychological distress were related to poorer health status. 4. Discussion To our knowledge, the present study is the first attempt in mainland China to assess anxiety and depression using C-HADS, and to report its psychometric properties from a dyadic perspective. For a better representation of the cancer population, participants were recruited from seven different administrative regions in China. The study results revealed that the C-HADS has satisfactory psychometric properties for use among both cancer patients and their family caregivers. The following discussion on the psychometric properties of the C-HADS mainly focuses on three aspects: construct validity, internal consistency reliability, and concurrent validity. Findings of the construct validity suggest the original two-factor model of C-HADS displays a best fit to the data for cancer patients, and an adequate fit to the data for family caregivers. These results support the basic assumption of two-factor construct of C-HADS. However, it should be noted there was a strong inter-correlation between anxiety and depression in both cancer patients (r ¼ 0.83) and family caregivers (r ¼ 0.84), indicating substantial covariance in the severity of anxiety and depression. This high inter-correlation between anxiety and depression was consistent with other studies using HADS in adult cancer patient populations (r ¼ 0.52e0.81) (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006), and family caregivers of cancer patients (r ¼ 0.63) (Gough and Hudson, 2009). As proposed by Saboonchi et al. (2013), “in order for HADS to demonstrate sound construct validity in cancer patients, such an inter-correlation
Fig. 1. The factor structure of the C-HADS for patients.
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
21
Fig. 2. The factor structure of the C-HADS for family caregivers.
Table 3 Cronbach's a for the subscales of C-HADS in cancer patients and family caregivers, total n ¼ 641.a Variables
Patients
Single-factor C-HADS total Original two-factor C-HADS-Anxiety C-HADS-Depression Modified two-factor C-HADS-Anxiety C-HADS-Depression
Family caregivers Percent
Cronbach's a
na
Percent
Cronbach's a
619
96.6
0.929
621
96.9
0.919
626 631
97.7 98.4
0.874 0.874
628 631
98.0 98.4
0.857 0.851
630 628
98.3 98.0
0.855 0.879
629 630
98.1 98.3
0.840 0.873
n
a
C-HADS¼Chinese version of hospital anxiety and depression scale. a The total n does not equal 641 because of missing values.
needs to be pre-specified” (pp. 2853). Zigmond and Snaith (1983) assume that a strong association between the subscales would indicate “they could be considered much the same thing, for example, emotional disturbance” (pp 364). This assumption resonates with the single-factor models of HADS from Razavi et al. (1990) and Smith et al. (2006). Our results further support the one-factorial measurement model of C-HADS by displaying a best fit to the data for cancer patients and an adequate fit to the data for family caregivers, hypothesizing a single factor of emotional distress. It is worth noting that the Chi-Square (c2), CFI, RMSEA and SRMR statistics revealed that the two-factor models tested offered acceptable fits to the data, with modified two-factor models (Smith et al., 2002; Moorey et al., 1991,Muszbek et al., 2006) offering a slightly better fit to the data than the original two-factor models (Saboonchi et al., 2013; Zigmond and Snaith, 1983; Mystakidou et al., 2004) in both cancer patients and family caregivers. This is a reminder that the modified two-factor model is the most suitable to use for Chinese cancer patients and their family caregivers. However, as the HADS remains a widely and easily used clinical instrument, it is suggested that HADS remain in its original twofactor model, to be scored and interpreted in the original
recommended manner (Zigmond and Snaith, 1983). This may benefit the intercultural adaptation of the instrument, and enable cross-cultural evaluation of psychological distress in cancer patients and their family caregivers. Regarding the high heterogeneity of the construct validity of HADS in a cancer population, considering the findings that both the two-factor and one-factor models offered an acceptable fit to the data in the present study, it is recommended that researchers incorporate both the total HADS score and the two subscale scores (Cosco et al., 2012). The HADS total, HADS anxiety and depression were reported in the present study. Study results indicated good internal consistency and reliability for the C-HADS with all of the Cronbach's a for the C-HADS total and two subscales greater than 0.851 in both cancer patients and family caregivers. The internal consistency of the C-HADS scale in both cancer patients and family caregivers corresponds well with those previously reported in other studies using the same instrument in adult cancer patient populations, with Cronbach's alpha for HADS anxiety varying from 0.73 to 0.89, and for HADS depression from 0.70 to 0.87 (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Mystakidou et al., 2004), and among cancer caregivers (Cronbach's alpha ¼ 0.89, 0.85, and 0.8 for HADS total, anxiety and depression respectively) (Gough and Hudson,
22
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
Table 4 Pearson correlations between C-HADS and C-SF-12 in patients and family caregivers (n ¼ 585e630).a C-SF-12 subscales of patients
C-HADS subscales of patients The original two-factor model
The modified two-factor model
Total
Anxiety
Depression
Anxiety
Depression
Physical Functioning Role Physical Bodily Pain General Health Vitality Role Emotional Social Functioning Mental Health PCS MCS
0.31b 0.34b 0.31b 0.25b 0.17b 0.44b 0.38b 0.52b 0.32b 0.55b
0.28b 0.32b 0.28b 0.24b 0.17b 0.40b 0.36b 0.50b 0.29b 0.53b
0.32b 0.34b 0.32b 0.25b 0.17b 0.42b 0.37b 0.50b 0.33b 0.52b
0.28b 0.32b 0.26b 0.22b 0.16b 0.40b 0.35b 0.48b 0.28b 0.51b
0.32b 0.34b 0.32b 0.27b 0.18b 0.42b 0.37b 0.51b 0.33b 0.53b
C-SF-12 subscales of family caregivers
C-HADS subscales of family caregivers
Physical Functioning Role Physical Bodily Pain General Health Vitality Role Emotional Social Functioning Mental Health PCS MCS
The original two-factor model
The modified two-factor model
Total
Anxiety
Depression
Anxiety
Depression
0.32b 0.45b 0.32b 0.29b 0.26b 0.51b 0.38b 0.45b 0.35b 0.53b
0.30b 0.42b 0.34b 0.25b 0.24b 0.49b 0.37b 0.45b 0.33b 0.51b
0.29b 0.44b 0.28b 0.30b 0.26b 0.48b 0.36b 0.42b 0.34b 0.50b
0.31b 0.41b 0.34b 0.24b 0.21b 0.49b 036b 0.44b 0.33b 0.50b
0.29b 0.44b 0.28b 0.30b 0.27b 0.47b 0.36b 0.43b 0.34b 0.51b
C-HADS¼Chinese version of hospital anxiety and depression scale; C-SF-12 ¼ Chinese version of the Medical Outcomes Study 12-item Short Form; PCS¼Physical Component Summary; MCS ¼ Mental Component Summary. a The total n does not equal 641 because of missing values in HADS and SF-12. b Correlation is significant at the 0.001 level (2-tailed).
2009). These findings highlight the satisfactory internal consistency of the HADS. Significant negative correlations between the C-HADS total, anxiety and depression scores and the scores of C-SF-12 subscales both in cancer patients and family caregivers confirmed the concurrent validity of the C-HADS. Similar results were reported in China (Li et al., 2015), as well as in other countries (Janda et al., ttir et al., 2011; Saevarsdottir et al., 2010). 2007; Fridriksdo
two-factor subscale scores be reported. Future research is needed to explore the relationship between the anxiety and depression of cancer patients and that of their family caregivers, and to systematically investigate the related factors or predictors of anxiety and depression for cancer patients and their family caregivers. This would benefit the interventions to be developed for relieving their psychological distress and improving QOL. Funding
5. Limitations The present study makes new contributions to our understanding of C-HADS in Chinese cancer patients and their family caregivers when coping with cancer as a dyad; however, it is essential to acknowledge several possible limitations. A limitation of this validation study was the inability to validate test-retest reliability, due to the cross-sectional study design. Moreover, the fact that the study population was only comprised of Chinese advanced cancer patients and their family caregivers could limit the generalizability of the results to other targeted populations in different cultures. Future studies targeting different demographic and medical cancer populations should be explored. 6. Conclusion Notwithstanding these limitations, the present study provides the satisfactory psychometric properties of C-HADS when applied to a sample of Chinese advanced cancer patients and their family caregivers. Findings indicate that the Chinese version of the HADS is a reliable and valid measure of anxiety and depression. The finding supports the basic assumption of two-factor construct of HADS. It is recommended that both the total HADS score and the
Financial support of this study was provided by the National Natural Science Foundation of China (No. 81573250). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflict of interest The authors declare no conflicts of interest to disclose. Acknowledgments: The authors gratefully acknowledge the support from all the related hospitals and all the participants for their sharing of their experience in this study. References Arving, C., Glimelius, B., Brandberg, Y., 2008. Four weeks of daily assessments of anxiety, depression and activity compared to a point assessment with the Hospital Anxiety and Depression Scale. Qual. Life Res. 17, 95e104. Bjelland, I., Dahl, A.A., Haug, T.T., Neckelmann, D., 2002. The validity of the Hospital Anxiety and Depression Scale: an updated literature review. J. Psychosom. Res. 52, 69e77. € de n, P., Sullivan, M., 1992. Anxiety Brandberg, Y., Bolund, C., Sigurdardottir, V., Sjo
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23 and depressive symptoms at different stages of malignant melanoma. Psycho Oncol. 1, 71e78. Carlson, L.E., Angen, M., Cullum, J., Goodey, E., Koopmans, J., Lamont, L., MacRae, J.H., Martin, M., Pelletier, G., Robinson, J., Simpson, J.S.A., Speca, M., Tillotson, L., Bultz, B.D., 2004. High levels of untreated distress and fatigue in cancer patients. Br. J. Cancer 90, 2297e2304. Cosco, T.D., Doyle, F., Ward, M., McGee, H., 2012. Latent structure of the hospital anxiety and depression scale: a 10-year systematic review. J. Psychosom. Res. 72, 180e184. Fletcher, B., Miaskowski, C., Given, B., Schumacher, K., 2012. The cancer family caregiving experience: an updated and expanded conceptual model. Eur. J. Oncol. Nurs. 16, 387e398. ttir, N., Sævarsdo ttir, P., Halfd ttir, S.I., Jo nsdo ttir, A., Fridriksdo anardo ttir, H., Olafsd ttir, K.L., Gudmundsdo ttir, G., Gunnarsdo ttir, S., 2011. Magnúsdo o Family members of cancer patients: needs, quality of life and symptoms of anxiety and depression. Acta Oncol. 50, 252e258. Gough, K., Hudson, P., 2009. Psychometric properties of the hospital anxiety and depression scale in family caregivers of palliative care patients. J. Pain Symptom Manag. 37, 797e806. Grassi, L., Indelli, M., Marzola, M., Maestri, A., Santini, A., Piva, E., Boccalon, M., 1996. Depressive symptoms and quality of life in home-care-assisted cancer patients. J. Pain Symptom Manag. 12, 300e307. Guillemin, F., Bombardier, C., Beaton, D., 1993. Cross-cultural adaptation of healthrelated quality of life measures: literature review and proposed guidelines. J. Clin. Epidemiol. 46, 1417e1432. Hagedoorn, M., Sanderman, R., Bolks, H.N., Tuinstra, J., Coyne, J.C., 2008. Distress in couples coping with cancer: a meta-analysis and critical review of role and gender effects. Psychol. Bull. 134, 1e30. Haley, W., 2003. The costs of family caregiving: implications for geriatric oncology. Crit. Rev. Oncol. Hematol. 48, 151e158. Herrmann, C., 1997. International experiences with the Hospital Anxiety and Depression Scaleea review of validation data and clinical results. J. Psychosom. Res. 42, 17e41. Hooper, D., Coughlan, J., Mullen, M., 2008. Structural equation modeling: guidelines for determining model fit. Electron. J. Bus. Res. Methods 6, 53e60. Hoyle, R.H., 2011. Structural Equation Modeling for Social and Personality Psychology, 1 ed. SAGE Publications Ltd. Janda, M., Steginga, S., Langbecker, D., Dunn, J., Walker, D., Eakin, E., 2007. Quality of life among patients with a brain tumor and their carers. J. Psychosom. Res. 63, 617e623. Kayser, K.P., Watson, L.E.M., Licsw, Andrade, J.T.M., 2007. Cancer as a “We-Disease”: examining the process of coping from a relational perspective. Fam. Syst. Health 25, 404e418. Kim, Y., Kashy, D.A., Wellisch, D.K., Spillers, R.L., Kaw, C.K., Smith, T.G., 2008. Quality of life of couples dealing with cancer: dyadic and individual adjustment among breast and prostate cancer survivors and their spousal caregivers. Ann. Behav. Med. 35, 230e238. Leung, C.M., Ho, S., Kan, C.S., Hung, C.H., Chen, C.N., 1993. Evaluation of the Chinese version of the hospital anxiety and depression scale. A cross-cultural perspective. Int. J. Psychosom. 40, 29e34. Li, Q., Lin, Y., Qiu, Y., Gao, B., Xu, Y., 2014. The assessment of health-related quality of life and related factors in Chinese elderly patients undergoing chemotherapy for advanced cancer: a cross-sectional study. Eur. J. Oncol. Nurs. 18, 425e435. Li, Q., Xu, Y., Zhou, H., Loke, A.Y., 2015. Testing a preliminary live with love conceptual framework for Cancer couple dyads: a mixed-methods study. Eur. J. Oncol. Nurs. 19, 619e628. Li, Q., Loke, A.Y., 2013. A spectrum of hidden morbidities among spousal caregivers for patients with cancer, and differences between the genders: a review of the literature. Eur. J. Oncol. Nurs. 17, 578e587. Lloyd-Williams, M., Friedman, T., Rudd, N., 2001. An analysis of the validity of the hospital anxiety and depression scale as a screening tool in patients with advanced metastatic Cancer. J. Pain Symptom Manag. 22, 990e996.
23
pez-Go mez, M., Malmierca, E., de Go rgolas, M., Casado, E., 2013. Cancer in Lo developing countries: the next most preventable pandemic. The global problem of cancer. Crit. Rev. Oncol. 88, 117e122. Mitchell, A.J., 2010. Short screening tools for cancer-related distress: a review and diagnostic validity meta-analysis. J. Natl. Compr. Canc Netw. 8, 487e494. Mitchell, A.J., Meader, N., Symonds, P., 2010. Diagnostic validity of the hospital anxiety and depression scale (HADS) in cancer and palliative settings: a metaanalysis. J. Affect. Disord. 126, 335e348. Moorey, S., Greer, S., Watson, M., Gorman, C., Rowden, L., Tunmore, R., Robertson, B., Bliss, J., 1991. The factor structure and factor stability of the hospital anxiety and depression scale in patients with cancer. Br. J. Psychiatry 158, 255e259. nszky, M., Ruzsa, A., Muszbek, K., Szekely, A., Balogh, E.M., Moln ar, M., Roha €llosi, M., Vada sz, P., 2006. Validation of the Hungarian translation Varga, K., Szo of hospital anxiety and depression scale. Qual. Life Res. 15, 761e766. Mystakidou, K., Tsilika, E., Parpa, E., Katsouda, E., Galanos, A., Vlahos, L., 2004. The hospital anxiety and depression scale in Greek cancer patients: psychometric analyses and applicability. Support. Care Cancer 12, 821e825. Northouse, L.L., Mood, D., Templin, T., Mellon, S., George, T., 2000. Couples' patterns of adjustment to colon cancer. Soc. Sci. Med. 50, 271e284. Portney, L.G., Watkins, M.P., 2009. Foundations of Clinical Research, 3 ed. Pearson/ Prentice Hall, London. Razavi, D., Delvaux, N., Farvacques, C., Robaye, E., 1990. Screening for adjustment disorders and major depressive disorders in cancer in-patients. Br. J. Psychiatry 156, 79e83. Rodgers, J., Martin, C.R., Morse, R.C., Kendell, K., Verrill, M., 2005. An investigation into the psychometric properties of the Hospital Anxiety and Depression Scale in patients with breast cancer. Health Qual. Life Outcomes 3, 41e41. Saboonchi, F., Wennman-Larsen, A., Alexanderson, K., Petersson, L.M., 2013. Examination of the construct validity of the swedish version of hospital anxiety and depression scale in breast cancer patients. Qual. Life Res. 22, 2849e2856. Saevarsdottir, T., Fridriksdottir, N., Gunnarsdottir, S., 2010. Quality of life and symptoms of anxiety and depression of patients receiving cancer chemotherapy: longitudinal study. Cancer Nurs. 33, E1eE10. Satin, J.R., Linden, W., Phillips, M.J., 2009. Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis. Cancer 115, 5349e5361. Schreier, A.M., Williams, S.A., 2004. Anxiety and quality of life of women who receive radiation or chemotherapy for breast cancer. Oncol. Nurs. Forum 31, 127e130. Smith, A.B., Wright, E.P., Rush, R., Stark, D.P., Velikova, G., Selby, P.J., 2006. Rasch analysis of the dimensional structure of the hospital anxiety and depression scale. Psycho Oncol. 15, 817e827. Smith, A.B., Selby, P.J., Velikova, G., Stark, D., Wright, E.P., Gould, A., Cull, A., 2002. Factor analysis of the hospital anxiety and depression scale from a large cancer population. Psychol. Psychother. 75, 165e176. Statistics. Health, 2015. Tools of the Trade: Strategy for Determing Sample Size. Retrieved from: http://www.statistics.health.pa.gov/StatisticalResources/ UnderstandingHealthStats/ToolsoftheTrade/Documents/Strategy_for_ Determining_Sample_Size.pdf. Wang, W., Chair, S.Y., Thompson, D.R., Twinn, S.F., 2009. A psychometric evaluation of the Chinese version of the Hospital Anxiety and Depression Scale in patients with coronary heart disease. J. Clin. Nurs. 18, 2436e2443. Ware, J.E., Kosinski, M., Turner-Bowker, D.M., Gandek, B., 2002. How to Score Version 2 of the SF-12 Health Survey (With a Supplement Documenting Version 1). QualityMetric Incorporated. Ware Jr., J., Kosinski, M., Keller, S.D., 1996. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med. Care 34, 220e233. WHO, 2015. Cancer Fact Sheet N 297. 2015. Retrieved from: http://www.who.int/ mediacentre/factsheets/fs297/en/index.html. Zigmond, A., Snaith, R., 1983. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 67, 361e370.
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
The Chinese version of hospital anxiety and depression scale: Psychometric properties in Chinese cancer patients and their family caregivers Qiuping Li a, Yi Lin a, Caiping Hu b, Yinghua Xu c, *, Huiya Zhou c, Liping Yang d, Yongyong Xu e, ** a
Wuxi Medical School, Jiangnan University, Wuxi, Jiangsu Province, China Shanxi Provincial Tomor Hospital, Taiyuan, Shanxi Province, China Wuxi People's Hospital, Wuxi, Jiangsu Province, China d Xijing Hospital, Xi'an, Shaanxi Province, China e Department of Health Statistics, Fourth Military Medical University, Xi'an, Shaanxi Province, China b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 30 January 2016 Received in revised form 29 August 2016 Accepted 9 September 2016
Purpose: The Hospital Anxiety and Depression Scale (HADS) acts as one of the most frequently used selfreported measures in cancer practice. The evidence for construct validity of HADS, however, remains inconclusive. The objective of this study is to evaluate the psychometric properties of the Chinese version HADS (C-HADS) in terms of construct validity, internal consistency reliability, and concurrent validity in dyads of Chinese cancer patients and their family caregivers. Methods: This was a cross-sectional study, conducted in multiple centers: one hospital in each of the seven different administrative regions in China from October 2014 to May 2015. A total of 641 dyads, consisting of cancer patients and family caregivers, completed a survey assessing their demographic and background information, anxiety and depression using C-HADS, and quality of life (QOL) using Chinese version SF-12. Data analysis methods included descriptive statistics, confirmatory factor analysis (CFA), and Pearson correlations. Results: Both the two-factor and one-factor models offered the best and adequate fit to the data in cancer patients and family caregivers respectively. The comparison of the two-factor and single-factor models supports the basic assumption of two-factor construct of C-HADS. The overall and two subscales of CHADS in both cancer patients and family caregivers had good internal consistency and acceptable concurrent validity. Conclusions: The Chinese version of the HADS may be a reliable and valid screening tool, as indicated by its original two-factor structure. The finding supports the basic assumption of two-factor construct of HADS. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Anxiety Depression Hospital anxiety and depression scale Psychometric properties Cancer Family caregivers Caregiver-patient dyads Chinese
1. Introduction As the leading cause of morbidity and mortality worldwide, cancers reached up to 14.1 million new cases, with 8.2 million cancer-related deaths in 2012. Of these, 21% (3 million) of new cancer cases, and 27% (2.2 million) of cancer deaths occurred in China (WHO, 2015). It is unfortunate that in developing countries,
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (Y. Xu), [email protected] (Y. Xu). http://dx.doi.org/10.1016/j.ejon.2016.09.004 1462-3889/© 2016 Elsevier Ltd. All rights reserved.
where most new cancer cases are frequently diagnosed at an advanced stage, treatment options are both limited and expensive pez-Go mez et al., 2013). With the increase in cancer cases, there (Lo is a need for a similar number of family caregivers, who are expected to provide care or support to cancer patients. Both cancer patients and their family caregivers need to cope together and adjust to the challenge of the profound emotional and social adversity imposed by a cancer diagnosis and its treatment (Kayser et al., 2007). Accumulating evidence has shown that psychological distress represents a significant adversity in cancer populations (Saboonchi
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
et al., 2013; Satin et al., 2009,Carlson et al., 2004) and their family caregivers (Li and Loke, 2013,Haley, 2003). Anxiety and depression - the most common presenting symptoms of psychological distress (Rodgers et al., 2005) - are reported to be prevalent in approximately one-third, or as many as 45% of cancer patients (Rodgers et al., 2005,Schreier and Williams, 2004,Grassi et al., 1996) and ttir et al., 2011). their family caregivers (Janda et al., 2007,Fridriksdo It has been reported that the psychological distress of family caregivers could be as high or even higher, than that of cancer patients themselves (Hagedoorn et al., 2008; Northouse et al., 2000). Anxiety and depression also exert a significant impact on quality of life (QOL) in both cancer patients (Grassi et al., 1996,Li et al., 2014; Saevarsdottir et al., 2010), and their family caregivers ttir et al., 2011; Kim et al., 2008; Northouse et al., 2000). (Fridriksdo Evidence from a meta-analysis concluded that depression, in particular, constitutes a predictor of mortality in cancer patients (Satin et al., 2009). Studies have also shown there is a mutual impact between dyads of cancer patients and family caregivers in terms of QOL and psychological distress (Kim et al., 2008; Northouse et al., 2000). Evidence also indicates that cancer affects caregiver-patient dyads as a unit, rather than as isolated individuals (Hagedoorn et al., 2008), leading to the primary focus of cancer care research to shift from the individual experiences of cancer patients or family caregivers, to the dyadic level of caregiver-patient dyads (Fletcher et al., 2012). For a better understanding of the related experiences of caregiver-patient dyads from the dyadic level, the participants in the present study included both cancer patients and their family caregivers. The growing recognition of the common prevalence of anxiety and depression, and the significant impact on the lives of cancer patients and their family caregivers, highlights the need for valid assessment and screening methods for anxiety and depression in cancer practice (Saboonchi et al., 2013). Self-report questionnaires appear to be specifically appropriate, and a practical tool in this context (Mitchell, 2010). The Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, 1983) stands out as one of the most frequently used self-reported measures, which is considered an effective screening measure for both anxiety and depression, and has been widely used across a variety of cancer populations and family caregivers (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Gough and Hudson, 2009; Mitchell et al., 2010). In terms of the instrument's psychometric properties, although internal consistency reliability with Cronbach's alpha for HADS anxiety varied from 0.68 to 0.93 (mean 0.83), and for HADS depression from 0.67 to 0.90 (mean 0.82), this suggests the instrument has good reliability and shows a capacity to consistently capture reliable data (Bjelland et al., 2002); however, the evidence for construct validity, which is based on analysis for instrument dimensionality, remains inconclusive. According to a 10-year systematic review of the latent structure of the HADS, the largest degree of heterogeneity of construct validity occurs in studies of cancer populations (Cosco et al., 2012). The heterogeneity of the factorial structure of HADS in cancer populations consists of singlefactor (Smith et al., 2006; Razavi et al., 1990), two-factor (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006), three-factor (Rodgers et al., 2005; Brandberg et al., 1992), and four-factor structure (Lloyd-Williams et al., 2001). Given that the original English version of HADS has been translated into and validated in many different languages (Herrmann, 1997), including Chinese (Leung et al., 1993; Wang et al., 2009), the measurement properties of the HADS, such as the inconclusiveness of the construct validity, could be affected by cross-lingual and cross-cultural adaptation processes (Guillemin
17
et al., 1993). Although several studies have been conducted to validate the factorial structure of the Chinese version of HADS (CHADS), including in a sample of medical students (Leung et al., 1993) and in coronary heart disease (Wang et al., 2009), no studies have been conducted in mainland China, to our knowledge, to validate its psychometric properties in cancer patients and their family caregivers. Consequently, this study's aim was to evaluate the psychometric properties of the C-HADS from a dyadic perspective. To be specific, to evaluate the psychometric properties of the C-HADS in terms of construct validity, internal consistency reliability, and concurrent validity in dyads of Chinese cancer patients and their family caregivers. This study provides the psychometric properties of the CHADS when applied to a sample of Chinese cancer patients and their family caregivers. 2. Methods 2.1. Study design and participants This was a cross-sectional study, conducted in multiple centers: one hospital in each of the seven different administrative regions in China. The seven administrative regions cover different areas of China, and one of the high-ranking hospitals within each of the administrative regions was recruited by convenience sampling. The seven regions were: (i) East China; (ii) Southern China; (iii) North China; (iv) Central China; (v) Northwest China; (vi) Southwest China; and (vii) Northeast China. Participants consisted of 641 cancer patient and their family caregiver dyads, and were recruited by convenience sampling from October 2014 to May 2015. The study criteria inclusions were as follows: (i) dyads of Chinese adult cancer patients and family caregivers (age >18 years old); (ii) a medical diagnosis of any type of cancer in patients, who had no other diseases, such as dementia, which could lead to unconsciousness; (iii) a primary family caregiver who provides informal care to cancer patients; (iv) both patients and their family caregivers could communicate in Mandarin-Chinese, and consent to take part in the study. Sample size calculation: the sample size in the HADS evaluation was calculated by n¼ (uas/d)2 and n¼ (uas/d)2z4 s2, given a ¼ 0.05 and d ¼ ±1, the error in estimation of population means for HADS -Total, HADS-Anxiety and HADS- Depression. s2 was estimated separately by deff s2 Total, deff s2 Anxiety and deff s2 Depression (Statistics. health, 2015). From our pilot study in a region, the estimated value of s2 Total was 9.5, and the value of s2 Anxiety and s2 Depression was 5. The design effect (deff) for sampling from different regions was given as 1.5. The sample size for estimating HADS -Total was 4 1.5 9.52 ¼ 541, and for estimating HADSAnxiety and HADS- Depression it was 4 1.5 52 ¼ 150. One hundred more cases were added to the actual survey in case of no response and missing data; the final sample size was 641. The sample sizes for each subgroup of HADS-Anxiety and HADSDepression should be no fewer than 150 cases. To simplify, the sample size of the present study is large enough to ensure the statistical power for doing confirmatory factor analysis (CFA). 2.2. Instruments Three groups of variables were collected: socio-demographic characteristics and clinical data, the Chinese version of HADS (CHADS) for anxiety and depression (Zigmond and Snaith, 1983), and the Chinese version of Medical Outcomes Study 12-item Short Form (C-SF-12) (version 2) for QOL. The purpose of using the C-SF-12 was to examine the concurrent validity of the C-HADS. Information on socio-demographic characteristics and clinical
18
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
data from both patients and their family caregivers was solicited through a demographic and background information sheet. The solicited socio-demographic and clinical data included age, gender, marital status, caregiver's relationship with the patient, education level, working status, cancer type, cancer stage, time since cancer diagnosis, and financial burden on the family due to medical treatment costs. The HADS contains 14 items making up two 7-item subscales, one measuring anxiety and the other depression (Zigmond and Snaith, 1983). A four-point Likert-type scale ranging from 0 (not a problem) to 3 (high level of problems) is used for scoring each item. The score for each subscale is obtained by simple summation of individual items with scores in each subscale (anxiety and depression) ranging from 0 to 21. A score of seven or less indicates a non-anxiety/depression case, 8e10 a borderline case of anxiety/ depression, and 11 or above a definite case of anxiety/depression (Arving et al., 2008). The SF-12 consists of 12 items making up eight scales that measure the following eight QOL domains (Ware et al., 1996): Physical Functioning (PF), Role-Physical (RP), Bodily Pain (BP), General Health (GH), Vitality (VT), Role-Emotional (RE), Social Functioning (SF), and Mental Health (MH). The eight SF-12 domains hypothetically form two dimensions of the Physical Component Summary (PCS) and the Mental Component Summary (MCS) (Ware et al., 1996). The eight scales and two dimensions were transformed and calculated according to the SF-12 (version 2) score manual to a scale with a theoretical range of 0e100 (Ware et al., 2002). A higher score indicates a higher level of QOL. The two dimensions (PCS and MCS) of the SF-12 version 2 achieved R squares of 0.905 with PCS and 0.938 with MCS when the SF-36 was used in a cross-validated Medical Outcomes Study. Test-retest (2-week) correlations of 0.89 and 0.76 were observed for the SF-12 PCS and MCS respectively, in the general U.S. population (n ¼ 232) (Ware et al., 1996). The internal consistency of the present study for the total SF-12 was Cronbach's a ¼ 0.821 (n ¼ 596) and 0.812 (n ¼ 593) for cancer patients and family caregivers respectively. 2.3. Data collection procedure Before study commencement, ethical approval was granted by the research ethics committee of Jiangnan University (HSEARS20140701001), and access approval was obtained from the related hospitals. Nurses working in the oncology unit in the related hospitals were provided with explanations of both the study and the instruments, prior to the study commencing. The research was conducted with utmost respect for participants, and the confidentiality of the collected data was held in accordance with the Helsinki Declaration of 1975, as revised in 2000. The hospital oncologists identified the cancer dyads in accordance with the eligibility criteria. Cancer patient-family caregiver dyads who met the criteria for inclusion were approached in the oncology wards when they were admitted for chemotherapy treatment during the recovery stage. After their written informed consent was obtained, the dyad participants were invited to complete the questionnaire separately, with the help of a nurse if required. According to their preference, they filled out the questionnaire in a private in-patient room or in a nurse's office, so their privacy could be protected and they would be removed from any possible disturbances. The questionnaire took approximately 10e15 min to complete. None of the study participants reported discomfort or distress. 2.4. Data analysis Data were analyzed using SPSS version 21.0. Descriptive
statistics such as frequencies, percentages, means, and standard deviations were used to describe subject characteristics. The C-HADS factor structure was determined using confirmatory factor analysis (CFA) based on Maximum Likelihood, using Amos version 21.0. Six models developed from HADS validity and psychometric studies in the sample of cancer population were tested. The six models were: (i) single-factor model: the same factor structure of single factor model from Razavi et al. (1990) and Smith et al. (2006); (ii) the original two-factor model (Saboonchi et al., 2013; Mystakidou et al., 2004); (iii) the modified two-factor model (Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Gough and Hudson, 2009); (iv) three-factor model 1 (Brandberg et al., 1992); (v) three-factor model 2 (Rodgers et al., 2005); (vi) and four-factor model (Lloyd-Williams et al., 2001). Four indices were used to evaluate the overall fit of CFA models, namely: Chi-Square c2, a confirmatory fit index (CFI), a root mean square error of approximation (RMSEA), and the standardized root mean squared residual (SRMR). Generally, an insignificant p value (p > 0.05) in the Chi-Square c2, CFI greater than 0.95, both a RMSEA and an SRMR of less than 0.08 are considered indicative of good model fit (Hooper et al., 2008). The competing models were compared by means of likelihood ratio (Dc2) test. The internal reliability was evaluated by Cronbach's coefficient alpha. Cut-offs for a good reliability for Cronbach's alpha was set at 0.75 (Portney and Watkins, 2009). The concurrent validity of the C-HADS was examined by comparing it with the C-SF-12 to determine whether it was psychometrically sound for measuring its respective attributes. A significant negative correlation between the measurements indicates good concurrent validity (Portney and Watkins, 2009). 3. Results 3.1. Participant characteristics Out of a total 750 approached eligible cancer patient and family caregiver dyads, 42 dyads declined to participate in the study. The remaining 708 dyads became our participants and completed the questionnaire (response rate ¼ 94.4%). Of those, 641 dyads provided valid data for study variables (valid rate ¼ 90.5%). Table 1 summarizes the dyad characteristics. In this study, about half of the patients (50.1%) and family caregivers (48.5%) were male. Approximately 90% of patients (91.4%) and family caregivers (89.1%) were married, and more than half of family caregivers (53.5%) were spouses of cancer patients. All patients were at an advanced stage of cancer, namely, stage III (n ¼ 364, 56.8%) or stage IV (n ¼ 277, 43.2%). Nearly half of the dyads (46.3%) were coping with digestive system cancer (e.g., esophageal, gastric, liver, or colorectal cancer). All patients were receiving chemotherapy, and 102 (15.9%) patients had had surgery. Approximately two-thirds of the families (64.6%) experienced serious financial burden due to the cost of treating the cancer. 3.2. Construct validity Of the six models included in the analysis, the solution for the two three-factor models is not admissible; no related c2 value was calculated; the four-factor model cannot be converged in either sample of cancer patients or family caregivers. The results of the CFA fit statistics of the remaining three models of single-factor and two-factor models are presented in Table 2. For cancer patients (the upper part of Table 2), both the single-factor and two-factor models offered the best fit for the data. In the three best fit models, although the chi-square (c2) was significant, other fit indices had acceptable values.
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
19
Table 1 Characteristics of patients and family caregivers (n ¼ 641). Characteristics
Patients [ n (%)]*
FC [ n (%)]*
Age (mean ± SD), years Gender Male Female Marital status Married Not married** FCs' relationship with patients Spouses Offspring Parents Sibling Others Education levels Primary school or less High school University or above Working status Working Not working Type of cancer*** Breast cancer Ovarian and cervical cancer Esophageal and gastric cancer Colorectal cancer Liver cancer Lung cancer Others Type of treatment Chemotherapy Chemotherapy þ surgery Chemotherapy þ radiotherapy Chemotherapy þ others Average time since diagnosis/duration in their role as a FC
54.6 ± 12.9(ranging from 18 to 88)
46.6 ± 13.2(ranging from 18 to 79)
321(50.1) 318(49.6)
311(48.5) 329(51.3)
586(91.4) 55(8.6)
571(89.1) 69(10.8)
357(55.7) 102(15.9) 114(17.8) 61(9.5) 13.3 ± 24.0 months (ranging from 1 to 228 months)
Financial burden on the family due to cancer treatment
Serious: 414 (64.6); Normal: 191 (29.8); Mild or None: 33 (5.1)
343(53.5) 215(33.5) 20(3.1) 41(6.4) 21(3.3) 356(55.5) 198(30.9) 85(13.3)
256(40.0) 222(34.6) 159(24.8)
373(58.2) 263(41.0)
411(64.1) 227(35.4)
70(10.9) 93(14.5) 152(23.7) 86(13.4) 59(9.2) 88(13.7) 76(11.9)
<6 months: 348(54.3) 6 monthse2 years: 192(30.0) >2 yearse5 years: 55 (8.6) >5 years: 39(6.1)
Note: FC¼ Family caregivers; SD ¼ standard deviation; * The total n does not equal 641 because of missing values; ** Not married include: Divorced, Widowed, and Never married; *** All of the patients had advanced cancer, were in stage III (n ¼ 364, 56.8%) and stage IV (n ¼ 277, 43.2%).
Table 2 Comparisons of goodness-ofefit indicators of nested single-factor and two-factor models of C-HADS. Model Patients Single-factor Original two-factor D Dc2a Modified two-factor D Dc2b Family caregivers Single-factor Original two-factor Dc2a Modified two-factor D Dc2b
FLI 1
FLI 2
All items 1,3,5,7,9,11,13
2,4,6,8,10,12,14
1,3,5,9,11,13
2,4,6,7,8,10,12,14
All items 1,3,5,7,9,11,13
2,4,6,8,10,12,14
1,3,5,9,11,13
2,4,6,7,8,10,12,14
c2
df
CFI
RMSEA
SRMR
283.11* 265.50* 17.61* 248.35* 34.76*
77 76 1 76 1
0.954 0.958
0.066 0.064
0.035 0.034
0.962
0.061
0.032
358.80* 358.00* 0.80 338.59* 20.21*
77 76 1 76 1
0.930 0.932
0.077 0.077
0.041 0.041
0.935
0.075
0.040
Note: Dc2a ¼ c2 (Original two-factor) -c2 (Single-factor), Dc2b ¼ c2 (Modified two-factor) -c2 (Single-factor). A significant change in chi-square (c2) indicates better model fit in the two-factor models than the single-factor model does. *P < 0.001. D A significant improvement in the model fit compared to the single-factor model.
Chi-square (c2) test statistics provide a measure of discrepancy between the measurement model and data, however, they are also overly sensitive to large sample sizes, sometimes even resulting in the rejection of models (Hoyle, 2011) with a good fit. CFI is a fit index that can be used in combination with chi-square (c2) test statistics to minimize this risk (Hoyle, 2011). Furthermore, a combination of the RMSEA and SRMR was used to minimize the risk of
rejection of well-fitting models. Therefore, although the chi-square (c2) was significant in this context, both the single-factor and twofactor models were considered the best fit to the data. The comparison of the two-factor and single-factor models revealed the two-factor model fits the data significantly better than the singlefactor model does, both in the original two-factor model (Dc2 17.61, df ¼ 1, P < 0.001) (Saboonchi et al., 2013; Zigmond and Snaith,
20
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
1983; Mystakidou et al., 2004) and the modified two-factor model (Dc2 34.76, df ¼ 1, P < 0.001) (Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006). Fig. 1 shows the factor structure of the CHADS for patients. The correlation between the C-HADS anxiety and C-HADS depression subscales indicated that the C-HADS subscales were positively correlated (r ¼ 0.83, and 0.82 for the original two-factor model and the modified two-factor model respectively). For family caregivers (the lower part of Table 2), the same profile of model fit was displayed in family caregivers as in cancer patients, except that the model fit in both the single-factor and two-factor models indicated adequate fit to the data. CFI values ranging from 0.93 to 0.94 were greater than the cutoff values of 0.90 for adequacy (Hoyle, 2011). The modified two-factor model also displayed significant improvement in fit (Dc2 20.21, df ¼ 1, P < 0.001) compared to the single-factor model. Fig. 2 shows the factor structure of the CHADS for family caregivers. The correlation between the C-HADS anxiety and C-HADS depression subscales also indicated that the CHADS subscales were positively correlated (r ¼ 0.84, and 0.81 for the original two-factor model and the modified two-factor model respectively). 3.3. Internal consistency reliability Cronbach's alpha coefficients for each subscale are displayed in Table 3. As is shown, the internal consistency of the C-HADS total, C-HADS anxiety and depression in both cancer patients and family caregivers displayed overall acceptable levels with all of the Cronbach's alpha 0.840. 3.4. Concurrent validity The concurrent validity of the C-HADS was established by the significant negative correlation with the relevant subscales of the SF-12 (Table 4). The overall significant correlations were found among the C-HADS, including C-HADS total, C-HADS anxiety and depression, and all the C-SF-12 subscales both in patients (the upper part of Table 4) and family caregivers (the lower part of Table 4) (all ps < 0.001). There were significantly moderate
correlations between the C-HADS total, C-HADS anxiety and depression, and the C-SF-12 subscales pertaining to mental health, such as role emotional (RE), mental health (MH), and mental component summary (MCS) both in cancer patients (r ¼ 0.40e0.55) and in family caregivers (r ¼ 0.41e0.53). The same correlation profile was displayed in the original two-factor model as in the modified two-factor model. The negative correlation coefficients indicated that higher levels of psychological distress were related to poorer health status. 4. Discussion To our knowledge, the present study is the first attempt in mainland China to assess anxiety and depression using C-HADS, and to report its psychometric properties from a dyadic perspective. For a better representation of the cancer population, participants were recruited from seven different administrative regions in China. The study results revealed that the C-HADS has satisfactory psychometric properties for use among both cancer patients and their family caregivers. The following discussion on the psychometric properties of the C-HADS mainly focuses on three aspects: construct validity, internal consistency reliability, and concurrent validity. Findings of the construct validity suggest the original two-factor model of C-HADS displays a best fit to the data for cancer patients, and an adequate fit to the data for family caregivers. These results support the basic assumption of two-factor construct of C-HADS. However, it should be noted there was a strong inter-correlation between anxiety and depression in both cancer patients (r ¼ 0.83) and family caregivers (r ¼ 0.84), indicating substantial covariance in the severity of anxiety and depression. This high inter-correlation between anxiety and depression was consistent with other studies using HADS in adult cancer patient populations (r ¼ 0.52e0.81) (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006), and family caregivers of cancer patients (r ¼ 0.63) (Gough and Hudson, 2009). As proposed by Saboonchi et al. (2013), “in order for HADS to demonstrate sound construct validity in cancer patients, such an inter-correlation
Fig. 1. The factor structure of the C-HADS for patients.
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
21
Fig. 2. The factor structure of the C-HADS for family caregivers.
Table 3 Cronbach's a for the subscales of C-HADS in cancer patients and family caregivers, total n ¼ 641.a Variables
Patients
Single-factor C-HADS total Original two-factor C-HADS-Anxiety C-HADS-Depression Modified two-factor C-HADS-Anxiety C-HADS-Depression
Family caregivers Percent
Cronbach's a
na
Percent
Cronbach's a
619
96.6
0.929
621
96.9
0.919
626 631
97.7 98.4
0.874 0.874
628 631
98.0 98.4
0.857 0.851
630 628
98.3 98.0
0.855 0.879
629 630
98.1 98.3
0.840 0.873
n
a
C-HADS¼Chinese version of hospital anxiety and depression scale. a The total n does not equal 641 because of missing values.
needs to be pre-specified” (pp. 2853). Zigmond and Snaith (1983) assume that a strong association between the subscales would indicate “they could be considered much the same thing, for example, emotional disturbance” (pp 364). This assumption resonates with the single-factor models of HADS from Razavi et al. (1990) and Smith et al. (2006). Our results further support the one-factorial measurement model of C-HADS by displaying a best fit to the data for cancer patients and an adequate fit to the data for family caregivers, hypothesizing a single factor of emotional distress. It is worth noting that the Chi-Square (c2), CFI, RMSEA and SRMR statistics revealed that the two-factor models tested offered acceptable fits to the data, with modified two-factor models (Smith et al., 2002; Moorey et al., 1991,Muszbek et al., 2006) offering a slightly better fit to the data than the original two-factor models (Saboonchi et al., 2013; Zigmond and Snaith, 1983; Mystakidou et al., 2004) in both cancer patients and family caregivers. This is a reminder that the modified two-factor model is the most suitable to use for Chinese cancer patients and their family caregivers. However, as the HADS remains a widely and easily used clinical instrument, it is suggested that HADS remain in its original twofactor model, to be scored and interpreted in the original
recommended manner (Zigmond and Snaith, 1983). This may benefit the intercultural adaptation of the instrument, and enable cross-cultural evaluation of psychological distress in cancer patients and their family caregivers. Regarding the high heterogeneity of the construct validity of HADS in a cancer population, considering the findings that both the two-factor and one-factor models offered an acceptable fit to the data in the present study, it is recommended that researchers incorporate both the total HADS score and the two subscale scores (Cosco et al., 2012). The HADS total, HADS anxiety and depression were reported in the present study. Study results indicated good internal consistency and reliability for the C-HADS with all of the Cronbach's a for the C-HADS total and two subscales greater than 0.851 in both cancer patients and family caregivers. The internal consistency of the C-HADS scale in both cancer patients and family caregivers corresponds well with those previously reported in other studies using the same instrument in adult cancer patient populations, with Cronbach's alpha for HADS anxiety varying from 0.73 to 0.89, and for HADS depression from 0.70 to 0.87 (Saboonchi et al., 2013; Smith et al., 2002; Moorey et al., 1991; Muszbek et al., 2006; Mystakidou et al., 2004), and among cancer caregivers (Cronbach's alpha ¼ 0.89, 0.85, and 0.8 for HADS total, anxiety and depression respectively) (Gough and Hudson,
22
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23
Table 4 Pearson correlations between C-HADS and C-SF-12 in patients and family caregivers (n ¼ 585e630).a C-SF-12 subscales of patients
C-HADS subscales of patients The original two-factor model
The modified two-factor model
Total
Anxiety
Depression
Anxiety
Depression
Physical Functioning Role Physical Bodily Pain General Health Vitality Role Emotional Social Functioning Mental Health PCS MCS
0.31b 0.34b 0.31b 0.25b 0.17b 0.44b 0.38b 0.52b 0.32b 0.55b
0.28b 0.32b 0.28b 0.24b 0.17b 0.40b 0.36b 0.50b 0.29b 0.53b
0.32b 0.34b 0.32b 0.25b 0.17b 0.42b 0.37b 0.50b 0.33b 0.52b
0.28b 0.32b 0.26b 0.22b 0.16b 0.40b 0.35b 0.48b 0.28b 0.51b
0.32b 0.34b 0.32b 0.27b 0.18b 0.42b 0.37b 0.51b 0.33b 0.53b
C-SF-12 subscales of family caregivers
C-HADS subscales of family caregivers
Physical Functioning Role Physical Bodily Pain General Health Vitality Role Emotional Social Functioning Mental Health PCS MCS
The original two-factor model
The modified two-factor model
Total
Anxiety
Depression
Anxiety
Depression
0.32b 0.45b 0.32b 0.29b 0.26b 0.51b 0.38b 0.45b 0.35b 0.53b
0.30b 0.42b 0.34b 0.25b 0.24b 0.49b 0.37b 0.45b 0.33b 0.51b
0.29b 0.44b 0.28b 0.30b 0.26b 0.48b 0.36b 0.42b 0.34b 0.50b
0.31b 0.41b 0.34b 0.24b 0.21b 0.49b 036b 0.44b 0.33b 0.50b
0.29b 0.44b 0.28b 0.30b 0.27b 0.47b 0.36b 0.43b 0.34b 0.51b
C-HADS¼Chinese version of hospital anxiety and depression scale; C-SF-12 ¼ Chinese version of the Medical Outcomes Study 12-item Short Form; PCS¼Physical Component Summary; MCS ¼ Mental Component Summary. a The total n does not equal 641 because of missing values in HADS and SF-12. b Correlation is significant at the 0.001 level (2-tailed).
2009). These findings highlight the satisfactory internal consistency of the HADS. Significant negative correlations between the C-HADS total, anxiety and depression scores and the scores of C-SF-12 subscales both in cancer patients and family caregivers confirmed the concurrent validity of the C-HADS. Similar results were reported in China (Li et al., 2015), as well as in other countries (Janda et al., ttir et al., 2011; Saevarsdottir et al., 2010). 2007; Fridriksdo
two-factor subscale scores be reported. Future research is needed to explore the relationship between the anxiety and depression of cancer patients and that of their family caregivers, and to systematically investigate the related factors or predictors of anxiety and depression for cancer patients and their family caregivers. This would benefit the interventions to be developed for relieving their psychological distress and improving QOL. Funding
5. Limitations The present study makes new contributions to our understanding of C-HADS in Chinese cancer patients and their family caregivers when coping with cancer as a dyad; however, it is essential to acknowledge several possible limitations. A limitation of this validation study was the inability to validate test-retest reliability, due to the cross-sectional study design. Moreover, the fact that the study population was only comprised of Chinese advanced cancer patients and their family caregivers could limit the generalizability of the results to other targeted populations in different cultures. Future studies targeting different demographic and medical cancer populations should be explored. 6. Conclusion Notwithstanding these limitations, the present study provides the satisfactory psychometric properties of C-HADS when applied to a sample of Chinese advanced cancer patients and their family caregivers. Findings indicate that the Chinese version of the HADS is a reliable and valid measure of anxiety and depression. The finding supports the basic assumption of two-factor construct of HADS. It is recommended that both the total HADS score and the
Financial support of this study was provided by the National Natural Science Foundation of China (No. 81573250). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Conflict of interest The authors declare no conflicts of interest to disclose. Acknowledgments: The authors gratefully acknowledge the support from all the related hospitals and all the participants for their sharing of their experience in this study. References Arving, C., Glimelius, B., Brandberg, Y., 2008. Four weeks of daily assessments of anxiety, depression and activity compared to a point assessment with the Hospital Anxiety and Depression Scale. Qual. Life Res. 17, 95e104. Bjelland, I., Dahl, A.A., Haug, T.T., Neckelmann, D., 2002. The validity of the Hospital Anxiety and Depression Scale: an updated literature review. J. Psychosom. Res. 52, 69e77. € de n, P., Sullivan, M., 1992. Anxiety Brandberg, Y., Bolund, C., Sigurdardottir, V., Sjo
Q. Li et al. / European Journal of Oncology Nursing 25 (2016) 16e23 and depressive symptoms at different stages of malignant melanoma. Psycho Oncol. 1, 71e78. Carlson, L.E., Angen, M., Cullum, J., Goodey, E., Koopmans, J., Lamont, L., MacRae, J.H., Martin, M., Pelletier, G., Robinson, J., Simpson, J.S.A., Speca, M., Tillotson, L., Bultz, B.D., 2004. High levels of untreated distress and fatigue in cancer patients. Br. J. Cancer 90, 2297e2304. Cosco, T.D., Doyle, F., Ward, M., McGee, H., 2012. Latent structure of the hospital anxiety and depression scale: a 10-year systematic review. J. Psychosom. Res. 72, 180e184. Fletcher, B., Miaskowski, C., Given, B., Schumacher, K., 2012. The cancer family caregiving experience: an updated and expanded conceptual model. Eur. J. Oncol. Nurs. 16, 387e398. ttir, N., Sævarsdo ttir, P., Halfd ttir, S.I., Jo nsdo ttir, A., Fridriksdo anardo ttir, H., Olafsd ttir, K.L., Gudmundsdo ttir, G., Gunnarsdo ttir, S., 2011. Magnúsdo o Family members of cancer patients: needs, quality of life and symptoms of anxiety and depression. Acta Oncol. 50, 252e258. Gough, K., Hudson, P., 2009. Psychometric properties of the hospital anxiety and depression scale in family caregivers of palliative care patients. J. Pain Symptom Manag. 37, 797e806. Grassi, L., Indelli, M., Marzola, M., Maestri, A., Santini, A., Piva, E., Boccalon, M., 1996. Depressive symptoms and quality of life in home-care-assisted cancer patients. J. Pain Symptom Manag. 12, 300e307. Guillemin, F., Bombardier, C., Beaton, D., 1993. Cross-cultural adaptation of healthrelated quality of life measures: literature review and proposed guidelines. J. Clin. Epidemiol. 46, 1417e1432. Hagedoorn, M., Sanderman, R., Bolks, H.N., Tuinstra, J., Coyne, J.C., 2008. Distress in couples coping with cancer: a meta-analysis and critical review of role and gender effects. Psychol. Bull. 134, 1e30. Haley, W., 2003. The costs of family caregiving: implications for geriatric oncology. Crit. Rev. Oncol. Hematol. 48, 151e158. Herrmann, C., 1997. International experiences with the Hospital Anxiety and Depression Scaleea review of validation data and clinical results. J. Psychosom. Res. 42, 17e41. Hooper, D., Coughlan, J., Mullen, M., 2008. Structural equation modeling: guidelines for determining model fit. Electron. J. Bus. Res. Methods 6, 53e60. Hoyle, R.H., 2011. Structural Equation Modeling for Social and Personality Psychology, 1 ed. SAGE Publications Ltd. Janda, M., Steginga, S., Langbecker, D., Dunn, J., Walker, D., Eakin, E., 2007. Quality of life among patients with a brain tumor and their carers. J. Psychosom. Res. 63, 617e623. Kayser, K.P., Watson, L.E.M., Licsw, Andrade, J.T.M., 2007. Cancer as a “We-Disease”: examining the process of coping from a relational perspective. Fam. Syst. Health 25, 404e418. Kim, Y., Kashy, D.A., Wellisch, D.K., Spillers, R.L., Kaw, C.K., Smith, T.G., 2008. Quality of life of couples dealing with cancer: dyadic and individual adjustment among breast and prostate cancer survivors and their spousal caregivers. Ann. Behav. Med. 35, 230e238. Leung, C.M., Ho, S., Kan, C.S., Hung, C.H., Chen, C.N., 1993. Evaluation of the Chinese version of the hospital anxiety and depression scale. A cross-cultural perspective. Int. J. Psychosom. 40, 29e34. Li, Q., Lin, Y., Qiu, Y., Gao, B., Xu, Y., 2014. The assessment of health-related quality of life and related factors in Chinese elderly patients undergoing chemotherapy for advanced cancer: a cross-sectional study. Eur. J. Oncol. Nurs. 18, 425e435. Li, Q., Xu, Y., Zhou, H., Loke, A.Y., 2015. Testing a preliminary live with love conceptual framework for Cancer couple dyads: a mixed-methods study. Eur. J. Oncol. Nurs. 19, 619e628. Li, Q., Loke, A.Y., 2013. A spectrum of hidden morbidities among spousal caregivers for patients with cancer, and differences between the genders: a review of the literature. Eur. J. Oncol. Nurs. 17, 578e587. Lloyd-Williams, M., Friedman, T., Rudd, N., 2001. An analysis of the validity of the hospital anxiety and depression scale as a screening tool in patients with advanced metastatic Cancer. J. Pain Symptom Manag. 22, 990e996.
23
pez-Go mez, M., Malmierca, E., de Go rgolas, M., Casado, E., 2013. Cancer in Lo developing countries: the next most preventable pandemic. The global problem of cancer. Crit. Rev. Oncol. 88, 117e122. Mitchell, A.J., 2010. Short screening tools for cancer-related distress: a review and diagnostic validity meta-analysis. J. Natl. Compr. Canc Netw. 8, 487e494. Mitchell, A.J., Meader, N., Symonds, P., 2010. Diagnostic validity of the hospital anxiety and depression scale (HADS) in cancer and palliative settings: a metaanalysis. J. Affect. Disord. 126, 335e348. Moorey, S., Greer, S., Watson, M., Gorman, C., Rowden, L., Tunmore, R., Robertson, B., Bliss, J., 1991. The factor structure and factor stability of the hospital anxiety and depression scale in patients with cancer. Br. J. Psychiatry 158, 255e259. nszky, M., Ruzsa, A., Muszbek, K., Szekely, A., Balogh, E.M., Moln ar, M., Roha €llosi, M., Vada sz, P., 2006. Validation of the Hungarian translation Varga, K., Szo of hospital anxiety and depression scale. Qual. Life Res. 15, 761e766. Mystakidou, K., Tsilika, E., Parpa, E., Katsouda, E., Galanos, A., Vlahos, L., 2004. The hospital anxiety and depression scale in Greek cancer patients: psychometric analyses and applicability. Support. Care Cancer 12, 821e825. Northouse, L.L., Mood, D., Templin, T., Mellon, S., George, T., 2000. Couples' patterns of adjustment to colon cancer. Soc. Sci. Med. 50, 271e284. Portney, L.G., Watkins, M.P., 2009. Foundations of Clinical Research, 3 ed. Pearson/ Prentice Hall, London. Razavi, D., Delvaux, N., Farvacques, C., Robaye, E., 1990. Screening for adjustment disorders and major depressive disorders in cancer in-patients. Br. J. Psychiatry 156, 79e83. Rodgers, J., Martin, C.R., Morse, R.C., Kendell, K., Verrill, M., 2005. An investigation into the psychometric properties of the Hospital Anxiety and Depression Scale in patients with breast cancer. Health Qual. Life Outcomes 3, 41e41. Saboonchi, F., Wennman-Larsen, A., Alexanderson, K., Petersson, L.M., 2013. Examination of the construct validity of the swedish version of hospital anxiety and depression scale in breast cancer patients. Qual. Life Res. 22, 2849e2856. Saevarsdottir, T., Fridriksdottir, N., Gunnarsdottir, S., 2010. Quality of life and symptoms of anxiety and depression of patients receiving cancer chemotherapy: longitudinal study. Cancer Nurs. 33, E1eE10. Satin, J.R., Linden, W., Phillips, M.J., 2009. Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis. Cancer 115, 5349e5361. Schreier, A.M., Williams, S.A., 2004. Anxiety and quality of life of women who receive radiation or chemotherapy for breast cancer. Oncol. Nurs. Forum 31, 127e130. Smith, A.B., Wright, E.P., Rush, R., Stark, D.P., Velikova, G., Selby, P.J., 2006. Rasch analysis of the dimensional structure of the hospital anxiety and depression scale. Psycho Oncol. 15, 817e827. Smith, A.B., Selby, P.J., Velikova, G., Stark, D., Wright, E.P., Gould, A., Cull, A., 2002. Factor analysis of the hospital anxiety and depression scale from a large cancer population. Psychol. Psychother. 75, 165e176. Statistics. Health, 2015. Tools of the Trade: Strategy for Determing Sample Size. Retrieved from: http://www.statistics.health.pa.gov/StatisticalResources/ UnderstandingHealthStats/ToolsoftheTrade/Documents/Strategy_for_ Determining_Sample_Size.pdf. Wang, W., Chair, S.Y., Thompson, D.R., Twinn, S.F., 2009. A psychometric evaluation of the Chinese version of the Hospital Anxiety and Depression Scale in patients with coronary heart disease. J. Clin. Nurs. 18, 2436e2443. Ware, J.E., Kosinski, M., Turner-Bowker, D.M., Gandek, B., 2002. How to Score Version 2 of the SF-12 Health Survey (With a Supplement Documenting Version 1). QualityMetric Incorporated. Ware Jr., J., Kosinski, M., Keller, S.D., 1996. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med. Care 34, 220e233. WHO, 2015. Cancer Fact Sheet N 297. 2015. Retrieved from: http://www.who.int/ mediacentre/factsheets/fs297/en/index.html. Zigmond, A., Snaith, R., 1983. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 67, 361e370.