Psychometric properties of the Chinese version of female self-advocacy in cancer survivorship scale.

Objective: To translate Female Self-Advocacy in Cancer Survivorship (FSACS) scale and evaluate the psychometric properties of Chinese version of FSACS scale among female cancer survivors in China.
Methods: This study employed a cross-sectional design. FSACS scale was forward-backward translated and cognitive interviews were conducted for cultural adaptation. The newly translated tool was distributed to female cancer survivors to test psychometric properties, including item analysis, content validity, construct validity, criterion validity, internal consistency, and test-retest reliability.
Results: A total of 436 female cancer survivors were recruited in a Chinese tertiary cancer center from May to August in 2021. Item analysis showed statistical significance (P < 0.05) for each one and no ceiling or floor effect. The item-level content validity index ranged from 0.86 to 1.00 and the scale-level content validity index was 0.98. Three factors were extracted based on parallel analysis, and confirmatory factor analysis proved a good model fit with the original 3-factor structure. Pearson's correlation coefficient showed acceptable criterion validity. The Cronbach's α of 0.880 demonstrated the scale's internal consistency reliability, and the Alpha coefficients were 0.826, 0.763, and 0.859 for its three dimensions, respectively. The interclass correlation coefficients for test-retest reliability was 0.904 (0.870-0.891, P < 0.01) which confirmed the external reliability of the scale. Conclusions: The Chinese version of FSACS scale proved to be a valid and reliable instrument that can be applied among Chinese female cancer survivors. Further research could be conducted in larger populations or people in different cancer stages.

Introduction

The last few years have witnessed an increase in the incidence of cancer and a decrease in death rate, particularly for a few female malignancies. This trend has led to a growing number of cancer survivors, and survivorship care has become one of the most important issues in healthcare research.

Cancer survivorship is defined as “the health and well-being of a person with cancer from the time of diagnosis until the end of life”. Good wellness of cancer survivors contains various dimensions, and healthcare providers are supposed to pay attention to each aspect. However, this group of people usually experienced negative emotions, persistent side-effects, and psychosocial issues. They still suffer from many unmet needs, which may pose challenges to their rehabilitation and quality of life in the long-term.

The Sustainable Development Goals of 2030 agenda proposed by the United Nations involve a series of plans for the purpose of addressing global challenges and setting action goals. Some key elements are included in these goals, such as gender equality, good health, and well-being for individuals of all ages. These elements are highly relevant to the females and often interact with each other in the 2030 agenda. Women's health is an essential component of Sustainable Development Goals, but attention has not been paid sufficiently to female patients who suffer from life-threatening diseases like cancer. Furthermore, it has been argued in the Lancet that deficiencies of research pertaining to girls and women may cause problems in health science, resulting in missed opportunities for health promotion and interventions.

There are obvious differences between males and females when considering symptoms, emotions, and behaviors. Evidence shows that women are more vulnerable to psychological problems such as anxiety, depression, and fatigue. At the same time, symptoms related to emotional needs and distress are more likely to be found among female patients because of their abilities to express feelings and seek support., Therefore, it is important for female cancer survivors to advocate for themselves so as to better participate in symptom management and clinical decision-making, thereby improving the quality of survivorship care. However, the lack of instruments in China to measure self-advocacy of cancer survivors limits our understanding of patients’ capacity to deal with cancer-related challenges, and hinders health professionals from providing better medical care.

Self-advocacy in the field of cancer survivorship is defined as a person's ability to cope with diseases through internalizing available resources into thoughts and actions, and finally rise up to cancer-related challenges. The Female Self-Advocacy in Cancer Survivorship (FSACS) scale is a valid and reliable instrument that can be used to evaluate the self-advocacy ability of women with cancer. This tool is currently solely available for its original English version, so the aim of this study is to translate and validate the FSACS scale among Chinese female cancer survivors.

Methods

Design and sample

This research was a cross-sectional study carried out in a tertiary Cancer Center in China. 436 female cancer survivors were recruited from May to August in 2021. According to the GLOBOCAN estimates of a new incidence for year 2020, the top ten tumor types of women were selected to represent the main group of female patients with cancer. Based on codes from the International Statistical Classification of Diseases 10th Revision, these specific types include cancer of breast, colorectum, lung, cervix uteri, thyroid, stomach, corpus uteri, liver, ovary, and pancreas.

It is suggested that 5–10 subjects are needed for each item in the process of exploratory factor analysis (EFA) and at least 200 participants for confirmatory factor analysis (CFA). Considering there are 20 items in FSACS scale, that we had to recruit additional 10 to 40 participants for the pilot study, as well as 10% withdrawal of participants was estimated, the reckoned sample size was 345–489 in total.

The inclusion criteria were: (1) ≥ 18 years old; (2) female survivors diagnosed with cancer; (3) being informed and able to complete questionnaires in Chinese. Participants with communicate impairments or mental illness were excluded. A cancer survivor was defined here as “a diagnosed individual who is still living as a survivor no matter when that diagnosis was made or whether it was successfully treated”. This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (Approval No. 2104234-26).

Instruments

The current study utilized one self-designed questionnaire and three scales. The questionnaire comprised participants’ demographic characteristics as well as clinical information. Following is the description of the scales utilized.

The Chinese version of FSACS scale

The Chinese version of FSACS (FSACS-C) scale consists of 20 items on a six-point Likert type scale ranging from 0 (strongly disagree) to 6 (strongly agree). Three dimensions are covered, namely application of information, connected strength, and communication with healthcare providers. Cronbach's alpha coefficients of the original scale are 0.82, 0.85, and 0.79 for each dimension.

Social support rating scale

The social support rating scale (SSRS) is a widely used scale developed by Xiao in 1989 which evaluate the extent and utilization of people's social support. It has three dimensions, including subjective support, objective support, and utilization of social support. This tool is appropriate for Chinese context, universally accepted by people living in China. The original scale's internal consistency ranges from 0.89 to 0.94, and its test-retest reliability is 0.9220.

General self-efficacy scale

General self-efficacy scale (GSES) is a unidimensional tool for the purpose of assessing the beliefs in general people. There are 10 items in total, using four-point Likert grades ranging from points 1 (completely incorrect) to 4 (completely correct). The scale has been translated into dozens of languages, and the Chinese version of GSES has proved to be good reliability and validity., This scale has been widely used for about two decades in China.

Procedure

The study consisted of two phases. First, the FSACS scale was forward-backward translated and followed by a cross-cultural adaptation using cognitive interviewing. Second, psychometric properties were evaluated through a cross-sectional survey. Paper questionnaires were distributed in person at the hospital. Participants were recruited and a signed informed consent form was required before completing these questionnaires.

Phase 1: translation and cross-cultural adaptation

Permission of translation and adaptation had been obtained from Dr. Hagan, the developer of FSACS scale. The process of forward and backward translation was guided by Brislin's translation model. The translation procedure was completed by 4 bilingual translators, two of whom translated the original scale from source language to Chinese, and the other two translators were English language speakers involved in the back translation process. Semantic and content equivalence could be realized through comparison and integration of these translated versions.

To make sure the items of this scale could be well-understood by the target population, cognitive interviews were performed for the purpose of cross-cultural adaptation. Face validity could also be assured during this procedure. These interviews were conducted face-to-face in quite rooms at the hospital, and the interviewing process was recorded with an audio apparatus. The privacy of participants was well-protected by researchers.

Phase 2: psychometric testing

Psychometric properties were tested based on the Classical Test Theory, including item analysis, validity, and reliability. The data analysis was performed by using IBM SPSS Statistics version 25.0 and AMOS Graphics version 23.0 for windows (IBM Corporation, the United States).

Item analysis

Item analysis was performed with the calculation of critical ratio. Items were ranked and divided into two groups according to the 25% for high and low scores, using independent t-test to examine the discrimination of each item. If the t value of an item failed to reach significance, this item would be deleted because of insufficient discrimination. Additionally, the ceiling and floor effect were tested by calculating the percentage of people with highest and lowest scores. A proportion of each group less than 15% represented that the scale had no ceiling or floor effect.

Validity

Content validity

The content validity was determined by an expert panel, which was composed of clinical practitioners and researchers who were familiar with the methodology of psychometric testing. The expert panel assessed the correlation with the concept actually measured, using a 4-point grade with 1 being not relevant to 4 being highly relevant and succinct. Both the item-level content validity index (I-CVI) and the average score of scale-level content validity index (S-CVI/Ave) were evaluated. I-CVI marked 0.78 or higher and S-CVI/Ave of 0.9 or above were considered of good content validity.

Criterion validity

The concurrent validity was evaluated with Pearson correlation coefficient to examine the correlation between FSACS scale and other relevant instruments. Since self-advocacy is a new concept in the field of cancer survivorship, a perfect gold standard for this concept did not exist; hence, instruments related to the defining characteristics of self-advocacy were used, including GSES and the subscales of SSRS. Hypotheses were made that self-advocacy was positively correlated with self-efficacy and social support utilization.

Construct validity

Since FSACS scale is a newly developed measurement instrument which requires cultural adaptation among Chinese population, EFA was performed to explore the latent construct of the translated scale. Based on the measurement model of self-advocacy, CFA was also applied to examine whether the construct of the instrument tested among Chinese population would be in accordance with the original scale.

Kaiser-Meyer-Olkin and Bartlett's test of sphericity were performed. Kaiser-Meyer-Olkin index ≥ 0.60 (preferably > 0.9) and a significance of Bartlett's test would suggest that data were suitable for factor analysis.,

The number of factors was determined by parallel analysis (PA), a procedure that was more appropriate than the criterion of eigenvalue ≥ 1 (Kaiser principle) and scree plot test., PA could calculate the average eigenvalue of a random correlation matrix and compared it with the eigenvalue obtained from the sample., Specifically, if the variance explained by a factor drawn from real data was smaller than the variance explained by the factor drawn from simulated random data, this factor would be valueless to be retained.

As for CFA, chi-square test was performed, and CFA was evaluated by a series of indexes, such as ratio of chi-square and degree of freedom (χ2/df), goodness-of-fit Index, adjusted goodness-of-fit, comparative fit index, incremental fit index, normed fit index and root mean square error of approximation. Modification could be appropriately used if necessary.

Reliability

Cronbach's alpha coefficient was calculated in order to evaluate the internal consistency of the scale and its subscales. Before this procedure, unidimensional feature was tested through factor analysis to meet the prerequisite of evaluating internal consistency for each dimension.

Since self-advocacy was considered a variable that does not change significantly over a short period of time, the stability was examined by calculating intraclass correlation coefficient , among 15–30 participants over a 2-week interval, so as to evaluate the test-retest reliability. In the process of filling in the questionnaires, participants were provided with almost the same conditions between their two tests, including testing spots, guidance of the researcher, their health status, and the treatments they received.

Results

We distributed 483 questionnaires and a total of 436 cancer survivors effectively completed these questionnaires, so the response rate was 90.27%. Participants’ demographic characteristics are showed in Table 1 and their clinical information is demonstrated in Table 2.

Table 1

Participants' demographic characteristics (n ​= ​436).

Variables	n (%)
Age (years)
18–39	98 (22.48)
40–59	239 (54.82)
≥ 60	99 (22.71)
Education level
Primary or below	70 (16.06)
Secondary	228 (52.29)
College	62 (14.22)
Bachelor	76 (14.43)
Master or above	8 (1.83)
Religion
Non	392 (89.91)
Buddhism	33 (7.57)
Christianism	10 (2.29)
Catholicism	1 (0.23)
Residence
Urban	279 (63.99)
Rural	157 (36.01)
Marital status
Single	21 (4.82)
Married	388 (88.99)
Divorced	16 (3.67)
Widowed	11 (2.52)
Occupational status
Employed	144 (33.03)
Unemployed	117 (26.83)
Retired	175 (40.14)
Monthly family income per person (yuan)
≤ 3000	119 (27.29)
3001–6000	156 (35.78)
6001–9000	80 (18.35)
> 9000	81 (18.58)
Source of medical expenses
Non-medical insurance	50 (11.47)
Medical insurance	386 (88.53)

Table 2

Participants' clinical information (n ​= ​436).

Variables	n (%)
Cancer type
Breast	48 (11.01)
Colorectum	47 (10.78)
Lung	36 (8.26)
Cervix uteri	50 (11.47)
Thyroid	48 (11.01)
Stomach	30 (6.88)
Corpus uteri	47 (10.78)
Liver	35 (8.03)
Ovary	46 (10.55)
Pancreas	32 (7.34)
Others	17 (3.90)
Metastasis
Non	213 (48.85)
Local	121 (27.75)
Distant	102 (23.39)
Recurrence
Yes	45 (10.32)
No	391 (89.68)
Treatment (in the past 1 year)
Surgery	275 (63.07)
Radiotherapy	2 (0.46)
Anticancer drug therapya	70 (16.06)
Surgery ​+ ​Radiotherapy	4 (0.92)
Surgery ​+ ​Anticancer drug therapy	74 (16.97)
Radiotherapy ​+ ​Anticancer drug therapy	2 (0.46)
More than three	9 (19.57)
Course of disease
0–6 months	351 (80.50)
∼2 years	50 (11.47)
∼5 years	23 (5.28)
> 5 years	12 (2.75)

Anticancer drug therapy includes chemotherapy, targeted therapy and immunotherapy.

Participants were from 18 to 79 years of age, with a mean of 49.7 (standard deviation [SD] ​= ​12.4) years old. About half of the participants received secondary education, one third attended college and 70 (16.06%) individuals got primary education or below. Most of the participants (89.91%) did not practice religion, 279 (63.99%) individuals live in cities, and 388 (88.99%) were in their marriage. Most women had at least one or two children, while a small number of them (7.24%) did not have any child. For working status, 144 (33.03%) participants were currently employed, and the others were unemployed or retired.

Our sample contained the main categories of female malignancies and the details are shown in Table 2. Nearly half of the patients did not have metastatic tumors, while 121 (27.75%) patients developed local metastasis and 102 (23.39%) patients suffered from distant metastasis. Most survivors (48.85%) did not experience the recurrence of cancer. 275 (63.07%) participants received treatment of surgery within 1 year, and 351 (80.50%) patients experienced course of disease less than 6 months.

Phase 1: translation and cross-cultural adaptation

We completed the translation process with the participation of four bilingual translators, and then a panel of experts held a meeting to discuss and integrate these translated versions. In the end, we formed an initial Chinese version of FSACS (FSACS-C) scale. For cultural adaptation, we conducted a three-round cognitive interviewing where 25 interviewees were involved. The interviewees were from 18 to 70 years of age, in different residential places, of various educational levels as well as employment status. We adapted the scale at the end of each interviewing round, and finally gained a well-understood instrument that could be accepted by people of diverse backgrounds.

Phase 2: psychometric testing

We examined psychometric properties of FSACS-C scale, and randomly divided the collected data into two portions for EFA and CFA, respectively. The split of data was performed in SPSS software and a sample size of 50% was set by using the ‘select cases’ menu. Item analysis showed statistical significance for each (P ​< ​0.05), suggesting fine discrimination so we did not delete any item. The proportion of those who gained the highest and lowest scores were both less than 15%, indicating no ceiling or floor effect in this scale.

Validity

We invited seven healthcare professionals to participate in the process of content validation. The results showed I-CVI ranged from 0.86 to 0.10 and S-CVI/Ave was 0.98, which indicated good content validity.

We examined concurrent validity of FSACS-C scale by testing the Pearson's correlation. Statistics showed the coefficient of 0.339 (P ​< ​0.01) between FSACS-C and GSES, and that of 0.193 (P ​< ​0.01) between FSACS-C and the subscale (social support utilization dimension) of SSRS, which is presented in Table 3.

Table 3

Correlation of FSACS with GSES and subscales of SSRS (n ​= ​436).

	SSRS			GSES
	S1	S2	S3
FSACS	0.155∗∗	0.143∗∗	0.162∗∗	0.339∗∗
F1	0.101∗	0.125∗∗	0.116∗	0.391∗∗
F2	0.096∗	0.105∗	0.043	0.147∗∗
F3	0.160∗∗	0.103∗	0.193∗∗	0.212∗∗

∗∗P ​< ​0.01; ∗P ​< ​0.05

Based on PA, we used principal components and generated random normal data at the same time. It was suggested by the mean and percentile of PA that the most appropriate number of factors was three (Table 4), so we finally choose to retain 3 factors. And then we set 3 as the fixed number of factors, by using “dimension reduction” of SPSS. The factor loadings with EFA are demonstrated in Table 5. The curve of scree plot with parallel analysis is showed in Fig. 1.

Table 4

Outcomes of parallel analysis (n ​= ​218).

Eigenvalues
Root	Raw data	Means	Percntyl
1	6.692,97	1.577,96	1.683,23
2	2.644,84	1.471,10	1.545,98
3	1.556,64	1.388,76	1.458,89
4	1.084,12	1.319,73	1.376,75
5	0.822,85	1.256,66	1.308,04

Table 5

Factor loadings with exploratory factor analysis (n ​= ​218).

	Component
	Factor 1	Factor 2	Factor 3
Item 2	0.769
Item 5	0.745
Item 3	0.731
Item 6	0.693
Item 11	0.672
Item 1	0.630
Item 4	0.594
Item 7		0.811
Item 10		−0.804
Item 8		0.756
Item 9		−0.725
Item 12		0.566
Item 13			0.779
Item 15			0.764
Item 16			0.757
Item 20			0.724
Item 17			0.718
Item 14			0.693
Item 18			0.684
Item 19			0.616

Extraction Method: Principal component analysis.

Rotation Method: Promax with Kaiser normalization.

Fig. 1

The scree plot with parallel analysis (n ​= ​218).

We examined the three-factor model in terms of the construct of the original scale by using CFA. The three-factor structure was evaluated by model fit indexes which were presented in Table 6, and the three-factor model was showed in Fig. 2.

Table 6

Model fit indexes of the three-factor structure (n ​= ​218).

Modification	χ2	χ2/df	GFI	AGFI	IFI	NFI	TLI	CFI	RMSEA
Before	334.557	2.003	0.869	0.836	0.898	0.815	0.882	0.897	0.068
After	244.571	1.510	0.901	0.872	0.950	0.865	0.940	0.949	0.048

AGFI, adjusted GFI; CFI, comparative fit index; GFI, goodness-of-fit index; IFI, incremental fit index; NFI, normed fit index; RMSEA, root mean square error of approximation.

Fig. 2

The three-factor model with confirmatory factor analysis (n ​= ​218).

Reliability

A Cronbach's α of 0.880 confirmed the internal consistency of the 20-item scale, and 0.826, 0.763, and 0.859 for each subscale. It was demonstrated in Table 7 that the Alpha coefficient did not increase with each item deleted, so we retained all the items in FSACS-C.

Table 7

Item scores and internal consistency of FSACS-C (n ​= ​436).

Items	Mean ​± ​SD	Alpha coefficient	Alpha coefficient after deleting the item
Factor 1	31.66 ​± ​5.03	0.826	–
Item 1	3.95 ​± ​1.21	–	0.817
Item 2	4.68 ​± ​1.05	–	0.794
Item 3	4.90 ​± ​0.92	–	0.793
Item 4	4.98 ​± ​0.73	–	0.805
Item 5	4.67 ​± ​1.04	–	0.797
Item 6	4.54 ​± ​0.87	–	0.800
Item 11	3.94 ​± ​1.27	–	0.817
Factor 2	23.88 ​± ​3.12	0.763	–
Item 7	5.09 ​± ​0.74	–	0.709
Item 8	4.71 ​± ​0.90	–	0.697
Item 9	4.50 ​± ​0.95	–	0.728
Item 10	4.75 ​± ​0.86	–	0.705
Item 12	4.84 ​± ​0.89	–	0.762
Factor 3	39.30 ​± ​4.79	0.859	–
Item 13	4.84 ​± ​0.77	–	0.833
Item 14	5.03 ​± ​0.76	–	0.850
Item 15	5.09 ​± ​0.72	–	0.842
Item 16	5.04 ​± ​0.69	–	0.839
Item 17	4.96 ​± ​0.84	–	0.840
Item 18	4.44 ​± ​1.10	–	0.845
Item 19	5.23 ​± ​0.65	–	0.848
Item 20	4.67 ​± ​1.10	–	0.840
Total	94.84 ​± ​10.15	0.880	–

As for test-retest reliability, we examined 21 participants over 14 days interval, providing them with the same environment of testing between these two measurements. The interclass correlation coefficients was 0.904 (0.870–0.891, P ​< ​0.01), indicating high external reliability of the scale.

Discussion

In this study, we translated and validated the FSACS scale and confirmed its reliability. We regard this scale as a useable tool to measure self-advocacy. Although the concept of self-advocacy is somewhat impenetrable before explaining, the evaluation of this concept can be quantified with psychometric scales to assess latent variables. Our study has demonstrated the acceptability and feasibility of applying this tool in the target population.

During the process of cultural adaptation, cognitive interviewing is considered a practical approach which helps to confirm the face validity of a cross-cultural psychometric instrument. The structure of FSACS-C performed by EFA using parallel analysis to determine the number of common factors by using this newly translated tool. PA is considered an effective method compared with certain traditional approaches. For example, the criterion of eigenvalue ≥ 1, may lead to over-extraction of the number of common factors and affect the outcome of construct validity. As we initially examined, it was slightly different from the original scale when employing eigenvalues for dimensionality assessment, but PA can solve this problem. Furthermore, CFA showed the model fit was acceptable based on the concept framework of self-advocacy, so we finally adopted the three-factor structure model for FSACS-C.

Among Chinese population, FSACS-C has assessed three dimensions with regard to expressing their needs, communicating with healthcare providers and gaining support from their social networks. It provides us with an opportunity to better understand how women diagnosed with cancer involve in clinical decisions and use available resources.

The translation and adaptation of this tool may lead to additional potential benefits for female cancer survivors. Specifically, this measurement tool can help medical staff assess cancer sufferers’ abilities to advocate for themselves and active participation in clinical decision-making when facing challenges during their cancer journey. It can also assist healthcare providers in identifying populations with low self-advocacy and discovering relevant facilitators and barriers. Furthermore, a reliable and valid psychometric tool plays an important role in evaluating intervention studies. If intervention programmes for cancer survivors pertaining to self-advocacy is implemented in the future, this scale, which has been applicable in Chinese context, can serve as a good psychometric tool.

Limitations of this study may be as follows. First, this cross-sectional study was conducted solely in one cancer center. The results may not apply to female cancer population of all regions in China, so this tool may not be valid for all female Chinese cancer survivors. There might also be selective bias in terms of resources available and information gained, as cancer survivors living in remote regions and undeveloped areas are less likely to enroll in large cancer centers particularly in first-tier cities. Second, most of the participants were in their active treatment when being surveyed, with a short-term experience of cancer since their initial diagnosis. We aligned with our participants with that of the original scale, which aimed at cancer survivors of all stages in terms of the definition by the National Cancer Institute. As a result, the recruitment of participants was not evenly distributed to different stages but largely involved those who received their initial therapy.

It is suggested that future study can be conducted in other regions of the country and incorporate participants with various course of disease, particularly survivors who receive extended period of a cancer treatment and suffer from long-term conditions.

Conclusions

FSACS-C scale is a valid and reliable instrument to evaluate how Chinese female cancer survivors advocate for themselves. Further research can be conducted among women with cancer in different stages of cancer.

Authors' contributions

Conceived and designed the analysis: Mingchun Deng, Zhenqi Lu, Anni Wang. Collected the data: Mingchun Deng, Zhenqi Lu, Yi Zhang, Yaqiong Chen, Jun Wang. Contributed data or analysis tools: Jiajia Qiu, Xiaoju Zhang, Anni Wang. Performed the analysis: Mingchun Deng, Anni Wang. Wrote the paper: Mingchun Deng, Zhenqi Lu, Anni Wang, Xiaoju Zhang, Jiajia Qiu, Yi Zhang, Yaqiong Chen, Jun Wang.

Ethics statement

This study was approved by the ​Ethics ​Committee of ​Fudan University Shanghai Cancer Center ​(Approval No. ​2104234-26).

Funding

This study was funded by School of Nursing.

Declaration of competing interest

None declared.