Validation of the Clinical COPD Questionnaire in Taiwan

ABSTRACT

Health status improvement is a critical treatment goal for physicians managing chronic obstructive pulmonary disease (COPD). Numerous instruments to measure the disease-specific health-related quality of life (HRQOL) for patients with COPD have been used in daily clinical practice. The Clinical COPD Questionnaire (CCQ) is one of these recommended by the Global Initiative for Chronic Obstructive Lung Disease (GOLD). This study examined the psychometric properties of the CCQ in patients with COPD in Taiwan. A descriptive, cross-sectional design was conducted. Data were collected in a secondary care unit. We administered the CCQ, the modified Medical Research Council (mMRC) dyspnea scale, and the 12-item Short Form Health Survey (SF-12) for patients with COPD. Reliability was assessed using Cronbach's alpha and item–total correlation coefficients. Construct validity was assessed using confirmatory factor analysis (CFA) and testing the hypothesis that severity of dyspnea measured using the mMRC dyspnea scale is associated with the CCQ scores. Convergent validity was assessed by testing the correlation between the CCQ and the SF-12. Discriminant validity was assessed to differentiate among the classifications of COPD Groups A to D. A total of 114 subjects were recruited in the study. Cronbach's alpha was high (0.90) for the total score of the CCQ. Significant correlations were found between the CCQ scores and those of the mMRC dyspnea scale (ρ = 0.67) and domains of the SF-12 (ρ = −0.44 to −0.75). Furthermore, the CCQ scores showed a significant difference among the classifications of COPD Groups A to D. CFA confirmed the construct validity, with a good model fit. Good to excellent psychometric properties of the Chinese Version CCQ were demonstrated in the study. Wide usage of the Chinese Version CCQ for Taiwanese COPD patients can be recommended in daily clinical practice or clinical trials.

KEYWORDS:

• Clinical COPD Questionnaire (CCQ)
• health status
• health-related quality of life (HRQOL)
• reliability
• validity

Background

Chronic obstructive pulmonary disease (COPD) is a systemic disease with a great impact on several dimensions of patients' lives (1). According to the World Health Organization, COPD is the third leading cause of death in the world (2). The mortality rate of COPD is ranked seventh among the top 10 causes of death in Taiwan, according to the Department of Health of the Executive Yuan (3). The number of patients with COPD increases annually and is responsible for the substantial economic burden of this disease (4). COPD is an irreversible disease characterised by airflow obstruction and cannot be completely cured by medication (5). As the disease progresses, the lung functions of patients deteriorate and cause the patients to suffer from dyspnea, cough, phlegm, exercise intolerance, and limited activities in daily life (4). Studies have revealed that constraints on daily life provoke negative psychological emotions, which affect the quality of life of patients with COPD (6, 7).

The severity of symptoms affects the quality of life of patients with COPD, and its complex mechanisms alternately provoke physiological, psychological, and emotional changes (7, 8). Empirical studies on health-related quality of life (HRQOL) in patients with COPD have revealed that the disease affects not only physiological status but also functional status. Most patients with COPD experience a negative impact on quality of life (9–11).

According to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) (4), the goals of COPD treatment are to reduce COPD symptoms, reduce the frequency and severity of exacerbations, and improve health status, quality of life, and exercise tolerance. Improving health status and quality of life are crucial goals of COPD management. Therefore, HRQOL is a suitable instrument to measure crucial representations of the impact of the disease for COPD patients (12).

Several health status and HRQOL questionnaire tools have been developed. Currently, the disease-specific instruments used to measure the HRQOL of patients with COPD include the St. George's Respiratory Questionnaire (SGRQ) (13), Quality-of-Life for Respiratory Illness Questionnaire (QOL-RIQ) (14), Chronic Respiratory Disease Questionnaire (CRQ) (15), COPD Assessment Test (CAT) (16), and Clinical COPD Questionnaire (CCQ) (17). Most COPD-specific HRQOL questionnaires such as SGRQ, QOL-RIQ, CRQ, and CCQ have similar basic content, mainly focusing on symptoms, physical activity functional states, and the impact of psychosocial factors (18). The SGRQ and QOL-RIQ have 76 and 55 items, respectively.

Although they are self-administered questionnaires, they are time consuming (13,14). In addition, the CRQ is a 20-item questionnaire administered by an interviewer and is not feasible for a busy clinic (15). Most COPD trials have used the SGRQ for measuring HRQOL and providing quantitative information regarding patients' subjective perceptions; it suitably reflects the health status of COPD (19). The SGRQ is a standardised, airway-disease-specific questionnaire. However, it is somewhat complex. Otherwise, the CCQ is concise and easy to use and can be completed in 2 minutes (17).

In addition, the CAT is an 8-item, one-dimension questionnaire that covers cough, phlegm, chest tightness, breathlessness going up hills and stairs, activity limitation at home, confidence leaving home, sleep, and energy. It is the newest one developed in 2009 (16). The CAT is simple and easy to administer and has a short time scale similar to the CCQ. Today, the CAT and CCQ recommended by GOLD (4) are used to assess the health status in patients with COPD. The International Primary Care Respiratory Group (IPCRG) developed a “users' guide to COPD wellness tools,” they analyzed 9 instruments and that revealed the CCQ as best. The CCQ has scored well on all criteria (validity/reliability, responsiveness, applicability to a primary care population, practicality/easy to administer, testing in practice and other language versions) for suitability for use in primary care. The CAT as second best also has scored well (20). However, the CAT is one-dimensional model which has not included mental health that the CCQ does (16,17).

The CCQ is a self-administered questionnaire developed to measure the health status of patients with COPD. It is also used to capture the effects of interventions for COPD patients. The CCQ measured the impact of COPD, not only regarding clinical status of the airways but also activity limitation and emotional dysfunction (17). The CCQ has been translated into over 60 languages for use by clinicians and is used to gain insight into a patient's health status and to measure the HRQOL of patients with COPD. Many studies have reported that the CCQ is valid and reliable (1,21–23).

Ringbaek and colleagues (24) revealed that the CCQ has the advantage of being easier and faster to complete than the SGRQ, especially for patients with low education levels. The CCQ has been suggested as a tool in daily clinical practice for measuring the health status of patients with COPD. An official translation of the Chinese Version CCQ for Taiwan is available. This translation has been extensively checked by the developer (17) and the translation office. There is no report of any validation of the Chinese Version CCQ. The specific aim of the study is to assess the reliability and validity of the Chinese Version CCQ in outpatients with COPD in Taiwan.

Methods

Study design and participant selection

A descriptive, cross-sectional design was conducted using structured questionnaires. This study was performed at the outpatient department of the Pulmonary Clinic of the Department of Internal Medicine at a medical centre in Central Taiwan. According to the patient's respiratory symptoms (dyspnea, chronic cough or sputum production), the history of exposure to risk factors of COPD, and the airflow limitation confirmed by the spirometry test, patients with a confirmed diagnosis of COPD based on the GOLD criteria made by a thoracic physician were selected using convenience sampling and enrolled after signing the informed consent form. The participants were enrolled from July 2014 to January 2015. The inclusion criteria included a spirometry test after bronchodilation with a FEV1/FVC ratio lower than 0.70. Pulmonary function predicted values were obtained from the multi-ethnic reference values for spirometry of the European Respiratory Society Global Lung Function Initiative (25). The exclusion criteria included asthma, heart failure, or other lung disease under active treatment, such as tuberculosis, lung cancer, or pneumonia.

Data collection

This study was approved by the Institutional Review Board of the Medical Centre. The researcher presented the purpose of the study, and the participants completed the informed consent forms before data collection.

After completing a section for collecting their demographic information and medical history, the patients completed the self-administered versions of 3 questionnaires including: the modified Medical Research Council (mMRC) dyspnea scale, the authorised Chinese Version CCQ provided by the developer (17), and the 12-item Short Form Health Survey (SF-12).

Instruments

CCQ

The CCQ consists of 3 domains and 10 items with an overall score: Symptoms (4 items), Functional State (4 items) and Mental State (2 items). All scores range from 0 (no impairment) to 6 (totally impairment); total scores range from 0 to 60, the main outcome measure of CCQ is the mean total score (divided with 10 items) with higher scores representing a worse health status and quality of life (17).

SF-12

The SF-12 was simplified from the 36-item Short Form Health Survey (SF-36) and developed for measuring generic HRQOL by Ware and colleagues. The 12-item health survey has physical and mental component summaries (PCS and MCS respectively) and consists of 8 domains: physical functioning, role limitation because of physical problems, bodily pain, general health, vitality, social functioning, role limitation because of emotional problems, and mental health. The scores range from 0 to 100, with higher scores indicating a more favourable health status (26).

mMRC dyspnea scale

The mMRC dyspnoea scale is a simple grading system for assessing dyspnea levels and is used for grading the effect of dyspnea on daily activities. There are 5 statements graded from 0 (I only get breathless with strenuous exercise) to 4 (I am too breathless to leave the house or I am breathless when dressing or undressing). Patients select the statement that most closely corresponds with their level of impairment (4). For assessing the severity of dyspnea, the GOLD (4) primarily recommends using the mMRC dyspnea scale. The mMRC scale is a reliable measure that correlates favourably with lung function measurements and is a tool for assessing symptoms in routine clinical practice (27).

Statistical analysis

The statistical analysis was performed using SPSS Version 18 (SPSS, Inc., Chicago, IL, USA). Data are expressed as means (SD) unless stated otherwise. Descriptive statistics were performed for continuous variables, and a data distribution assessment was conducted. Analysis of floor and ceiling effects in all domains in both the CCQ and the SF-12 were conducted. This was done by calculating the proportion of subjects that had highest possible score and the proportion of subjects that had lowest possible score in each domain.

The internal consistency of the CCQ was evaluated by calculating the Cronbach's alpha coefficient. Item–total correlations were evaluated using Spearman's rank correlation coefficient. The construct validity was established by testing the hypothesis that the severity of dyspnoea measured with the mMRC dyspnea scale is at least modestly associated with the CCQ scores. The hypothesis testing was examined using Spearman's rank correlation to determine the relationship between the CCQ scores and the mMRC scores. The convergent validity was examined using Spearman's rank correlation coefficient between the CCQ and the SF-12. The Kruskal–Wallis one-way analysis of variance (ANOVA) by ranks was used to determine the discriminant validity of the CCQ to differentiate among the classifications of COPD Groups A to D, according to the mMRC dyspnea scale, the spirometric classification and the exacerbations rate (4). Subsequently, the Mann–Whitney U-test was used to compare specific groups. Statistical significance was set at p < 0.05. In addition, confirmatory factor analysis (CFA) with Linear Structural Relationships (LISREL) was applied to examine the construct validity. LISREL is an iterative software package designed to evaluate the validity of structural equation models that consist of measurement and structural models (28).

Results

Score distributions and floor and ceiling effects

The distributions of all domain scores and the total score of the CCQ were skewed. In the study population (N = 114), 1.8% scored optimally ( = 0) in the symptom domain, 9.6% scored optimally in the functional state domain, and 46.5% scored optimally in the mental state domain; 50.9% of the study population showed a total score of less than 1.

Distributions in the all domain scores of the SF-12 were skewed, also. 12.3% scored optimally ( = 100) in the physical functioning domain, 68.4% scored optimally in the bodily pain domain, 74.6% scored optimally in the role limitation because of physical problems domain, 70.2% scored optimally in the role limitation because of emotional problems domain, 45.6% scored optimally in the vitality domain, 51.7% scored optimally in the mental health domain, 62.3% scored optimally in the social functioning domain, and 1.8% scored optimally in the general health domain.

Participant characteristics

A total of 114 participants completed the study. The participants ranged in age from 45–94 years, with a mean age of 71.83 (SD = 10.58); 94.7% of the participants were male. The demographics of this study population are displayed in Table 1.

Table 1. Characteristics of the subjects (N = 114).

Characteristics	n	(%)	Mean (SD)
Age (year)			72 (10)
BMI (kg/m^2)			24 (4)
Sex
Male	108	(95)
Female	6	(5)
FVC/FEV1			57 (11)
FEV1%			60 (19)
mMRC			1 (0.9)
Grade 0	18	(15)
Grade 1	49	(43)
Grade 2	36	(32)
Grade 3	11	(10)
GOLD Groups
Group A	52	(46)
Group B	24	(21)
Group C	17	(15)
Group D	21	(18)
Smoking habits
Ex-smokers	65	(57)
Smokers	35	(31)
Non-smokers	14	(12)

SD, standard deviation; BMI, body mass index; FVC, forced vital capacity; FEV₁, forced expiratory volume in 1 second; mMRC, modified Medical Research Council dyspnea scale; GOLD, Global Initiative for Chronic Obstructive Lung Disease; Group A: Low risk and less symptoms; Group B: Low risk and more symptoms; Group C: High risk & less symptoms; Group D: High risk and more symptoms.

Reliability

The Cronbach's alpha was 0.90 for the total score of the CCQ. The internal consistencies of the symptom, functional state, and mental state domains were 0.75, 0.87, and 0.82, respectively. The item–total correlation ranged from 0.459 to 0.877 (Table 2). This result indicated that all items contributed to the scale.

Table 2. Descriptive statistics and item-total correlations of the CCQ (N = 114).

Item	Mean	SD	Correlation with total^†
Shortness of breathing at rest	0.62	0.67	0.753^***
Shortness of breathing doing physical activities	1.39	1.44	0.877^***
Concerned about getting a cold or your breathing getting worse	0.96	1.32	0.761^***
Depressed because of your breathing problems	0.63	1.15	0.705^***
Did you cough	1.33	0.93	0.459^***
Did you produce phlegm	1.27	0.95	0.637^***
Strenuous physical activities	2.05	1.52	0.870^***
Moderate physical activities	1.44	1.45	0.807^***
Daily activities as home	0.74	1.11	0.733^***
Social activities	0.40	0.58	0.634^***

CCQ, Clinical COPD Questionnaire;

^†Spearman's rank correlation coefficient,

^***p < 0.001.

Construct validity

Hypothesis testing: Relationship between the CCQ and the mMRC

There was a moderate correlation between the CCQ scores and the mMRC dyspnea scale, particularly in the score of the functional state domain (ρ = 0.665, p < 0.001), providing support for the construct validity of the CCQ (Table 3).

Table 3. Correlations between CCQ with SF-12 and mMRC (N = 114).

	Clinical COPD Questionnaire
	Symptom	Functional state	Mental state	Total
SF-12
Physical functioning	−0.59^**	−0.80^**	−0.53^**	−0.75^**
Social functioning	−0.47^**	−0.60^**	−0.57^**	−0.59^**
Role physical	−0.49^**	−0.42^**	−0.37^**	−0.49^**
Role emotional	−0.49^**	−0.40^**	−0.41^**	−0.48^**
Mental health	−0.36^**	−0.40^**	−0.53^**	−0.45^**
Vitality	−0.57^**	−0.58^**	−0.64^**	−0.64^**
Pain	−0.40^**	−0.40^**	−0.42^**	−0.44^**
Health perceptions	−0.43^**	−0.48^**	−0.54^**	−0.53^**
mMRC dyspnea scale	0.594^**	0.665^**	0.474^**	0.668^**

^**p < 0.01; SF-12, 12-item Short Form Health Survey (higher score indicates better health status); mMRC, modified Medical Research Council dyspnea scale (higher grade indicates more severity of dyspnea).

Convergent validity

The CCQ scores showed significant correlations with all components of the SF-12 (Table 3). The functional state domain of the CCQ strongly correlated with the physical functioning component of the SF-12 (ρ = −0.80, p < 0.001). The total score of the CCQ also strongly correlated with the physical functioning component (ρ = −0.754, p < 0.001) and moderately correlated with the vitality component of the SF-12 (ρ = −0.64, p < 0.001). Correlation between the CCQ and the SF-12 overall scores were strong (ρ = −0.80, p < 0.001) for the entire population (Figure 1).

Figure 1. Relationship between CCQ scores and SF-12 scores.

Discriminant validity

The Kruskal–Wallis one-way ANOVA by ranks showed that the CCQ scores were significantly different among Groups A to D of the COPD patients (Table 4). The classification of the COPD group was based on combined assessments according to GOLD that included symptoms, degree of airflow limitation, and risk of exacerbations. Groups A and C had fewer symptoms, low- and high-risk of exacerbation, respectively. Groups B and D had a greater number of symptoms, low- and high-risk of exacerbation, respectively (4). In addition, the Mann–Whitney U-test revealed a significant difference in the CCQ scores between the groups. The symptom domain scores of the patients with COPD in Groups B and D were significantly higher (both p < 0.001) than those of patients in Groups A and C (p = 0.021 and p = 0.009, respectively).

Table 4. Comparison of CCQ scores in subgroups.

	Combine assessment of COPD
	Group A	Group B	Group C	Group D
	n = 52	n = 24	n = 17	n = 21
CCQ	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	X^2	P
Symptom	0.77 (0.51)	1.52 (0.81)	1.14 (0.96)	1.67 (0.69)	35.29	< 0.001
Functional state	0.60 (0.48)	1.70 (0.88)	1.00 (0.75)	2.00 (1.47)	35.57	< 0.001
Mental state	0.37 (0.59)	1.10 (1.31)	0.69 (0.85)	1.57 (1.59)	18.87	< 0.001
Total	0.62 (0.41)	1.51 (0.73)	0.99 (0.77)	1.78 (1.03)	37.57	< 0.001

CCQ, Clinical COPD Questionnaire; Group A: Low risk and less symptoms; Group B: Low risk and more symptoms; Group C: High risk and less symptoms; Group D: High risk and more symptoms; X², chi-square.

The functional state scores of patients with COPD in Groups B and D were significantly higher (both p < 0.001) than those of patients in Groups A and C (p = 0.011 and p = 0.035, respectively). The mental state scores of patients with COPD in Groups B and D were significantly higher than those of patients in Group A (all p < 0.001) but showed no significant difference with those of Group C. The CCQ total scores of patients with COPD in Groups B and D were significantly higher (both p < 0.001) than those of patients in Groups A and C (p = 0.023 and p = 0.015, respectively). There were no significant differences in the CCQ scores between Groups A and C or in the CCQ scores between Groups B and D.

Confirmatory factor analysis

The fit of the model was evaluated using structural equation modelling, and the CFA models were evaluated using unweighted least squares (ULS) estimation because the data violated multivariate normality. The goodness of fit was estimated using the chi-squared goodness-of-fit test, chi-squared/degrees of freedom ratio (χ²/df ratio), Goodness-of-Fit Index (GFI), Adjusted Goodness-of-Fit Index (AGFI), Comparative Fit Index (CFI), Incremental Fit Index (IFI), Normed Fit Index (NFI), Non-Normed Fit Index (NNFI), Parsimony Normed Fit Index (PNFI), and the Parsimony Goodness-of-Fit Index (PGFI). The results are summarised in Table 5. A nonsignificant result for the chi-squared goodness-of-fit test, a χ²/df ratio of 3 or less, PNFI and PGFI values of 0.5 or higher, and other indices of 0.9 or higher were considered indicators of the model's good fit (28,29). The CFA revealed the good fit of the model. The results confirmed the 3-factor model of the CCQ: the symptom, functional state, and mental state domains (Figure 2).

Table 5. Construct validity of the CCQ: Goodness of fit statistics

Model	χ^2 / df	GFI	AGFI	CFI	IFI	NFI	NNFI	PNFI	PGFI
CCQ (3-factor model)	4.18	0.99	0.98	0.99	0.99	0.97	0.99	0.69	0.58

CCQ, Clinical COPD Questionnaire; χ²/df, chi-square/degrees of freedom ratio; GFI, Goodness of Fit Index; AGFI, Adjusted Goodness of Fit Index; CFI, Comparative Fit Index; IFI, Incremental Fit Index; NFI, Normed Fit Index; NNFI, Non-Normed Fit Index; PNFI, Parsimony Normed Fit Index; PGFI, Parsimony Goodness of Fit Index.

Figure 2. Measurement model of the CCQ.

Discussion

Our study is the first evaluation of the Chinese Version CCQ and Taiwanese patients with COPD. This study indicated that the Chinese Version CCQ was a valid and reliable instrument for measuring the health status and HRQOL of patients with COPD. It exhibited good to excellent reliability, good convergent, and discriminate validity. Overall, the correlation between the CCQ scores and the mMRC dyspnea scale were moderate, between the CCQ scores and all components of the SF-12 were moderate to good. In addition, the CCQ scores were significantly different among Groups A to D of the COPD patients. Especially, the symptom domain scores and functional state scores of the patients with COPD in Groups B and D were significantly higher than those of patients in Groups A and C. Furthermore, we also found a moderate correlation between the symptom domain scores and the functional states domain scores of CCQ (ρ = 0.71). This finding supported the notion that dyspnea influenced physical activities and impaired HRQOL (30).

In cross-cultural comparisons, the original English version of the CCQ has excellent psychometric properties for measuring the health status of patients with COPD, and these properties are evident when the CCQ is used in various countries. Comparing our data with the results of the original CCQ development and validation study (17) showed similar internal consistency and validity. Comparing the severity groups of patients with COPD was impossible because of the original CCQ development and validation study (17). The classification of COPD patients into various groups has been changed from a degree of airflow limitation to a combined assessment (symptoms, degree of airflow limitation, and risk of exacerbations), according to GOLD (4). Therefore, this study used the new classification.^{Table 4Table 5}

Reliability

This study showed that the CCQ is a reliable questionnaire regarding internal consistency for measuring the health status of patients with COPD. The high Cronbach's alpha value (α = 0.90) indicated homogeneity among the individual items of the questionnaire. These results were similar to those of the Greek version by Tsiligianni et al. (31), Swedish version by Ställberg et al. (32), and Italian version by Damato et al. (33). The item–total correlations also supported the good internal consistency of the CCQ (their correlations were greater than 0.30).

Validity

Regarding the construct validity, there were moderate and positive correlations between the scores of all of the CCQ domains and the mMRC, particularly for the functional state and total scores. This result indicated the higher score of dyspnea following worsening quality of life that agreed with the previous studies (21,32,33). In other words, patients with more severe dyspnea had worse HRQOL. This result was similar to those of previous studies in which dyspnea affected the HRQOL values of patients with COPD (9,11,30).^{Figure 2}

In addition, the validated Chinese Version SF-12 was used as an instrument to measure the convergent validity of the CCQ. A strong correlation between the functional state domain of CCQ and the physical functioning component of the SF-12 was found. And there was a moderate correlation between the mental state domain of CCQ and the mental health component of the SF-12. Overall, moderate to high correlations were found in this study that were highly indicative of convergent validity, reflecting the original English validation study (18) and showing a result similar to the previous studies (31,33). This study also revealed that a moderate correlation between the symptom domain of CCQ and the vitality component of the SF-12.

This finding supported the previous studies that respiratory symptoms such as dyspnea were closely related to fatigue (34,35). Furthermore, psychological distress such as anxiety or depression was closely related to dyspnea and impact the overall health status in patients with COPD (7, 36). However, this study showed that moderate correlation between the mental state domain of CCQ and the all components of the SF-12. This finding indicated the mental health status was an influential factor of HRQOL in COPD patients and showing a result similar to the previous studies (7, 30,36).

Regarding discriminant validity, the CCQ showed a tendency to reflect differences in COPD severity. Patients with more COPD symptoms reported worse health status and HRQOL. The results showed that the scores of patients with COPD in Groups D and B were higher than those of Groups C and A in all domains of the CCQ. In this study, we also found significant differences in CCQ scores and various parts of the mMRC dyspnea scale (not shown in data). However, previous studies have assessed the discriminant validity to differentiate among the stages of COPD disease severity based on airflow limitation according to GOLD (17,31,33).

In recent years, the GOLD (4) has recommended using a combined assessment, not just the airflow limitation, for classifying COPD groups. In this study, we also analyzed the discriminant validity among the old GOLD COPD stages that revealed significant differences only in the symptom domain scores and the total scores of CCQ (not shown in data). To our surprise, the highest CCQ scores was stage IV COPD patients, but the second highest was stage II COPD patients. The lung function (FEV₁%) correlated poorly with each domain scores of the CCQ (not shown in data). However, the results of the present study support the good discriminant validity of the CCQ. This finding indicated the symptoms of COPD were more closely related to functional state than lung function. HRQOL was affected more by symptoms than by changes in lung function.

In model fit analysis, one of the most commonly used statistical methods for estimating parameters in the CFA is maximum likelihood (ML). An ML estimation is performed by assuming a continuous and multivariate normal distribution of observed variables. This assumption is not supported when the observed data are obtained using ordinal scales (e.g., Likert scale). Otherwise, using ML to estimate parameters with violations of multivariate normality might result in biased parameters and ML estimates with standard errors (28,37). Because the score distributions did not show normality in this study, a ULS estimation was used, which did not require assuming multivariate normality. In analysing the model fit of the CCQ, we compared the model fit indices of the ULS and ML estimations to determine whether the model fit indices of the ULS estimation were more favourable than those of the ML estimation (not shown in data). The results indicated that the scores of the CCQ showed similar variation in the participants of this study.

Regarding the goodness of fit assessment of the CCQ model, the results of the present study showed that the chi-squared value was statistically significant, indicating some differences between the data and the model. A nonsignificant chi-squared value indicates a data–model fit. The chi-squared goodness-of-fit test was considerably influenced by the sample size and violations of multivariate normality (37). Therefore, it is critical not to rely on only one goodness-of-fit test in determining the model fit. In the present study, the results of other fit indices (i.e., GFI, AGFI, CFI, IFI, NFI and NNFI) were higher than 0.90, indicating a good model fit. The PNFI and PGFI values were higher than 0.50, which also supported a good model fit. Furthermore, the t value (range: 10.50–29.08) of each observed variable was high (absolute values were higher than 2.58; not shown in data), which indicated that the estimated parameters reached a significant level. In other words, all of the standardised coefficients showed a high level of significance, which indicated that the items could be effective indices of their respective dimensions and used to reflect the dimensions. Overall, the CFA (assessing model fit) results indicated that the CCQ measurement model exhibited a good model fit.

Limitations

This study had several limitations. The data was collected in the outpatient department of a medical centre, and differences in the pattern of medical care may have influenced the study results. Our participants were mainly low-risk COPD patients with few symptoms (Group A). Because our study setting was an outpatient department, and the patients showed mild symptoms and were relatively stable, the subgroup totals were not equal. However, this factor did not influence the results. Another limitation of this study might be the subject selection; the subjects of this study were predominantly men (94.7%). This was a limit to generalization of this research results to all COPD patients if gender difference. In addition, this study was a cross-sectional design that was unable to perform the stability (test-retest reliability) of the CCQ; therefore the reliability test was uncompleted. However, the findings of this study support the CCQ was a valid and reliable instrument for measuring the health status in patients with COPD.

Conclusions

Our results show that the Chinese Version CCQ has good-to-excellent measurement properties. The CCQ is a valid and reliable instrument for assessing the health status and HRQOL of patients with COPD and is able to discriminate the disease severity of COPD patients (group level). Furthermore, the CCQ is short, easy to score, and allows data to be collect quickly. It is suitable for use in daily clinical practice or clinical trials.

Acknowledgments

We are grateful for the cooperation of the participants in this study and for the assistance from the staff at the Pulmonary Clinic of the Department of Internal Medicine at Chung Shan Medical University Hospital.

Declaration of interest statement

The authors of this paper declare no competing interests. The authors alone are responsible for the content and writing of the paper.