Development of a Diagnostic Questionnaire for Damp Phlegm Pattern and Blood Stasis Pattern in Coronary Heart Disease Patients (CHD-DPBSPQ)

Background The aim was to develop a diagnostic questionnaire for damp phlegm pattern and blood stasis pattern in coronary heart disease patients (CHD-DPBSPQ). Methods The standard procedures of questionnaire development were carried out to develop and assess CHD-DPBSPQ. The patients were assessed using the CHD-DPBSPQ, CHD-DPPQ, and CHD-BSPQ. Four methods were used to select the items on the CHD-DPBSPQ in a pilot study based on data from a Guizhou tertiary grade A hospital. Cronbach's alpha and the split-half reliability, test-retest reliability, content validity, criterion validity, construct validity, and convergent validity were determined in a validation study using a nationwide sample. Results After item selection, the CHD-DPBSPQ contained 15 items in two domains: the phlegm domain (9 items) and the blood stasis domain (6 items). For the CHD-DPBSPQ, the alpha coefficient was 0.88, the split-half coefficient was 0.90, and the intraclass correlation coefficient was 0.83. The range of the item-level content validity index (I-CVI) was 0.71 to 1.0 and that of the scale-level content validity index/average (Scale-CVI/Ave) was 0.97. The domain scores on the CHD-DPBSPQ were in close relation to the scores on a questionnaire for damp phlegm pattern in coronary heart disease patients (CHD-DPPQ) and a questionnaire for blood stasis pattern in coronary heart disease patient (CHD-BSPQ) (P < 0.01). The root mean square error of approximation (RMSEA) was equal to 0.05 (90% CI: 0.044, 0.059). Convergent validity was demonstrated with a moderate correlation. Conclusion The CHD-DPBSPQ is a reliable and valid instrument.


Introduction
Currently, coronary heart disease (CHD) is the most common type of cardiovascular disease (CVD) and is a major cause of death and disability among adults worldwide [1,2]. CHD causes approximately one-third of all deaths in people older than 35 years in Western countries [3]. CHD still affects more than 10 million people in China, and an estimated 7.4 million people suffer from CHD every year. In particular, approximately 3 million Chinese individuals die of CHD each year; the mortality rate is second only to that of cerebrovascular disease [4,5]. is may be related to increased serum cholesterol levels caused by smoking and dietary changes [6]. Under the influence of diabetes, hypertension, and hyperlipidaemia, the incidence of CHD increases annually.
Damp phlegm pattern and blood stasis pattern (DPBSP) is commonly found in CHD patients. Along with changes in lifestyle habits, DPBSP is becoming increasingly common [7][8][9]. Turbid phlegm and blood stasis are two important risk factors in the development of DPBSP. According to Chinese Medicine (CM), turbid phlegm is formed by body fluids, and its clinical manifestations include cough with sputum, chest tightness, dizziness, body fat accumulation, atherosclerosis, and hyperlipidaemia. Blood stasis is caused by illiquidity, and its clinical manifestations include tingling, localized pain, enclosed masses, dark purple tongue, and hemorheology [10]. Patients with CHD-DPBSP tend to have both phlegm and blood stasis pathological manifestations [11]. Bi et al. reported that DPBSP is found in 73.42% of CHD patients, and it is a primary syndrome in these patients [12].
TCM syndromes can be studied using measures such as scales. In recent years, some researchers have developed diagnostic questionnaires for CHD with stable angina (syndrome involving both phlegm and blood stasis) [13][14][15]. However, these questionnaires have some shortcomings. First, the process of developing these scales is not standardized. e method of determining construct validity is not suitable, as construct validity should be determined through confirmatory factor analysis (CFA) rather than exploratory factor analysis (EFA) [16]. Second, the items contained in these questionnaires are largely divergent; thus, the diagnosis of CHD is not systematic or consistent. For example, among the four diagnostic methods of TCM (inspection, listening and smelling, inquiry, and palpation), palpation was the main item in one diagnosis questionnaire [15], but we cannot find this aspect in others [13,14]. Some experts insist that tongue and pulse information are also very important for the diagnosis of DPBSP and can increase the diagnostic accuracy [17][18][19][20][21]. erefore, we aimed to develop and assess a damp phlegm pattern and blood stasis pattern questionnaire for patients with coronary heart disease (CHD-DPBSPQ) based on a series of standard and systematic procedures of instrument development.

Methods
To develop the CHD-DPBSP diagnostic questionnaire, we used standard procedures for developing and validating the CHD-DPBSPQ [22][23][24][25][26][27][28][29]. e procedures included the following steps: construct definition, item generation, a pilot study, and a validation study. Construct definition and item generation were used to define the structure and generate the initial items for item selection. As this methodology has been reported in other papers [30][31][32][33], it will not be reported in detail in this study. e pilot study was used to select the items for the CHD-DPBSPQ, and the validation study was applied to assess the reliability and validity of the CHD-DPBSPQ.

Item Source.
All the items on the CHD-DPBSPQ were drafted from systematic reviews, the Delphi method and analytic hierarchy process (AHP) [31][32][33]. Ultimately, 20 items were generated for CHD-DPBSPQ20 [34]. According to the theory of TCM, (1) chest distress, sleepiness, physical heaviness, obesity, sticky mouth, greasy tongue fur, slippery pulse, wiry pulse, abdominal fullness, anorexia, viscous stool, and taut and slippery pulse were used to assess the phlegm pattern. (2) Chest pain, cyanotic lips, dim complexion, dark purple tongue, petechiae or ecchymosis on the tongue, sublingual vein cyanosis, uneven pulse, and taut and uneven pulse were used to assess the blood stasis pattern in CHD patients.

Samples.
A pilot study was adopted to identify items for the CHD-DPBSPQ. Recruited from a Guizhou tertiary grade A hospital, all of the CHD patients with DPBSP provided informed consent prior to their participation. e eligibility criteria are as follows: (1) diagnosis of CHD according to the Chinese guidelines for diagnosed with CHD based on the Chinese guidelines for the diagnosis and treatment of patients with chronic stable angina published in 2007 [35]; (2) diagnosis of DPBSP by 2 experienced experts [36]; (3) ≥40 years of age; and (4) provided informed consent for participation. Patients diagnosed with other syndromes or other diseases were excluded.

Analytical Methods.
Four methods were employed to select the items for the CHD-DPBSPQ: (1) the coefficient of variation: if the standard deviation (SD) of every item was less than 0.9, the item was discarded; (2) EFA: if the correlation coefficient was less than 0.4 after factor rotation, the item was deleted; (3) alpha reduction: if Cronbach's alpha coefficient was apparently increased if one item was removed, that item was marked; and (4) correlation analysis: if the item had a proportion less than 0.4, the item was abandoned [37][38][39]. If an item met three or more of the abovementioned criteria, it was removed. Two TCM doctors with over 20 years of experience in CVD performed the assessments, and the doctors explained the questionnaire to each patient. e questionnaire included a demographic portion, followed by the CHD-DPBSPQ, CHD-DPPQ, and CHD-BSPQ. e demographic portion collected data on age, sex, ethnicity, marital status, occupation, and other diseases. e CHD-DPPQ and CHD-BSPQ were used to evaluate the criterion validity. e CHD-DPBSPQ was used for diagnosis by two experienced TCM doctors, and these two doctors could independently differentiate between the syndromes. If their diagnoses were inconsistent, diagnosis was made by a third doctor (associate professor or higher). Terwee et al. [40] believed that more than 50 samples should be assessed to determine test-retest reliability. At least 50 patients in ward settings were assessed by applying the CHD-DPBSPQ within 1-7 days, which is an applicable period to assess test-retest reliability.

Methods
Used to Evaluate the Scale. SPSS version 22.0 and Amos software 22.0 were used for data analysis [41,42].
is scale was evaluated using classical test theory (CTT), including reliability and validity measures.
Internal consistency reliability, split-half reliability, and test-retest reliability were all evaluated [16,43]. e internal consistency reliability was estimated using Cronbach's α. After the arrangement of odd and even numbers, split-half reliability was calculated using Pearson's correlation coefficients. Test-retest reliability was assessed by adopting the intraclass correlation coefficient (ICC) and its 95% confidence interval.
Validity was evaluated as content validity, criterion validity, construct validity, and convergent validity. e content validity of the questionnaire was assessed using the item-level content validity index (I-CVI) and the scale-level content validity index (S-CVI). When more than 6 experts assessed this questionnaire, the value of the I-CVI was not less than 0.78 [44][45][46]. We had 7 experts assess this questionnaire, and the value of the S-CVI/Ave exceeded 0.90, which indicated a high degree of content validity. e correlation coefficients between the CHD-DPPQ and the CHD-BSPQ were calculated to assess the criterion validity. CFA was used to evaluate whether the theoretical model was suitable for the data [28]. e goodness of fit index (GFI), adjusted goodness of fit index (AGFI), comparative fit index (CFI), and normed fit index (NFI) were greater than 0.9, indicating that the model was suitable. e root mean square error of approximation (RMSEA) was 0.05, indicating that the fit was close to good [47]. e correlation coefficients of the subscales and items were equal to or greater than 0.4, suggesting good convergence validity [48].

Pilot Study.
Ultimately, 103 CHD patients with DPBSP were involved in the pilot study. e age of the patients ranged from 46 years to 91 years (mean age: 68.8, Table 1); 60.2% of the sample was male, and 91.3% of the sample were retirees. Table 1 presents a brief demographic summary of the sample. e Cronbach's alpha coefficient of the CHD-DPBSPQ20 was 0.835. e results of item selection are shown in Table 2.
Of the 20 items, 5 items were removed from the original pool based on the selection criteria (n ≥ 3), namely, taut and slippery pulse, viscous stool, wiry pulse, uneven pulse, and taut and uneven pulse.
us, CHD-DPBSPQ ultimately contained 15 items in 2 domains-the phlegm syndrome domain and the blood stasis syndrome domain. (1) e phlegm syndrome domain included chest distress, sleepiness, physical heaviness, obesity, sticky mouth, abdominal fullness, anorexia, greasy tongue fur, and slippery pulse. (2) e blood stasis syndrome domain included chest pain, cyanotic lips, dim complexion, dark purple tongue, petechiae or ecchymosis on the tongue, and sublingual vein cyanosis.

Validation Study.
To further improve the demographic composition of the sample, data were collected from the clinical population of 8 hospitals in China. e demographic characteristics of the 729 total participants are shown in Table 1. Statistically, the mean age was 67.5 ± 10.6 years (range, 32.0-91.0 years), and a total of 460 male and 269 female patients were included. Among these patients, 99.3% were of Han ethnicity, 67.5% were retirees, 99.5% were married, and 88.0% had another disease. e sample (N � 729) was measured for the first time, with 81 of the patients measured a second time. Table 2 shows the item distributions, which were assessed using means, standard deviation (SD), and missing data. In total, 729 patients were enrolled and completed the questionnaire. All the item scores were between 0 and 4 ( Table 3). Chest distress had the highest score (1.94), while sleepiness had the lowest (0.45). Chest pain, sticky mouth, and cyanotic lips had missing values.

Reliability.
e CHD-DPBSP had high internal consistency and retest reliability. e mean score on the CHD-DPBSPQ was 17.4, the mean score on the phlegm domain was 9.7, and the mean score on the blood stasis domain was 7.7 (Table 4). ① e Cronbach's alpha values for both domains were greater than 0.75 (Table 4). ② e split-half coefficients of the domains were greater than 0.80 (Table 4). ③ e interitem correlation coefficient (ICC) of the domains were 0.78, 0.83, and 0.83, which were more than 0.8, and the retest reliability coefficient was high (Table 4).

Validity.
e CHD-DPBSPQ had good validity. ① For content validity, the range of the I-CVI was 0.71 to 1.0, and the S-CVI/Ave was 0.97 (Table 5). ② e criterion validity was checked by comparing the scale with the CHD-DPPQ and the CHD-BSPQ. e correlation coefficient between the phlegm domain and the CHD-DPPQ was 0.76, and the correlation coefficient between the blood stasis domain and the CHD-BSPQ was 0.96 (P < 0.01). ③ e model fit for the scale was tested using CFA (Figure 1). e GFI, AGFI, NFI, IFI, TLI, and CFI were all greater than 0.90, and the RMSEA was 0.05 (90% CI: 0.044, 0.059) ( Table 6). All these indexes indicated that the model fit was good.④ Convergent validity was demonstrated by a moderate correlation (0.423-0.796), as shown in Table 7.

Discussion
At present, the specific scales used in patients with CHD abroad are the Seattle angina scale (SAQ) [49] and the cardiovascular limitations and symptoms profile (CLASP) [50]. Nevertheless, owing to cultural differences, we researched and developed the CHD-DPBSPQ to diagnose DPBSP in Chinese CHD patients.
Previous instruments for the assessment of CHD-DPBSP have not been widely adopted; our group formulated and validated TCM outcomes on the basis of standard development and validation procedures [22,28,40]. e US FDA and the WHOQOL group proposed the establishment of a conceptual framework for questionnaires [23,26]. According to TCM theories, the framework for DPBSP should be classified into damp phlegm and blood stasis patterns [10,11]. e diagnosis of DPBSP requires the simultaneous diagnosis of sputum and blood stasis [11,20,51,52]. However, these questionnaires [13][14][15] all contain different frameworks. For example, one questionnaire involved the division of dimensions into phlegm and blood stasis [14], but the other two did not [13,15]. Moreover, some important validity coefficients, such as convergent validity, were not evaluated. e method of determining content validity was not suitable, as content validity should be determined according to the CVI [45], but none of the three questionnaires used the CVI [13][14][15]. e method of construct validity was also not suitable, as  construct validity should be determined via CFA rather than EFA [16]. e 15 items on the CHD-DPBSPQ were empirically selected by four methods using a sample population from a Guizhou tertiary grade A hospital. e CHD-DPBSPQ was verified with national data, showing that it is a reliable and valid tool for research and clinical trials. (1) e internal consistency of the domains was expressed by Cronbach's α coefficients (0.77-0.88) and split-half reliability coefficients (0.81-0.90). Terwee et al. considered that the Cronbach's α showed a good internal consistency range, from 0.70 to 0.95 [40], and the CHD-DPBSPQ had good internal consistency. Moreover, eighty-one patients completed the CHD-DPBSPQ a second time, and the ICC was between 0.78 and 0.83, indicating good reproducibility [40].
(2) e content validity of the questionnaire was evaluated by seven experts. A standard value of I-CVI greater than 0.78 was recommended by Lynn [53]; the score for sticky mouth was 0.71, but the Kappa-like index (K) was 0.65 (>0.60), which showed good validity [54]. e S-CVI for the   scale was 0.97, which meets the requirements of the standard value (0.90) recommended by Polit and Beck [55]. e correlation coefficient between the phlegm domain and the CHD-DPPQ was 0.76, and the correlation coefficient between the blood stasis domain and the CHD-BSPQ was 0.96. eir values were greater than 0.7, which indicated positive criterion validity [40]. e RMSEA was equal to 0.05 (90% CI: 0.044, 0.059), and the NFI, NNFI, and CFI were greater than 0.90. ese results indicated that the scale had good construct validity [47]. e correlations of the CHD-DPBSPQ between the phlegm domain and our hypothesized components (chest distress, sleepiness, physical heaviness, obesity, sticky mouth, abdominal fullness, anorexia, greasy tongue fur, and slippery pulse) were greater than 0.4, as were the correlations of the CHD-DPBSPQ between the blood stasis domain and our hypothesized components (chest pain, cyanotic lips, dim complexion, dark purple tongue, petechiae or ecchymosis on the tongue, and sublingual vein cyanosis), indicating moderate convergence validity [48]. In short, these findings are encouraging and support the structural integrity of the CHD-DPBSPQ. e purpose of our study is to develop and validate the CHD-DPBSPQ, a questionnaire intended for use in the clinical practice of TCM. However, the questionnaire is not applicable for diagnosing CHD and all TCM syndromes of CHD. us, the diagnosis of CHD was first confirmed in patients recruited for our study according to the Chinese guidelines for the diagnosis of CHD, which were based on the Chinese guidelines for the diagnosis and treatment of patients with chronic stable angina published in 2007 [35]. e CHD-DPBSPQ was developed on the basis of the theory of TCM syndromes. Meanwhile, structural equation modelling is a valid method for testing TCM syndromes, as described in other studies [23,29]. Up to now, other researchers have formulated and exerted similar methods to explore TCM syndromes [16,24]. For example, Chen et al. developed the LIDHS questionnaire and evaluated its reliability and validity as a "TCM Syndrome Questionnaire of Ulcerative Colitis" [16]. e CHD-DPBSPQ, as an effective and reliable pattern of TCM questionnaire, can also be used to diagnose CHD-DPBSP in patients.
e CHD-DPBSPQ should be further assessed throughout every province in the country. (4) e gold standard adopted was the CHD-DPPQ and CHD-BSPQ developed in China, which do not include the internationally recognized scale for CHD.
is questionnaire also has some strengths. First, compared with one published scale [13], our questionnaire was more convincing in terms of verification. Second, the development of a diagnostic CHD-DPBSPQ is rare. ird, 100% of the patients completed the questionnaire. Fourth, the content of our scale was easy to understand, and the average completion time for the questionnaire was 5.4 minutes. Finally, it was confirmed that the results were generalizable by using a multicentre clinical investigation.

Conclusions
A DPBSP questionnaire was developed for patients with CHD. e 20 items were successfully reduced to 15 items during the selection process, a quantitative stage of questionnaire development.
e CHD-DPBSPQ showed good reliability and validity and is feasible for diagnosing DPBSP in CHD patients. We recommend the application of the CHD-DPBSPQ for diagnosing DPBSP in Chinese CHD patients.

CHD-DPBSPQ:
Questionnaire of damp phlegm pattern and blood stasis pattern for coronary heart disease CVI: Content validity index I-CVI: Item-level content validity index S-CVI: Scale-level content validity index CHD-DPPQ: Questionnaire of damp phlegm pattern for coronary heart disease CHD-BSPQ: Questionnaire of blood stasis syndrome for coronary heart disease CM: Chinese medicine CFA: Confirmatory factor analysis QOL: Quality of life SD: Standard deviation CTT: Classical test theory ICC: Intraclass correlation coefficient RMSEA: Root mean square error of approximation DPBSPQ: Damp phlegm pattern and blood stasis pattern GFI: Goodness of fit index AGFI: Adjusted goodness of fit index CFI: Comparative fit index NFI: Normed fit index RMSEA: Root mean square error of approximation SAQ: Seattle angina scale CLASP: Cardiovascular limitations and symptoms profile.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Ethical Approval
e Research Ethics Committee of Tianjin University of Traditional Chinese Medicine provided approval for this study (no. TJUTCMEC2015000).

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.

Authors' Contributions
Ge Fang participated in the study design, analysed the data, and wrote and modifed the manuscript. Ling-lin Zhang participated in the study design and drafted the manuscript. Qi Ren and Xiao-wen Zhou helped to modify the manuscript. Bin Wang and Xuan Zhou participated in the study design and data collection. Xiao-qi Liu and Dan-hong Peng participated in the data collection. Xin-lin Chen greatly contributed to the data analysis and manuscript review. Xian-tao Li greatly contributed to the conception of the study and study design. All authors read and approved the final manuscript.