Validated questionnaire on diagnosis and symptom severity for functional constipation in the Chinese population

Aliment Pharmacol Ther 2005; 22: 483 488. doi: 10.1111/j.1365-2036.2005.02621.x Validated questionnaire on diagnosis and symptom severity for functional constipation in the Chinese population A. O. O. CHAN*, K. F. LAM, W.M.HUI*,W.H.HU*,J.LI*,K.C.LAI*,C.K.CHAN*,M.F.YUEN*, S.K.LAM*&B.C.Y.WONG* Departments of *Medicine and Statistics and Actuarial Science, University of Hong Kong, Hong Kong Accepted for publication 5 July 2005 SUMMARY Background: Functional constipation is a common problem in clinical practice. No validated questionnaire is available in Chinese. Aim: To develop a validated questionnaire for diagnosis and symptom assessment in functional constipation for the Chinese population. Methods: One hundred and eleven patients with constipation and 110 healthy controls were presented with a 24-item constipation questionnaire in the Chinese language. Quality of life in constipation patients was assessed by Short Form-36. Polyethylene glycol was prescribed, for 4 weeks, to 20 patients with newly diagnosed constipation. The questionnaire was administered before and 4 weeks after treatment. Concept, content, construct, discriminant validity and reliability of the questionnaire were assessed. Results: Six items were selected by logistic regression to account for most of the differences between controls and constipated patients with a good reproducibility and internal consistency. A cut-off score of 5 was determined to discriminate between controls and constipated patients with a sensitivity of 91% and a specificity of 91%. The constipation questionnaire correlated negatively with seven domains of the Short Form-36 and discriminated between constipated patients who reported symptomatic improvement during polyethylene glycol treatment. Conclusions: The Chinese constipation questionnaire could be used in epidemiological studies to assess the frequency and severity of constipation in patient populations and in interventional studies of constipation. INTRODUCTION Functional constipation is a common problem in clinical practice. In the western population, the prevalence was reported to be as high as 24% in elderly subjects and more commonly among women. 1 We observed a prevalence of 14% of constipation in Asian population. 2 Most current clinical definitions of constipation are based on a measure of frequency of defecation alone, or in combination with other symptoms such as discomfort or straining to defecate. 3 Stool frequency alone is not Correspondence to: Dr A. O. O. Chan, Department of Medicine, University of Hong Kong, Queen Mary Hospital, Hong Kong. E-mail: aoochan@hku.hk generally used for diagnosis because of its poor correlation with objective measures. 4, 5 The Rome II criteria were developed to assist diagnosis of functional constipation. 6 However, these criteria may not apply when the patient is taking laxatives. In addition, the Rome II criteria do not assess the severity of constipation. Supplementary to the Rome II criteria, there were several clinician rating scales for assessing the severity of constipation, such as the Constipation Scoring System 7 and the Constipation Assessment Scale. 8 Owing to differences in language and perception of illness, individual symptom scales may not be valid across different cultures. No validated constipation symptom score exists in the Chinese language. Before a symptom assessment tool can be used clinically, various Ó 2005 Blackwell Publishing Ltd 483

484 A. O. O. CHAN et al. questions on reliability and validity of the instrument need to be addressed. Thus, the aim of this study was to develop and validate a questionnaire for Chinese subjects with constipation for both diagnosis and severity, and assessment of symptoms with treatment. METHODS Patients and controls Controls were recruited from healthy subjects, aged 18 years and over. Those under active medical care for gastrointestinal complaints or taking regular medication for other medical illness were excluded. The constipated subjects had been assessed by a gastroenterologist and had chronic constipation, defined by the Rome II criteria as follows: at least 12 weeks, which need not be consecutive, in the preceding 12 months with two or more of the following conditions. Less than three bowel movements per week, straining at stool more than 25% of the time, passage of lumpy or hard stools more than 25% of the time, sensation of incomplete evacuation for more than 25% of the time, sensation of anorectal obstruction/blockage for more than 25% of the time and manual manoeuvres to facilitate more than 25% of the time of defecation. In addition, loose stools are not present, and there were insufficient criteria for the diagnosis of irritable bowel syndrome. 6 Additional exclusion criteria were inability to understand Chinese, cognitive impairment affecting understanding of the questionnaire, history of malignancy and significant systemic disease. The study was approved by the Research and Ethics Committee of the Faculty of Medicine, University of Hong Kong. Procedures: questionnaire administration The questionnaire was administered at the clinic during the first visit with informed consent. The questionnaire consisted of 30 gastrointestinal questions; and 36 questions of a previously validated Chinese version of the Short Form (SF)-36 quality of life scale. 9 The questionnaire on gastrointestinal symptoms was developed from the reviews of published papers and interviews with patients. The most frequently used and salient symptom of constipation was identified from the Rome II criteria. They were graded according to a 5-point Likert scale, 1 5 from asymptomatic to very severe symptoms), 10 plus questions on the use of laxatives and satisfaction of bowel motions. These questions were discussed among the members of the research team comprising of gastroenterologists and family doctors to form the initial questionnaire. The gastrointestinal symptoms in the preceding 2 weeks and the preceding 3 months were asked. The questions were translated into Chinese and back-translated and compared with the original questions. Differences were modified accordingly. A batch of 28 patients with constipation was interviewed to discuss these symptoms and to volunteer further symptoms that they encountered. The response and wording were discussed, and the questionnaire was modified accordingly. The final questionnaire contains 24 questions and both frequency and severity of the symptoms were assessed. In addition, the patients were asked whether the wording of the individual items were easily understood. The questionnaire was administered again at 2 weeks after completion of the first questionnaire. Comparison between the two responses was used to determine test retest reproducibility of individual responses. Drug administration and follow-up Forlax (polyethylene glycol 4000) (Ipsen, Paris, France) two sachets per day for 4 weeks was given to the constipated patients. Forlax has been shown to be efficacious in constipated subjects with a response rate of 70%. 11 Statistics Statistical analysis was performed using spss (SPSS Inc., Chicago, IL, USA) and jmp (SAS Institute, Cary, NC, USA) statistical software. Discriminatory power and aspects of questionnaire reliability and validity were assessed. Discriminatory power of the questionnaire. In performing cross-section comparisons, an instrument will need to have good discriminative power to detect differences between individuals and groups at a given point in time. 12 The ability of the questionnaire to distinguish between normal controls and patients with constipation was tested using a multiple logistic regression model with group as the dependent variable and the items as the explanatory variables. The independent variables are highly correlated and hence the problem of multicollinearlity exists and the usual variable selection

CHINESE CONSTIPATION QUESTIONNAIRE 485 procedure would be extremely difficult to be implemented. A two-stage procedure was adopted. As the questionnaire was divided into four major sections, in the first stage, a nominal logistic regression analysis was carried out for each section, with the variable group being the response variable and the independent variables were the questionnaire items in the section. By the backward elimination method that the variable with the largest P-value will be eliminated from the working model at each step until all P-values are <0.15. The independent variables that remained in the final model of each section will enter the second stage as an independent variable to start with. Then the same backward elimination procedure is performed again with the set of independent variables being those selected variables in the first stage. A composite score was generated by principal component analysis (PCA). A cut-off score that best discriminated between constipation and healthy controls were determined by plotting a receiver operating characteristic (ROC) curve. Sensitivity and specificity were determined. Reliability and internal consistency. Test retest reliability of a measurement tool refers to the stability of measurement under different conditions. In the present study, test retest reliability was assessed between baseline and retesting after an interval of 10 14 days. In the present study, test retest reliability was determined by the intraclass correlation coefficient between the first and second responses to questionnaire administration. An intraclass correlation of 0.7 or above is generally considered to be sufficient to demonstrate reliability. Internal consistency of the questionnaire items was assessed by the Cronbach s-a coefficient. Cronbach s-a value of 0.7 is generally considered to be sufficient to demonstrate internal consistency. 13 Content validity. Content validity relates to the choice, appropriateness, importance and representativeness of the instrument s content. 12 The individual components in the questionnaire must be seen as being comprehensible and relevant to the multiple gastrointestinal symptoms suffered by constipated patients. In the present study, subjects were asked if the individual items assessed in the questionnaire were easily understood. Relevance of symptoms to disease was assessed in the constipation patients. We arbitrarily defined a question to have poor face validity if <80% of subjects rated the item as being comprehensible and relevant. Construct validity. A construct is a theoretical idea developed to explain and to organize some aspects of existing knowledge. 14 Construct validity, in this aspect, refers to evaluation of a test by looking at the relationship between the test and the various phenomena which the theory predicts. In the present study, it was assumed that patients with more severe constipation symptoms would have a lower quality of life. The total constipation score calculated from the items was correlated with SF-36 quality of life domains by the Kendall s tau coefficient. Discriminant validity. Discriminant validity is a measure of the tool s ability to distinguish clinically significant differences in therapeutic responses between-patients and within-patients over time. 15 In the present study, we assessed discriminant validity in constipated patients by administering the questionnaire before and after 4 weeks of Forlax treatment. Patients reporting an improvement in constipation were expected to have a lower symptom severity score. The symptom score between patients who reported improvement in symptoms was compared with the score in patients who reported no change or worsening in symptoms using the Kendall s tau coefficient. RESULTS A total of 221 subjects were recruited, including 111 patients (106 female, age 38.5 ± 12 years, range: 8 74) and 110 controls (70 female, age 34.2 ± 12 years, range: 18 77). All patients completed the study with no defaulters. Discriminatory power Finally six items were left in the final model. The questions were: (i) severity of false alarm (feeling like that you have to pass a stool but you cannot), Table 1. The sensitivity, specificity and area under the receiver operating characteristic (ROC) curve using the 6-item model and the composite score Sensitivity (%) Specificity (%) Area under the ROC curve 6-item model 92 94 0.9870 Composite score model 91 91 0.9776 (sum of the six items)

486 A. O. O. CHAN et al. Table 2. Questionnaire items In the past 2 weeks In the past 3 months Patients (%) Controls (%) Patients (%) Controls (%) Assessment of bowel function Frequency of false alarm 45.9 0.9 48.6 3.6 Severity of false alarm* 73.7 2.5 75.5 5.3 <3 defecations/week* 61.2 7.3 73.8 6.3 Frequency of incomplete evacuation 60.9 2.7 70.2 6.4 Severity of incomplete evacuation* 67.0 4.2 71.6 5.4 Frequency of anorectal obstruction/blockade 39.5 0 54.9 0.9 Severity of anorectal obstruction/blockade 67.9 0 74.1 0 Frequency of manual manoeuvres to facilitate defecations 17.1 0 22.5 0 Severity of manual manoeuvres to facilitate defecations 67.8 0 76.5 0 Frequency of straining 70.2 6.5 76.5 19.1 Severity of straining 73.2 9.7 82.3 10 Frequency of painful (at the abdomen) bowel movement? 38.7 3.6 41.4 6.3 Severity of painful (at the abdomen) bowel movement? 48.8 4.0 57.2 7.4 Assessment of stool and rectal symptoms Frequency of lumpy or hard stools 65.5 1.8 70.2 9.1 Severity of lumpy or hard stools* 77.0 5.5 82.3 6.4 Frequency of stools that were too small 47.7 1.8 55.8 3.6 Severity of stools that were too small 63.9 0.9 71 2.7 Frequency of rectal burning during or after a bowel movement 32.6 1.8 33.3 2.7 Severity of rectal burning during or after a bowel movement 60.6 1.8 60.3 2.7 Frequency of rectal bleeding or tearing 21.6 1.8 29.7 1.8 Severity of rectal bleeding or tearing 41.1 0.9 52.8 4.5 Frequency of mucus per rectum 31.6 0.9 34.8 0.9 Severity of mucus per rectum 46.9 0.9 50.9 2.7 Time taken to pass stool 50.5 0 49.5 0 Assessment of laxatives used Number of laxatives used* 53.0 2.3 53.2 2.7 Number of enema/supportive used 28.7 1.5 30.6 1.8 Assessment of abdominal symptoms Frequency of abdominal pain 35.9 4.5 41.4 4.5 Severity of abdominal pain 58.9 4.5 63.0 10.9 Frequency of abdominal bloating 70.2 6.4 75.6 12.7 Severity of abdominal bloating* 73.0 3.6 79.4 10.9 Frequency of stomach discomfort 25.2 3.6 29.7 8.2 Severity of stomach discomfort 53.4 3.6 54.8 13.6 Frequency of stomach pain 11.7 2.7 14.4 5.5 Severity of stomach pain 42.6 4.5 49.3 10 Frequency of stomach bloating 30.6 3.6 34.2 6.4 Severity of stomach bloating 63.1 3.6 64 7.3 Frequency of stomach cramps 1.8 0.9 2.7 1.8 Severity of stomach cramps 10.1 2.7 12.7 4.5 You have abdominal discomfort, you have more bowel motion 16.1 9.1 15.5 11.8 You have abdominal discomfort, you have less bowel motion 17.1 1.8 22.9 5.4 You have abdominal discomfort, you have loose stool 28.8 13.6 31.8 21.8 You have abdominal discomfort, you have hard stool 29.7 2.7 31.9 3.6 Pain relieved by defecation 65.6 50.9 64.9 55.5 Assessment of satisfaction Satisfied with your bowel habits 39.6 100.0 31.5 100.0 Your constipation was bothersome 81 0 88.3 0 How bothersome was your abdominal bloating 74.7 4.5 81.9 6.4 How bothersome was your abdominal pain 61.2 6.4 67.5 8.2 * Final six questions were selected. All the above responses were statistically different between patients and controls (i.e. P 0.05), except those indicated by the symbol.

CHINESE CONSTIPATION QUESTIONNAIRE 487 (ii) frequency of <3 defecations/week, (iii) severity of incomplete evacuation, (iv) severity of lumpy or hard stools, (v) number of laxatives used and (vii) severity of abdominal bloating. Using this model, the probability of misclassification as a control is 0.08 and the probability of misclassification as a constipated subject is 0.06. Using PCA, the relative weightings of the six items were 0.3569, 0.4344, 0.3680, 0.3282, 0.5600 and 0.3567, respectively which are quite close to each other. There is strong evidence to use a simpler composite score instead of the six items individually. Hence a composite score, which is the sum of the above 6-item scores, is proposed. The correlation coefficient between the composite score and the first PC score is 0.9870 that is extremely high. A nominal logistic regression analysis was carried out with the only explanatory variable being the composite score just proposed. Based on this model, subjects will be classified as constipated if the composite score (sum of the six items) exceeds 5. The sensitivity, specificity and area under the ROC curve is summarized in Table 1. Content validity Summary of the responses of the subjects are tabulated in Table 2. All symptoms assessed were regarded as comprehensive by more than 88% of subjects (range: 88.7 97.7). The number of subjects who understand the six questions chosen were: 204 (92%), 201 (91%), 208 (94%), 202 (91%), 214 (97%) and 210 (95%), respectively. Test retest reliability The correlation coefficient for the composite scores in the two trials is 0.70. This is considered to be reasonably high. Internal consistency Cronbach s-a coefficient for the six items was 0.792 which is moderately high particularly when the number of items was not too large. Construct validity In the 221 subjects, there was good correlation between the composite constipation score with physical functioning (PF), role physical (RP), bodily pain (BP), Table 3. Correlation coefficient and significance (Kendall s tau) between the constipation score and various domains of the SF-36 quality of life scale in the 221 subjects at 2 weeks and 3 months (significance levels are two-tailed) Domain general health (GH), social functioning (SF), role emotional (RE) and mental health (MH) domains of the SF-36 scale in the 2-week and 3-month questionnaires, but not in vitality in both cases, while the score according to the 3-month questionnaire is also insignificantly correlated with the domain SF probably due to the changes or improvement in the constipation condition over the 3 months time (Table 3). Discriminant validity In the 20 patients who received Forlax treatment for constipation, 13 (65%) reported subjective improvement in constipation symptoms. There was significant correlation between the constipation severity score after 4 weeks of treatment and whether the patients reported symptomatic improvements (Kendall s tau ¼ 0.69, P ¼ 0.001). DISCUSSION Constipated and control subjects (n ¼ 221) 2 weeks 3 months Physical functioning )0.33 (0) )0.34 (0) Role physical )0.41 (0) )0.39 (0) Bodily pain 0.49 (0) 0.48 (0) General health 0.18 (0.005) 0.17 (0.01) Vitality )0.02 (0.78) )0.02 (0.72) Social functioning )0.15 (0.03) )0.12 (0.08) Role emotional )0.36 (0) )0.37 (0) Mental health )0.14 (0.04) )0.13 (0.05) Several constipation symptom severity scores have previously been published. 7, 8, 10 However, this is the only constipation questionnaire specifically developed and validated for Chinese-speaking patients. It can be used both to detect clinically apparent constipation and to grade disease severity. There was good internal consistency and reproducibility and good discriminatory power to distinguish between patients with constipation and controls. As only six questions are selected for the final model, this provides a simple and reliable questionnaire for future use, especially in epidemiological studies to assess the prevalence of constipation in the

488 A. O. O. CHAN et al. community and also as an end point in therapeutic trials of constipation. We have specifically assessed the symptoms in constipated subjects in the past 3 months and the past 2 weeks. For 2 weeks symptoms, it is used for assessing recent symptoms and therefore more appropriate for assessment of response to treatment. For the 3 months, like the Rome II criteria, the questions are more for the diagnosis of constipation. However, we find in the present study the patients responses are quite the same according to the interview. Hence, symptoms of constipation in the past 2 weeks were used in our questionnaire. The finally selected questions in our study were quite similar to those in the PAC-SYM. But our questions, in addition to the severity of symptoms, had also included both the frequency of constipation symptoms, i.e. frequency of <3 times bowel motion per week, and also the use of laxatives. Hence, the current final questionnaire was in fact a combination of the Rome II criteria (frequency), and the PAC-SYM (severity) and the use of laxatives. 10 The use of laxative has seldom been included in previous questionnaire. However, it is a consistent complaint elicited during the development of this questionnaire and in the pilot study. With its present validation, its inclusion in the instrument is important. This study was mainly conducted in an ethnic Chinese population using the Chinese language. Caution should be taken when applying our findings to other areas with different racial/ethnic populations. Nevertheless, we think that our questionnaire will be useful in studying the epidemiology of constipation in greater China, for which current data are scanty. It should also be noted that constipated subjects in the present study were recruited from those who were motivated to doctor s consultation at primary and referral centres and further study is needed for use at the population level. In conclusion, the Chinese constipation questionnaire was easy to understand, internally consistent and reproducible. It was the first questionnaire assessing frequency, symptom severity and laxatives used in constipated subjects. The symptom severity scores correlated negatively with quality of life. It could be used in epidemiological studies to assess the frequency and severity of constipation in patient populations and in interventional studies of constipation. ACKNOWLEDGEMENT This study was funded by a grant from the Peptic Ulcer Research Fund. REFERENCES 1 Talley NJ, O Keefe EA, Zinsmeister AR, et al. Prevalence of gastrointestinal symptoms in the elderly: a population-based study. Gastroenterology 1992; 102: 895 901. 2 Cheng C, Chan AO, Hui WM, et al. Coping strategies, illness perception, anxiety, and depression of patients with idiopathic constipation: a population-based study. Aliment Pharmacol Ther 2003; 18: 319 26. 3 Schaefer DC, Cheskin LJ. Constipation in the elderly. Am Fam Physician 1998; 58: 907 14. 4 Ashraf W, Park F, Lof J, et al. An examination of the reliability of reported stool frequency in the diagnosis of idiopathic constipation. Am J Gastroenterol 1996; 91: 26 32. 5 Koch A, Voderholzer WA, Klauser AG, et al. Symptoms in chronic constipation. Dis Colon Rectum 1997; 40: 902 6. 6 Drossman DA, Thompson WG, Talley NJ, et al. Rome II: multinational consensus document on functional gastrointestinal disorders. Gut 1999; 45 (Suppl. II): II1 1181. 7 Agachan F, Chen T, Pfeifer J, et al. A Constipation Scoring System to simplify evaluation and management of constipated patients. Dis Colon Rectum 1996; 39: 681 5. 8 McMillan SC, Williams FA. Validity and reliability of the Constipation Assessment Scale. Cancer Nurs 1989; 12: 183 8. 9 Lam CLK, Gandek B, Ren XS, et al. Tests of scaling assumptions and construct validity of the Chinese (HK) version of the SF-36 health survey. J Clin Epidemiol 1998; 51: 1139 47. 10 Frank L, Kleinman L, Farup C, et al. Psychometric validation of a constipation symptom assessment questionnaire. Scand J Gastroenterol 1999; 34: 870 7. 11 Chaussade S, Minic M. Comparison of efficacy and safety of two doses of two different polyethylene glycol-based laxatives in the treatment of constipation. Aliment Pharmacol Ther 2003; 17: 165 72. 12 Hutchinson A, Bentzen N, Konig-Zahn C. Cross Cultural Health Outcome Assessment: A User s Guide. European research group of health outcomes, 1994: 16. 13 Nunnally JC. Psychometric Theory, 3rd edn. New York, USA: McGraw Hill, 1994. 14 Silva F. Psychometric Foundations and Behavioural Assessment. Newbury Park, California, USA: Sage Publications, 1993. 15 Bombardier C, Tugwell P. A methodological framework to develop and select indices for clinical trials: statistical and judgmental approaches. J Rheumatol 1982; 9: 753 7.