	New content
Chronic Respiratory Disease

Free access

Research article

First published online June 6, 2014

Reliability and validity of Leicester Cough Questionnaire Korean version

Jong-Min Han, In-Chul Jung, Weechang Kang, Seung-Su Kim, Yoon Yeo, and Yang-Chun Park [email protected]View all authors and affiliations

All Articles

https://doi.org/10.1177/1479972314536206

Abstract

The Leicester Cough Questionnaire (LCQ) is a self-administered questionnaire developed in England and validated for reliability. We developed a Korean translation of this questionnaire by applying a sequential forward and backward translation approach. The purpose of this study is to validate the Korean version of the LCQ (LCQ-K) in Korean patients with chronic cough. A multicenter prospective study was undertaken with 100 chronic cough patients who consented to participate in the study. The LCQ-K includes eight physical items, seven psychological items, and four social items. Visual analog scale (VAS) of cough, Borg Cough Scale (BCS), and Short Form-36 (SF-36) were used as external comparators. Participants included 52 women and 48 men with ages ranging from 18 years to 69 years. The concurrent validity comparing LCQ-K to VAS, BCS, and SF-36 yielded statistically significant Pearson correlation coefficients. The LCQ-K showed good reliability in three domains, with Cronbach’s α coefficients ranging from 0.84 to 0.87 (total: 0.91). Test–retest reliability was investigated with single measure intraclass correlation coefficients, which were found to be practically and statistically significant (p = 0.005). Responsiveness was validated by effective size ranging from 1.16 to 1.40 in each domain. LCQ-K is a reliable, valid, and responsive disease-specific questionnaire for assessing symptoms and quality of life of Korean patients with chronic cough.

Introduction

Cough is the most frequent symptom, which if lasts more than 8 weeks is considered chronic. The cause of chronic cough is usually found with little difficulty; however, it remains uncertain in approximately 20% of patients in spite of thorough study and empirical treatment.¹ Chronic cough interferes with social life, work, and sleep and induces physical symptoms, such as anxiety about serious disease and complications like chest pain, nuchal pain, urine incontinence, and syncope.²

Chronic cough has been measured simply by asking patients the frequency, severity, and extent that it affects their daily living and activities. Chronic cough: scores, diaries, symptom questionnaires, and visual analog scale (VAS) have been developed to quantify the measurements; however, they have not been thoroughly validated.³ Therefore, it is highly desirable to establish quantitative questionnaires on chronic cough-specific health-related quality of life and evaluate their reliability and validity.

Existing clinical studies on chronic have frequently used a symptom rating scale such as Borg Cough Scale (BCS), which scores the degree of symptoms from 0 to 10 by patient’s subjective feeling.⁴ Its reliability and concurrent validity with VAS have been studied.⁵ However, components of these studies only concern symptoms but not the psychological or social disturbances, and the quality of life is not considered. In 2001, Birring et al. developed Leicester Cough Questionnaire (LCQ) consisting of three parts,⁶ namely physiological, psychological, and social, which score patient’s quality of life. In 2007, LCQ Dutch version (LCQ-D) was developed and assessed with respect to the reliability and validity.⁷ In 2009, LCQ Chinese version was evaluated with respect to the validity and clinical implications in Chinese patients with chronic cough.⁸ LCQ has been used in many recent clinical studies since then,^9–11 but it has not been appropriately evaluated as a tool to assess chronic cough in Korea. Here, we report the development of Korean version of LCQ (LCQ-K) and assessment of its reliability and validity.

Methods

LCQ and elaboration of Korean version

LCQ includes 19 questions consisting of 8 physical items, 7 psychological items, and 4 social items, given in 7-point Likert-type frequency scales.⁶ One bilingual physician translated the English version of LCQ into Korean, and the translation panel reviewed the translation and prepared for a Korean draft (version 1). The other bilingual physician, blind to the original English version, back translated the Korean draft into English. The back-translated English version was compared with the original version to detect the ambiguity or inadequacy in the choice of vocabulary. The translation panel elaborated on the Korean draft (version 1.1), which was examined by 10 patients of various ages and genders for comprehensibility. For example, the patients were asked whether they could describe Korean words for “coughing bouts” in layman terms. The finalized LCQ-K was submitted to the validation process.

Study design

The Daejeon Korean Medicine Hospital institutional review board approved the protocol (authorization number: DJOMC-79). All patients understood the purpose and method of this study and agreed to participate with written informed consents. Participants were recruited from December 3, 2011 to July 26, 2012 in the Daejeon Korean Medicine Hospital of the Daejeon University and the Daejeon St. Mary’s Hospital.

Patients with symptoms lasting more than 8 weeks, diagnosed as chronic cough by research physicians, were included in the study. Exclusion criteria included age younger than 18 years or older than 70 years and diseases that seriously affected participants’ quality of life, such as liver cirrhosis, chronic renal failure, and chronic heart failure.

The sample size was calculated so that the number of cases who have a global rating of change (GRC) score of 0 in 2 weeks, indicating “no change” in their cough symptoms, could be at least 30 in the study. The information needed to estimate the sample size was based on the analysis of test–retest reliability in the LCQ-D, where the “no change rate” was reported to be 31.6%.⁷ Consequently, a total of 100 patients were enrolled (>30/0.316). Here, the GRC scores were determined to evaluate self-perceived changes in the symptoms since the first visit. Responses were scored from +7 (a very great deal better) to −7 (a very great deal worse), where 0 indicated no change.¹² Participants were asked to complete LCQ-K, cough VAS, BCS, and Short Form-36 (SF-36) at the first visit, and LCQ-K after 2 weeks and after 2 months. The participants were allowed to take any cough medicines during this study (Figure 1).

Figure 1. Flow of study. LCQ-K: Leicester Cough Questionnaire Korean version; VAS: Visual Analog Scale; BCS: Borg Cough Scale; GRC: global rating of change; SF-36: Short Form 36.

Statistical analysis

Data were reported as means ± standard deviation (SD) unless specified otherwise. The level of significance for all tests was set at p < 0.05. Concurrent validity was evaluated by estimating the Pearson correlation coefficient between LCQ-K and other instruments (cough VAS, BCS, and SF-36). Internal reliability of physiological, psychological, and social domain was determined by calculating Cronbach α coefficient of each part in primary written LCQ-K. Test–retest reliability was estimated by calculating the intraclass correlation coefficient in 17 people with GRC scores of 0 in 2 weeks, indicating no change in their symptoms. The responsiveness was assessed with the effective size (ES), estimated as the difference between mean LCQ-K scores of the first and the third visits divided by the SD of the first LCQ-K scores. Seven patients with a GRC score of 4 were analyzed for responsiveness. Here, the GRC score of 4 was the minimum value considered as significant improvement.⁷ Patients with a GRC score of 5 or higher were excluded from the analysis to avoid overestimation of the responsiveness.

Results

Demographics of participants

One hundred participants were initially enrolled in the study and filled out all four questionnaires in the first visit. Of these, 12 participants did not respond to the second or third time questionnaires. All participants who replied once or more were included in the analysis. Ages of the participants ranged from 18 to 69 (mean 47.78). The numbers of female and male participants were 58 and 42, respectively. Of these, two patients reported that chronic cough was induced by an angiotensin-converting enzyme inhibitor. Also, 11 patients had more than two causes of chronic cough, including 7 cases of upper airway cough syndrome (UACS) + cough variant asthma (CVA), 3 cases of CVA + gastroesophageal reflux disease (GERD), 1 case of UACS + CVA + GERD, and 8 patients reported unknown causes (Table 1).

Table 1. Participant characteristics.

Variable	Value
Age (years)
Mean	47.78 ± 13.01
Range	18–69
Gender (number/total (%))
Female	58/100 (58.0)
Male	42/100 (42.0)
Duration of cough (years)	1.38 ± 1.60
FEV₁ (%predicted)	94.41 ± 7.70
Causes of cough (number (%))
UACS	53 (53%)
CVA	22 (22%)
GERD	4 (4%)
ACE-I use	2 (2%)
More than two causes	11 (11%)
Unknown	8 (8%)
LCQ
Physical	3.93 ± 1.17
Psychological	3.45 ± 1.24
Social	3.82 ± 1.41
Cough VAS	55.64 ± 23.36
BCS	4.55 ± 2.23
SF-36
General health	49.02 ± 20.37
Smoking (number/total (%))
Smoking	15/100 (15.0)
Nonsmoking	85/100 (85.0)

FEV₁: forced expiratory volume in 1 second, UACS: upper airway cough syndrome, CVA: cough variant asthma, GERD: gastroesophageal reflux disease, ACE-I: angiotensin-converting enzyme inhibitor, VAS: Visual Analog Scale; LCQ: Leicester Cough Questionnaire; BCS: Borg Cough Scale; SF-36: Short Form 36.

Concurrent validity

The concurrent validity of LCQ-K was evaluated by correlating with cough VAS, BCS, and SF-36 at the first visit. Pearson’s correlation coefficients, determined with 88 participants, are summarized in Table 2. All outcomes, except for one, were statistically significant. The correlation coefficient of LCQ-K total with BCS was lower than that of LCQ-D, but the correlation coefficient of LCQ-K total with SF-36 general health was higher than that of LCQ-D.

Table 2. Concurrent validity of the LCQ-K.^a

Validated outcome scale	LCQ-K
Validated outcome scale	Physical	Psychological	Social	Total
Cough VAS	−0.35	−0.29	−0.26	−0.32
BCS	−0.28	−0.28	−0.20 (NS)	−0.28
SF-36 general health	0.55	0.42	0.41	0.50
SF-36 physical functioning	0.45	0.39	0.37	0.44
SF-36 role physical	0.46	0.43	0.38	0.46
SF-36 role emotional	0.47	0.37	0.38	0.44
SF-36 social functioning	0.48	0.40	0.46	0.49
SF-36 body pain	0.54	0.43	0.43	0.51
SF-36 vitality	0.60	0.47	0.43	0.54
SF-36 mental health	0.56	0.46	0.41	0.51

LCQ-K: Leicester Cough Questionnaire Korean version; VAS: Visual Analog Scale; BCS: Borg Cough Scale; SF-36: Short Form 36; NS: not significant.

^aPearson’s correlation coefficients between scores on validated questionnaires (cough VAS, BCS, and SF-36) and the domain scores and the total score of LCQ-K. All correlation coefficients were significant at p < 0.05, unless otherwise described.

Internal reliability

The Cronbach’s α coefficients of three domains (total 19 questions) ranged from 0.84 to 0.91. According to the guidelines for interpreting coefficient α, 0.7–0.8 is considered respectable, and >0.80 very good.¹³ The seeing result expressed 0.84 in physical domain, 0.86 in psychological, 0.87 in social, and 0.91 in total score, which means very good internal reliability in all domains (Table 3).

Table 3. Internal reliability of LCQ-K.

Domain	Reliability^a
Domain	LCQ-K	LCQ^b
Physical score	0.84	0.79
Psychological score	0.86	0.89
Social score	0.87	0.85
Total score	0.91	0.92

LCQ: Leicester Cough Questionnaire; LCQ-K: Leicester Cough Questionnaire Korean version.

^aCronbach’s α.

^bData from Birring et al.⁶

Test–retest reliability

Intraclass correlation coefficient of the test–retest reliability was significant and over fair agreement (≥0.4) consentaneity.¹⁴ However, the overall consentaneity was lower than that of original LCQ (Table 4).

Table 4. Test–retest reliability of LCQ-K.

Domain	Intraclass correlation coefficient		p Value^b
Domain	LCQ-K	LCQ^a	p Value^b
Physical score	0.53	0.93	0.010
Psychological score	0.70	0.90	0.001
Social score	0.44	0.88	0.030
Total score	0.58	0.96	0.005

LCQ: Leicester Cough Questionnaire; LCQ-K: Leicester Cough Questionnaire Korean version.

^aData from Birring et al.⁶

^bTest for true value in zero.

Responsiveness test

Responsiveness was evaluated using the ES from seven participants who had the GRC score of 4 between the first and third visits. The ES was estimated to be 1.40 in physical, 1.38 in psychological, 1.16 in social domain, and 1.34 in total, greater than those of original LCQ in physical and social domains (Table 5).

Table 5. Responsiveness of LCQ-K.

LCQ-K domain	ES		p Value^b
LCQ-K domain	LCQ-K	LCQ^a	p Value^b
Physical score	1.40	1.00	0.019
Psychological score	1.38	1.75	0.052
Social score	1.16	0.84	0.063
Total score	1.34	1.68	0.040

LCQ: Leicester Cough Questionnaire; LCQ-K: Leicester Cough Questionnaire Korean version; GRC: global rating of change; ES: effective size.

^aData from Birring et al.⁶

^bPaired t test for the changes of the LCQ-K scores between the first and the third visits in 2 months values in seven people with the GRC score of 4.

Discussion

LCQ-K is composed of 19 items, same as LCQ. The 19 items are divided into 3 domains: physiological 8 items, psychological 7, and social 4. The physiological domain focuses on chronic cough degree and symptom patterns, psychological on anxiety about lasting chronic cough and mental distress, and social on discomfort in working and social life, relationship with colleagues.

The validation test estimates exactness and appropriateness of the scale. We calculated concurrent validity comparing LCQ-K with another scales such as chronic cough VAS, BCS, and SF-36. LCQ-K showed significant correlation with cough VAS in all domains, and so with SF-36 in eight parts. On the other hand, the correlation with social part of LCQ-K and BCS was at a boundary level (p = 0.056). The concurrent validity of LCQ-K with SF-36 (Table 2) was higher than LCQ-D.⁷ (LCQ-K (physical): 0.45–0.60, LCQ-D (physical): 0.11–0.61; LCQ-K (psychological): 0.37–0.47, LCQ-D (psychological): 0.22–0.41; LCQ-K (social): 0.37–0.46, LCQ-D (social): 0.16–0.45; LCQ-K (total): 0.44–0.54, and LCQ-D (total): 0.22–0.55). LCQ-K showed relatively lower correlation with VAS and BCS than with SF-36. It is likely because VAS and BCS only addressed the severity of cough unlike SF-36, which addressed specific items for quality of life.

The reliability test examines whether the same object produces consistent outputs in repeated measurements, that is, whether we can trust the score from one test.¹⁵ Cronbach’s α coefficient was calculated to test internal reliability, which was >0.8 in all domains, superior to LCQ. Test–retest reliability was checked in 17 participants with the GRC score of 0 between the first and the second time. Although LCQ-K showed lower intraclass correlation than original LCQ, it was statistically significant and considered a reasonable agreement according to the standard by Portney and Watkins.¹⁴

Responsiveness was tested using the ES estimated from seven people with the GRC score of 4 between first and third questionnaires, considering noticeable improvement in their symptoms. Physical, psychological, social, and overall domain showed ES values of 1.40, 1.38, 1.16, and 1.34, respectively. LCQ-K showed greater ES than LCQ in physical and social domains.

LCQ-K is a well-validated, reliable, and responsive scale. The testing results were similar to LCQ in statistical validation. In testing responsiveness, due to the less number of participants, statistical significance was not observed in two domains. Despite these limitations, LCQ-K was the first well-validated scale in Korea for quantitative evaluation of chronic cough and the quality of life. This may be used to reckon the influence of psychological and social sector on chronic cough patients in clinical studies. We hope LCQ to be verified multilaterally and lay the groundwork for the development of assessment tools that better reflect different aspects of chronic cough.

Acknowledgment

The authors thank Surinder S Birring, MD, King’s College Hospital, London, UK, for allowing us to use LCQ.

Conflict of interest

The authors declared no conflicts of interest.

Funding

This study was supported by a grant from the Oriental Medicine R&D Project (B110021) of the Ministry of Health, Welfare and Family Affairs, Republic of Korea. This work was also supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2012R1A1A2041904).

References

1. McGarvey LP, Heaney LG, Lawson JT, et al. Evaluation and outcome of patients with chronic non-productive cough using a comprehensive diagnostic protocol. Thorax 1998; 53: 738–743.

Sign in

Abstract

Introduction

Methods

LCQ and elaboration of Korean version

Study design

Statistical analysis

Results

Demographics of participants

Concurrent validity

Internal reliability

Test–retest reliability

Responsiveness test

Discussion

Acknowledgment

Conflict of interest

Funding

References

Cite article

Cite article

Download to reference manager

Share

Share this article

Share with email

Share access to this article

Information

Published In

Keywords

Rights and permissions

History

Authors

Affiliations

Notes

Metrics

Journals metrics

Article usage*

Articles citing this one

Figures & Media

Tables

View options

PDF/ePub

Get access

Access options

Sign in

Also from Sage

Article usage^*