Validation and exploratory factor analysis of Urdu Version of Chronic Liver Disease Questionnaire

Objective: To present and validate psychometric properties of Urdu version CLDQ, yet another objective was to do exploratory factorial analysis (EFA) of CLDQ Urdu version. Methods: This Cross-sectional Analytical Study was conducted at Dr. Ruth K.M. Pfau Civil Hospital Karachi during the period Nov. 15, 2021 to Jan. 30, 2022. CLDQ Urdu questionnaire was self-administered by the patients. The questionnaire consisted of 29 items and responses were recorded on 7-point Likert type scale. Reliability testing was done by Cronbach’s α, test value of >0.7 is taken as reliable. Exploratory factor analysis (EFA) was conducted with principal component analysis with varimax rotation. Adequacies for conduction of EFA depended on Kaiser-Meyer-Olkin (KMO) value of ≥0.5 and Bartlett’s Test of Sphericity (BTS) of ≤0.05. Mean CLDQ Urdu scores were also compared with Child Class using ANOVA and post-hoc analysis was done. Results: A total of 320 patients were selected after informed consent. All conditions for adequate EFA were met (Cronbach’s α =.949; KMO = .846; BTS ≤.001). Mean CLDQ Urdu score was 156.74 in male and 133.27 in female (p<.001). Child Class-A had best quality of life with score of 186.63 ±6.91 and Child Class-C had the worst with scores of 109.78 ±21.33. EFA resulted in reduction of domains to 4 (Muscular Symptoms, Emotional Symptoms, Abdominal Symptoms & Somnolence) & reduced the number of items from 29 to 11. Conclusion: Urdu CLDQ version is validated in our settings. EFA resulted in reductions in number of domains and items. CLDQ Urdu showed that quality of life decreases significantly with Child Class.


INTRODUCTION
In the past twenty years, there has been an increasing apprehension that the traditional assessment of medical outcomes after medical interventions is inadequate. Therefore, health-related quality of life (HRQOL) gained significance as an outcome measure in clinical and epidemiological studies. 1 Many generic and specific instruments are in vogue to measure HRQOL. Amid generic instruments, Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), 2 has been used widespread to postulate a global assessment of a given disease and allow its appraisals with general population and other diseases. Generic instruments are unable to evaluate disease-specific symptoms of liver disease, e.g., pruritus. They are less receptive to small, nevertheless clinically important, variations. Diseasespecific instruments were developed to assess specific characteristics of a disease and provide an accurate instrument for clinical studies. Chronic Liver Disease Questionnaire (CLDQ) is disease-specific HRQOL

Original Article
Validation and exploratory factor analysis of Urdu Version of Chronic Liver Disease Questionnaire instrument to evaluate different stages of liver diseases. 3 Chronic liver disease (CLD) is a global health problem and is the 11 th leading cause of death per year. It badly affects the quality of life of CLD patients and accounts for one of the top 20 causes of global morbidity. 4 Patients of chronic liver disease not only suffer from disease and its complications, therapeutic procedure related morbidity, side effects of drugs but also from emotional and psychological issues. Overall survival of CLD patients has increased due to advances in medical sciences. So along with survival, improvement in HRQOL and periodic assessment and monitoring of disease impact on patients health, physiological and socioeconomic status should also be prioritized. 5 HRQOL includes physical, social, psychological, and functional aspect of disease and has influence on patient's behavior and quality of life. It is categorized into generic and disease specific which provide health utilities and deals with single disease respectively. 6 Chronic Liver Disease Questionnaire (CLDQ) is a disease specific instrument that was developed by Younossi et al in English language for the assessment of health related quality of life of chronic liver disease patients. 3 It was later translated into many regional and international languages of world including Spanish, Italian, Greek, German, Swedish, Persian, Thai, Indian, Tamil, Japanese, Chinese, Bengali and Sri Lankan languages. 7-9 Validity and reliability of CLDQ has been well established in many studies. The English version of this questionnaire has been validated in Pakistani population previously. 10 Urdu is national language of Pakistan but Urdu version of CLDQ has not been developed yet for better understanding and assessment of HRQOL in Urdu speaking population. So, it is high time to formulate CLDQ in Urdu language for the assessment of healthrelated quality of life. Although CLDQ has been translated into many languages but only two, Spanish 11 and Italian 12 reported its factorial analysis which were much divergent with inconsistent results. While the Exploratory Factorial Analysis (EFA) of Spanish version was much near to original version of Younossi et al; the EFA of Italian version was much different from that of original.
There is imminent need for development of Urdu version of CLDQ and its EFA that would provide the assessment of HRQOL of chronic liver disease patients with items that are more lucid for our population. 5 This study will help in developing an Urdu version of CLDQ questionnaire, its validation and EFA will yield important factors in Urdu version of questionnaire leading to better assessment of quality of life in local population. The objective if this study was to present and validate psychometric properties of Urdu version CLDQ, yet another objective was to do exploratory factorial analysis (EFA) of CLDQ Urdu version. All patient of chronic liver disease visiting outpatient's department and admitted in medical wards of Dr Ruth KM Pfau Civil Hospital Karachi fulfilling the inclusion criteria were enrolled in the study during the period 15-11-2021 & 30-01-2022. Detail physical examination was carried out. Baseline characteristics and demographic data of each patient were collected on pre-designed proforma. CLDQ Urdu questionnaire was selfadministered by the patients, one of the authors was available for resolution of any problem in filling out the questionnaire. The questionnaire consists of 29 items and response was recorded on seven point Likert type scale, where one was worst and seven was best. Blood sample of 10 ml was withdrawn by phlebotomist for estimation of Serum Albumin, INR, HBs Ag & HCV Ab. Child Class of all patients was calculated. Data analysis: Demographics and etiology of liver disease etiology was described. Item analysis was based on calculated: means, ±SDs, and item discrimination. The proportion of patients who did not answer an item (missing) was used as a consideration of ''acceptance.'' Subscale level analysis was done. For each subscale, means and ±SDs were calculated. Reliability testing was done by internal consistency Cronbach's α test, value of ≥0.70 was taken as acceptable. Exploratory factor analysis (principal component analysis with varimax rotation) was conducted. Only factors with an Eigenvalue ≥1 were interpreted. List wise exclusion of cases was done to decrease impact of missing data. Child Class of selected patients was calculated. Mean CLDQ Scores were compared with gender using Student's t-test. Mean CLDQ Urdu scores were also compared with Child Class using ANOVA. Post-hoc analysis was also done to assess inter-category comparisons of CLDQ Urdu Scores. Significance level was ≤.05 and SPSS 26.0 was used for analysis.

RESULTS
A total of 320 patients fulfilling inclusion/exclusion criteria were selected after informed consent, these  Table-I) Reliability Analysis: The reliability analysis of all 29 items by Cronbach's α gave highly significant value of .949. Two of 29 items (Itching & availability of liver for transplant) had communalities value of <0.5 and they were excluded, analysis repeated after exclusion resulted same Cronbach's α value of .949 thus there was no effect on overall reliability. Factor Analysis: Factor analysis was done after excluding two items that had poor communalities value of <0.5. The correlation matrix value was a positive value (.002). It must be positive value for proper factor analysis. Kaiser-Meyer-Olkin (KMO) and Bartlett's Test of Sphericity (BTS) were done to see the fitness of data for factor analysis. KMO Measure of Sampling Adequacy was 0.846 which was more than the required value of 0.5. Bartlett's Test of Sphericity showed p-value of <0.001 indicating adequacy for conducting the factor analysis. Factor analysis was done using Principal Component Extraction method for Correlation Matrix. Factors with Eigenvalue <1 was excluded. Varimax rotation was used for final factor reduction. KMO Measure of Sampling Adequacy was 0.846 which was more than the required value of 0.5. Bartlett's Test of Sphericity showed p-value of <0.001 indicating adequacy for conducting the factor analysis. Table-II initial results of total variance identified four factors that were significant based on Eigenvalues of ≥1 these are detailed in Table-III and graphically in Fig.1.

Details of correlation matrix of all questions of CLDQ Urdu are given in
Initial Rotated Component Matrix showed many communalities having loadings on multiple factors. All items with cross loadings were excluded, and analysis was run again. The results obtained are shown in Table-IV.
Exploratory factor analysis after factor reduction identified four main components, which were named on basis of factors included,

Limitations:
The exploratory factor analysis of our Urdu version of CLDQ showed high internal validity and consistency and reduced factor items from 29 to 4 which explained a79.3 % of total variance. This being single center study need to be validated in multi center design.

DISCUSSION
Chronic liver disease carries a huge burden at health care in Pakistan leading to many complications like ascites, encephalopathy, and varices. 14-16 Assessment of quality of life in these patients is of outmost importance. There are many modalities and different ways for treatment of physical ailments of a patient but that is not all needed for complete recovery of patient from any illness. Instead of only physical health, global assessment of heath is necessary including psychological, social and mental health. Tools have been developed by which severity and progression and quality of life of CLD patients can be assessed but unfortunately none is available in our national language, i.e., Urdu. HRQOL has been assessed and reported in many diseases like dental, ischemic vascular, liver and neurological diseases in Pakistan but they were all done in English language and were done with the assistance of the investigator. 10,[17][18][19] Up till now none was done with native language spoken in Pakistan.
Our Urdu CLDQ version is easier, simple, reliable, and excellent way for the assessment of health-related quality of life of CLD patients as demonstrated in current study. This study demonstrates good reliability and statically It not only provides overall well-being of CLD patient but also helps in assessment of disease progression and effect of management on patient health with minimum items. Though the validity of CLDQ has been confirmed by many studies but exploratory factor analysis was done in Spanish, 11 Greek 20 and Italian 12 studies only. There is remarkable change in the structure of CLDQ after exploratory factor analysis in our study. We are left with only four main domains instead of 6 in original and Italian 12    This type of study has never been conducted in Pakistan. Many versions of CLDQ have been published in different languages of the world but this is first time that CLDQ is translated in Urdu language which help us in evaluation of health-related quality of life of CLD patient in much better way in our population and that is the main and the major strength of this study. Contrary to some studies in which CLDQ is applied only to of CLD patients of selected etiologies, [21][22][23][24] our study has included all the patient of CLD so is not limited to certain group of CLD patient. Urdu CLDQ version enables us to effectively assess the quality of CLD patient even without applying any scoring system. Limitations: One of the limitations of current study was its being a single center study. As we have excluded patients with hepatic encephalopathy and HCC from the study, we are not able to comment on health-related quality of life of patient with hepatic encephalopathy and HCC. Data of CLD patients who had undergone liver transplantation or waiting for transplantation is also lacking because of non-availability of transplantation facility at place where study is conducted.

CONCLUSION
Urdu CLDQ version is a practical and effective tool for evaluation of health-related quality of CLD patients. It is recommended that Urdu CLDQ version be included in routine assessment of all CLD patient as a general predictor of their quality of their health and response to treatment. Validity of Urdu CLDQ version is established and confirmed in our study. Exploratory factor analysis resulted in a simplified version of original CLDQ version with only eleven items. This will ease the assessment of quality of life in patients with chronic liver disease