A disease- specific quality of life instrument for non- alcoholic fatty liver disease and non- alcoholic steatohepatitis: CLDQ- NAFLD

Transcription

1 Received: 10 November 2016 Accepted: 8 February 2017 DOI: /liv METABOLIC LIVER DISEASE A disease- specific quality of life instrument for non- alcoholic fatty liver disease and non- alcoholic steatohepatitis: CLDQ- NAFLD Zobair M. Younossi 1,2,3 Maria Stepanova 3 Linda Henry 3 Andrei Racila 1,3 Brian Lam 1 Huong T. Pham 1 Sharon Hunt 3 1 Center For Liver Disease, Department of Medicine, Inova Fairfax Hospital, Falls Church, VA, USA 2 Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, VA, USA 3 Center for Outcomes Research in Liver Diseases, Washington, DC, USA Correspondence Zobair M. Younossi, MD, MPH, Betty and Guy Beatty Center for Integrated Research, Claude Moore Health Education and Research Building, Falls Church, VA, USA. Zobair.younossi@inova.org Handling Editor: Helena Cortez-Pinto Abstract Background: Non- alcoholic fatty liver disease and non- alcoholic steatohepatitis are the most common causes of chronic liver disease with known negative impact on patients health- related quality of life. Our aim was to validate a disease- specific healthrelated quality of life instrument useful for efficacy trials involving patients with non- alcoholic fatty liver disease and non- alcoholic steatohepatitis. Methods: From a long item selection questionnaire, we selected relevant items which, by factor analysis, were grouped into domains constituting Chronic Liver Disease Questionnaire Non- Alcoholic Fatty Liver Disease version. The developed instrument was subjected to internal validity, test- retest reliability and construct validity assessment using standard methods. Results: For development of the Chronic Liver Disease Questionnaire Non- Alcoholic Fatty Liver Disease version instrument, a 75- item- long item selection questionnaire was administered to 25 patients with non- alcoholic fatty liver disease. After item reduction, factor analysis found that 98.7% of variance in the remaining items would be explained by six factors. Thus, the resulting Chronic Liver Disease Questionnaire Non- Alcoholic Fatty Liver Disease version instrument had 36 items grouped into six domains: Abdominal Symptoms, Activity, Emotional, Fatigue, Systemic Symptoms, and Worry. The independent validation group included another 104 patients with nonalcoholic fatty liver disease. The Cronbach s alphas of suggested good to excellent internal consistency of the domains. Furthermore, the presence of obesity and history of depression were discriminated best by Chronic Liver Disease Questionnaire Non- Alcoholic Fatty Liver Disease version scores (P<.05). The domains correlations with the most relevant domains of Short Form- 36 exceeded Test- retest reliability in a subgroup of patients (N=27) demonstrated no significant within- patient variability with multiple administrations (all median differences were zero, all P>.15, intraclass correlations ). Conclusion: The Chronic Liver Disease Questionnaire Non- Alcoholic Fatty Liver Disease version is a disease- specific health- related quality of life instrument Abbreviations: CLDQ, Chronic Liver Disease Questionnaire; CLDQ-HCV, Chronic Liver Disease Questionnaire Hepatitis C Virus version; CLDQ-NAFLD, Chronic Liver Disease Questionnaire Non-Alcoholic Fatty Liver Disease version; DM, diabetes mellitus; HCC, hepatocellular carcinoma; HRQL, health-related quality of life; ICC, intraclass correlation; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; PBC-40, Primary Biliary Cirrhosis-40; PRO, patient-reported outcome; SF-36, Short Form-36. Liver International. 2017;37: wileyonlinelibrary.com/journal/liv 2017 John Wiley & Sons A/S Published by John Wiley & Sons Ltd

2 1210 developed and validated using an established methodology and useful for clinical trials of non- alcoholic fatty liver disease. KEYWORDS non-alcoholic fatty liver disease, non-alcoholic steatohepatitis, patient-reported outcome, quality of life 1 INTRODUCTION Health- related quality of life (HRQL) has become an essential component of evaluating the outcomes of patients in the clinical setting or in the setting of clinical trials. 1,2 In fact, HRQL and other patientreported outcomes (PROs) are increasingly required to show the value of treatment intervention for Medicare and other payers in the USA. In addition, HRQL assessment is important when determining the comprehensive burden of a chronic disease. In addition to HRQL, health utilities are frequently used in economic analysis determining the cost of the disease or the cost- effectiveness of a treatment. 3 6 Non- alcoholic fatty liver disease (NAFLD) is an extremely common chronic condition with an estimated 24% of the adult population in the world having NAFLD Some patients with NAFLD progress to cirrhosis, hepatocellular carcinoma (HCC) and liver- related mortality. 8 10,14 17 In fact, NAFLD is currently among the top three leading causes of HCC in the USA Furthermore, NAFLD is the second most common indication for liver transplantation in the USA. 18,19 Over the past two decades, it has become increasingly clear that NAFLD is also closely associated with other chronic diseases, namely, type 2 diabetes mellitus (DM), hypertension, metabolic syndrome, cardiovascular disease, hypertriglyceridemia and hypercholesterolaemia; 7,20 22 notably, worldwide prevalence of all these conditions is growing fuelled by the epidemic of obesity in both the developed and developing countries. 23,24 Moreover, in the presence of DM, non- alcoholic steatohepatitis (NASH), the advanced and progressive subtype of NAFLD, is more prevalent. 21,22 In addition to the tremendous clinical consequences and the enormous economic burden associated with NAFLD and NASH, both are known to negatively impact HRQL and other PROs Therefore, it is has become increasingly apparent that, to fully appreciate the total burden of NAFLD as a chronic disease, one needs to accurately assess the impact of the disease itself and also of concurrent treatment on all relevant outcomes including HRQL and other PROs. 6,25 31 Currently, there exist a number of generic HRQL instruments which could be used to assess the overall impact of a disease on patients wellbeing. At present, the most commonly used tool is the Short Form- 36 (SF- 36) However, in a population with a chronic disease, it is particularly important to measure the impact of the most frequent problems and concerns caused by that particular condition, since some of the disease- specific issues may be overlooked by a generic instrument. For patients with chronic liver disease, a fully validated disease- specific instrument (Chronic Liver Disease Questionnaire or CLDQ) has been developed and is widely used. 35 There is also a hepatitis C- specific Key points Non-alcoholic fatty liver disease is a highly prevalent chronic disease in the USA and worldwide. Although there is no approved treatment at present, emerging treatments for advanced forms of non-alcoholic fatty liver disease are being tested in clinical trials. Patients health-related quality of life needs to be assessed in order to appreciate the total disease burden and also the comprehensive benefit of treatment. The CLDQ-NAFLD is a disease-specific instrument developed and validated for assessment of health-related quality of life in patients with non-alcoholic fatty liver disease and non-alcoholic steatohepatitis. version, or CLDQ- HCV, which is also fully validated and widely used. 36 Finally, PBC- 40 is a disease- specific instrument for patients with primary biliary cholangitis. 37 At the same time, despite multiple studies suggesting the negative impact of NAFLD on patients HRQL, there is no validated disease- specific instrument for NAFLD The aim of this paper is to present development and systematic validation of a disease- specific HRQL instrument for patients with NAFLD. 2 METHODS 2.1 Development of the CLDQ- NAFLD instrument Patients Patients with an established diagnosis of NAFLD were enrolled. Included NAFLD patients (biopsy- proven or imaging) were at least 18 years of age, did not have any other major chronic diseases which might bias their HRQL (such as congestive heart failure, active cancer, human deficiency virus infection etc.), and were willing to give an informed consent. Patients were excluded if they had any condition which, in the opinion of the investigator, would make the subject unsuitable for enrolment, or which could interfere with the subject s participation in and/or completion the protocol (such as psychiatric or emotional problems, language or cognitive difficulties); had history of bariatric surgery; had any other cause of chronic liver disease; were unwilling or unable to provide an informed consent.

3 Item selection A variety of sources (available generic and liver- specific instruments, patient interviews, expert input and focus groups 32,35,38,39 ) were used to develop a long item selection questionnaire. In the questionnaire, patients were asked whether the problem was important to them. The questions were close- ended, and the answers ranged from This is not a problem for me to This bothers me a lot on a 1-5 Likert scale. The item selection questionnaire was then administered to 25 patients by mail, phone or face- to- face Item reduction To reduce the number of items from the long item selection questionnaire, the proportion of patients who indicated that a problem bothered them at least a little was calculated for each item. The items which had been reported as not bothering them at all by more than 2/3 of patients were removed as non- relevant Factor analysis To group the items into easily interpretable domains, we conducted exploratory factor analysis using the items which remained in the questionnaire after the item reduction step. The only restriction to the factor analysis output was that each domain must include at least three items, to reduce the effect of sporadic haphazard responses to individual questions. The number of factors was not predefined but rather determined from the proportion of variance accounted for by the factors (95% or more, but not exceeding 100%). Items were presumed to be the most related to a factor with the greatest loading after varimax rotation. Internal consistency of the resulting groups of items (domains) was assessed by a Cronbach s alpha coefficient Final CLDQ- NAFLD instrument The distribution of the items into domains was subjected to content validity assessment, checked for consistency with other chronic liver disease- related HRQL instruments, and for redundancy. Additionally, unlike the item selection questionnaire, in the final instrument all the questions were formulated as How much of the time/how often you <experienced a problem>?, and a 1-7 Likert scale was introduced for the responses (the score of 1 would correspond to All of the time, and the score of 7 to None of the time ). The final scoring scheme suggested that the scores were to be calculated separately for each domain as an average of the domain s items. All final questions were designed so that no item score reversal would be necessary, and, thus, in all domains, greater scores reflect a better health. 2.2 Validation of CLDQ- NAFLD instrument In the second phase of the study, a standard HRQL instrument validation pipeline was used for validation of CLDQ- NAFLD in a separate group of patients. 1,36, Patients Similarly, after informed consent, patients with an established diagnosis of NAFLD were enrolled in the validation phase; the exclusion criteria were similar to above. Patients enrolled in the validation phase were required not to have been enrolled in the development phase. All patients enrolled in the validation phase completed CLDQ- NAFLD along with SF- 36 by mail Internal consistency Internal consistency of CLDQ- NAFLD was assessed by calculation of Cronbach s alpha coefficients for the domains, and by calculation of item- to- own- domain correlations after adjustment for overlap for all items. The distributions of the domain scores were also qualitatively evaluated for discriminatory power and for skewness Validity Validity was assessed by evaluation of the association of CLDQ- NAFLD scores with potentially relevant demographic and clinical parameters. The validation parameters were age, gender, obesity (body mass index 30.0 kg/m 2 ), metabolic syndrome (at least three out of obesity, hypertension, hyperlipidemia, and type 2 diabetes, as documented in patient s medical history) and its components, severity of liver disease and history of depression. Wilcoxon non- parametric test was used to compare the CLDQ- NAFLD domain scores between the groups of patients listed above. Additionally, for validity assessment, correlations between the most related HRQL domains of CLDQ- NAFLD and SF- 36 were expected to be the highest Test- retest reliability For test- retest reliability assessment, a subgroup of patients completed CLDQ- NAFLD twice a few weeks apart. Correlations between two administrations were calculated. The distributions of differences in scores between multiple administrations were evaluated, and the median differences were compared to zero by a non- parametric test for matched pairs. Additionally, intraclass correlation coefficients (ICCs), which is another indicator of a repeated measure reliability, were calculated for the CLDQ- NAFLD domains as the ratios of between- subjects variance to the total sample variance. In order to calculate the required variances, a general linear model was used for each domain separately, with subject ID and the administration ID (first or second) as two predictors of an outcome (the domain score). Furthermore, for reliability assessment, it was expected that P- values for the administration ID parameter were notably insignificant suggesting that the outcome is driven solely by the subject. All analyses were in SAS 9.3 (SAS Institute, Cary, NC). All participants of both development and validation phases gave an informed consent. The study was approved by the Institutional Review Board of Inova Health System (Falls Church, VA, USA).

4 1212 TABLE 1 Factor analysis of CLDQ- NAFLD items remaining after item reduction step (removed items are shown in Table S1) (N=25) Question text ( How much does this problem bother you? ) Factor 1 (Abdominal symptoms) Important to % of patients (bothers a little or more) Importance (0-5) Factor loading Being troubled by a feeling of abdominal bloating ± Experiencing abdominal pain ± Troubled by a feeling of abdominal discomfort ± Factor 2 (Activity) Unable to eat as much as you would like ± Bothered by a limitation on your diet ± Having trouble walking several blocks or climbing a few flights of stairs ± because of your health. b Having trouble bending, lifting or stooping. b ± Had trouble lifting or carrying heavy objects Factor 3 (Emotional) not tested, inherited from CLDQ Feeling anxious ± Feeling unhappy ± Feeling irritable ± Having difficulty sleeping ± Having mood swings ± Not enjoying life. b ± Feeling depressed ± Having problems concentrating ± Unable to fall asleep at night ± Factor 4 (Fatigue) Being tired or fatigued ± Feeling sleepy during the day ± Feeling drowsy ± Taking naps (5 min. or longer) during the day. b ± Decreased strength ± Decreased level of energy ± Factor 5 (Systemic symptoms) Experiencing bodily pain ± Having muscle cramps ± Experiencing joint pain. b ± Itching ± Shortness of breath is a problem in daily activities Having dry mouth Factor 6 (Worry) not tested, inherited from CLDQ not tested, inherited from CLDQ Feeling like you may die earlier because of your fatty liver. b ± Worried about the impact your liver disease has on your family a ± Worried that your symptoms will develop into major problems ± Worried about your condition getting worse ± Worried about never feeling better ± Concerned about the availability of a liver if you need a liver transplant ± Feeling frustrated by having fatty liver. b ± a included from the original CLDQ. b new in CLDQ- NAFLD in comparison to the original CLDQ.

5 RESULTS 3.1 Development of CLDQ- NAFLD At the development step of CLDQ- NAFLD, 25 patients completed a 75- item- long item selection questionnaire (Figure S1). The cohort was 54.1±14.3 years old, 36% male, 72% obese, 32% with type 2 diabetes, 32% with history of depression, 20% with histological cirrhosis (no clinical or decompensated cirrhosis) and 80% had NASH. As a result of the item reduction step, 38 items were further fed to factor analysis, while 37 items were reported to be not important by at least 67% of patients (Table 1, Table S1). Interestingly, out of 26 items from the original CLDQ instrument 35 which were also included in the long item selection questionnaire in this study, 25 remained in the questionnaire after the item reduction step. Exploratory factor analysis of the remaining 38 items resulted in 55.0% of variance in those explained by one factor, 71.2% by two factors, 82.8% by three factors, 89.4% by four factors, 94.7% by five factors, 98.7% by six factors and 101.7% by seven factors. Since the proportion exceeding 100% indicates over- fitting, we chose the resulting number of factors and, thus, domains in CLDQ- NAFLD to be six. Varimax rotation was used to calculate the resulting factor loadings for all items. Again, 25 items overlapping with the original CLDQ instrument were found to be loading to the six factors similarly to the six original CLDQ domains. Given that, for CLDQ- NAFLD, the domain names were chosen to be the same as in the original CLDQ. 35 After an additional assessment of content validity, six items were removed as non- relevant or redundant (Table S1), resulting in a 32- item- long preliminary questionnaire with 25 questions overlapping with the original CLDQ. Furthermore, in order to retain validity of the original CLDQ questionnaire, we elected to add the remainder of items from original CLDQ into the newly developed CLDQ- NAFLD. Thus, the final CLDQ- NAFLD instrument has 36 items (29 items from original CLDQ and 7 new items) grouped into 6 HRQL domains: Abdominal Symptoms, Activity, Emotional, Fatigue, Systemic Symptoms and Worry (Table 1). The total CLDQ- NAFLD score can be calculated as an average of the six domain scores. 3.2 Validation of CLDQ- NAFLD The validation cohort included 104 patients with NAFLD who completed CLDQ- NAFLD and SF- 36. Patients were, on average, 57.2±12.1 years, 35% male, 63% obese, 26% with type 2 diabetes, 21% with history of depression, 15% compensated cirrhosis, 50% histological NASH. Of the study cohort, 27 patients without any evidence of clinical change had CLDQ- NAFLD administered twice 4-20 weeks apart. All questionnaires were fully completed without any items missing Internal consistency Both first and second administrations of CLDQ- NAFLD were used for assessment of internal consistency (N=131 entries). An acceptable to excellent internal consistency was detected in all six CLDQ- NAFLD domains (Table 2). Additionally, after sequential one- item exclusions, the resulting Cronbach s alphas did not change substantially (Table 2); this suggests that the items are neither too correlated with, nor too different from other items belonging to the same domains. Notably, the greatest variability in Cronbach s alphas over the course of oneitem exclusions was found in the Activity domain which indeed consists of the items measuring two potentially distinct constructs, namely, physical strength and diet restrictions. Despite this, the itemto- own- dimension correlations were above 0.4 for all but two items (Table 2). The distribution of the CLDQ- NAFLD domain scores suggested some skewness towards higher values (Figure 1A- G). Indeed, the prevalence of values of 3.0 or less did not exceed 10% for all six domain scores, and did not exceed 5% for all but Abdominal Symptoms and Fatigue domains. Overall, the Emotional and Fatigue domains were the least skewed suggesting a potentially greater discriminatory power in those Validity Validity of CLDQ- NAFLD and its discriminant function with reference to demographic and clinical parameters is summarized in Table 3. Expectedly, older patients had lower Activity and higher Emotional scores, and female patients also had lower Activity scores (all P<.05). Additionally, obesity was associated with significantly lower Activity, Fatigue and Systemic Symptoms scores (all P<.05). Patients with metabolic syndrome had lower Activity scores, while patients with type 2 diabetes also had lower Fatigue scores (all P.05). Furthermore, patients with history of depression had lower scores in nearly all CLDQ- NAFLD domain, with the greatest impairment observed in the Emotional domain (P<.0001) (Table 3). Correlations of CLDQ- NAFLD domains scores with BMI were significant for Activity (R=.27, P=.0063), Fatigue (R=.26, P=.008), Systemic Symptoms (R=.34, P=.0004) and the total CLDQ- NAFLD score (R=.26, P=.0068). Since all cirrhotic patients had only histological cirrhosis without clinical signs of cirrhosis or decompensated cirrhosis, we did not observe any impairment in CLDQ- NAFLD scores in those patients (all P>.05). From the magnitude of difference in the CLDQ- NAFLD scores observed in this study between a few clinically meaningful groups, we suggest the minimal clinically important difference (MCID) for CLDQ- NAFLD scores to be 0.5 until a systematic MCID determination study for this instrument is performed. Correlations of CLDQ- NAFLD domains with the domains of SF- 36 are summarized in Table 4. As expected, the highest correlated domain for Activity of CLDQ- NAFLD was Physical Functioning of SF- 36, for Emotional of CLDQ- NAFLD it was Mental Health of SF- 36, for Fatigue of CLDQ- NAFLD Vitality of SF- 36, for Systemic Symptoms of CLDQ- NAFLD Bodily Pain of SF- 36. On the other hand, there were no markedly highly correlated domains of SF- 36 for the domains of Abdominal Symptoms and Worry of CLDQ- NAFLD, which is expected given the disease- specific nature of these two domains (Table 4).

6 1214 Domain Cronbach α Item- to- own- dimension correlations Cronbach α with one item removed TABLE 2 Internal consistency of CLDQ- NAFLD (N=131) Abdominal symptoms Activity Emotional Fatigue Systemic symptoms Worry FIGURE 1 The histograms of the domain scores of CLDQ- NAFLD Test- retest reliability For this analysis, 27 patients completed the CLDQ- NAFLD instrument for the second time. The median time between administrations was 11 weeks (minimum 5 weeks, maximum 19 weeks). In test- retest analysis of CLDQ- NAFLD items, all but one item was highly significantly correlated between the two administrations (all P<.03); the outlier item was Q7 ( < > not been able to eat as much as you would like? ) which was also linked to the poorest internal consistency, as reported above. The correlations between repeated measures of the domain scores ranged from 0.61 (Worry) to 0.76 (Activity, Emotional, Fatigue) (all P<.0007) (Table 5). An analysis of the distribution of changes in the item responses revealed that none of them were significantly different from zero (all median changes equal to 0, and all P>.05), although in some patients their responses changed between the two instrument administrations by as much as six points. The changes in the domain scores were also non- significant (all P>.15), while the magnitudes of those changes were substantially smaller in comparison to those in their constituent items, with the size of interquartile ranges not exceeding 1.2 points and the maximum change not exceeding three points (Table 5). This suggests that the impact of aleatory responses to individual questions can indeed be alleviated by the inclusion of a number of items measuring the same construct into one domain. The general linear models for the domain scores also confirmed highly insignificant impact of repeated administration: all P>.35 for the repeated measure predictor (Table 5). The ICC values ranged from 0.76 to 0.88 for the domain scores, again suggesting high retest reliability of the CLDQ- NAFLD instrument (Table 5). 4 DISCUSSION The CLDQ is a widely used, fully validated disease- specific instrument used to assess HRQL in patients with chronic liver disease. Since its development and validation, CLDQ has been used worldwide in numerous studies and translated into dozens of languages. 41 In addition, a hepatitis C version of CLDQ (CLDQ- HCV) has been developed and validated as a simple and accurate tool to measure HRQL in patients with hepatitis C. Currently CLDQ- HCV is being used in a number of international clinical trials of treatments for patients with HCV. 30,31 However, unlike HCV, for which a number of highly effective treatment options are available, there are no effective or approved treatments for NAFLD. 42 In the context of better appreciation of the burden of NAFLD and to develop effective therapeutic regimens for NAFLD, it has become critical develop a validated disease- specific HRQL instrument for NAFLD. This is especially important as large numbers of efficacy trials are being designed for

7 1215 TABLE 3 Validity and discriminatory power of CLDQ- NAFLD (N=104) Item Age 60 y Age<60 y P N Abdominal symptoms 6.01± ± Activity 5.34± ± Emotional 5.59± ± Fatigue 4.94± ± Systemic symptoms 5.21± ± Worry 5.80± ± Total CLDQ- NAFLD 5.48± ± Item Male Female P N Abdominal symptoms 5.78± ± Activity 5.99± ± Emotional 5.35± ± Fatigue 4.79± ± Systemic symptoms 5.40± ± Worry 5.95± ± Total CLDQ- NAFLD 5.55± ± Item Obesity (BMI 30) N No obesity (BMI<30) Abdominal symptoms 5.57± ± Activity 5.42± ± Emotional 5.10± ± Fatigue 4.51± ± Systemic symptoms 4.89± ± Worry 5.75± ± Total CLDQ- NAFLD 5.21± ± Item Metabolic syndrome N No metabolic syndrome Abdominal symptoms 5.57± ± Activity 5.21± ± Emotional 5.19± ± Fatigue 4.39± ± Systemic symptoms 4.86± ± Worry 5.74± ± Total CLDQ- NAFLD 5.16± ± Item N Type 2 diabetes Type 2 diabetes P Abdominal symptoms 5.23± ± Activity 5.14± ± Emotional 5.12± ± Fatigue 4.40± ± P P (Continues) TABLE 3 (Continued) Item NAFLD in which capturing PROs as important secondary endpoints will be critical. 43 Type 2 diabetes Type 2 diabetes P Systemic symptoms 4.85± ± Worry 5.70± ± Total CLDQ- NAFLD 5.07± ± Item Depression No depression P N Abdominal symptoms 5.17± ± Activity 5.14± ± Emotional 4.13± ± Fatigue 4.14± ± Systemic symptoms 4.80± ± Worry 5.56± ± Total CLDQ- NAFLD 4.82± ± In this study, we report the development and validation of NAFLD- specific version of CLDQ. We followed a well- established process for the development of disease- specific instruments. 1,32,40 In the development phase of our study, we found that 98.7% of variance in patients responses was captured by six factors, and those factors were identical to those of the original CLDQ instrument; the latter could serve as an indirect evidence of validity of the instrument. The seven items which were new in CLDQ- NAFLD in comparison to the original CLDQ seemed to reflect a greater influence of fatigue in the NAFLD population. Furthermore, we believe that our internal consistency testing and test- retesting of CLDQ- NAFLD during the validation phase provides sufficient evidence for face validity and reliability. High correlations between presumably related domains of CLDQ- NAFLD and the widely used and extensively validated Short Form- 36 (SF- 36) confirm convergent validity of the developed instrument. Also, in an additional round of construct validation, female patients and patients of older age were found to have lower scores, primarily in the domain of physical activity. Furthermore, patients with comorbidities were found to have impairment in their HRQL as captured by CLDQ- NAFLD, particularly in the domains which were expected to be the most relevant to their conditions, such as significantly lower Activity scores in patients with obesity and notably lower Emotional score in patients with depression. All these serve as an evidence for construct validity of the studied instrument. However, in addition to a relatively small sample size, one major limitation of the study was that our validation sample lacked patients with very advanced liver disease. In fact, our cirrhotic patients were well- compensated and only had evidence of histological cirrhosis. Given that the majority of items in CLDQ- NAFLD overlaps the original CLDQ, we suspect it will retain the discriminate ability of CLDQ vs severity of liver disease. Additionally, we are also testing the responsiveness of CLDQ- NAFLD as a part of large clinical trials which involve patients with NASH. On the other hand, the prevalence of NASH in our training and testing sample was greater than usually seen in NAFLD patients in

8 1216 TABLE 4 Validity of CLDQ- NAFLD: correlations between domains of CLDQ- NAFLD and domains of SF- 36 (Spearman correlation [P- value]) SF- 36 domain Abdominal symptoms Activity Emotional Fatigue Systemic symptoms Worry Total CLDQ- NAFLD Physical Functioning.19 (.0559).71 (<.0001).16 (.11).47 (<.0001).47 (<.0001).39 (<.0001).53 (<.0001) Role Physical.22 (.0228).64 (<.0001).30 (.0022).57 (<.0001).59 (<.0001).44 (<.0001).61 (<.0001) Bodily Pain.35 (.0003).58 (<.0001).29 (.0031).39 (<.0001).65 (<.0001).34 (.0004).56 (<.0001) General Health.28 (.0042).41 (<.0001).42 (<.0001).54 (<.0001).38 (<.0001).44 (<.0001).55 (<.0001) Vitality.41 (<.0001).45 (<.0001).69 (<.0001).79 (<.0001).46 (<.0001).43 (<.0001).73 (<.0001) Social Functioning.41 (<.0001).52 (<.0001).60 (<.0001).58 (<.0001).47 (<.0001).45 (<.0001).69 (<.0001) Role Emotional.44 (<.0001).36 (.0002).66 (<.0001).48 (<.0001).36 (.0002).45 (<.0001).60 (<.0001) Mental Health.41 (<.0001).26 (.0075).79 (<.0001).54 (<.0001).31 (.0014).40 (<.0001).60 (<.0001) Physical summary.19 (.0561).66 (<.0001).08 (.4035).43 (<.0001).58 (<.0001).37 (.0001).50 (<.0001) Mental summary.44 (<.0001).20 (.0397).78 (<.0001).56 (<.0001).29 (.0028).35 (.0002).59 (<.0001) TABLE 5 Test- retest reliability of CLDQ- NAFLD (N=27; median 11.1 (min 5.0, max 18.9) wk between administrations) CLDQ- NAFLD domain Abdominal symptoms Activity Emotional Fatigue Systemic symptoms Worry Total CLDQ- NAFLD Correlation between administrations Pearson R P- value.0002 <.0001 <.0001 <.0001 < <.0001 Difference between administrations Median Interquartile range Range (min to max) P Generalized repeated measures model Intraclass correlation (95% CI).79 ( ).86 ( ).77 ( ).85 ( ).76 ( ).84 ( ).88 ( ) P for repeated measure

9 1217 a population- based or community setting, 7 9 which might introduce an additional bias. Nevertheless, this was done purposefully to make the questionnaire more relevant to the suit most clinical trials that enrol NASH patients. Further work will be done to continue validation of this tool as it is used in more diverse samples, different countries and settings. This is especially necessary as the number of patients used in this study was relatively small, although our results suggested good stability of the domains. In summary, the CLDQ- NAFLD is developed to provide researchers with a valid and reliable instrument to measure HRQL in patients with NAFLD. As this patient population continues to grow, the clinical, economic, PRO burden and future treatments for NAFLD will be more accurately and easily assessed. CONFLICT OF INTEREST ZMY is a consultant to Intercept, Abbvie, Gilead Sciences, BMS, Janssen, GSK and Salix. Other co- authors have no conflict of interest to disclose. REFERENCES 1. Cohen RD. Validation of health- related quality of life instruments. Hepatology. 1999;29(6 Suppl):7S 8S. 2. Nanda U, Andresen EM. Health- related quality of life. A guide for the health professional. Eval Health Prof. 1998;21: Committee for Economic Development. Adjusting the Prescription- Recommendations for Health Care Reform March. Obtained from: Accessed September 12, Claxton G, Cox C, Gonzales S, Kamal R, Levitt L. Kaiser Family Foundation. Measuring the quality of healthcare in the U.S. Obtained from Accessed September 12, Oliver A. Incentivising improvements in health care delivery Health Economics. Policy Law. 2015;10: Horsman J, Furlong W, Feeny D, Torrance G. The Health Utilities Index (HUI ): concepts, measurement properties and applications. Health Qual Life Outcomes. 2003;1: Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease- Meta- analytic assessment of prevalence, incidence, and outcomes. Hepatology. 2016;64: Younossi ZM, Blissett D, Blissett R, et al. The economic and clinical burden of nonalcoholic fatty liver disease in the United States and Europe. Hepatology. 2016;64: Sayiner M, Koenig A, Henry L, Younossi ZM. Epidemiology of Nonalcoholic Fatty Liver Disease and Nonalcoholic Steatohepatitis in the United States and the Rest of the World. Clin Liver Dis. 2016;20: Fazel Y, Koenig AB, Sayiner M, Goodman ZD, Younossi ZM. Epidemiology and natural history of non- alcoholic fatty liver disease. Metabolism. 2016;65: Lazo M, Clark JM. The epidemiology of nonalcoholic fatty liver disease: a global perspective. Semin Liver Dis. 2008;28: Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology and natural history of non- alcoholic fatty liver disease and non- alcoholic steatohepatitis in adults. Aliment Pharmacol Ther. 2011;34: Loomba R, Sanyal AJ. The global NAFLD epidemic. Nat Rev Gastroenterol Hepatol. 2013;10: Zoller H, Tilg H. Nonalcoholic fatty liver disease and hepatocellular carcinoma. Metabolism. 2016;65: Margini C, Dufour JF. The story of HCC in NAFLD: from epidemiology, across pathogenesis, to prevention and treatment. Liver Int. 2016;36: Piscaglia F, Svegliati-Baroni G, Barchetti A, et al. Clinical patterns of hepatocellular carcinoma in nonalcoholic fatty liver disease: a multicenter prospective study. Hepatology. 2016;63: Walker M, El-Serag HB, Sada Y, et al. Cirrhosis is under- recognised in patients subsequently diagnosed with hepatocellular cancer. Aliment Pharmacol Ther. 2016;43: Wong RJ, Aguilar M, Cheung R, et al. Nonalcoholic Steatohepatitis Is the Second Leading Etiology of Liver Disease Among Adults Awaiting Liver Transplantation in the United States. Gastroenterology. 2015;148: Pais R, Barritt AS, Calmus Y, et al. NAFLD and liver transplantation: current burden and expected challenges. J Hepatol. 2016;65: Kotronen A, Yki-Jarvinen H. Fatty liver: a novel component of the metabolic syndrome. Arterioscler Thromb Vasc Biol. 2008;28: Leite NC, Salles GF, Araujo AL, et al. Prevalence and associated factors of nonalcoholic fatty liver disease in patients with type- 2 diabetes mellitus. Liver Int. 2009;29: Adams LA, Waters OR, Knuiman MW, et al. NAFLD as a risk factor for the development of diabetes and the metabolic syndrome: an elevenyear follow- up study. Am J Gastroenterol. 2009;104: Younossi ZM, Stepanova M, Afendy M, et al. Changes in the prevalence of the most common causes of chronic liver diseases in the United States from 1988 to Clin Gastroenterol Hepatol. 2011;9: e Ng M, Fleming T, Robinson M, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during : a systematic analysis for the Global Burden of Disease Study Lancet. 2014;384: Younossi ZM, Henry L. Economic and Quality- of- Life Implications of Non- Alcoholic Fatty Liver Disease. Pharmacoeconomics. 2015;33: Dan AA, Kallman JB, Wheeler A, et al. Health- related quality of life in patients with non- alcoholic fatty liver disease. Aliment Pharmacol Ther. 2007;26: Afendy A, Kallman JB, Stepanova M, et al. Predictors of health- related quality of life in patients with chronic liver disease. Aliment Pharmacol Ther. 2009;30: Golabi P, Otgonsuren M, Cable R, et al. Non- alcoholic Fatty Liver Disease (NAFLD) is associated with impairment of Health Related Quality of Life (HRQOL). Health Qual Life Outcomes. 2016;9: David K, Kowdley KV, Unalp A, et al. Quality of life in adults with nonalcoholic fatty liver disease: baseline data from the nonalcoholic steatohepatitis clinical research network. Hepatology. 2009;49: Younossi ZM, Stepanova M, Henry L, et al. Minimal impact of sofosbuvir and ribavirin on health related quality of life in chronic hepatitis C (CH- C). J Hepatol. 2014;60: Younossi Z, Henry L. Systematic review: patient- reported outcomes in chronic hepatitis C the impact of liver disease and new treatment regimens. Aliment Pharmacol Ther. 2015;41: Ware JE Jr, Sherbourne CD. The MOS 36- item short- form health survey (SF- 36). I. Conceptual framework and item selection. Med Care. 1992;30: McHorney CA, Ware JE Jr, Raczek AE. The MOS 36- Item Short- Form Health Survey (SF- 36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care. 1993;31: McHorney CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36- item Short- Form Health Survey (SF- 36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;32:40 66.

10 Younossi ZM, Guyatt G, Kiwi M, Boparai N, King D. Development of a disease specific questionnaire to measure health related quality of life in patients with chronic liver disease. Gut. 1999;45: Younossi ZM, Stepanova M, Henry L. Performance and Validation of Chronic Liver Disease Questionnaire- Hepatitis C Version (CLDQ- HCV) in Clinical Trials of Patients with Chronic Hepatitis C. Value Health 2016;19: Jacoby A, Rannard A, Buck D, et al. Development, validation, and evaluation of the PBC- 40, a disease specific health related quality of life measure for primary biliary cirrhosis. Gut. 2005;54: Webster K, Cella D, Yost K. The Functional Assessment of Chronic Illness Therapy (FACIT) Measurement System: properties, applications, and interpretation. Health Qual Life Outcomes. 2003;16: Bergner M, Bobbitt RA, Carter WB, Gilson BS. The Sickness Impact Profile: development and final revision of a health status measure. Med Care. 1981;19: Brazier JE, Harper R, Jones NM, et al. Validating the SF- 36 health survey questionnaire: new outcome measure for primary care. BMJ. 1992;305: Two R, Verjee-Lorenz A, Clayson D, Dalal M, Grotzinger K, Younossi ZM. A methodology for successfully producing global translations of patient reported outcome measures for use in multiple countries. Value Health. 2010;13: Younossi ZM, Reyes MJ, Mishra A, Mehta R, Henry L. Systematic review with meta- analysis: non- alcoholic steatohepatitis - a case for personalised treatment based on pathogenic targets. Aliment Pharmacol Ther. 2014;39: Sayiner M, Stepanova M, Pham H, Noor B, Walters M, Younossi ZM. Assessment of health utilities and quality of life in patients with nonalcoholic fatty liver disease. BMJ Open Gastroenterol. 2016;3:e SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article. How to cite this article: Younossi ZM, Stepanova M, Henry L, et al. A disease- specific quality of life instrument for nonalcoholic fatty liver disease and non- alcoholic steatohepatitis: CLDQ- NAFLD. Liver Int. 2017;37:

11 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具