Development and validation of a primary sclerosing cholangitis-specific health-related quality of life instrument: CLDQ-PSC

Objectives:

To understand the full impact of primary sclerosing cholangitis (PSC) on patients’ health, it is important to assess their health-related quality of life (HRQL). Using the Chronic Liver Disease Questionnaire (CLDQ), we aimed to develop and validate a PSC-specific HRQL instrument.

Methods:

Previously collected clinical and patient-reported outcome data from PSC patients were used. The original CLDQ with 29 items was subjected to item reduction, followed by factor analysis. A standard HRQL instrument validation pipeline was then applied to the new CLDQ-PSC.

Results:

There were 100 PSC patients (44±13 y, 32% male, 79% college educated, 39% cirrhosis, 67% inflammatory bowel disease, 66% ulcerative colitis, and 50% on ursodeoxycholic acid After item reduction and exploratory factor analysis, there were 24 items and 5 factors left; based on factor loadings, the factors were named emotional function, fatigue, symptoms, worry, and sleep. Internal consistency assessment returned Cronbach alpha 0.85–0.94, item-to-own domain correlations >0.66 for 22/24 items. Known-groups validity suggests discrimination between PSC patients with and without cirrhosis or its complications, obesity, history of depression, weight loss, and PSC patients on versus not on ursodeoxycholic acid (p<0.05 for all or select CLDQ-PSC domains). Relevant items of Short Form-36 and CLDQ-PSC were highly correlated (all p<0.0001). Matching with items of another PSC-specific instrument (PSC-patient-reported outcome; 42 items) for relevance and redundancy suggests that CLDQ-PSC is a relevant, comprehensive, and short HRQL instrument, which can be used for patients with PSC.

Conclusions:

The CLDQ-PSC is a PSC-specific HRQL instrument that was developed using an established methodology and demonstrated good psychometric characteristics.

INTRODUCTION

Patient-reported outcomes (PROs) are surrogates of patient’s experiences with their disease and treatments.1 By definition, PROs must come from a patient without any modification or influence by their health care provider or anyone else.1 The PRO is an umbrella term that includes health-related quality of life (HRQL), as well as other PROs related to function, work productivity, and other aspects of general well-being.1,2

Assessment of HRQL is increasingly important in the management of chronic diseases such as chronic liver disease (CLD).3,4 In this context, generic HRQL instruments are applicable to various clinical populations, which allow the comparison of scores across different diseases or to the general population norms.5,6 In contrast, disease-specific instruments are designed to be specifically tailored to the aspects of well-being affected by a specific disease and, therefore, tend to be more responsive to changes occurring during the disease natural history or as a result of treatment.5,6 For patients with CLD, several validated disease-specific instruments currently exist, which include the Chronic Liver Disease Questionnaire (CLDQ)7 and its subtypes according to the specific etiology of liver disease (CLDQ-HCV, CLDQ-HBV, and CLDQ-NAFLD/NASH,8–10) as well as SF-LDQOL,11 PSC-PRO,12 and NASH-Check.13

Primary sclerosing cholangitis (PSC) is an autoimmune liver disease, which causes chronic injury to the biliary system resulting in obstructive cholangitis, advanced liver disease, and is associated with an increased risk of hepatobiliary malignancies including cholangiocarcinoma.14 PSC is also strongly associated with inflammatory bowel disease and can lead to a variety of symptoms that can affect the patients’ HRQL. We have previously validated a 42-item PSC-PRO instrument.12 However, a shorter PSC-specific HRQL instrument can still be useful to supplement PSC-PRO and other PROs to assess various aspects of patients’ quality of life. Therefore, our aim was to develop and validate a PSC-specific CLDQ-PSC using an established methodology for developing disease-specific HRQL instruments.

METHODS

Patient sample

This study was designed and received Institutional Review Board approval with exempt status to use historic deidentified data collected in 2016-2017 for a study that assessed PROs of subjects with an established diagnosis of PSC (age 18–70 years, of both sexes, all ethnicities, with or without inflammatory bowel disease, with or without cirrhosis, and without any other potentially confounding comorbidities, enrolled from multiple institutions).12 Included subjects had completed medical history questionnaires, which asked about complications of liver disease and select nonhepatic comorbidities; this self-reported medical history was further used without additional verification through medical records or case note review. The subjects also completed PRO questionnaires: the Short Form-36 version 2 (SF-36v2), the CLDQ, Primary Biliary Cirrhosis-40 (PBC-40), and the PSC-PRO.12

Development of the CLDQ-PSC instrument

The original CLDQ instrument was developed using exhaustive collection of items from patient focus groups, literature search, and experts. The long questionnaire underwent item reduction step to develop CLDQ, followed by factor analysis to determine domains. The final CLDQ instrument includes 29 items each of which is scored on a Likert scale from 1 to 7, with higher scores indicating better health status. Its items belong to 6 nonoverlapping HRQL domains: Abdominal symptoms, Activity/Energy, Emotional, Fatigue, Systemic symptoms, and Worry, which are calculated as an average of their constituent items; the total CLDQ score is an average of the 6 domains. The CLDQ has been widely used and extensively validated in both clinical trial and real-world settings since its development in 1999.7,15–19 It also has been used as a source for developing other etiology-specific instruments such as CLDQ-HCV, CLDQ-HBV, and CLDQ-NAFLD/NASH.8–10 In this study, we developed an etiology-specific version of CLDQ for patients with PSC (CLDQ-PSC) using a similar methodology.

Selection of items for CLDQ-PSC

For the identification of items that would be relevant specifically for PSC patients, the proportions of PSC patients who indicated that they never or rarely had a particular problem over the recall period of 2 weeks by means of choosing the highest scores of 6 or 7 for the respective question [worded as “None of the time” or “Hardly any of the time” (have a problem/experience a symptom)] were calculated. Items that had been reported as never/rarely a problem by >65% of patients in the training set were removed as nonrelevant (item reduction).

Domains of CLDQ-PSC

To group the items into easily interpretable domains, we conducted exploratory factor analysis using the items that remained after the item reduction step. The number of factors was not pre-defined but rather determined from the proportion of variance accounted for by the factors (≥95% but <100%). Items were presumed to be the most related to a factor with the greatest loading after varimax rotation. The resulting distribution of the items into domains was subjected to content validity assessment and checked for consistency with the original CLDQ. The same Likert scale ranging 1–7 was also supposed to be applied to the items of newly developed CLDQ-PSC.

Validation of the CLDQ-PSC instrument

Primary validation

At the first step, a standard HRQL instrument validation pipeline8–10 was used for validation of CLDQ-PSC using deidentified data previously collected from PSC subjects.12

Internal consistency of CLDQ-PSC was assessed by calculation of Cronbach alpha coefficients for its 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.

Known-group validity was assessed by evaluation of the association of CLDQ-PSC domain scores with potentially relevant demographic and clinical parameters. The validation parameters were age, sex, obesity (body mass index ≥30 kg/m²), severity of liver disease (the presence of cirrhosis and its complications such as ascites, HE, or esophageal varices), receiving treatment with ursodeoxycholic acid (UDCA), and having depression, as reported in medical history collected using the study questionnaires. Wilcoxon nonparametric test was used to compare the CLDQ-PSC domain scores between the groups of patients listed above.

For the purpose of convergent validity assessment, correlations between the most related domains of CLDQ-PSC and SF-36 or PBC-40 were expected to be the highest. In addition, the items of PSC-PRO were assessed for relevance and redundancy to identify the aspects of PSC potentially missed by the CLDQ-PSC.

Independent validation

At the second step of validation, historic data collected from a different sample of PSC patients (1997-1998) were used.7 A similar validation pipeline was applied to this sample using clinical and PRO (CLDQ, SF-36) data; known-groups validity used only available clinical parameters (the presence of cirrhosis and Child-Pugh score).

All analyses were in SAS 9.4 (SAS Institute, Cary, NC). The study was approved by the Western Institutional Review Board and was approved for a waiver of consent, as all data were analyzed as deidentified data.

RESULTS

There were 100 PSC patients used in this study: mean (SD) age 44±13 years old, 32% male, 79% college educated, 39% with cirrhosis, 67% with inflammatory bowel disease, 66% with ulcerative colitis, and 50% on UDCA (Table 1).

TABLE 1.

Demographics and clinical characteristics of PSC patients

Parameter	N (%) or mean±SD
N	100
Age, y	43.6±13.1
Male sex	32 (32.0)
Married	75 (75.0)
College degree	79 (79.0)
Employed	73 (73.7)
BMI, kg/m2	25.6±4.8
Type 2 diabetes	1 (1.0)
Depression	37 (37.4)
Duration of PSC, years	7.57±7.72
Cirrhosis	36 (38.7)
Inflammatory bowel disease (IBD)	67 (67.0)
Ulcerative colitis	66 (66.0)
Crohn disease	15 (16.0)
On UDCA	50 (50.0)
History of liver cancer	1 (1.0)
History of cirrhosis complications	14 (14.7)
History of ascites	7 (7.2)
History of HE	7 (7.1)
History of esophageal varices	9 (9.0)
History of weight loss or malnutrition	26 (26.3)
History of colon cancer	6 (6.1)
History of colon resection	23 (23.0)
History of any other abdominal surgery	50 (50.5)
HRQL scores	Mean±SD
SF-36
Physical Functioning	82.3±20.9
Role Physical	69.7±28.3
Bodily Pain	71.0±21.7
General Health	44.6±24.2
Vitality	46.9±25.2
Social Functioning	72.3±27.2
Role Emotional	75.0±26.9
Mental Health	68.0±21.0
Physical Component Summary (PCS)	48.0±8.2
Mental Component Summary (MCS)	45.0±11.3
PBC-40
Cognitive	1.92±0.96
Emotional	2.58±1.15
Fatigue	2.42±0.99
Itch	1.53±1.19
Social	2.35±0.90
Symptoms	1.84±0.64
Total PBC-40 score	2.11±0.74

Abbreviations: PBC-40, Primary Biliary Cirrhosis; SF-36, PSC, primary sclerosing cholangitis; SF-36, Short Form-36.

Development of CLDQ-PSC

The 29 items of CLDQ were used for item reduction. As a result, 24 items were further fed to factor analysis (Table 2) while 5 items were reported as not important (a problem or symptom is never or rarely experienced) by at least 65% of PSC patients (Supplemental Table 1, http://links.lww.com/HC9/A108) and, thus, were excluded from CLDQ-PSC. Interestingly, no items belonging to the CLDQ domains of Activity/Energy passed this item reduction step, suggesting that this HRQL domain is not severely affected in most PSC patients; in addition, 2 CLDQ Systemic symptoms items (asking about having dry mouth and shortness of breath) were also eliminated as not relevant to PSC (Supplemental Table 1, http://links.lww.com/HC9/A108).

TABLE 2.

Description and factor analysis of CLDQ-PSC items remaining after item reduction step; removed items are shown in Supplemental Table 1( http://links.lww.com/HC9/A108)

Question text (“How much does this problem bother you?”)	Proportion with the highest possible scores (6 or 7), n (%)	Mean (SD) score	Factor loading
Factor 1 (Fatigue)
Being tired or fatigued	17 (17)	3.63±1.73	0.84
Feeling sleepy during the day	15 (15)	3.76±1.63	0.77
Having decreased strength	54 (54)	5.27±1.80	0.49
Feeling a decreased level of energy	24 (24)	3.93±1.81	0.85
Feeling drowsy	25 (25)	4.36±1.62	0.78
Having problems concentrating	41 (41)	4.89±1.57	0.54
Factor 2 (Worry)
Worried about the impact the liver disease has on the family	30 (30)	4.22±1.85	0.82
Worried that symptoms will develop into major problems	24 (24)	4.10±1.85	0.83
Worried about the condition getting worse	26 (26)	4.19±1.77	0.88
Worried about never feeling any better	48 (48)	4.85±1.95	0.62
Concerned about availability of a liver transplant	49 (49)	5.08±1.85	0.82
Factor 3 (Symptoms)
Feeling of abdominal bloating	41 (41)	4.86±1.72	0.66
Bodily pain	50 (50)	4.93±1.72	0.72
Abdominal pain	51 (51)	5.35±1.49	0.81
Feeling of abdominal discomfort	57 (57)	5.36±1.59	0.84
Muscle cramps	63 (63)	5.74±1.43	0.49
Itching	55 (55)	5.32±1.81	0.33
Factor 4 (Emotional)
Feeling anxious	47 (47)	4.85±1.74	0.47
Feeling unhappy	43 (43)	5.08±1.36	0.65
Being irritable	30 (30)	4.72±1.32	0.66
Having mood swings	43 (43)	5.08±1.46	0.72
Feeling depressed	54 (54)	5.25±1.73	0.71
Factor 5 (Sleep)
Having difficulty sleeping	31 (31)	4.21±1.95	0.54
Unable to fall asleep at night	44 (44)	4.84±1.89	0.55

Exploratory factor analysis of the remaining 24 items resulted in 95% of variance explained by 5 factors. Varimax rotation was used to calculate the resulting factor loadings for the items (Table 2). Three out of 5 factors were found to be structured similarly to the CLDQ domains of Worry, Emotional, and Fatigue, with the exception of 1 item (asking about problems concentrating), which originally belonged to the CLDQ Emotional domain but was found to be primarily loading onto the Fatigue factor for PSC patients. In addition, the items from the CLDQ domains of Abdominal symptoms and Systemic symptoms were all found to be loading on the same factor, and this sustained in an additional round of exploratory factor analysis with forcing a sixth factor into the model. Finally, 2 items (asking about having difficulty sleeping and being unable to fall asleep) originally from the CLDQ Emotional domain were found to be loading on the fifth factor for CLDQ-PSC. As a result, for CLDQ-PSC, the 3 domain names were chosen to be the same as in the original CLDQ (Worry, Emotional, and Fatigue), the combined domain of Abdominal symptoms and Systemic symptoms was named Symptoms, and the domain of Sleep was added to the instrument.

The final CLDQ-PSC instrument has 24 items grouped into 5 domains: Emotional, Fatigue, Sleep, Systemic Symptoms, and Worry (Table 2, Supplemental Table 2, http://links.lww.com/HC9/A108). The total CLDQ-PSC score can be calculated as an average of the 5 domain scores.

Validation of CLDQ-PSC

Good to excellent internal consistency was detected in all 5 CLDQ-PSC domains: all Cronbach alphas >0.85 (Table 3). In addition, after sequential 1-item exclusions, the resulting Cronbach alphas did not change substantially; this indicates that the items are neither too correlated with, nor too different from other items belonging to the same domains. The item-to-own-dimension correlations were >0.5 for all but 1 item, >0.66 for 22/24 items (Table 3). The distribution of the CLDQ-PSC domain scores suggested some skewness toward higher values, with the Fatigue domain being the least skewed (Figure 1A–F).

TABLE 3.

Internal consistency of Chronic Liver Disease Questionnaire-primary sclerosing cholangitis

Domain	Cronbach alpha	Item-to-own-dimension correlations	Cronbach alpha with 1 item removed
Emotional	0.90	0.69–0.83	0.87–0.90
Fatigue	0.94	0.72–0.91	0.92–0.94
Sleep	0.85	0.75	Not available
Symptoms	0.88	0.47–0.81	0.83–0.89
Worry	0.93	0.74–0.90	0.90–0.93

FIGURE 1.

The histograms of the domain and total scores of CLDQ-PSC. Abbreviations: CLDQ-PSC, Chronic Liver Disease Questionnaire-primary sclerosing cholangitis; IQR, interquartile range.

Known-group validity of CLDQ-PSC is summarized in Table 4. The instrument is able to discriminate between PSC patients with and without cirrhosis (p<0.05 for 3 domains and the total score) or its complications (p<0.01 for 3 domains and the total score), with and without obesity (p<0.05 for 3 domains and the total score), with and without history of depression (all p<0.05), on versus not on UDCA (p<0.05 for 2 domains, p<0.10 for 3 domains and the total score), and with versus without history of weight loss or malnutrition (all p<0.05) (Table 4).

TABLE 4.

Known groups validity and discriminatory power of CLDQ-PSC

Item	Age ≥55 y	Age <55 y
N	20	80
Emotional	5.47±1.44	4.88±1.23	0.0251
Fatigue	4.77±1.62	4.19±1.44	0.11
Sleep	5.03±1.72	4.40±1.80	0.19
Symptoms	5.43±1.34	5.22±1.26	0.49
Worry	5.01±1.48	4.36±1.66	0.13
Total CLDQ-PSC	5.14±1.37	4.61±1.19	0.07
Item	Male	Female	p
N	32	68	—
Emotional	5.17±1.31	4.91±1.28	0.31
Fatigue	4.95±1.38	4.00±1.45	0.0028
Sleep	4.97±1.63	4.32±1.84	0.11
Symptoms	5.49±1.43	5.15±1.19	0.09
Worry	4.61±1.64	4.43±1.65	0.65
Total CLDQ-PSC	5.04±1.29	4.56±1.19	0.05
Item	Obesity (BMI ≥30)	No obesity (BMI <30)	p
N	19	81	—
Emotional	4.49±1.67	5.11±1.16	0.13
Fatigue	3.57±1.41	4.48±1.46	0.0122
Sleep	3.74±1.90	4.71±1.72	0.0320
Symptoms	4.52±1.38	5.43±1.19	0.0145
Worry	4.07±1.86	4.59±1.58	0.27
Total CLDQ-PSC	4.08±1.43	4.86±1.15	0.0328
Item	History of depression	No history of depression	—
N	37	62	—
Emotional	4.08±1.36	5.54±0.90	0.0000
Fatigue	3.45±1.47	4.83±1.26	0.0000
Sleep	4.00±2.00	4.82±1.61	0.0327
Symptoms	4.86±1.31	5.50±1.21	0.0134
Worry	3.68±1.71	4.96±1.41	0.0003
Total CLDQ-PSC	4.01±1.30	5.13±1.00	0.0000
Item	On UDCA	Not on UDCA	—
N	50	50	—
Emotional	5.34±1.07	4.66±1.41	0.0200
Fatigue	4.57±1.43	4.04±1.51	0.07
Sleep	4.78±1.90	4.27±1.65	0.14
Symptoms	5.25±1.17	5.27±1.39	0.74
Worry	4.87±1.55	4.10±1.65	0.0335
Total CLDQ-PSC	4.96±1.13	4.47±1.30	0.08
Item	Cirrhosis	No cirrhosis	—
N	36	57	—
Emotional	4.86±1.54	5.17±1.03	0.53
Fatigue	3.94±1.62	4.61±1.34	0.0401
Sleep	4.08±1.85	4.84±1.62	0.0442
Symptoms	4.74±1.33	5.62±1.10	0.0018
Worry	4.24±1.78	4.69±1.36	0.22
Total CLDQ-PSC	4.37±1.37	4.99±1.01	0.0307
Item	Cirrhosis complications	No cirrhosis complications	—
N	14	81	—
Emotional	4.36±1.61	5.20±1.12	0.10
Fatigue	3.37±1.54	4.56±1.38	0.0082
Sleep	3.11±1.30	4.86±1.70	0.0006
Symptoms	4.23±1.25	5.52±1.17	0.0010
Worry	4.17±1.58	4.68±1.57	0.24
Total CLDQ-PSC	3.85±1.24	4.97±1.10	0.0038
Item	Weight loss or malnutrition	No weight loss or malnutrition	—
N	26	73	—
Emotional	4.52±1.45	5.15±1.20	0.0395
Fatigue	3.64±1.50	4.53±1.42	0.0116
Sleep	3.79±1.87	4.78±1.71	0.0139
Symptoms	4.69±1.27	5.45±1.22	0.0102
Worry	3.68±1.75	4.79±1.51	0.0076
Total CLDQ-PSC	4.06±1.33	4.94±1.13	0.0025

Abbreviations: BMI, body mass index; CLDQ-PSC, Chronic Liver Disease Questionnaire-primary sclerosing cholangitis.

The correlations of CLDQ-PSC domains with domains of other validated PRO instruments are summarized in Table 5. For the domains of CLDQ-PSC, the highest correlated domains of SF-36 were as follows: Social Functioning and Mental Health for Emotional Function (both rho>0.70), Vitality for Fatigue (ρ=0.88), General Health for Sleep (ρ=0.66), Bodily Pain for Symptoms (ρ=0.83), and Mental Health for Worry (ρ=0.69) (all p-values <0.0001). Furthermore, the Fatigue domain of CLDQ-PSC was highly correlated with the Fatigue domain of PBC-40 (ρ=−0.91), and Symptoms of CLDQ-PSC was the highest correlated with Symptoms of PBC-40 (ρ=−0.73) (all p<0.0001) (Table 5).

TABLE 5.

Validity of CLDQ-PSC: correlations between domains of CLDQ-PSC and domains of SF-36 and PBC-40 (Spearman correlation ρ; all p<0.01 for absolute ρ>0.25)

SF-36 domain	Emotional	Fatigue	Sleep	Symptoms	Worry
Physical Functioning	0.40	0.72	0.62	0.58	0.33
Role Physical	0.59	0.83	0.61	0.66	0.47
Bodily Pain	0.53	0.70	0.60	0.83	0.45
General Health	0.58	0.74	0.66	0.62	0.53
Vitality	0.65	0.88	0.62	0.59	0.54
Social Functioning	0.71	0.72	0.56	0.59	0.63
Role Emotional	0.66	0.81	0.54	0.60	0.50
Mental Health	0.86	0.62	0.48	0.49	0.69
PBC-40 domain
Symptoms	−0.57	−0.60	−0.55	−0.73	−0.44
Itch	−0.27	−0.38	−0.41	−0.50	−0.23
Fatigue	−0.67	−0.91	−0.73	−0.63	−0.49
Cognitive	−0.58	−0.76	−0.60	−0.53	−0.42
Emotional	−0.68	−0.60	−0.45	−0.50	−0.73
Social	−0.66	−0.78	−0.63	−0.60	−0.62

Abbreviations: PBC-40, Primary Biliary Cirrhosis; SF-36, Short Form-36.

Bold values indicate the strongest correlations for each domain of CLDQ-PSC.

Assessment of the items of PSC-PRO for the aspects of PSC potentially missed by CLDQ-PSC returned its 39/40 items being either not relevant to PSC (>65% report a PSC-PRO item being never or rarely a problem; 26/40 items) or highly correlated with at least 1 item of CLDQ-PSC (absolute ρ>0.65; 13/14 remaining items) (Supplemental Table 3, http://links.lww.com/HC9/A108).

Validation of the CLDQ-PSC in an independent sample

There were 45 PSC patients in the independent historic data set for whom CLDQ-PSC scores could be calculated: age 49±12 years, 71% male, 62% cirrhosis, and 27% Child-Pugh class B or C (significant or decompensated cirrhosis).

In this validation sample, the CLDQ-PSC scores of patients with Child-Pugh class B or C were significantly lower than those of patients with no cirrhosis or Child-Pugh class A (all but 1 p<0.05, all p<0.07) (Supplemental Table 4, http://links.lww.com/HC9/A108). On the other hand, there was no difference in the scores between PSC patients with no cirrhosis and patients with Child-Pugh class A (Supplemental Figure 1, http://links.lww.com/HC9/A108). Furthermore, correlations with the domains of SF-36 were similar to those observed in the first validation sample (all p<0.01) (Table 5, Supplemental Table 5, http://links.lww.com/HC9/A108). In addition, the level of serum albumin was significantly correlated with CLDQ-PSC Fatigue (ρ=+0.46, p=0.01) and CLDQ-PSC Symptoms (ρ=+0.45, p=0.01) scores. The same domains also negatively correlated with the level of bilirubin (ρ=−0.31 and −0.35, respectively) but did not reach statistical significance (p<0.10).

DISCUSSION

HRQL assessment has increasingly become important for patient-centered care and clinical trials of patients with liver disease. Historically, HRQL assessments for patients with CLD has been carried out using generic instruments such as SF-36 or liver disease-specific instruments such as the CLDQ (5–7). However, generic instruments may include items that are not applicable to patients with PSC while missing items that are not as relevant for the general population but are commonly affected in patients with this disease. At the same time, the validated disease-specific instrument PSC-PRO is relatively long (42 items) that includes questions not specifically related to HRQL (7, 12). In contrast, the newly developed CLDQ-PSC is a short and easy to use instrument with only 24 items, which covers 5 domains. Given the importance of questionnaire burden, a shorter HRQL instrument may be preferred in some clinical or research settings.

Our validation process found that all domains of CLDQ-PSC were highly correlated with relevant domains of the widely used generic SF-36 and of another biliary disease-specific instrument (PBC-40). Also, the CLDQ-PSC has excellent internal consistency with Cronbach alpha ranging from 0.85 to 0.94. In addition, CLDQ-PSC was able to show sufficient discriminatory power based on a number of important clinical parameters such as age, sex, comorbidities, and presence of cirrhosis and its severity. Finally, the instrument was found to be comprehensive when compared with the longer instrument, PSC-PRO. As such, the CLDQ-PSC may be preferred especially when PRO instrument burden is an issue.

In addition to the original validation cohort, we used a separate cohort to validate the CLDQ-PSC instrument. Again, evidence supported the discriminatory capability of the instrument, its high correlations with relevant domains of SF-36, and its internal consistency. These data support validity of the newly developed CLDQ-PSC. However, further work needs to be done to continue validation of this tool when used in different countries and settings, to determine whether there exists an optimal mode of administration (eg, paper vs. electronic, during visit vs. between visits), to prospectively assess its responsiveness to changes in patients’ status over time, to appreciate its test-retest reliability, and to determine the minimal clinically important difference.

The limitations of this study that need to be addressed in future studies include the assessment of correlations of CLDQ-PSC scores with laboratory tests such as alkaline phosphatase, albumin, and bilirubin, and markers of disease severity such as MELD scores; given the previous data obtained from other similar CLDQ instruments and other liver diseases (CLDQ-HCV), we expect that CLDQ-PSC should correlate with markers of disease severity. Furthermore, most of the patients in the study sample were female, employed, and college educated and these characteristics could have potentially introduced a bias; validation of the instrument in more diverse and representative populations is needed. The final version of the instrument was also not subjected to additional content or face validity assessments in this study.

In summary, CLDQ-PSC has been developed and validated as a short instrument to measure HRQL in patients with PSC. Although further longitudinal studies are needed, CLDQ-PSC could be used in clinical trials and observational studies of patients with PSC to allow comprehensive assessment of the disease burden and treatment benefits.

Supplementary Material

hc9-7-e0049-s001.docx ^{(44.6KB, docx)}