This was a two-phase cross-sectional study involving translation and adaptation into the target language, followed by validation. Thirty-seven patients with SSc were consecutively recruited between July and October 2017. The patients were diagnosed using the 2013 American College of Rheumatology SSc criteria in the SSc Clinic of Soonchunhyang University Seoul Hospital [17]. For proper sample size calculation, the number of subjects was determined according to a previous study [5]. Two-sided tests (alpha = 0.05, power = 80% [beta = 0.2]) were performed with the correlation coefficient rho set to 0.2 to 0.6. A rho value of 0.2 indicated a minimum number of subjects of 30. Taking into consideration a 20% dropout rate, 37 subjects were included in this study.

SSc is classified as limited cutaneous (lc) or diffuse cutaneous (dc) systemic disease according to the distribution of skin involvement [18]. For each patient, the following data were collected: demographic characteristics, body mass index (BMI), Raynaud’s phenomenon, and the extent of skin thickening. The extent of skin thickening was evaluated using the modified Rodnan total skin score (mRSS). Laboratory findings, including serum levels of anti-nuclear, anti-centromere, anti-topoisomerase, and anti-RNP antibodies, were also recorded. We investigated internal organ involvement including digital ulcers, SSc-associated interstitial lung disease (as represented by pulmonary function), gastroesophageal reflux disease (GERD; by esophagogastroduodenoscopy, EGD) [19], scleroderma renal crisis (defined as malignant hypertension with acute renal insufficiency occurring during the course of SSc) [20], and heart involvement including pulmonary arterial hypertension (defined as a mean pulmonary artery pressure ≥ 20 mmHg, pulmonary artery wedge pressure < 15 mmHg, and pulmonary vascular resistance > 3 Wood units measured during right-heart catheterization at rest according to the Sixth World Symposium on Pulmonary Hypertension) [21]. Pulmonary function was assessed by determining the functional vital capacity (FVC; % predicted) and diffusion capacity of the lung for carbon monoxide (DLCO; % predicted) (corrected for hemoglobin). Patients with an incomplete clinical assessment were excluded. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board for Human Research of Soonchunhyang University Seoul Hospital (IRB no. SCH 2017-07-029). Written informed consent was obtained from all participants.

The English-language version of the instrument is available at https://umich.qualtrics.com/jfe/form/SV_3eBP4A4umBwnSvj/ and was used in this study with the permission of Professor Khanna. Translation and cross-cultural adaptation of the instrument were performed according to the published recommendations [22,23]. The proposed steps were synthesis of translation, back translation, expert committee review, and pre-testing. First, three people of different backgrounds (e.g., one medical and two non-medical) independently performed a forward translation of the instrument from the English-language to the Korean language. The adjustments made were combined by consensus among the translators. Then back translation was performed by two independent bilingual native speakers of English, who were blinded to the original English-language version. The two (forward and back) non-medical translators are very familiar with cross-cultural adaptation. The final version (Supplementary Table 1) was, with the approval of the Expert Committee, preliminarily administered to patients with SSc attending the hospital. After patients answered, each question was discussed with them to check whether all items had been fully understood.

The UCLA SCTC GIT 2.0 is a 34-item questionnaire with seven subscales (reflux, distension/bloating, fecal soilage, diarrhea, social functioning, emotional well-being, and constipation). Each item scores the frequency of symptoms over a recall period of 7 days from 0 to 3, where 0 indicates no GI problems and three indicates worse health. The exceptions are questions 15 (diarrhea subscale) and 31 (constipation subscale), which are scored from 0 (no matter) to 1 (worse health). The score of six subscales (excluding constipation) is summed as the total score, which captures the overall burden of disease (the possible scores range from 0 to 2.83).

The Short Form Health Survey version 2 (SF-36v2) is the most widely used measurement of general health and QoL [24]. It is based on norm-based scoring algorithms standardized by the same average (i.e., 50) and the same standard deviation. The survey has eight aspects: physical functioning (PF), physical role (RP), bodily pain (BP), general health (GH), vitality (VT), social role functioning (SF), emotional role functioning (RE), and mental health (MH). The physical (PCS) and mental component scores (MCS) are calculated based on those aspects; higher scores indicate a higher QoL [25].

The EQ-5D questionnaire is one of the most widely used instruments for measuring utility, and is available in several languages including Korean [26]. It consists of five health dimensions: mobility, self-care, usual activities, pain or discomfort, and anxiety or depression. The questionnaire also comprises a visual analogue scale domain, which measures the overall HRQoL on a scale of 0 (worst health state) to 100 (best health state).

The reliability of the UCLA SCTC GIT 2.0 was assessed by the test–retest method at a 1-week interval in all 37 patients. This yielded an estimate of the instrument’s reproducibility over time as measured by intraclass correlation (ICC). The test–retest reliabilities were interpreted as appropriate if the ICCs were > 0.7 and 0.9 for group and individual comparisons, respectively.

Internal consistency was evaluated using the Cronbach’s α method for the seven subscales and the total score of the Korean version of the UCLA SCTC GIT 2.0 questionnaire. A Cronbach’s α coefficient > 0.7 indicates a good correlation. Floor and ceiling effects were defined as the proportion of patients with the minimum (floor, representing absence of symptoms) and maximum (ceiling, representing the worst symptoms) possible scores. These proportions were calculated for each scale and an effect was considered present if more than 15% of patients had the maximum or minimum score. We assessed the construct validity of the UCLA SCTC GIT 2.0 questionnaire by examining its correlations with the SF-36v2 and EQ-5D scales by Spearman correlation [8–10,12]. Correlations (rho) ≤ 0.29 were considered small, 0.30 to 0.49 were moderate, and those ≥ 0.50 were considered large. SPSS version 26.0 software (IBM Co., Armonk, NY, USA) was used for statistical analyses. Continuous variables were analyzed using the Wilcoxon signed-rank test and categorical variables were assessed using the chi-square or Fisher exact tests. Simple relationships were assessed using Spearman rank-correlation analyses. Differences were considered statistically significant at p < 0.05.

Patients with an incomplete clinical assessment were excluded from the analyses. Complete data were available for 37 patients (12 lcSSc and 25 dcSSc) with consecutive scheduled study instruments. Those patients’ characteristics are listed in Table 1. Most of the patients with SSc (n = 37) were women (89.19%), and they had a mean age of 52.2 years, BMI of 22.1 kg/m², median disease duration (interquartile range [IQR]) of 24 months, and median mRSS of 4 points. The lcSSc subtype was diagnosed in 12 patients (32.4%) and the dcSSc was documented in 25 (67.6%). The patients with dcSSc were more frequently anti-Scl70 positive; had higher GERD on EGD, and had lower FVC, forced expiratory volume in 1 second (FEV1), and DLCO than those with lcSSc. The HRQoL instruments in patients with SSc are shown in Table 2. The median total GIT score was 0.3 and did not differ between the lcSSc and dcSSc subtypes.

The Korean UCLA SCTC GIT 2.0 questionnaire showed good and reproducible reliability; the ICC values were close to 0.7 and statistically significant for all subscales (Table 3). For most subscales, Cronbach’s α was ≥ 0.7; the exceptions were the diarrhea and social function subscales (Cronbach’s α = 0.551, α = 0.421), suggesting good internal consistency. There was floor effect for both the scale itself and all of the subscales, ranging from 16% (total GIT score) to 89% (fecal soilage); however, there was no ceiling effect (Table 4).

To assess construct validity, the UCLA SCTC GIT 2.0 was compared to the SF-36v2 and EQ-5D. The reflux domain of the UCLA SCTC GIT 2.0 was moderately correlated with the SF-36v2 (PF, RP, BP, VT, SF, RE, MH, and MCS). The emotional well-being domain of the was negatively correlated with that of the SF-36v2, except the PCS. The total GIT score of the UCLA SCTC GIT 2.0 was negatively correlated with that of the SF-36v2, except RP, BP, and GH (Table 5). However, most domains of the UCLA SCTC GIT 2.0 were not correlated with those of the EQ-5D. The emotional well-being domain was significantly correlated with that of the EQ-5D, except usual activities and pain. The total GIT score of the UCLA SCTC GIT 2.0 was significantly correlated with that of the self-care domain of the EQ-5D (rho = 0.385, p = 0.019) (Table 5).