Utility and Its Relationship With Disease Activity and Physical Disability of Patients With Psoriatic Arthritis in Thailand

Physical disability and severity of skin disease are important factors for evaluating healthy utility in psoriatic arthritis (PsA) patient. This study aims to estimate health utility and its related factors incorporating both rheumatologic and dermatologic factors in PsA. This cross-sectional study performed in PsA patients in Thailand. EQ-5D-5L, Health Assessment Questionnaire Disability Index (HAQ-DI), and Psoriasis Area and Severity Index (PASI) tools were used to collect health utility, physical disability, and severity of psoriatic skin lesions by face-to-face interviewed. Linear regression analysis was performed to relate health utility and its related factors. Of 84 patients enrolled, 49 (58.3%) were female, 67 (79.8%) had remission or low disease activity. Eighty-three patients (98.8%) received conventional synthetic disease-modifying antirheumatic drugs (csDMARDs) or biologic DMARDs (bDMARDs). The mean±SD overall utility was 0.87±0.15. The utility score of patients with low disease activity was signicantly greater than those with moderate to severe disease activity (0.89±0.12 vs 0.72±0.19, p<0.001). The HAQ-DI (unstandardized β = -0.167, 95%CI; -0.218 to -0.116, p<0.001) and PASI (unstandardized β = -0.006, 95%CI; -0.009 to -0.003, p<0.001) were found to be signicant related factors for utility. These results indicated that Thai patients with PsA had relatively high health utility. However, most included patients were in remission or had low disease activity. The HAQ-DI and PASI showed a strong predictors of patients’ health utility.


Introduction
Psoriatic arthritis (PsA) is a chronic in ammatory musculoskeletal disease associated with psoriasis, manifesting most commonly as peripheral arthritis, dactylitis, enthesitis, spondylitis, and nail plate abnormalities 1 . The incidence of PsA is 3.6-7.2 per 100,000 person-years, while the prevalence ranges from 6-41% in patients with psoriasis 2 . PsA has a negative impact on health-related quality of life (HRQoL) and healthcare resource utilization 3 . Early identi cation and early treatment are important for improving long-term outcomes 4 .
Patients with PsA could experience severe impairment of physical function, occupational incapability and negative psychosocial affects. PsA with mild disease, usually involves a few joints, has minimal impact on the patients' HRQoL. Moderately severe and severe PsA shows a signi cant impact on the daily tasks of living and physical functions 5,6 . It may also impact the patients' physical and mental well-being and limit treatment responses 7 . Many patients could hardly perform the simple tasks of living due to severe pain. Physical disability is substantial in these patients 6 .
Treatment of PsA should aim to ameliorate the activity and severity of both joint and skin in ammation.
Medical treatments of PsA include non-steroidal anti-in ammatory drugs (NSAIDs), intra-articular corticosteroids, and disease-modifying anti-rheumatic drugs (DMARDs). DMARDs can reduce joint and/or skin symptoms and prevent disease progression. These agents may be classi ed as conventional synthetic DMARDs (csDMARDS) and biologic DMARDs (bDMARDs) 8 . Evidence indicates that such treatments can improve the physical signs and symptoms of PsA, as well as the patients' HRQoL 9 .
In economic evaluation, HRQoL may be measured as utility, which refers to the individual preference to his/her health status. For PsA, estimation of health utility is valued from both physical disability and severity of skin disease. A previous systematic review 10 has shown that the Health Assessment Questionnaire Disability Index (HAQ-DI) and the Psoriasis Area and Severity Index (PASI) are important tools for estimating health utility in patients with PsA. A previous study in Thai patients with PsA did not include PASI in the analysis model, causing incomprehensive estimation of utility and related factors 11 . Therefore, this study aimed to estimate health utility and its related factors which incorporating both rheumatologic and dermatologic clinical factors in Thai patients with PsA.

Material And Methods
Study design and patient enrollment. This cross-sectional study was conducted to determine the health utility and its relationship with disease activity, HAQ-DI and PASI scores in Thai patients with PsA. All PsA patients diagnosed according to the classi cation criteria for PsA (CASPAR) criteria 12 were invited to participate in this study. These patients had their outpatient regular rheumatology clinics visit at three university a liated hospitals between January and April 2020. Those who were unable to communicate or denied participating were excluded. Study protocol. All the subjects received detailed information about the purpose and procedure of the study and agreed to participate in writing. All patients gave their written informed consents before entering the study. Approval to conduct this cross-sectional study was obtained from the Central Research Ethics Committee (CREC) of Thailand in 2019 (certi cate number: COA-CREC004/2020). All the methods were performed in accordance with the Declaration of Helsinki, the relevant guidelines and regulations. Sample size estimation. A previous study on health utility in PsA patients reported mean and SD of 0.5 and 0.3, respectively 13 . Using the above mean and SD, and the 95%CI of true mean of 0.5 ± 0.05 (error = 0.05), the sample size of this study was calculated to be 138. Data collection. All eligible patients were invited to participate in this study. Baseline characteristics of participants were collected by a medical record review performed by rheumatologists. Participants were face-to-face interviewed for their current clinical status including their health utility, physical disability, and skin lesion activity by a trained research nurse or a rheumatologist. A structured data collection and interview form was developed. It consisted of four parts including 1) baseline characteristics; sex, age, health insurance, types of PsA, current disease activity (measured by clinical Disease Activity Index for Psoriatic Arthritis (cDAPSA)) 14 , deformity, co-morbidity, and PsA treatment, 2) the Thai version of the EQ-5D-5L for patients' health utility, 3) the Thai version of HAQ (Thai HAQ) for physical disability, and 4) PASI for the severity of psoriatic skin lesions. All collected data was veri ed by RS, PD, and UP.
Rheumatologists who were responsible for data collection were asked for any incomplete data to ensure the data validity.
Outcome Measures. Health utility and physical disability scores were determined by the Thai version of the EQ-5D-5L, and Thai HAQ, respectively 15 . The PASI questionnaire was used to determine the severity of the psoriatic skin lesions, the PsA disease activity, and the global assessment of the disease activity by the physicians, respectively.
The EQ-5D-5L 16 was used for assessing the patients' health-related utility. It consists of ve dimensions including mobility, self-care, usual activities, pain/ discomfort, and anxiety/depression. Each dimension has one question with ve response levels. Patients' responses from the EQ-5D-5L questionnaire were converted to the utility scores based on the Thai algorithms, which had been elicited from 1,207 general population living in 12 provinces from all regions of Thailand 17 .
The Thai HAQ was used for determining functional disability. It consists of eight domains including dressing and grooming, arising, eating, walking, hygiene, reach, grip, and activities 15 . The 4-point di culty level of the individual items from each domain is ranked from 0 (no dif culty) to 3 (unable to do). The highest score from each domain was then summed and averaged to a single disability index. The HAQ-DI scores range from 0-3, with higher score re ects greater disability.
The PASI questionnaire 18 was used for determining the severity of psoriatic skin lesions in such patients. It consisted of four domains including percentages of skin involvement and the severity of skin lesions assessed by 3 clinical signs: erythema score, in ltration score, and desquamation score 18 . The PASI score was calculated using an equation reported in the original study. For each body section; (head, upper limb, lower limb and trunk), the percentage of affected skin area by psoriasis was estimated on a scale from 0 to 4 according to erythema score, in ltration score and desquamation score, then converted to a grade, ranged from 0 to 6 18 . Statistical analyses. Baseline characteristics were presented as frequencies for categorical variables and means ± standard deviation (SD) for continuous variables. Kruskal-Wallis test was used to test the differences of the HAQ-DI, PASI, and health utility among disease activity (remission, low, and moderateto-severe). Dunn's pairwise comparison with Bonferroni correction was used to test the differences of each comparison when signi cant difference was observed among disease severity. Spearman rho correlation was performed to relate utility score and VAS.
Two steps regression approach was applied to determine relationship of health utility and it related factors. Univariate linear regression was performed to relate patients' health utility and its factors. Factors with p-value < 0.10 in the univariate analysis and variables of interests from the literature relating to health utility (age, cDAPSA, HAQ-DI, and PASI) were selected for multivariate linear regression using a forward selection. Factors with p-value < 0.05 in the multivariate analysis indicated signi cant relationship with patients' health utility. Variance in ation factor (VIF) was used to assess multicollinearity among independent variables, while adjusted R 2 was used to determine model's goodness-oft. All analyses were performed using STATA version 15.0.   Table  2). Additional subgroup analyses demonstrated the signi cant differences between low and moderate to severe disease activity (p<0.001), and between remission and moderate-to-severe disease activity (p<0.001). However, the difference was not observed between patients with low disease activity and those in remission (p=0.301).   (Table  4). Spearman rho correlation analysis indicated that utility score and VAS are signi cantly correlated (p<0.001, rho = 0.548; Fig. 1). Abbreviations: EQ-5D-5L: EuroQoL; SD: Standard deviation. *Tested by Kruskal-Wallis Subgroup analysis found that the utility was signi cantly different between patients with low and moderate-to-severe disease activity (p<0.001) and between those in remission and moderate-to-severe disease activity (p<0.001). However, the difference was not signi cant between patients with low disease activity and those in remission (p=0.306).

Results
Health utility and its related factors. Univariate analysis indicated that cDAPSA, HAQ-DI score, and PASI score were signi cantly related factors of health utility (p<0.05). No signi cant relationship were found  (Table 5). According to standardized beta coe cient, HAQ-DI score was the most in uencing factor for patients' health utility (standardized β = -0.664) followed by PASI score (standardized β = -0.299). Adjusted R 2 of the nal model was 0.5899 with VIF ranged from 1.09 -2.45.

Discussion And Conclusions
This study found that PsA patients with remission or low disease activity signi cantly had lower HAQ-DI scores compared to those with moderate-to-severe disease activity. The majority of PsA patients had low PASI scores meaning that our samples had slightly skin lesion. The HAQ-DI score and PASI score had reverse relationship to patients' health utility based on the negative sign of beta coe cient.
Our ndings showed relatively high utility possibly due to 80% of PsA patients in this study being in remission or having low disease activity. This was in line with a Thai study which reported high utility in patients with low cDAPSA 11 .
Our ndings were similar to a previous study from the United Kingdom 19 which found that patients' health utility was associated with HAQ-DI and PASI scores. However, the magnitudes of effect were slightly different. The UK study reported unstandardized β of HAQ-DI as -0.298, while that of PASI was -0.004, while those for our study were -0.167 and -0.006, respectively. The differences meant that decreases in HAQ-DI and PASI score in Thai patients had less impact on overall health utility than patients in the United Kingdom.
The magnitude of reduction in HAQ-DI score in this study was also in line with a report from Canada 20 , which found that reduction of the HAQ-DI by approximately 0.13 point resulted in a clinically signi cant perception of clinical improvement by the patients. This nding re ected a meaningfully better HR-QoL.
The HAQ-DI might vary in patients with different joint diseases. The average HAQ-DI in Thai PsA in this study (0.49±0.60) was lower than that reported in Spanish population (0.76±0.67) 21 . This might be due to several reasons where differences in patients' characteristic were the main reason. Difference in cultures might in uence in the perception of the disease. In Thailand, family supports are very common, and patients are looked after by their own family members, which might indirectly affect the patients' self-care activities, resulting in lower HAQ-DI scores. Lastly, the Thai healthcare systems allow patients to easily get access to the primary care units upon their needs, for example, the village health volunteers could reach out the patients at home.
Our study was different from a previous Thai study 11 which reported the association of health utility and its related factors in Thai patients with PsA. The previous study was conducted in a university hospital in Bangkok, a capital of Thailand, while our study was conducted in three university hospitals. Two hospitals are located in Bangkok and another hospital is located in the northern part of Thailand. We believe that our study was better represent characteristics of Thai PsA patients. Another important difference was that the previous study did not incorporate PASI score in the study which was found to be signi cant related factors for patients' health utility in this study. This study was more comprehensive in terms of the factors used in the analysis because we incorporated factors from both rheumatologic and dermatologic clinical factors.
Since this study found a strong relationship between the HAQ-DI, PASI, and health utility, we suggest that the HAQ-DI and PASI instruments should be incorporated in routine monitoring practice. The HAQ-DI and PASI can be used to estimate health utility score which is an important input for further cost-effectiveness study.
Several limitations should be addressed in this study. Although this study planned to collect the data from the three large university hospitals located in Thailand, but the number of the patients could not be Figures Figure 1 A correlation between utility and visual analog scale (VAS).