The purpose of this study was to determine the prevalence, incidence, and mortality rates of autoimmune rheumatic diseases (ARDs) by using a population-based database.
The purpose of this study was to determine the prevalence, incidence, and mortality rates of autoimmune rheumatic diseases (ARDs) by using a population-based database.
We used the longitudinal health insurance database (comprising 1,000,000 beneficiaries) of the Taiwan National Health Insurance from 2000 to 2008 and the National Death Registry of Taiwan from 2000 to 2008.
The overall prevalence of major ARDs was 101.3 (95% confidence interval [95% CI] 27.5–107.9) per 100,000 populations; the prevalence was 165.1 (95% CI 44.8–177.1) in women and 40.1 (95% CI 10.9–46.1) in men. The prevalences of rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), Sjögren's syndrome, progressive systemic sclerosis, polymyositis/dermatomyositis, vasculitis, and Behçet's disease were 52.4 (95% CI 14.2–57.2), 37.0 (95% CI 10.0–41.0), 16.0 (95% CI 4.3–18.7), 3.8 (95% CI 1.0–5.3), 2.9 (95% CI 0.8–4.2), 5.7 (95% CI 1.6–7.4), and 1.4 (95% CI 0.4–2.3) per 100,000 persons, respectively. Between 2001 and 2008, the incidence rates (per 100,000 person-years) for these diseases were 17.3, 8.4, 10.6, 1.5, 1.5, 1.2, and 0.8, respectively. The incident cases with ARDs had a higher risk of mortality, with the standardized mortality ratio (SMR) ranging from 1.3 to 3.7.
In 2000, the prevalence of major ARDs was 1.4–52.4 per 100,000 persons in Taiwan. Between 2000 and 2008, the incidence rates of various ARDs were 0.8–17.3 per 100,000 person-years. The prevalence and incidence of RA were the highest, followed by SLE and Sjögren's syndrome, and those of Behçet's disease were the lowest. Patients with different types of ARDs had higher mortality and SMR than those of the general population.
Autoimmune rheumatic diseases (ARDs) such as rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), Sjögren's syndrome, systemic sclerosis (SSc), polymyositis (PM), dermatomyositis (DM), vasculitis (including polyarteritis nodosa, Kawasaki disease, hypersensitivity angiitis, granulomatosis with polyangiitis [Wegener's] [GPA], giant cell arteritis, Takayasu disease, etc.), and Behçet's disease affect the immune system and are a common cause of disability. ARDs are generally thought of as being relatively rare, but their effects on mortality and morbidity are quite high. Over the past 2 decades, few population-based epidemiologic studies on ARDs have been reported (1–7). Most of these studies were conducted using defined populations such as insurance- or community- based populations. Moreover, there is a lack of current population-based prevalence data from areas other than Europe and North America for many of the autoimmune diseases (8). The present study is based on a nationwide population from Taiwan, which can provide the most reliable estimate of the prevalence and incidence of ARDs.
Taiwan encompasses an area of 35,563 km2 (13,869 square miles) and has a population of approximately 23 million. In 1995, Taiwan established a compulsory single-payer National Health Insurance (NHI) system for nearly the entire nation. The population coverage had reached 99% by the end of 2004 (9). The Taiwan National Health Insurance Research Database (NHIRD) is an excellent database from which to obtain a representative and reliable estimate of the prevalence and incidence of ARDs in Taiwan. This study aimed to obtain the estimates of prevalence, incidence, and mortality rates associated with major ARDs in Taiwan by using the NHIRD and National Death Registry (NDR).
This nationwide population-based epidemiologic study on autoimmune rheumatic disease (ARDs) in Taiwan might allow for global comparison in the future.
This is the first nationwide population-based epidemiologic study on ARDs in Asia Taiwan.
Estimates of the prevalence and incidence rate of individual ARDs provide data needed for health policy discussions and setting research priorities, and can add on health services planning.
Compared with the Taiwanese population having the same age composition, the incident cases with ARDs had a higher risk of mortality, with the standardized mortality ratio ranging from 1.3 to 3.7.
This study had a population-based longitudinal design. Primary data were from the Taiwan NHIRD and the NDR. This study was approved by the Institutional Review Board of the Chang Gung Memorial Hospital. Due to the anonymous nature of the data, informed patient consent was not required.
The NHI system is a single-payer social health insurance system. All residents of Taiwan are required to participate. The Taiwan NHIRD releases sets of sampling files, called the Longitudinal Health Insurance Database (LHID), for research purposes. The LHID 2000 contains the entire original claims data for 1,000,000 beneficiaries randomly sampled from the entire population of NHI beneficiaries in the year 2000. The source population was the entire population covered by the NHI system (23,753,407 individuals) in the year 2000. Each individual was assigned a serial number ranging from 1 to 23,753,407. By means of a linear congruential random number generation method, 1,000,000 nonduplicated random numbers were generated. Individuals with serial numbers corresponding to these numbers were selected. All registration and claims data for these 1,000,000 individuals, recoded for the period of 1996 to 2008, were collected and distributed as LHID 2000. The population of the present study is representative of the Taiwanese population. There were no statistically significant differences in age, sex, or health care costs between the sample group and all enrollees, as reported by the Taiwan NHIRD (10). In Taiwan, the issuance of a Catastrophic Illness Certificate to each ARD patient with a catastrophic illness requires a thorough clinical and laboratory evaluation, the fulfillment of the appropriate classification criteria, and a review by rheumatologists commissioned by the Bureau of the NHI system. Therefore, the catastrophic illness patient data are highly accurate and reliable.
Survival status and date of death from 2000–2008 were ascertained by using the Taiwan NDR, which records the cause of death for all deceased citizens. The accuracy of this coding system has been validated in previous studies (11).
Records with the following International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes were identified as follows: 714.0 for RA, 710.0 for SLE, 710.2 for Sjögren's syndrome, 710.1 for SSc, 710.3 for PM, 710.4 for DM, 446.0–446.7 for vasculitis (including polyarteritis nodosa, Kawasaki disease, hypersensitivity angiitis, GPA, giant cell arteritis, and Takayasu disease), and 136.1 for Behçet's disease.
Because the LHID 2000 contains individuals from the 2000 registry of NHI beneficiaries, the prevalence of ARDs was computed for the year 2000 only. Although 1,000,000 NHI beneficiaries were included in the LHID 2000, only 963,355 individuals (96.3%) were used in this study because the sex-related data for some patients were missing, or some patients had died before the end of 2000. The prevalence of ARDs was calculated as the number of ARD patients divided by the total NHI beneficiaries in the same year. Incidence was calculated as the numbers of new cases from 2000 through 2008, dividing by person-years involved. The 95% confidence intervals (95% CIs) of incidence density were calculated assuming a Poisson distribution. The Kaplan-Meier method was used to compute the survival rate for various ARDs. Cox proportional hazard model was used to obtain the unadjusted and adjusted hazard ratios of sex for varied ARDs. The standardized mortality ratio (SMR) was calculated by dividing observed deaths by expected deaths using the sex- and age-related death rate of the Taiwanese population in 2004. The 95% CIs for SMRs were derived on the basis of the assumption of a Poisson distribution of the observed deaths (12). All statistical analyses were performed by using SAS statistical software, version 9.1. The significance level was set at 0.05.
In 2000, 1,143 (932 women and 211 men) patients with major ARDs were identified. The prevalences of RA, SLE, Sjögren's syndrome, SSc, PM/DM, vasculitis, and Behçet's disease were 52.4, 37.0, 16.0, 3.8, 2.9, 5.7, and 1.4 per 100,000 populations, respectively (Table 1). The overall prevalence of major ARDs, including RA, SLE, SSc, Sjögren's syndrome, PM, DM, vasculitis, and Behçet's disease, was 101.3 (95% CI 27.5–107.9) per 100,000 populations; the prevalence per 100,000 populations was 165.1 (95% CI 44.8–177.1) in women and 40.1 (95% CI 10.9–46.1) in men (data not shown). In general, the prevalence of ARDs was higher in women than in men, with an overall women to men ratio of 4.1 (95% CI 3.5–4.8). Except for vasculitis and Behçet's disease, the female to male ratio was significant and was the highest for Sjögren's syndrome, followed by SLE, SSc, RA, and PM/DM (Table 1).
|No.||Rate (95% CI)||No.||Rate (95% CI)||No.||Rate (95% CI)||W:M ratio (95% CI)|
|RA||505||52.4 (14.2–57.2)||408||86.5 (23.5–95.3)||97||19.7 (5.4–24.1)||4.4 (3.5–5.5)|
|SLE||356||37.0 (10.0–41.0)||314||66.6 (18.1–74.0)||42||8.5 (2.3–11.6)||7.8 (5.6–10.8)|
|Sjögren's syndrome||154||16.0 (4.3–18.7)||136||28.8 (7.8–34.1)||18||3.7 (1.0–5.8)||7.9 (4.8–12.9)|
|SSc||37||3.8 (1.0–5.3)||30||6.4 (1.7–9.1)||7||1.4 (0.4–2.9)||4.5 (2.0–10.2)|
|PM/DM||28||2.9 (0.8–4.2)||20||4.2 (1.2–6.6)||8||1.6 (0.4–3.2)||2.6 (1.1–5.9)|
|Vasculitis†||50||5.7 (1.6–7.4)||17||3.6 (1.0–5.8)||33||6.7 (1.8–9.4)||0.5 (0.3–1.0)|
|Behçet's disease||13||1.4 (0.4–2.3)||7||1.5 (0.4–3.1)||6||1.2 (0.3–2.7)||1.2 (0.4–3.6)|
Between 2000 and 2008, 3,312 new cases of major ARDs were recorded in 8,026,058 person-years. The sex-specific incidences of major ARDs are shown in Table 2. There were 1,390 RA incidence cases during the 8-year study period, comprising 1,033 women and 357 men. The incidence rate of RA during 2000–2008 was 17.3 per 100,000 person-years, which was 26.0 in women and 8.8 in men. There were 671 SLE incidence cases during the 8-year study period, including 595 women and 76 men. The female/male ratio was 8.0. The incidence rate of SLE was 8.4 per 100,000 person-years, which was 15.0 in women and 1.9 in men. The incidence of other ARDs is shown in Table 2.
|No.||Rate (95% CI)||No.||Rate (95% CI)||No.||Disease||W:M ratio (95% CI)|
|RA||1,390||17.3 (16.4–18.2)||1,033||26.0 (24.4–27.6)||357||8.8 (7.9–9.7)||3.0 (2.6–3.3)|
|SLE||671||8.4 (7.7–9.0)||595||15.0 (13.8–16.2)||76||1.9 (1.5–2.3)||8.0 (6.3–10.2)|
|Sjögren's syndrome||855||10.6 (9.9–11.4)||736||18.5 (17.2–19.9)||119||2.9 (2.4–3.5)||6.3 (5.2–7.7)|
|SSc||118||1.5 (1.2–1.7)||101||2.5 (2.1–3.0)||17||0.4 (0.2–0.6)||6.1 (3.6–10.1)|
|PM/DM||117||1.5 (1.2–1.7)||76||1.9 (1.5–2.3)||41||1.0 (0.7–1.3)||1.9 (1.3–2.8)|
|Vasculitis||97||1.2 (1.0–1.5)||51||1.3 (0.9–1.6)||46||1.1 (0.8–1.5)||1.1 (0.8–1.7)|
|Behçet's disease||64||0.8 (0.6–1.0)||40||1.0 (0.7–1.3)||24||0.6 (0.4–0.8)||1.7 (1.0–2.8)|
Mortality rate (per 100,000 person-years), survival, and unadjusted and age-adjusted female/male ratios for mortality in cases with ARDs are shown in Table 3. The overall mortality rate of major ARDs ranges from 943.6 to 3,185.8 per 100,000 person-years. In the female ARD cases, the PM/DM group had the highest mortality rate (3,187.1 per 100,000 person-years) and the vasculitis group had the lowest mortality rate (370.0 per 100,000 person-years). In the male ARD cases, the RA group had the highest mortality rate (4,130.7 per 100,000 person-years), and the Behçet's disease group had the lowest morality rate (991.3 per 100,000 person-years). The 8-year survival rate in the cases with ARDs ranged from 83.9% to 94.1%. Notably, the mortality rate was significantly higher in male than in female cases of RA, SLE, and Sjögren's syndrome, even after age adjustment (Table 3).
|Disease||Deaths, no.||Mortality||Survival, %|
|1 year||2 years||5 years||8 years||Unadjusted W:M ratio (95% CI)||Age-adjusted W:M ratio (95% CI)|
|Women||72||1,639.1||98.2||97.3||92.0||87.9||0.4 (0.3–0.6)||0.5 (0.4–0.7)|
|Women||28||1,019.5||98.8||97.6||94.8||92.4||0.3 (0.1–0.5)||0.4 (0.2–0.8)|
|Women||29||987.4||99.4||97.7||94.8||92.8||0.3 (0.2–0.5)||0.4 (0.2–0.7)|
|Women||7||1,590.8||99.0||91.0||91.0||91.0||0.6 (0.1–2.9)||0.9 (0.2–4.6)|
|Women||10||3,187.1||94.5||91.5||83.0||83.0||1.0 (0.3–2.8)||1.0 (0.3–3.0)|
|Women||1||370.0||98.0||98.0||98.0||98.0||0.2 (0.0–2.1)||0.3 (0.03–2.3)|
|Women||2||1,240.1||100.0||100.0||93.2||93.2||1.1 (0.1–12.6)||0.8 (0.1–9.7)|
Compared with the Taiwanese population having the same age composition, the incident cases with ARDs had a higher risk of mortality, with SMRs ranging from 1.3 to 3.7 (Table 4). Except for male and female cases of Sjögren's syndrome; male cases of SSc, Behçet's disease, and PM/DM; and female cases of vasculitis, significant SMRs were noted in all other cases. For RA, the overall SMR was 1.6 (1.7 for women and 1.5 for men). The SMR for SLE was 2.9 (2.7 for women and 3.7 for men). The SMR for Sjögren's syndrome was 1.3 (1.2 for women and 1.5 for men). The SMR for SSc was 2.2 (2.4 for women and 1.6 for men). The SMR for PM/DM was 2.5 (4.6 for women and 1.3 for men). The SMR for vasculitis was 2.6 (1.2 for women and 3.7 for men). The SMR for Behçet's disease was 3.7 (6.5 for women and 2.0 for men) (Table 4).
|Disease||No.||E||SMR (95% CI)|
This is the first nationwide population-based epidemiologic study of ARDs in Asia Taiwan. This study provides an opportunity to conduct an international comparison of prevalence and incidence rates of ARDs in the future. The prevalence and incidence of RA were the highest, followed by SLE and Sjögren's syndrome, and the rates for Behçet's disease were the lowest. The findings on the incidence of RA and SLE in the present study are similar to those reported in previous studies (13, 14). There are notable differences in the sex distribution among ARDs. Almost all ARDs disproportionately affect women. In some diseases (e.g., SLE, RA, Sjögren's syndrome, and SSc), 75% or more of patients are women. The disparity by sex is small in other diseases (e.g., inflammatory myositis, Behçet's disease). A relatively equal risk between women and men is seen in some ARDs (overall vasculitis). The different types of ARDs, excluding vasculitis and Behçet's disease, were more predominant in women; this finding is consistent with those reported in previous studies (1–3). In general, patients with different types of ARDs had higher mortality and SMRs than those of the general population. However, the mortality rate was higher in male incident cases than in female incident cases. The accumulative global burden of ARDs is substantial. Studying the distribution of ARDs across various global regions and ethnic groups may provide epidemiologic data and also advance our understanding of the genetic and environmental influences. In this study, we examined large studies pertaining to the prevalence and incidence of a broad array of ARDs. Below, the epidemiologic data for major ARDs investigated in the present study have been compared to those of large epidemiology studies conducted in other countries.
The incidence and prevalence of RA vary across populations, statistical methods, and disease definition (15). Geographic and ethnic factors also influence incidence and prevalence data. The prevalence of RA is approximately 0.5–1.1% in whites (13, 16, 17); 0.1% in rural Africans; and 5% in Pima, Blackfeet, and Chippewa Indians (13, 17). Some studies showed declining incidences and prevalences after the 1960s (15). In China, the RA prevalence ranged from 0.2% to 0.37% (5). In the present study, the prevalence of RA was 0.05%, which is lower than that reported in previous studies conducted in Taiwanese (4), Chinese (5), and white (13) populations. Similar prevalence and incidence rates of RA and SLE in Taiwanese populations have been reported using the same database (18, 19). The reason for this discrepancy remains unclear and mandates further investigation. The reported annual incidence rates of RA were higher in North American and North European countries (20 and 50 per 100,000 persons, respectively) than in Southern European countries (10–20 per 100,000 persons) (13). In the present study, the incidence rate of RA was 17.3 per 100,000 person-years, which was similar to that reported in previous studies from Southern European countries, but lower than that reported in studies conducted in North American and North European countries (13). The SMRs varied from 1.29–2.98 in previous reports (7); this value is similar to that found in the present study (1.6).
The reported prevalence of SLE in the population is 20–150 cases per 100,000 cases across a broad diversity of geographic regions (1, 14). In a recent comprehensive review, the incidence rates of SLE ranged from approximately 1–10 per 100,000 person-years, and the prevalence generally ranged from 20–70 per 100,000 person-years (14, 20, 21). Surveys conducted in China showed that the prevalence of SLE was 31.1 to 70.1 per 100,000 persons (22, 23), which is similar to the prevalence rates reported for US white populations (14). In the present study, the prevalence rate of SLE in Taiwan in 2000 was 37.0 (which was 66.6 in women and 8.5 in men) per 100,000 population, and the incidence rate was 8.4 per 100,000 person-years, which was 15.0 and 1.9 per 100,000 person-years in women and men, respectively; this rate was similar to that reported previously in China and the US (14, 20–23). The SMR of SLE patients in this study (2.9) was similar in magnitude to those reported in Western countries (2.70–5.25) and Hong Kong (24–26).
SSc has an annual incidence of 1–2 per 100,000 individuals in the US (27). In a population-based study of SSc in southeast Michigan, 706 SSc cases were identified and extrapolated to the US population, yielding a prevalence of 24.2 per 100,000 adults (95% CI 21.3–27.4) (28). According to a Quebec physicians billing and hospitalization database, the prevalence of SSc in 2003 was 74.4 cases per 100,000 women and 13.3 cases per 100,000 men (29). In the present study, the prevalence of SSc was 3.8 per 100,000 persons, which was much lower than that reported in previous studies conducted in other countries (28, 29). The SMR of SSc patients varied from 1.2 to 4.1 in previous reports (30, 31); this is similar to the SMR found in the present study (2.2).
The prevalence estimates of primary Sjögren's syndrome range from 0.05% to 4.8% across international communities (2). The prevalence of Sjögren's syndrome among women from 2 primary care practices in the UK ranged from <0.1% to 0.4% (32). In a population-based study from Olmsted County, Minnesota, the average annual age- and sex-adjusted incidence of physician-diagnosed Sjögren's syndrome was estimated to be 3.9 (95% CI 2.8–4.9) per 100,000 populations (33). In a study from China (34), the prevalence rate was 0.77% when the Copenhagen criteria were used and 0.33% when the San Diego criteria were used. In the present study, the prevalence of primary Sjögren's syndrome was 0.016%, which is lower than those reported in whites (32, 33). The SMR of Sjögren's syndrome was reported to be 1.02–2.07 (35), which is similar to that found in the present study (1.3).
In Canada, Bernatsky et al found that the prevalence of PM/DM was 21.5 per 100,000 persons (95% CI 19.4–23.9) (36). The estimated annual incidence rate of PM and DM varies between 1.9 and 7.7 per million (37). Between 1976 and 2007, a study in Olmsted County, Minnesota, showed that the overall age- and sex-adjusted incidence of DM, including all subtypes, was 9.63 per million persons (38). In the present study, the prevalence of PM and DM was 1.5 per 100,000 persons and 1.6 per 100,000 persons, respectively, which was lower than that reported in a previous study (36). The SMRs of PM/DM were reported to be 1.75–2.92 (39, 40), which is similar to that found in the present study (2.5).
Systemic vasculitis comprises a heterogeneous group of diseases characterized by a common histopathologic feature of inflammation and necrosis of blood vessels (41). Studies from England, Spain, and Scandinavia have shown that the overall annual incidence of primary systemic vasculitis is approximately 20 cases per million (42–44). In various populations, certain types of vasculitis appear more frequently, suggesting that either genetic or environmental factors may be associated with the prevalence of vasculitis in certain populations. For example, the incidence of polyarteritis nodosa and microscopic polyangiitis among the ethnic Kuwaiti population was found to be 16 per million and 24 per million, respectively (45). Giant cell arteritis is the most common type of vasculitis, and its incidence is the highest in the populations of Scandinavian descent, where the annual incidence reaches 15–35 per 100,000 persons in individuals ages >50 years (46). Takayasu arteritis has a relatively uniform global incidence of 1–2 per million populations, and antineutrophil cytoplasmic antibody–associated vasculitides have an overall incidence rate of 20 per million populations, with a peak age at onset of 65–74 years (46). GPA appears to be more common in northern Europe than microscopic polyangiitis, which seems to be more common in southern Europe (46). While Kawasaki disease is the most common type of vasculitis in Southeast Asian children, in Japan the annual incidence among children ages <5 years is approximately 137.7 per 100,000 persons per year (47). In the present study, the overall prevalence of vasculitis was 5.7 per 100,000 persons, and Kawasaki disease was detected in more than half of the cases (data not shown). The SMR of certain vasculitides was reported to be 2.77 (48), which is similar to that found in the present study (2.6).
The prevalence of the HLA–B51 allele is high among patients with Behçet's disease who live in areas along the Silk Road, but not among white patients who live in Western countries (13%) (49, 50). Turkey has the highest prevalence at 80–370 cases per 100,000 populations. The prevalence in Japan, Korea, China, Iran, and Saudi Arabia ranges from 13.5–20 cases per 100,000 persons, whereas it is lower in Western countries at 0.64 per 100,000 persons in the UK and 0.12–0.33 per 100,000 persons in the US (50, 51). In the present study, the prevalence of Behçet's disease was 1.4 per 100,000 populations. The SMR has been reported to vary from 1.13–10.12 (52, 53), which is similar to that found in the present study (3.7).
Our investigation was subject to the following potential limitations. One of the drawbacks of the study is the presence of coding and keying errors in the administrative database and death registry of Taiwan. Such errors, although inevitable, should be minimal since the quality of these 2 data sets is monitored routinely. Second, the dates of diagnoses may vary over several months, usually not more than half a year, compared to the dates of disease onset. However, this study includes a large, representative, and longitudinal sample. The quality of each ARD case has been ascertained because the board reviewed the samples before issuing the catastrophic illness certificates. Third, the comparison figures included in the discussion are mostly taken from studies of the adult population only. The prevalence of several conditions was lower than previously reported, but incidence was not so different. All cases of these diseases followed up in Taiwan were defined by ICD-9-CM coding, and some patients not regularly seen or followed up may have been missed. New cases are more likely to present clinically, which may be why incidence rates are more similar to other studies and prevalence rates are lower than previously reported. Moreover, a delay can exist between the diagnosis and the registration in the database of these different diseases. Cases of mild ARDs may escape detection more easily in developing countries, where access to medical care is limited, leading to the underestimation of the prevalence of ARDs. Furthermore, our estimates are based on more restrictive criteria and thus might underestimate the prevalence of RA and other ARDs. Further variation occurs as a result of differences in statistical methods and case-ascertainment criteria. When the results are compared with different studies, differences in case finding and case definition need to be considered. The strengths of the present study lie in the fact that it is population-based and that there is a vigorous system of case verification.
In conclusion, in 2000, the prevalence rates of major ARDs were 1.4–52.4 per 100,000 persons in Taiwan. Between 2000 and 2008, the incidence rates of various ARDs were 0.8–17.3 per 100,000 person-years. Women were at greater risk than men of acquiring an ARD. Except for vasculitis and Behçet's disease, women exhibited the highest incidence of ARDs, but male incident cases had higher mortality. Prevalence and incidence rates vary among the ARDs. Although there are commonalities, there are also important demographic differences between ARDs. Although considerable variation in the prevalence for most diseases has been seen (e.g., RA), the prevalence for most diseases for which data are from many geographic regions span overlapping ranges. Furthermore, in a setting of increasing treatment costs, managed care, and economic constraints, data about the prevalence and incidence of different ARDs might help health care providers to develop specific plans for the care for a given disease.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Yu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Yu, See, Kuo, I-Jun Chou.
Acquisition of data. Yu, See, Kuo, I-Jun Chou, Meng-Jiun Chou.
Analysis and interpretation of data. Yu, See, Kuo, Meng-Jiun Chou.