INTRODUCTION: Biliary atresia is a major cause of extrahepatic obstructive jaundice in neonates. Early Kasai operation is the gold standard of treatment. In this study, we evaluated the effectiveness of stool color card screening by using claims data from the National Health Insurance Research Database.
METHODS: This was a retrospective cohort study. Data from medical charts of all inpatients who were diagnosed with biliary atresia from 1996 to 2008 were collected from Taiwan's National Health Insurance Research Database. Patients who received a Kasai operation or liver transplant were identified by the Operation code. The patients' gender, age at admission, and type of operation were collected and analyzed.
RESULTS: From 1996 to 2008, the overall incidence of biliary atresia was 1.48 per 10 000 live births. The median age at first admission for patients with suspected biliary atresia decreased after the implementation of stool color card screening (47 vs 43 days). The proportion of very late referral decreased from 9.5% to 4.9%. The median age of Kasai operation advanced from 51 to 48 days. The proportions of Kasai operation within 60 days of age were 68.9% before and 73.6% after screening program.
CONCLUSION: Stool color card screening seemed to increase parents' and physicians' awareness of biliary atresia. It also was associated with a decline in the proportion of late referral. Thus, screening might be especially effective in areas with high a proportion of late referral. Improvements in the speed of workup and the operation room should be the focus of education and training in the future.
WHAT'S KNOWN ON THIS SUBJECT:
The Kasai operation is regarded as the gold standard of treatment for biliary atresia, and early operation is associated with a better outcome. Using stool color cards have been described as a cost-effective, simple, and convenient screening method.
WHAT THIS STUDY ADDS:
Stool color card screening was associated with a decline in the proportion of late referral. Improvements in the speed of workup and the operation should be the focus of education and training in the future.
Biliary atresia is a major cause of extrahepatic obstructive jaundice among neonates,1 but the etiology is still unclear. It is also the leading indication of liver transplantation in children.2,3 There is a higher incidence in east Asian countries.1 In Taiwan, the annual incidence is ∼1.46 cases per 10 000 newborn infants.4 The Kasai operation (hepatoportoenterostomy) is still regarded as the gold standard of treatment.5,–,9 Early operation was proved to be associated with a better jaundice-free rate and overall survival.5,–,7,9 Thus, the age of infants at the time of operation has traditionally been regarded as the key to their survival. Only early diagnosis of biliary atresia can lead to early operation. Several screening methods have been proposed for neonates.10 Among them, stool color card has been described as cost-effective, simple, and convenient.11,12 However, whether stool color card screening can lead to earlier diagnosis remains controversial.13,–,15
Nationwide stool color card screening for neonates was launched in Taiwan since 2004 (Supplemental Fig 4).16 The aim of this study was to analyze the effectiveness of the stool color card using claims data from Taiwan's National Health Insurance (NHI) system.
METHODS AND MATERIALS
This was a retrospective cohort study. Although the NHI of Taiwan was initiated in 1995, the database for that year was incomplete. So, we searched Taiwan's National Health Insurance Research Database (NHIRD) for patients' medical charts from 1996 to 2008. Files of inpatient expenditures by admissions were the main resource for data analysis. In total, 312 patients and 160 patients were identified before and after stool color card screen program, respectively.
The medical charts of all inpatients whose first, second, or third diagnosis was biliary atresia (International Classification of Diseases [ICD] 751.61) were retrieved. If the ICD Operation (OP) codes of the patients included anastomosis of the hepatic duct to the gastrointestinal tract (ICD 51.37), they were identified as having received a Kasai operation. If the ICD OP codes contained other transplant types of liver (ICD 50.59), they were identified as having received a liver transplant. The patients' gender, age at admission, and type of operation were collected. The population statistics were obtained from the Web site of the Ministry of the Interior (Taiwan).
In accordance with privacy rules, data obtained from the NHIRD that could be used to identify patients or care providers, including medical institutions and physicians, are scrambled before being sent to the National Health Research Institutes for use in a database, and is further scrambled before being released to individual researchers. The computer-processed personal data protection law of Taiwan was strictly followed.
SAS 9.0 (SAS Institute, Inc, Cary, NC) was used for data retrieval and data analysis. The 1-tail Wilcoxon signed-rank test was used for comparing patients' age at admission and operations performed before and after the implementation of stool color card screening. The χ2 test was applied for testing the proportion of delayed surgery before and after use of stool color card screening. The annual quartiles data were calculated by Tukey's hinges. A P level of <.05 was defined as statistically significant.
Epidemiology of Biliary Atresia in Taiwan
From January 1996 to December 2008, 1133 patients were admitted under the suspicion of biliary atresia. Among them, 472 patients underwent the Kasai operation or liver transplantation (Kasai operation for 440 patients and liver transplantation for the remaining 32 patients). These infants were regarded as patients with a confirmed diagnosis of biliary atresia. The female-to-male ratio was 1.06 (243 vs 229) (Table 1). The numbers of patients with confirmed diagnosis of biliary atresia varied from 17 to 50 cases each year. The average number was 36.3. The annual incidence was 0.74 to 1.99 per 10 000 live births. The overall incidence was 1.48 per 10 000 live births during the study period (Fig 1).
Effectiveness of Stool Color Card Screening
The median age at first admission for patients with suspected biliary atresia decreased after the implementation of stool color card screening from 47 days to 43 days (P = .028). The median ages for patients receiving a Kasai operation were 51 days and 48 days before and after the use of stool color card, respectively (P = .051) Exclusion of late referral patients (>90 days) from the data set resulted in median ages of 49 vs 47 days (P = .18), and exclusion of patients older than 75 days resulted in median ages of 48 vs 45 days for before and after use of stool color card, respectively (P = .10). However, the differences between the 2 groups did not show statistical significance. The median ages for both kinds of treatment (Kasai operation or liver transplantation) before and after use of the stool card were 52 vs 50 days, respectively (P = .27) (Table 2). The median age of patients at the first admission ranged from 39.5 to 53 days (Fig 2). The median age of patients who received a Kasai operation ranged from 46 to 58 days (Fig 3).
Timing of the Kasai Operation
Before implementation of stool color card screening, 296 patients underwent a Kasai operation. After the screening program commenced, 144 case-patients received a Kasai operation. Among them, the proportions of patients receiving a Kasai operation who were 60 days of age or younger were 68.9% before and 73.6% after use of screening (P = .31) (Table 3). The proportion of patients who were 91 days of age or older when they underwent a Kasai operation decreased from 9.5% to 4.9% after implementation of screening (P = .094). However, the differences between the 2 groups did not reach statistical significance in either of the 2 abovementioned situations. Because of the limited number of patients, the statistical power was only 0.43, so a type 2 error may exist (Supplemental Fig 5). A statistically significant result may be obtained if a larger sample size could be used.
Only 32 patients in this study received liver transplant without previous Kasai operation. The age of such patients ranged from 182 to 693 days. The median age was 317 days. The annual distribution is shown in Table 3.
In this study we evaluated the effects of the nationwide application of stool color card screening using claims data from Taiwan's NHIRD. We found a 46% decrease in very late referrals, patients undergoing a Kasai operation after the age of 90 days (9.5% before to 4.9% after the stool card screening program). Although this was not statistically significant, we believed it was limited by the case numbers. We also found the median ages of neonates at first admission for suspected biliary atresia decreased after commencement of the stool color card screening program (47 vs 43 days, before and after screening, respectively; P = .028). The median ages of patients receiving the Kasai operation advanced a little after the stool color card era, although it only reached borderline statistical significance (51 vs 48 days, before and after screening, respectively; P = .051). After elimination of late referral patients from the data set (>90 or >75 days), the results were similar. In addition, we observed that there was a slight increase in the proportion of Kasai operations performed in patients 60 days of age or younger (from 68.9% before to 73.6% after stool color card screening, respectively).
In our study, the overall incidence of biliary atresia was 1.48 per 10 000 live births in the 13-year study period (1996–2008), ranging from 0.74 to 1.99 per 10 000 live births. The incidence rate was the highest when compared with rates reported in Japan (1.04 per 10 000 live births),5 the United States (0.73 per 10 000 live births),17 Canada (0.53 per 10 000 live birth),18 France (0.51 per 10 000 live births),19 and Sweden (0.70 per 10 000 live births).20 We also observed that there was a higher incidence during 2003–2004 (1.85–1.99 per 10 000 live births); the highest incidence occurred in 2004 (1.99 per 10 000 live births). The incidence has decreased gradually in recent years. The case definition did not change before and after stool color card screening. The decline in incidence is consistent with the unpublished registry data of the Taiwan Children Liver Foundation. The cause of the decline has yet to be clearly established. In addition, we found an overall decrease in the total number of admissions for suspected biliary atresia. It might be the benefit of the screening program contributing for overall health care cost by selecting out the jaundiced infants with normal color stool from being admitted for suspected biliary atresia investigation.
In comparison with the registry data from the Taiwan Children Liver Foundation reported by Hsiao et al,16 our study had 3 more patients in 2004 and 1 more patient in 2005. The NHIRD is the largest and most comprehensive medical database in Taiwan, so this was used as the source of data in the present study. The 4 abovementioned patients might not have signed up on the Taiwan Children Liver Foundation registry. However, data from these 2 sources were still quite similar. This phenomenon further confirms the validity of case definition in this study.
Only 32 patients in this study received liver transplant without a previous Kasai operation. The age of these patients ranged from 182 to 693 days. The median age was 317 days. We did not include them in the late referral group because we believed patients who underwent primary liver transplant might be because of liver decompensation at the moment of diagnosis, presence of comorbidities, and families' attitude toward this disease. We used 3 different approaches including first time of admission, time of Kasai operation, and time of any intervention in Table 2 to evaluate the effectiveness of stool color card screening. In Table 3, if we only include the patients who had a Kasai operation, there was a reduction in the number of very late referral cases (≥90 days) after initiation of the screening program. Although these results did not reach statistical significance, they may have been subject to type 2 error because of the limited case sample size. At least 3 times more patients were needed if we wanted 80% power to detect such differences (Supplemental Fig 4).
In a previous study in the United States it was reported that the mean age of patients at admission was 53 (±28) days of age, and the mean age at Kasai operation was 61 (±26) days.21 In our study it was shown that infants in Taiwan were younger at diagnosis and during management of biliary atresia than those in the abovementioned study, regardless of the implementation of stool card screening. We think the reason may be that a routine visit to health care providers is recommended for infants at 1 month of age, which is the typical timing of acholic stool presentation for biliary atresia.22 Because pediatricians are well-trained in Taiwan, stool color screening might only contribute slightly to the diagnosis of biliary atresia.
Many screening methods have been adopted for the purpose of early diagnosis of biliary atresia.10 Stool color card screening was first described in Japan in 1987.13 However, the proportion of patients who received an operation before 60 days old did not significantly differ before and after this intervention (39.8% vs 34.0%). Mass screening has been implemented in Tochigi Prefecture, Japan; the authors claimed that early diagnosis and an increase in the rate of Kasai operations performed before 60 days were achieved.14 However, there were only 5 infants in the control group. In Taiwan, the stool color card for biliary atresia screening has been performed since 2004.16 In a pilot study, sensitivity, specificity, and positive predictive values of 89.7%, 99.9%, and 28.6%, respectively, were reported for detection of biliary atresia among patients younger than 60 days of age.12,15 In a 2-year observational study using the registry data in Taiwan, the authors reported the rate of Kasai operations performed in patients younger than 60 days of age increased from 72.5% in 2004 to 97.1% in 2005. Meanwhile, no patient was operated on after reaching 91 days of age.16 In our study, we found a transient change in 2005. However, the effect did not persist as expected.
Lien et al23 reported a significant improvement in the proportion of operation rate younger than age of 60 days, 65.7% vs 49.4%, before and after implementation of the stool color card screening program, on the basis of registry data in Taiwan. The authors found that the increased proportion of patients who underwent surgery before the age of 60 days contributed to a better 3 months-jaundice-free rate, as well as improved 3-year and 5-year native liver survival rates, jaundice-free-native-liver-survival rate and overall survival rate. In our study, the improvement in the proportion of patients who underwent the Kasai operation younger than 60 days old was not significant, 68.9% vs 73.6%. The reasons for the difference between these 2 studies might be the choice of the control group. Lien et al23 selected the control group from a single tertiary medical center in the period from 1990 to 2000. The control group of this study comprised data obtained from the NHIRD from 1996 to 2002. We postulate that the control group selected in the present study was better matched with patients in the study groups in regard to their baseline characteristics. Improvements in surgical technique, post operative care, and medical support in Taiwan might also have contributed to the increased survival rates.
Early Kasai operation is generally accepted as the first choice of management for biliary atresia.5,–,9 The 10-year survival rates for the Kasai operation were 57%, 41%, 30%, and 13% for those who underwent surgery at <60, 61–70, 71–90, and >90 days, respectively.6 Higher postoperative jaundice-free rates and better native liver survival rates were also associated with early operation.5,7,9,15,16,22 In a Canadian cohort study, the native liver survival rate for those who received the Kasai operation before 30 days of age was better through 10 years of observation.18 In a French cohort study, higher 5-year native liver survival rate also was reported if the operation was performed within 30 days. A 12.1% difference (22.8% vs 34.9%) in 15-year native liver survival rate was also reported if the Kasai operation was performed before the age of 45 days.7 Thus, earlier operation before the age of 45 or even 30 days has been advocated.22 Trying to advance the age of operation was also a part of the initial goals in this screening program. Therefore, we chose the median age of operation as 1 of the primary end points. We also found that the stool color card screening program did bring forward the date of admission for biliary atresia investigation, although the median age of surgical intervention was not significantly affected. The average time for workup to confirm the diagnosis of biliary atresia has been reported to be 8 days.15 We think the reason operation time was not significantly brought forward may be that the course of diagnosis is prolonged in Taiwan. A 3-day protocol, as noted by Chang et al,24 should be considered in the future for improvement of outcome in these patients.
In addition, when we compared the rate of liver transplantation before and after the use of stool color card screening, a gradual increase was found (5.1% vs 10.0%, not shown in Table 3). With the improvement in surgical techniques for liver transplantation, the introduction of new immunosuppressive agents, and the availability of living related donors for liver transplantation, some studies advocated primary liver transplantation without a Kasai operation for treatment of biliary atresia.25,26 Greater intraoperative blood loss and longer operative time were identified as the most serious factors that affect the surgical results of liver transplantation among recipients who had undergone the Kasai operation previously.27 In our study, we could not identify the reasons why patients had undergone liver transplantation for the primary management of biliary atresia. However, excluding the possibility of living with a native liver, prolonged use of immunosuppressive agents and the shortage of donor organs are problems that still have to be faced when considering living transplantation as the primary management of biliary atresia.26 Reducing the age at which patients receive the Kasai operation is still the leading goal in the treatment.
In this study we used the research database of Taiwan's NHI program, which includes almost the entire population of children in Taiwan.28 We believe it offers accurate data for evaluation of the incidence and epidemiologic characteristics of a disease in 1 country in a defined period of time. However, there were still some limitations in this study. First, this is a claims data study, so we did not have data at the individual level. Second, we were not able to determine the reasons for delayed diagnosis in infants. Furthermore, we could not analyze the proportion of infants with biliary atresia who were diagnosed by stool card screening. Also, it was not possible to distinguish definite from suspected biliary atresia using the coding system. We used the operation code for Kasai operation or liver transplantation to confirm the diagnosis of biliary atresia. Therefore, we believe the validity of this study is acceptable because the incidence was similar to that of previous reports using national registry data.4,16
The stool color card screening seemed to increase parents' and physicians' awareness of biliary atresia. It also demonstrated a decline in the proportion of very late referral patients (>90 days of age). Thus, it may be useful in areas with a high proportion of late referrals. Improvements in the speed of workup and speed to the operating room should be the focus of education and training in the future.
This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance and the Department of Health and managed by the National Health Research Institutes of Taiwan (registration No. 98138).
- Accepted August 14, 2011.
- Address correspondence to Ming-Chih Lin, MD, Taichung Veterans General Hospital, No. 160, Sec. 3, Taichung-Kang Road, Taichung City, Taiwan 407. E-mail:
The interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, the Department of Health, or the National Health Research Institutes.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- NHI —
- National Health Insurance
- NHIRD —
- National Health Insurance Research Database
- ICD —
- International Classification of Diseases
- OP —
- Serinet MO,
- Wildhaber BE,
- Broue P,
- et al
- Wadhwani SI,
- Turmelle YP,
- Nagy R,
- Lowell J,
- Dillon P,
- Shepherd RW
- Chen SM,
- Chang MH,
- Du JC,
- et al
- Yoon PW,
- Bresee JS,
- Olney RS,
- James LM,
- Khoury MJ
- Sokol RJ
- Copyright © 2011 by the American Academy of Pediatrics