Time Trend and Geographic Distribution of Treated Patients with

Congenital Hypothyroidism Relative to the Number of Available

Endocrinologists in Japan
Yan-Hong Gu, MD, PhD, Tadaaki Kato, MD, PhD, Shohei Harada, MD, PhD, Hiroaki Inomata, MD, and Kikumaro Aoki, MD
Objective To investigate the time trend and geographic distribution of treated patients with congenital hypothyroidism (CH) and explore their possible relationship to the availability of endocrinologists in Japan.

Study design The 2-source capture-recapture method was used to estimate the total number of patients. The
ratio of the total estimated number of patients with CH to the number of endocrinologists and Spearman correlation
coefficients were calculated. Curve fitting for changes in incidence or prevalence was estimated.
Results The incidence and prevalence of CH exhibited upward trends, with linear slopes of increase during the
period 1994-2002. A statistically significant positive correlation was observed between the prevalence of CH and
the ratio of the number of patients with CH to the number of endocrinologists in the 10 regions studied. The prevalence of CH was significantly higher in the regions with a higher ratio of patients with CH to endocrinologists, and
also in younger patients.
Conclusions A shortage of endocrinologists may be one reason for the upward trend in the incidence and prevalence of treated patients with CH. (J Pediatr 2010;157:153-7).

nationwide neonatal screening program for congenital hypothyroidism (CH) was implemented in Japan in 1979. Since
1984, nearly 100% of all neonates in Japan have been screened.1 In 1989, a highly sensitive method for detecting CH was
introduced, and by 1992, enzyme-linked immunosorbent assay had completely replaced older methods such as radioimmunoassay and enzyme immunoassay. Nationwide follow-up surveys to evaluate the status of neonatal screening were conducted every year during the periods 1979-1987 and 1994-2003.1,2 The 1994-2003 follow-up survey was conducted by the Aiiku
Maternal and Child Health Center, Tokyo (Aiiku), and had a response rate of 54.2%-91.5%.1,2
In Japan, treatment of patients with CH is supported by 3 categories of medical subsidies provided by the national and/or
local government: medical expenses for infants, the Medical Aid Program for Chronic Pediatric Diseases of Specified Categories
(MAPChD), and Medical Aid for Immature and Premature infants.3 Only 1 of these 3 subsidies may be used. MAPChD was
implemented in 1968. The number of diseases eligible under this program has expanded annually, and currently stands at 514. A
national registry of MAPChD beneficiaries was started in 1998. In 2005, the MAPChD was legalized and was included under the
Child Welfare Law in Japan.3 Beneficiaries of the MAPChD should be registered every year, and beneficiaries who sign a document of informed consent may be enrolled as subjects for MAPChD study projects. The registered data used for the study
projects do not include any personal data, such as names, home addresses or phone numbers. Thus, the list of patients with
CH is incomplete in both the follow-up survey at the Aiiku center and the registered data in the MAPChD.
The 2-source capture-recapture method (CRM) is widely used to estimate the total number of patients with CH. With CRM,
an estimated whole number of patients can be calculated using information from overlapping lists of patients from 2 distinct,
independent, incomplete sources. This statistical method, originally developed to estimate the size of a closed animal population, is used increasingly in epidemiologic studies to assess the completeness of cancer registries, to evaluate the incidence of
many diseases and health problems (including diabetes), and in governmental disease surveillance.4-9
Obtaining an accurate estimate of the number of treated patients with CH in
Japan is important. The present study was performed to estimate the total number of patients treated for CH and further explore the time trend and geographic
From the Department of Health Policy, National
distribution of CH by CRM, using lists from the Aiiku follow-up survey and the
Research Institute for Child Health and Development,
Tokyo, Japan (Y.-H.G., T.K., S.H.); Inomata Childrens
registered data in the MAPChD.
Clinic, Chiba, Japan (H.I.); and Department of Research
and Development, Aiiku Maternal and Child Health
Center, Tokyo, Japan (K.A.)

Aiiku
CH
CI
CRM
MAPChD

Aiiku Maternal and Child Health Center, Tokyo

Congenital hypothyroidism
Confidence interval
Capture-recapture method
Medical Aid Program for Chronic Pediatric Diseases of Specified Categories

Supported by a Health Labor Sciences Research Grant

(Research on Child and Families) from the Japanese
Ministry of Health, Labor and Welfare (A Study of Information Offer, Assessment, Management, and Registration on Medical Aid Program for Chronic Pediatric
Diseases of Specified Categories After Legislation and
A Study on the Construction of Comprehensive Database About Chronic Diseases in Children). The authors
declare no conflict of interest.
0022-3476/$ - see front matter. Copyright 2010 Mosby Inc.
All rights reserved. 10.1016/j.jpeds.2009.12.051

153

154

230, 30.0%
237, 32.9%
237, 25.7%
273, 32.6%
354, 20.9%
348, 24.1%
373, 17.2%
438, 11.6%
357, 14.8%
2847, 21.9%

179
203
150
211
193
195
171
124
153
1579

179, 38.5%
203, 38.4%
150, 40.7%
211, 42.2%
193, 38.3%
195, 43.1%
171, 37.4%
124, 41.1%
153, 34.6%
1579, 39.5%

69
78
61
89
74
84
64
51
53
623

597 (504-689)
617 (529-705)
583 (486-680)
647 (563-731)
923 (776-1070)
808 (694-921)
997 (821-1172)
1065 (854-1276)
1031 (824-1238)
7216 (6826-7605)

41.9
40.5
43.7
45.9
40.6
44.9
39.6
41.1
36.6
41.9

1 233 072
1 187 067
1 206 551
1 194 510
1 199 183
1 184 302
1 190 547
1 170 662
1 153 855
10 719 749

4.8 (4.5-5.2)
5.2 (4.8-5.6)
4.8 (4.4-5.2)
5.4 (5.0-5.8)
7.7 (7.2-8.2)
6.8 (6.4-7.3)
8.4 (7.9-8.9)
9.1 (8.6-9.6)
8.9 (8.4-9.5)
6.8 (6.6-6.9)

1 209 000
1 205 000
1 186 000
1 192 000
1 198 000
1 170 000
1 168 000
1 171 000
1 168 000
10 667 000

4.9 (4.5-5.3)
5.1 (4.7-5.5)
4.9 (4.5-5.3)
5.4 (5.0-5.8)
7.7 (7.2-8.2)
6.9 (6.4-7.4)
8.5 (8.0-9.1)
9.1 (8.5-9.6)
8.8 (8.3-9.4)
6.8 (6.6-6.9)

8
7
6
5
4
3
2
1
0

www.jpeds.com

250
250
255
297
375
363
395
461
377
3023

Total
number
Birth year
Total
(April-March) number

1994
1995
1996
1997
1998
1999
2000
2001
2002
Total

Population (P)
(JanuaryDecember)

Prevalence of CH
in each age group
per 10 000
children, N/P
Age,
(95% CI)
years
Estimated total
number of
Registration rate
Incidence of CH
Eligible number
in infants
(N2), comparable Patients with patients with of patients with CH
CH from both CH calculated by with the MAPChD Live births per 10 000 live
rate in CRM
(N1/N, %)
(April- March) births (95% CI)
(N12/N2  100%) sources (N12) CRM, N (95% CI)
Eligible number
(N1), comparable
rate in CRM
(N12/N1  100%)

By age group
as of October 1, 2002
CH registered
with the MAPChD (source 1) CH patients in Aiiku (source 2)

Table I. Estimated total number of patients with CH, rate of registration with the MAPChD, the annual incidence from April 1994 to March 2003, and the prevalence
of CH by age group as of October 1, 2002

THE JOURNAL OF PEDIATRICS

Vol. 157, No. 1

Methods
Data Sources
Data source 1 included data from the electronic registry of
MAPChD beneficiaries born between April 1994 and March
2003. This database comprised patients with both primary
and central CH, either detected by neonatal screening or
diagnosed later after false-negative results on neonatal
screening.
Data source 2 was lists from the Aiiku nationwide followup survey for the neonatal screening. This database comprised 1579 patients with primary CH born between April
1994 and March 2003 who were started on treatment with
L-thyroxine. They were first identified as having CH by neonatal screening and then referred for further examination to
a hospital or clinic in any of the various study regions, except
the Hokkaido region1,2 (Tables I and II).
Data Matching
In accordance with the Japan Personal Information Protection Act,10 personal information about patients, such as
names, was not available from the MAPChD electronic registry. The identifying variables were date of birth, sex, location
of hospital or clinic, date of diagnosis, prefecture of birth, and
prefecture of current residence. A matching software for
2-source CRM developed by our department and Mitsubishi
Electric Business Systems Co. Ltd., (Tokyo, Japan) was used.
Two-Source CRM Analysis
A 2-source CRM model (Lincoln method) was used to estimate the total number and 95% confidence interval (CI) of
patients with primary and central CH.8,9 There are 3 fundamental assumptions in 2-source CRM: a closed population,
independent sources, and an equal probability of individual
cases being captured within any source. The formulas are
as follows:
Total number = (eligible number of patients in source 1
[N1])  (eligible number of patients in source 2 [N2])/ number of patients in sources 1 and 2 [N12]); 95% CI of estimated
total number of patients = N  1.96 {Ovar (N)}, var (N) =
{(N1 - N12)  (N2 - N12)  N1  N2} / {N12  N12 
N12}.
The number of patients captured by both sources was determined by data matching. The number of patients missed
by both sources was estimated based on the assumption of
an equal probability of ascertainment from both sources.
Stratified analysis was performed by birth year, age as of
October 2002, and location of hospital or clinic, excluding
the Hokkaido region (Tables I-III).
Other Statistical Analyses
The number of nationwide live births and the mortality of
children were obtained from Japans Ministry of Health,
Labor, and Welfare.11 The nationwide number of children
under age 9 years in each prefecture as of October 2002
was obtained from the Ministry of Internal Affairs and
Gu et al

ORIGINAL ARTICLES

14.9
4.0
7.0 (5.7-8.3)
5.4 (5.3-5.6)
119 (106-133)
5848 (5452-6012)
40
671

Comparable Data and Probability of Ascertainment

of a Case From Source 1 and Source 2
The comparable rates were 21.9% (range, 11.6%-32.9%)
from source 1 (N12/N1) and 39.5% (range, 34.6%-43.1%)
from source 2 (N12/N2) (Table I). The comparable rates
were 27.5% (range, 19.2%-44.4%) from source 1 and
42.5% (range, 26.9%-75.5%) from source 2 (Table II). The
probabilities of ascertainment of a case shown in Table I
were 52.7% from both sources {(N1 + N2 - N12)/N},
39.5% from source 1 (N1/N), and 21.9% from source 2
(N2/N). The probabilities of ascertainment of a case shown
in Table II were 57.2% from both sources, 41.7% from
source 1, and 27.0% from source 2.

*No data available.

90, 44.4%
2436, 27.5%

53, 75.5%
1579, 42.5%

324 (232-415)
1188 (1061-1315)
25
137
87, 28.7%
539, 25.4%

93, 26.9%
302, 45.4%

1044 (893-1195)
490 (391-589)
89
42
392, 22.7%
219, 19.2%

237, 37.6%
94, 44.7%

158 (129-187)
440 (362-518)
33
50
62, 53.2%
110, 45.5%
84, 39.3%
200, 25.0%

102
140
13
202, 50.5%
397, 35.3%
29, 44.8%

Tohoku region
Kanto region
Shinetsu region (not including
Niigata Prefecture*)
Hokuriku region
Tokai region (not including
Gifu Prefecture*)
Kinki region
Chugoku region (not including
Tottori Prefecture*)
Shikoku
Kyushu (not including
Saga Prefecture*)
Okinawa
Total

250, 40.8%
508, 27.6%
67, 19.4%

The number of
patients with CH
in both sources
(N12)
Location of
hospital or clinic
that the CH
patients visited

Regions of Japan
Japan is divided into 47 prefectures (todofuken) or, alternatively, into 11 regions.

75.6
41.7

170 000
10 794 000

8
1467

5.2
8.1
8.8 (7.8-9.7)
10.0 (9.4-10.6)
26.9
45.4

370 000
1 190 000

62
147

3.2
7.8
5.3 (5.0-5.6)
7.4 (6.8-8.1)
37.5
44.7

1 981 000
661 000

325
63

3.3
2.6
5.3 (4.5-6.2)
3.5 (3.1-3.8)
53.2
45.5

296 000
1 271 000

48
172

6.9
2.6
6.0
50.5
35.3
45.0
495 (443-547)
1441 (1277-1604)
149 (95-204)

5.5 (4.6-5.6)
3.9 (3.7-4.1)
6.9 (5.8-8.0)

Registration rate
of patients with
CH with the MAPChD
(N1/N, %)

900 000
3 740 000
215 000

72
545
25

Ratio of patients
with CH to
endocrinologists
Number of
endocrinologists
(pediatrics and
internal medicine)
as on February 28,
2007
Prevalence of
patients with CH in
each region, per
10 000 children
(N/P) (95% CI)
Population of
children under
9 years of age in
each region as on
October 1,
2002 (P)
Estimated total
number of patients
with CH calculated
by CRM,
N (95% CI)

Communications.12 The number of society-certified endocrinologists in each region as of February 2008 was obtained
from the Japan Endocrine Society.13 In Japan, endocrinologists usually care for both pediatric and adult patients in
clinics, especially in rural areas.
Spearman correlation coefficients (2-tailed) were calculated to determine the association between the ratio of the
number of patients with CH to the number of endocrinologists and the prevalence of CH in each region. The prevalence
of CH was calculated assuming that all of the neonates with
CH were alive throughout the study period, because (1) the
mortality rate of neonates less than 28 days old, as well as
that of infants under 1 year old, is extremely low in Japan
(1.3 and 2.6 per 1000 live births, respectively, in 2006);11
(2) the mortality rate of infants with CH was very low
(1.34%) during the period 1985-1987;14 and (3) the mortality
rate of children under age 19 years also was very low (3.3-4.4
per 10 000) during the period 1994-2002.11
Curve fitting for the changes of the incidence or prevalence
was estimated using SSPS version 15.0 (SSPS Inc., Chicago,
Illinois).

Results

Eligible number of
patients with CH
in Aiiku (source 2,
N2), comparable
rate (N12/N 2 
100%)
Eligible number of
patients with CH
with the MAPChD
(source 1, N1),
comparable rate
(N12/N1  100%)

Table II. Estimated total number of patients with CH, rate of registration with the MAPChD, prevalence of CH and the ratio of the number of patients with CH to
the number of endocrinologists in each region of Japan

July 2010

Time Trend and Geographic Distribution of CH

Table I shows the estimated total number of patients with
CH, the annual incidence (and 95% CI) for 1994-2002, and
the prevalence of CH by age group as of October 1, 2002.
An upward trend of incidence and a higher prevalence of
younger patients with CH are evident. The slope of the
increase in incidence (R = 0.939; adjusted R2 = 0.865;
P <.001) and prevalence (R = 0.936; adjusted R2 = 0.877;
P <.001) during the study period was linear. Table II
shows the estimated total number of patients with CH, the
prevalence (with 95% CI) of patients with CH, and the
ratio of the number of patients with CH to the number of
endocrinologists in the 10 regions studied. A regional
variation in the prevalence of CH is significant (P <.05).
Moreover, Table III shows an upward trend of the

Time Trend and Geographic Distribution of Treated Patients with Congenital Hypothyroidism Relative to the Number
of Available Endocrinologists in Japan

155

THE JOURNAL OF PEDIATRICS

www.jpeds.com

Vol. 157, No. 1

Table III. Prevalence of patients with CH in regions where the ratio of patients with CH to endocrinologists was above or
below the nationwide level (4.0)
Prevalence of patients with CH (95% CI)  10 000

Age group as of October 1, 2002 (birth year)

Group 1: Kanto, Tokai, and Kinki regions

(not including Gifu Prefecture)*

Group 2: Tohoku, Shinetsu, Hokuriku, Chougoku,

Shikoku, Kyushu, and Okinawa regions (not including
Niigata, Tottori, and Saga Prefectures)

3.3 (3.0-3.6)
4.0 (3.7-4.3)
4.5 (4.1-4.8)
5.5 (5.1-5.9)
4.3 (4.1-4.5)

7.3 (6.7-7.9)
7.6 (7.0-8.3)
10.2 (9.5-10.9)
13.6 (12.7-14.4)
9.6 (9.3-10.0)

7-8 (April 1994 to March 1996)

5-6 (April 1996 to March 1998)
3-4 (April 1998 to March 2000)
1-2 (April 2000 to March 2002)
Total
*CH patient-to-endocrinologist ratio below the nationwide value of 4.0.
CH patient-to-endocrinologist ratio above the nationwide value of 4.0.

prevalence of CH in both groups 1 and 2 (P <.05), with

regional variation between the 2 groups (P <.001).
The rate of registration with the MAPChD for each year
from 1994 to 2002 ranged from 36.6% to 45.9% (Table I).
The rate of MAPChD registration varied among regions,
ranging from 26.9% to 75.6% (Table II).
Ratio of the Number of Patients with CH to the
Number of Endocrinologists in Each Region
Table II shows the ratio of the number of patients with CH to
the number of endocrinologists in each region. Obvious
differences in the ratio among the regions can be seen.
Time Trend and Geographic Distribution of CH
Cases Relative to the Number of Endocrinologists
A highly statistically significant positive correlation was
observed between the patient with CH-to-endocrinologist
ratio and the prevalence of CH in each region (Table II).
The Spearman correlation coefficient was 0.81, and the
corresponding P value for the correlation was .004.
The nationwide patient with CH-to-endocrinologist ratio
was 4.0 (Table II). Thus, we divided the 10 regions of Japan,
excluding the Hokkaido region, into 2 groups: a group in
which the ratio was <4.0 and a group in which the ratio
was >4.0. Because of the small number of patients with CH
in some regions within a single year, we investigated these
2 groups in 2-year time frames (Table III). There was
a significant difference in the prevalence of CH between the
2 groups during the study period (P <.05).

Discussion
The MAPChD database includes treated patients with both
primary and central CH, and the data in the Aiiku survey
comprise treated patients with primary CH only. Thus, the
comparable rate from source 1 (MAPChD) was lower than
that from source 2 (Aiiku). Notifications in both sources
may originate from the same doctor. These 2 conditions
could result in underestimation; however, as pointed out
by Brenner,9 this underestimation is much less severe than
that associated with the traditional registration method.
156

Moreover, according to Tables I and II, the probability of

ascertainment of a case from both sources and from source
1 was approximately $40%, suggesting that our results are
credible.
In Japan, neonatal screening for CH was instituted in 1979.
At present, nearly 100% of Japanese neonates are screened;1
nonetheless, the total number of patients treated for CH in
Japan remains unclear. Using 2-source CRM, the total number of treated patients with CH nationwide during the study
period could be calculated, and the incidence and prevalence
of CH could be estimated. Thus, 2-source CRM is an effective, low-cost tool for accurately estimating the number of
patients with CH.4-9 According to Ministry of Health, Labor,
and Welfare data for 1994-2002, the fiscal year positive rates
of CH (number of suspected cases/total number of tested
neonates  100%, including primary CH and a small percentage of central CH) were 3.0, 3.6, 3.8, 3.8, 4.3, 4.3, 4.3,
4.5, and 4.8 per 10 000, respectively. This positive rate of
CH demonstrates the validity of our results, especially the
incidence data shown in Table I. The MAPChD
beneficiaries were registered every year, allowing for an
accurate estimate of the current number of beneficiaries.
Thus, we also used the estimated number of patients with
CH to calculate the prevalence of CH in each age group for
the year 2002 (Table I).
With regard to the annual incidence of CH, an upward
trend in incidence was obvious (Tables I and III). The
prevalence of CH calculated by age showed a higher
prevalence of CH in younger patients (Tables I and III). A
guideline for neonatal CH screening in Japan has
recommended examinations for differential disease-type
diagnoses, such as scintigraphy and radioactive iodine
uptake test, after age 3 years in children with CH;15 such
examinations are usually performed at age 5-6 years in
Japan. Thus, considering the timing of the clinical
classification of the diagnosis, patients with treated infantile
transient hyperthyrotropinemia also may be counted,
resulting in a higher incidence in recent birth years or
a higher prevalence in younger patients with treated CH in
Japan, despite the lower patient with CH-to-endocrinologist
ratio (Table III). Consequently, correctly diagnosing CH is
not an easy task in infants, requiring a high level of
Gu et al

July 2010
professional skill. Moreover, in a previous study, we found
that an upward trend in the incidence of CH was
significantly associated with environmental temperature,2
suggesting that environmental factors also may be
responsible for this upward trend in the incidence of CH. It
is clear that prenatal exposure to high levels of iodine or
other substrates can lead to persistent or transient CH in
neonates. Unfortunately, no maternal data are available to
allow evaluation of this possibility in the present study. A
previous study from the United States found similar results,
that disease diagnosis, environmental conditions, and
prenatal exposure might contribute to the estimated
incidence of CH.16
In our previous survey based on a list of positive cases of
CH from Japanese screening centers, only 81.7% of CH cases
detected by neonatal screening were diagnosed as CH by clinical examination in a clinic or a hospital.1 The prevalence of
CH varied among the 10 regions of Japan and showed a positive correlation with the patient with CH-to-endocrinologist
ratio in the regions, suggesting that the uneven nationwide
geographic distribution of endocrinologists may be one reason for the regional variations in the distribution of patients
with CH (Table II). The prevalence of CH was significantly
higher in the group with a higher patient with CH-toendocrinologist ratio in each region, suggesting that higher
prevalence of CH in regions with fewer endocrinologists
may reflect a higher rate of incorrect diagnosis (Table III).
Some of the nonspecialists may have, for example, accepted
the positive screening test as a definitive diagnostic study
instead of viewing the screening result as an indication for
a repeat assessment of thyroid function. Thus, the
endocrinologists may play an important role in the
diagnosis, treatment, consultation, and management of
patients with CH, especially younger patients.
The use of 2-source CRM also helped clarify the rate of
MAPChD registration by birth year or region. The variable
rate of registration with the MAPChD in each region suggests
variation in the medical aid support provided by local governments. In some regions (eg, Okinawa), patients with
CH preferred to register with the MAPChD than with other
medical aid support systems (Table II).
Although neonatal screening for CH in Japan has been
performed for 30 years and is well supported by medical
aid systems, there are some serious pediatric medical care
problems in Japan, including a shortage of specialists in
endocrinology for the treatment and follow-up of patients
with CH. The shortage of pediatricians and overworking
of pediatricians in Japan are reaching crisis levels. The
pediatrician-to-patient ratio in Japan is reportedly about
half that in the United States.17 Such a situation demands
urgent attention. One solution might be to train young
doctors for the profession, especially in rural areas.
Basically, an increased number of pediatricians and other
specialists is needed in some regions of Japan. This study

ORIGINAL ARTICLES
provides evidence supporting the need for improvement
of local medical care. n
Submitted for publication May 25, 2009; last revision received Nov 23, 2009;
accepted Dec 29, 2009.
Reprint requests: Yan-Hong Gu, Department of Health Policy, National
Research Institute for Child Health and Development, 2-10-1 Okura, Setagayaku, Tokyo, 157-8535, Japan. E-mail: gyh@nch.go.jp.

References
1. Inomata H, Aoki K. Nationwide follow-up results of neonatal mass
screening for congenital hypothyroidism in 1994-1999. J Jpn Soc Mass
Screen 2003;13:27-32.
2. Gu YH, Kato T, Harada S, Inomata H, Saito T, Aoki K. Seasonality in the
incidence of congenital hypothyroidism in Japan: gender-specific patterns and correlation with temperature. Thyroid 2007;17:869-74.
3. Gu YH, Kato T, Harada S, Sato Y, Kakee N. Medical aid program for
chronic pediatric diseases of specified categories in Japan: current status
and future prospects. Pediatr Int 2008;50:376-87.
4. Gu YH, Kodama H, Ozawa H, Harada S, Kato T. Re-estimation of incidence of Menkes disease in Japan using the two-source capture-recapture method. Jpn J Inhert Metab Dis 2006;22:84-7.
5. Whitfield K, Kelly H. Using the two-source capture-recapture method to
estimate the incidence of acute flaccid paralysis in Victoria, Australia.
Bull World Health Organ 2002;80:846-51.
6. Blanchard JF, Dean H, Anderson K, Wajda A, Ludwig S, Depew N.
Incidence and prevalence of diabetes in children aged 0-14 years in Manitoba, Canada, 1985-1993. Diabetes Care 1997;20:512-5.
7. Kozinetz CA, Skender ML, MacNaughton N, Almes MJ, Schultz RJ,
Percy AK, et al. Epidemiology of Rett syndrome: a population-based
registry. Pediatrics 1993;91:445-50.
8. Chao A, Huggins R. Classical closed-population capture-recapture
models. In: Amstrup SC, McDonald TL, Manly BFJ, editors. Handbook
of Capture-Recapture Analysis. Princeton, NJ: Princeton University
Press; 2005. p. 22-35.
9. Brenner H. Use and limitations of the capture-recapture method in
disease monitoring with two dependent sources. Epidemiology 1995;6:
42-8.
10. Japan Personal Information Protection Act. Available from: http://www.
kantei.go.jp/jp/it/privacy/houseika/hourituan/. Accessed February 3, 2009.
11. e-Stat Portal Site of Official Statistics of Japan. Vital Statistics of Japan.
http//www.e-stat.go.jp/SG1/estat/GL02010101.do. Accessed February
3, 2009.
12. 2005 Population Survey, Japan. Ministry of Internal Affairs and Communications of Japan. Available from: http://www.stat.go.jp/data/
kokusei/2005/index.html. Accessed February 3, 2009.
13. List of endocrinologists of the Japan Endocrine Society. Japan Endocrine
Society. Available from: http://square.umin.ac.jp/endocrine/senmon_i/
index.html. Accessed February 3, 2009.
14. Sato H, Nakajima H. The 7th and 8th national studies of patients with
congenital hypothyroidism and its related diseases found on neonatal
screening. J Jpn Pediatr Soc 1989;93:1152-8.
15. Working Group on Congenital Hypothyroidism of the Japanese Society
for Pediatric Endocrinology and the Japanese Society for Mass Screening. Guideline for neonatal mass screening for congenital hypothyroidism. Clin Pediatr Endocrinol 1999;8:51-5.
16. Harris KB, Pass KA. Increase in congenital hypothyroidism in New York
State and in the United States. Mol Genet Metab 2007;91:268-77.
17. A comparison of the numbers of medical doctor between Japan and USA.
Ministry of Health, Labor, and Welfare. Available from: http://www.
mhlw.go.jp/shingi/2005/03/s0311-5a4.html. Accessed February 3, 2009.

Time Trend and Geographic Distribution of Treated Patients with Congenital Hypothyroidism Relative to the Number
of Available Endocrinologists in Japan

157