Using health care service administrative
data to improve national vital statistics:
Thailand experiences
United Nations Expert Group Meeting on International Standards
for Civil Registration and Vital Statistics Systems
Session 1.7 Use of health services records for vital statistics
28 June 2011
United Nation North Lawn Building, New York City
Boonchai Kijsanayotin MD., PhD.
Health Systems Research Institute, Thailand
Outline




Thailand Civil Registration System
Thailand Vital statistics System
Health care service administrative data
Using administrative data to improve
vital statistics
Civil Registration System
 1909: Monarchy (King Rama V); Western
colonization period
 the first registration law was enacted specified
the preparation and maintenance of population
registration and creation of birth and death
registration.
 1917, enforced births and deaths
registration throughout the kingdom.
Civil Registration System
 1936: The democratic revolution
 creation of a network of local registry office,
acting registrars, laid down clear guidelines for
the registration of birth, death, fetal death (for
the first time)
 1956 Comprehensive civil registration act
 increased number of registration offices
spreading over the entire country
Civil Registration System
 1972 Revise civil registration act (Military
government)
 appointment hierarchy of registration authorities
 laid down clear procedures for vital registration
 specified types of informants place and time for
registration, forms, functions, responsibility of
registrars, and other requirement for registration.
 1982 the Population Identification Number
Project. Start of the computerized population
database.
MOI Central Data Processing
<<BACK
Civil Registration System
 National Authority: Bureau of Registration
Administration Department of Local
Administration, Ministry of Interior
 Operates and maintains the data base for the
entire population of the country, and issues
identity cards and household booklets
Citizen Identification Number
 Citizen ID issues at the of birth registration
 Called 13 Digits number
 The number is needed when transact with
government and many non government
agencies:
 getting free healthcare services
 enrolling children to free education
 asking for bank account.
Citizen Identification Card
1969-1986 Typing machine 1987 – 1996 Computer
ID card
Printout ID card
Current Smart card type
ID card
Vital statistics system
 The compilation of vital statistics was initiated in
1920
 Vital Statistics Division was established in MOPH
in 1942.
 1993 Health Information Center, Bureau of the
Health Policy and Strategy, the Office of the
Permanent secretary of the MOPH.
Vital statistics Before 1996
Ministry
of Interior
Central
Registry
Copy
birth/death
list for
validation
Compile, code, validate and
process statistics
Aggregated
report by
Province
MOI Local
Register
Printed Annual
Vital statistics
report
Ministry of Public Health
Copy Birth/
Death list
Provincial
Health
offices
Send copy of
death certificates
Hospitals
Weakness of the systems
 Discrepancy of birth/death counts between
report from provincial health offices and MOI
central registry
 Not timely: report lag time 2-3 years
 Cannot validate cause of death
 Need a lot of data entry personnel to enter
data from hospital’s death certificate
Streamline vital event data
between MOI & MOPH
1996, The Ministry of Public Health has signed an agreement
with the Department of Local Administration, Ministry of
Interior, regarding utilization of data in the central
registration database of the administration, in order to
reduce redundancy of birth, death data.
The Department of Local Administration, Ministry of Interior,
will send the data relevant to death of all the persons who
have been registered birth, dead from the central
registration database to the Office of Permanent
Secretary of the Ministry of Public Health.
After 1996 - Present
Ministry
of Interior
Central
Registry
Electronic
files
Compile, code, validate
and process statistics
Ministry of Public Health
Report
statistics back
to provinces
Provincial
Health
offices
Printed/ Web
Vital statistics
report
Web entry death
certificates (start 2006)
Hospitals
Health care service administrative data
 2001 implemented Universal care program
 3 Major health insurance schemes
 Civil Servant Medical Benefit Scheme (CSMBS)
5.6 million civil servants and their dependants (9% Pop.)
 Social Security Scheme (SSS)
8.8 million private employees (14% Pop.)
 Universal Health-care Coverage Scheme (UCS)
45 millions (77 % of Pop)
Casemix Information System
 Diagnosis Related Groups (DRGs) : acute in-patient
financing tool
 Insurance schemes reimburse hospitals using DRGs
 5 millions in-patients (7 millions admission)/year
 Standard data sets for health insurance




Patient demography
Citizen ID
Diagnosis and Procedure (ICD Coded)
Patient discharge status (e.g. recovery, death, transfer
etc.)
Flow of vital event registration
Electronic file transfer monthly
Ministry of Interior
ICD-10 coding
at MoPH (BPS)
On-line system
Death registration at district office or municipality
Medical death certificate form
(In-hospital) with COD
Death certificate by doctor
Death inside hospital
Death notification form
(Non-hospital) with COD
Death notification by village head
40%
Death outside hospital
60%
Death counts and Cause of Death (COD)
 Death data processed
 Total death ~400,000 / year
 Code cause of death with ICD 10
 Compile and Analyzed
 Annual reports in print and web format
 Provide provinces population, birth, death information
every month
 Decrease ill defined COD (R0-R99) from 45% to 37%
Problems of Causes of death data
 High % of ill-defined causes of death
 Non-hospital death (relatives provide general symptoms or
senility as the causes of death)
 In-hospital death (doctors use mode of death instead of the
causes)
Problems of Causes of death data
 Misclassification of causes of death
 In-hospital death (doctors select and write the wrong COD)
 Non-hospital death (relatives bias the causes of death to
avoid stigma such as HIV)
Ill-defined causes of death
%Ill-defined
100
80
60
40
20
0age0_1
age5_9
age15_19 age25_29 age35_39 age45_49 age55_59 age65_69 age75_79 age85_89 age>=95
age1_4
age10_14 age20_24 age30_34 age40_44 age50_54 age60_64 age70_74 age80_84 age90_94
Age group
year 2000
%illdefined %illdefinedmale %illdefinedfemale
Analyzed from Vital registration databases 2000, MoPH
Improving COD statistics
 Improving COD data for in-hospital death
 Improving COD data for non-hospital death
 Validation for COD statistics through Verbal
Autopsy
Improving in-hospital COD data
Training for medical doctors to define actual COD in
medical death certificate
 Reduce mode of death, un-specified causes, injury code
 Reduce misclassification, wrong selection
Medical records and coding audit for better quality of
diagnosis of diseases and causes of death
Improving non-hospital COD
 Using electronic in-patient records (administrative
data – reimbursement data) for defining outside
hospital COD
 Match death outside hospital with recent
hospitalization using citizen identifier number
(13 digit number)
 Using local health personnel to provide COD
(accompanied with VA tool and medical history)
Improving non-hospital COD
 Verbal autopsy (VA) questionnaire is another tool
that can use to improve Cause of Death (COD)
information.
 Thailand is piloting a project: using a customized
VA tool by rural health personal to interview the
deceased’s relative before registering to civil
registration system
 Preliminary results: Ill-defined COD (ICD 10 code
R00-R99) decrease from 41%(2007) to 27%(2008)
and 25%(2009)
Improve accuracy of Maternal Mortality Ratio
(MMR) Calculation
 Thailand faces unreliable MMR.
 Many MMR reporting figures that are not
consistent depend on the sources of data.
 MMR calculation
MMR from different sources
Identify Maternal death who gave live birth
from Birth-Death Registration
From: Chandoevwit W & et al, Using Multiple Data for Calculating the Maternal Mortality
Ratio in Thailand. Thai Development Research Institute Quarterly Review. 2007
Identify Maternal death with still birth & neonatal death
From death registration & heath service records
From: Chandoevwit W & et al,Using Multiple Data for Calculating the Maternal Mortality
Ratio in Thailand. Thai Development Research Institute Quarterly Review. 2007