International Health Policy Program -Thailand International Health Policy Program (IHPP), Thailand 13 March 2013 Post Enumeration Survey (PES) of the household.
Post on 19-Jan-2018
222 Views
Preview:
DESCRIPTION
Transcript
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
International Health Policy Program (IHPP), Thailand13 March 2013
Post Enumeration Survey (PES) of the household Socio-
Economic Survey (SES) in Thailand
An adjustment of household out-of-pocket payments
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Outline of presentation1. What is PES ?2. Problems in HH out-of-pocket payments3. Objectives of the study4. Methods and conceptual framework5. Results6. Discussion and policy recommendations
2
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
landWhat is PES ?
• An independent survey replicating a census or a national HH survey aiming to evaluate quality and accuracies of the census or the national survey data,
• PES results can be compared with the results of census or national HH surveys, and permit estimates to be made of:– Coverage errors;– Content errors;
• PES is worth conducting if it is carefully planned and well implemented.
3
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Problems of HH out-of-pocket payments
• OOP, a major component of total health expenditure (THE) in developing countries,
• Potential under-reporting by households due to:– Sampling bias;– Non-sampling bias;
• Respondents • Proxy respondents tend to be under-reporting in
morbidity, utilization, and health expenditure, • The use of proxy respondents instead of
interviews of all eligible household members. • Interviewers
4
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Objectives• To identify discrepancy among morbidity (OP and
IP), health service utilization, and out-of-pocket payments (OOP) between NSO and IHPP interviewers in three options:
• Full proxy situation • The most knowledgeable HH members provide
information on behalf of all respondents, • Real life situation
• Interviews of all members on the interview day and proxy respondents for non-presence members,
• Gold standard • Interviews of all eligible members, until
exhausted in three visits, • Eligible members, age 15-60, proxy allowed
proxy respondents for those <15 and >60 years.
• To consider applying discrepancies indexes for the adjustment of OOP reported by SES
5
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
landMethodology 1
• NSO SES 2002 • National representative, a 12-months HH survey • Health module available• Interviewers were instructed to use gold
standard approach, • However 2.5 hours to complete one SES
Q’naire, might slip to real life or full proxy. • IHPP-PES
• Post-enumeration survey in 23 provinces in 3 months (Mar-May 03) in the randomly selected households, two weeks after ending SES interviewed by NSO field staffs
• IHPP 20 trained interviewers, instructed to do
gold standard interviews of all eligible respondents until exhausted in the 3rd visit.
6
Methodology 2• IHPP methodological testing
• Three methods were applied in the same household by the same IHPP interviewer:
• Full Proxy by interviews of the most knowledgeable person;
• Real life, interviews all eligible present on the first visit and allow proxy respondents for non-presence members;
• Gold standard, three repeated visits to interview the eligible member. Last visit allow proxy respondents.
• PES of the SES completed households in 5 provinces in 1 month (July 03)
• To produce 3 adjusting factors • IHPP Gold standard to NSO real life• IHPP Gold standard to NSO Gold standard • IHPP Gold standard to NSO full proxy
7
Conceptual framework of PES SES 2002
Household sampleheadmember < 15member > 15
IHPP Interviewanother set ofsample HHFull proxyReal lifeGold std
PQ approach to estimate national household spendingon H=62 mil pop x morbid rate x choice x baht OOP/visit or admission
Discrepancy index II
Discrepancy index III
Three factor of morbid rate, choice and hh spending OOPFactor 1: Gold std / Full proxyFactor 2: Gold std / Real lifeFactor 3: Real life / Full proxy
Use factor2 adjustNSO-Real life
versionTo be gold standard
Use 1/factor3 adjust
NSO-Real life version
To be Full proxy
2 wk time lagsame recall periodOP: last mo IP: last year
Morbid•OP•IPChoices•OP•IPHH spending•OP•IP
Argument for NSO “Real life” InterviewKI: Key InformantTraining NSO staff 2003, it must be gold standard slip Real life.Nosy interviewees join the interview.
Household sampleheadmember < 15member > 15
IHPPGold standardA billion baht
NSOReal lifeB billion baht
NSOGold standardC billion baht
NSOFull proxyD billion baht
Discrepancy index I
Morbid•OP•IPChoices•OP•IPHH spending•OP•IP
NSO“Real life”Interview
IHPP “Gold std”Interview
8
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Results
• PES-SES• Total 855 households in 23 provinces, in three
months, • 559 municipality and 296 non-municipality
households,• IHPP methodological testing
• Total 150 households in 5 provinces in 1 month • 80 municipality and 70 non-municipality
households
9
PQ approach
Discrepancy index II
1.05Discrepancy index III
1.14Discrepancy index I
1.13
A B C DIHPP NSO NSO NSO
"Gold std" "Real life" "Gold std" "Full proxy" OP (mil baht) 40548, 37499, 40089, 37780,
IP (mil baht) 12374, 9,264
10109, 87,41 Total (mil baht) 52921, 46764, 50198, 46521,
ratio of OP 77 80% 80% 81% ratio of IP 23% 20% 20% 19%
HH spending on health expenditure
10
Final adjustment for SES2002Percentage of OOPs increased from 32.7% to 33.8% and 35.7% respectively.
11
SES 2000
Adjust with 1.05
Adjust with 1.14
New ratio (min)
New ratio (max)
OOP 32.74 34.38 37.32 33.83 35.69Other 67.26 67.26 67.26 66.17 64.31THE 100% 101.64% 104.58% 100% 100%
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Discussion
• OOPs reported by biennial SES (even years) which discrepancy index should be used to adjust OOP: – Depends on NSO interviewers behaviors
(real life or full proxy), – NSO gold standard is unlikely – costly,
lengthy q’naire of SES, average 2.5 hours
• We decided not use this correcting factor for OOP adjustment due to small sample size of SES.
12
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Policy recommendations• PES is quite expensive, and needs
carefully plan and well implementation,
• Discussion with Secretary General of NSO – Decided to conduct national PES for the
next population and housing census.
13
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Acknowledgements• National Statistical Office (NSO)• National Health Security Office (NHSO)• Ministry of Public Health• Health Systems Research Institute
(HSRI)
14
Inte
rnat
iona
l Hea
lth P
olic
y Pr
ogra
m -T
haila
ndIn
tern
atio
nal H
ealth
Pol
icy P
rogr
am -
Thai
land
Sawasdee Thank you for your attention
15
top related