Transcript Health Sector PERs: Ukraine Case Study: Fiscal

Health Sector PERs: Ukraine
Case Study
Fiscal, Efficiency, and Equity
Issues in the Health Sector
Adam Leive
Human Development Network
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers and
Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Conclusions and Recommendations
Core Objectives – Fiscal Issues and
Health Sector
• Examine the intergovernmental fiscal and
administrative issues affecting efficiency and
equity of public spending in the health sector.
• Identify other social/economic trends affecting
the efficiency of public spending in the sector
• Link efficiency issues to quality of service
delivered, performance, and equity.
• To provide recommendations to improve the
financing, efficiency, and equity of spending of
the health sector.
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers and
Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Conclusions and Recommendations
Health Outcomes – Low Life Expectancy
• Life expectancy has not rebounded as quickly as other ECA
countries and was higher 40 years ago
80
TRENDS IN LIFE EXPECTANCY, 1960-2004
UKRAINE VS COMPARATOR COUNTRIES/REGIONS
75
High-Income OECD
EU
70
ECA
Ukraine
65
Russia
1960
1970
Source: World Development Indicators
1980
Year
1990
2000
Health Outcomes – Differences in Life
Expectancy Between Males and Females
• Life expectancy is 12 years higher for females than for males, on
average
LIFE EXPECTANCY VS INCOME, 2004
FEMALES
75
Ukraine
65
60
70
65
Ukraine
Life expectancy at birth
70
80
75
85
80
LIFE EXPECTANCY VS INCOME, 2004
MALES
100
250
1000 2500
1000025000
GDP per capita
Source: World Development Indicators
Note: GDP per capita in constant 2000 US$; Log scale
100
250
1000 2500
1000025000
GDP per capita
Source: World Development Indicators
Note: GDP per capita in constant 2000 US$; Log scale
Demographics – Aging Population
Structure
2025
2005
75+
70-74
65-69
60-64
55-59
50-54
45-49
40-44
35-39
30-34
25-29
20-24
15-19
10-14
5-9
0-4
Males
Males
6
Females
Females
4
2
00
Percent
Source: World Bank staff calculations
22
44
66
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers and
Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Conclusions and Recommendations
International Comparisons of Total and
Public Health Spending
Country/Region
Total health spending per
capita (US $)
Total health spending (%
GDP)
Public health spending (%
GDP)
Ukraine
90
6.5
3.7
ECA
318
6.4
3.9
3,365
9.3
6.7
High-income OECD
15
PUBLIC HEALTH EXPENDITURE VS INCOME
10
Public health expenditure as share
10 of GDP
5
15
HEALTH EXPENDITURE VS INCOME
5
Ukraine
0
0
Ukraine
100
250
1000 2500
1000025000
GDP per capita
Source: World Development Indicators & WHO
Note: Log scale
100
250
1000 2500
1000025000
GDP per capita
Public Health Expenditure by Economic
and Functional Classification
• Wages and other current expenditures constitute the majority of
public health spending (usually over 90%)
• Hospitals capture a large share of spending while polyclinics and
preventive facilities receive less
Government Health Expenditure by Functional Classification, 2006
Hospitals and
sanatoriums
70%
Polyclinics,
ambulatories, firstaid, and
emergency
14%
Preventive and
anti-pandemic
institutions and
activities
5%
Health protection
research
1%
Other health
protection
facilities
10%
Health Care Delivery
• The Ukrainian health system has evolved from a centrally-planned
Soviet-style model to one where responsibilities have been
delegated to lower tiers of government
• Most facilities are owned and managed at the regional (oblast) and
sub-regional (rayon) levels
• There is negligible participation by the private sector in service
delivery
• Ukraine has slightly fewer physicians per 100,000 (301) than in EU
(348) and CIS (378)
• However, it has about the same hospital beds per 100,000 (872) as
CIS (866) and more than EU (591)
• Ukraine also has a similar level of hospitals per 100,000 (5.6) as CIS
(5.9) but more than EU (3.1)
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers
and Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Discussion of Policy Reforms and Recommendations
Intergovernmental Fiscal Issues
•
•
•
The budget reform in 2001 radically changed the
intergovernmental finances in the country
Funding from the central government for health is part of a
horizontal equalization formula
In Ukraine the most basic shape of the equalization formula for
the allocation to a local government is the following:
Ti   (Vi  Di )
•
where Ti is the transfer to local government i; Vi are the estimated
expenditure needs for local government i and Di its estimated
revenue capacity
The estimated expenditure needs for health are calculated based
on population size
Ministry of Health (MoH) “Norms”
•
•
Despite decentralization in service delivery there remains extensive
regulation by the center
The MoH norms (mandatory guidelines) dictate to all health facilities across
the country how they should allocate resources (particularly staffing, but
also others) based mostly on bed size or another indicator
Examples of Norms for the functioning of health facilities
1 Infection disease doctor per 25 beds in outpatient aid in Rayon hospitals;
1 Surgeon per 20 beds (adults) and 15 beds (children) in Rayon hospitals
1 Nurse (gynecology) per 25 beds in Rayon hospitals;
1 Nurse per 20 beds in Children’s hospitals
1 Dietarian nurse per 500 portions served a day
1 Obstetrician-gynecologist per 20 beds in Rayon hospitals; 1 post per 15 beds in
District hospitals;
0.5 Statisticians per 20 posts of doctors in polyclinics
1 Cook per 30 beds in a health facility
1 Cleaner per 500 square meters (0.5 per each 250 square meters)
Source: Extracted from Order No. 33 Ministry of Health
•
Facilities prepare budgets fulfilling the “norms”, which leads to current
spending over 90% of total spending in most cases and little autonomy is
left to local governments
The Inefficient Process of Budgeting by Input
Norms Dictated from the Fiscal Framework
Ministry of Finance
Budget Allocation Through:
(1) Shared Revenues (PIT, Land)
(2) Equalization Transfer (populationbased for health)
(3) Local Taxes
Ministry of Health
Input Norms to
form health facility
budgets, which
constrain budget
flexibility (Order
No 33)
Mismatch
Local Budget Submission
complying with Norms
Rayon
Health Facility
“Norms” fulfillment +
Large network of facilities
= non-flexible local
budgets crowded by high
recurrent spending little
spending autonomy is left to
local governments.
Budget Formation:
Prepare budgets
fulfilling the “norms”,
which lead to
extremely high
current spending
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers and
Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Conclusions and Recommendations
The Effect of Intergovernmental Fiscal
Transfers on Public Health Expenditure
•
In 2001, public health spending varied across oblasts and was correlated
with income
180
HEALTH EXPENDITURE VS INCOME BY OBLAST, 2001
80
100
120
140
160
Kyiv city
AR Crimea
Sevastopol city
Zaporizhzhia
Donetsk Dnipropetrovsk
Cherkasy Kyiv
Sumy
Chernihiv
Kharkiv
Luhansk
Kherson
Kirovohrad
Poltava
Zhytomyr
Mykolayiv
Rivne
Volyn
Lviv
Vinnytisa
Khmelnytsky
OdesaIvano-Frankivsk
Zakarpattia
Ternopil
Chernivtsi
3000
3500
4000
4500
Oblast GDP per capita (hryvnas)
5000
Source: Ministry of Finance; State Treasury; and Bank Staff estimations
•
Despite a major change in the way transfers are allocated, differences in
absolute public health spending per capita across oblasts remain
Labor vs. Capital Across Oblasts
• Personnel highly correlated with public health exp. across oblasts while
there is little correlation with public expenditure and the number of
health facilities
• This is most likely due to rigidities created by health norms which
specify strict requirements for staffing levels and must be followed in
order to receive budgetary allocations at the oblast and rayon levels
1500
HEALTH FACILITIES VS EXPENDITURE
0
0
500
1000
Oblast total number of health facilities
50000
100000
HEALTH PERSONNEL VS EXPENDITURE
0
200000
400000
600000
800000
Oblast public health expenditure (hryvnas)
0
200000
400000
600000
800000
Oblast public health expenditure (hryvnas)
Source: Ministry of Finance; State Treasury; and Bank Staff estimations
Health facilities include general hospitals, polyclinics, medical assistance, and obstetric centers
Variation in Average Length of Stay
(ALOS) Across Oblasts
• ALOS in Ukraine was about 15 days, higher than the CIS average of
13 days and EU average of 9.
• ALOS also varies widely across oblasts (see Annex 2 for probit
regression)
• ALOS is higher in regions with lower inpatient utilization rates and may
be compensating to maintain budgetary allocations
Average length of stay by oblast
Mykolaevska
Luganska
Ternopilska
Zaporizhska
Vinnitska
Odeska
Sebastopol city
Kiev city
Rivnenska
Zhytomirska
Khersonska
Poltavska
Kharkivska
Zakarpatska
Sumska
AR Crimea
Kiev
Donetska
Chernigivska
Cherkaska
Lvivska
Dniepropetrovska
Khmelnitska
Ivano-Frankivska
Kirovogradska
Chernivetska
Volynska
10
15
20
Average length of stay (days)
Source: World Bank Health and Education Survey 2004
25
Budget Rigidities and Inefficiencies
Translate into Inequities - OOP Expenditure
Across Oblasts
• Insufficient public financing translates into the need for households to purchase drugs
•
•
•
outside the health facility and into high rates of informal payments at health facilities
Extent of OOP expenditure is large, although estimates vary (40% to 60% of total health
expenditure depending on source)
As a share of total income, OOP expenditures range widely across oblasts
Drugs constitute the largest share of OOP spending and households finance 99.9% of
total drug costs in Ukraine (National Health Accounts)
8.0%
7.0%
6.0%
5.0%
4.0%
3.0%
2.0%
1.0%
0.0%
Drugs
Travel
Travel Cost
Outpatient
Inpatient
Zaka
rpats
AR C ka
rim
Ivano
-F ran ea
kivsk
a
Seba
stapo
Done l
tska
Cher
nigiv
Zapo ska
rizhs
ka
Odes
Luga ka
nska
Lvivs
Dnie
k
prop
etrov a
ska
Kie
Cher
nivet v
s
Khar ka
kivsk
Voly a
Myko nska
laev
Cit y ska
of K
Rivn iev
ensk
a
Sum
ska
Polt a
vska
Khm
elnit s
Kirov
k
ogra a
dska
Kher
so n s
Cher ka
ka
Zhyto ska
mirsk
Vinn a
it
Tern ska
opils
ka
OOP Share of Income
Out of Pocket Expenditures by Type
Oblast
Source: Authors’ calculations from World Bank Health and Education Survey 2004
High Levels of Catastrophic OOP
Spending
• Compared to other ECA countries, Ukraine has the highest
proportion of households facing catastrophic OOP spending (OOP
spending greater than 40% non-subsistence spending. See Annex
3 for methodology)
• 10 % of households face catastrophic spending due to drug
spending alone, suggesting little financial protection
Percent of households spending more than 40%
of non-subsistence expenditure on OOP
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
Slovakia
Croatia
Slovenia
Hungary
Latvia
Estonia
Russia
Ukraine
Outpatient Facility Use by Type of Facility
and Economic Status
• The poorest quintile is significantly less likely to use outpatient services
• Public polyclinics and rayon hospitals constitute the largest share of
outpatient visits and these are used most by the three richest quintiles
Outpatient visits by facility type and economic status
Public polyclinic
Rayon or town hospital
Med Asst/Obstetric care center
Village hospital
Private polyclinic
Oblast hospital
Other
Specialized facility
0
500
1,000
Sum of outpatient visits
Bottom Quintile
Quintile 3
Top Quintile
1,500
Quintile 2
Quintile 4
Source: W orld Bank Health and Education Survey 2004
Inpatient Facility Use by Type of Facility
and Economic Status
• The richest and poorest quintile use inpatient services the least and
their utilization rates are statistically significantly less than the
median and 2nd richest quintiles
• Across income quintiles, the majority of inpatient visits occur at
rayon hospitals
Inpatient visits by facility type and economic status
Raion or town hospital
Oblast hospital
Maternity home
Other
Sanatorium
Village hospital
Departmental hospital
Academic Medical Clinic
Private hospital
0
200
400
600
800
Sum of inpatient visits
Bottom Quintile
Quintile 3
Top Quintile
1,000
Quintile 2
Quintile 4
Source: W orld Bank Health and Education Survey 2004
Outline
•
•
•
•
Objectives of PER
Health outcomes and demographics
Description of health financing and delivery system
Background on intergovernmental fiscal transfers and
Ministry of Health norms in Ukraine
• Analysis of efficiency and equity of health sector
• Conclusions and Recommendations
Conclusions and Policy
Recommendations of PER
• Reformulate the MoH norms to eliminate the rigidities preventing
efficient resource use (necessary condition but it may not be
sufficient)
• Attempt to reduce underlying inefficiencies through:
– Basing transfers on outputs rather than inputs
– Gearing regulations more towards accreditation of facilities
rather than towards the control of their production function
• Purchase drugs, especially for the poor, if physician expenditures
can be reduced by reforming the norms
• Target expenditures on facilities that poor prefer to use (public
polyclinics and rayon/town hospitals)
• Provide incentives to shift a larger proportion of doctors into primary
care
• Regarding HIV/AIDS and TB, improve training of public health
professionals and improve coordination between levels of
government
• In the medium to long term, introduce prospective payment into
provider reimbursement at the level of the hospital or the individual
provider
Annex 1 - Probit regression of a higher
length of stay than average
Variable
Economic Region
Eastern
Donetsk
Prechornomorsk
Podilia
Central
Predniprovsk
Carpathian
Coefficient
Z-statistic
0.252
0.241
0.291
0.109
0.074
0.235
0.014
(1.94)
(1.79)
(2.36)*
(0.87)
(0.55)
(1.71)
(0.11)
1,688 Observations * = significant at 5%
Other control variables: age, sex, body mass index, marital status, income, education level,
inpatient facility type, and geographical location of facility
Average Partial Effects
Economic Region
Mean
Standard deviation
Min
Max
Eastern
0.085
0.016
0.024
0.100
Donetsk
0.081
0.016
0.023
0.096
Prechornomorsk
0.099
0.019
0.029
0.116
Podilia
0.036
0.007
0.010
0.043
Carpathian
0.005
0.001
0.001
0.005
Central
0.024
0.005
0.006
0.029
Predniprovsk
0.079
0.015
0.022
0.094
•
Probit regression of a higher length of stay than average
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Variable
Economic Region
Eastern
Donetsk
Prechornomorsk
Podilia
Central
Predniprovsk
Carpathian
Socioeconomic and demographic characteristics
19 < age < 30
30 < age < 40
40 < age < 50
50 < age < 60
60 < age
High education
No education
2nd Income quintile
3rd Income quintile
4th Income quintile
5th Income quintile
Male
Body mass index
Single
Divorced
Widow
Location and type of inpatient facility
Oblast hosp.
Rayon hosp.
Maternity hosp.
Sanatorium
Other hosp.
Clinic
City
Town
Constant
Observations
1,688
* significant at 5%; ** significant at 1%
Ramsey’s RESET Test: p-value = 0.09
Reference individual is female, under 19, lives in a rural area of the Polisya
secondary education, and was hospitalized in a village hospital.
Coefficient
Z-statistic
0.252
0.241
0.291
0.109
0.074
0.235
0.014
(1.94)
(1.79)
(2.36)*
(0.87)
(0.55)
(1.71)
(0.11)
0.291
0.217
0.549
0.586
0.571
0.107
-0.016
-0.131
-0.267
-0.185
-0.289
0.141
155.31
0.083
-0.103
0.074
(1.59)
(1.02)
(2.60)**
(2.73)**
(2.79)**
(1.38)
(0.09)
(1.13)
(2.23)*
(1.54)
(2.09)*
(1.96)
(1.81)
(0.57)
(0.74)
(0.71)
0.305
-0.027
-0.595
1.649
0.950
0.573
0.056
0.022
-1.537
(1.61)
(0.15)
(2.11)*
(6.29)**
(3.99)**
(2.64)**
(0.60)
(0.25)
(4.70)**
economic region, married, lowest income quintile, has
Annex 2 – Definition of Catastrophic
Spending
Catastrophic out-of-pocket (OOP) spending is defined for a household h if OOP spending exceeds
40% of the household’s capacity to pay. The definition of capacity to pay is constructed in the
following way and closely follows Xu, K. (2005). "Distribution of health payments and catastrophic
expenditures: Methodology." World Health Organization Health Systems Financing Discussion
Paper, Number 2.
First, the food expenditure share of total household expenditure is constructed by dividing the
household's food expenditure by its total expenditure.
FOOD h
FOODEXPh = TEXP
h
The household equivalence scale is used instead of the actual household size to account for
economies of scale in household consumption. The equivalence scale is defined as:
EQSIZEh = HHSIZEhB
Previous research from household surveys in 59 countries indicates that the B = 0.56 (Xu et al.
2003). The equivalised food expenditure share is generated by dividing food expenditure by the
equivalent household size.
FOOD h
EQFOODh = EQSIZE
h
The poverty line is defined as the food expenditure of the household with food expenditure share
of total household expenditure at the 50th percentile in the country. Average food expenditure of
the households with food expenditure shares between the 45th and 55th percentiles of the total
sample are used to minimize measurement error. The percentiles consider the household
weighting variable of the survey.
 wh  EQFOODh
PL =
 wh
The subsistence expenditure of each household is then calculated as the poverty line multiplied
by the equivalent household size of each household.
SEh = PL X EQSIZEh
Capacity to pay is defined as a household's non-subsistence spending. Additionally, for those
households reporting food expenditure lower than the level of subsistence spending, non-food
expenditure is used as capacity to pay.
CTPh = TEXPh - SEh if SEh ≤ FOODh
CTPh = TEXPh - FOODh if SEh > FOODh
Annex 3 - Decomposition of the
Redistributive Effect (RE) of OOP health
payments
•
In addition to access, the effect of financing source on the income distribution is an important
element of equity. The RE, which is the difference between pre- and post-payment Gini
coefficients, is affected by: 1) degree of progressivity 2) OOP share of income 3) degree of
horizontal inequity 4) reranking due to payment. Possible to think broadly of two components of
RE: vertical (V) and horizontal (H+R), where V is a combination of 1) and 2) and (H+R) is the sum
of 3) and 4).
G is the Gini coefficient before payment, G is the
RE  G X  G X  P
G X P    x GF ( x)  GB  G X P  C X P
H    x GF ( x)
R  G X P  C X P
 g 
 K
V  G X  G B  
1  g 
G X  G X P  V  H  R
X
X-P
Gini coefficient after payment,GF(x) is the Gini
coefficient for post-payment income for households
with pre-payment income x, αx are weights equal to
the product of the population share squared and the
post-tax income share of households with income x,
GB is the Gini coefficient of the income distribution
after payment that would exist if all members of each
equal pre-payment income group paid the same
amount, and CX-P is the concentration index after
payment that is obtained by ranking households first
by their income before payment and then within each
group of pre-payment equals according to their
income after payment
References: Aronson, J.R. and P. Lambert. (1994). “Decomposing the Gini Coefficient to Reveal the Vertical, Horizontal, and
Reranking Effects of Income Taxation.” National Tax Journal. 47(2): 273-94.
van Doorslaer, E. et al. (1999). “The redistributive effect of health care finance in twelve OECD countries.” Journal of
Health Economics, 18: 291-313.
Annex 3 - OOP Payments are Regressive
in Ukraine
• The redistributive effect, which is the difference between pre- and post-payment
•
Redistributive Effect of OOP Financing
0.020
Hungary
0.015
Redistributive Effect
•
Gini coefficients, indicates the effect of a financing source on the income
distribution
Compared to other ECA countries, OOP payments are most regressive (RE is
lowest) in Ukraine
Increase in income inequality is also greater than of that of direct payments in 10
of 12 OECD countries estimated in van Doorslaer et al (1999).
0.010
Slovenia
0.005
Croatia
Estonia
0.000
Slovakia
-0.005
-0.010
Russia
Latvia
Ukraine
-0.015
Annex 3 - Decomposing the Redistributive
Effect: Horizontal Effect Dominates
Vertical
Effect
• RE is affected by: 1) degree of progressivity 2) OOP share of income 3) degree
•
Decomposition of Redistributive Effect of OOP Financing
0.030
Redistributive Effect
•
of horizontal inequity 4) reranking due to payment (See Annex 4)
Differential treatment of households of similar income levels is very large in
Ukraine
While possibly due to random nature of illness, most likely greater indicator of
regional differences in OOP payment rates and variation in informal payments
across facilities, regions, and type of care
0.020
0.010
0.000
HungarySlovenia Croatia
Estonia Slovakia Russia
-0.010
-0.020
-0.030
Vertical Effect
Horizontal Effect
Latvia
Ukraine