Maternal Mortality& the MDGs
Deborah MaineProfessor, International Health
Boston University, School of Public Health
MDG Goal: Improve maternal health
Target: Reduce the MM Ratio by 3/4 by 2015
Indicators: Maternal mortality ratio Proportion of births attended by skilled
health personnel
The MDG for MM
Is it realistic ?
History
MMRs
Sri Lanka: 1947 -- 1500
1960 -- 250
1980 -- 100
Malaysia: 1950 > 500
1975 < 100
To reduce MM …
Need to understand the epidemiology of maternal mortality [MM]
A counterintuitive phenomenon Many “obvious” approaches don’t work,
e.g. risk screening, training TBAs
Maternal Mortality
Region MM Ratio Lifetime Risk
1 in …
Africa 830 20
Asia 330 94
Latin America 190 160
North America 17 2,500
World 400 74
Causes of Direct Obstetric Deaths
The “Big 5” Hemorrhage Infection Hypertensive diseases Obstructed labor Unsafe induced abortion
The Way Programs Should WorkEvidence
Interventions
Indicators
Strategy
Interventions
IndicatorsStrategy
Assumptions
The way it often works
Assumption
If we just take very good care
of pregnant women,
few will develop serious
obstetric complications.
History: Prenatal Care
1910-15 first clinics in UK (and US) By 1930, 80% pregnant women in UK
have prenatal care But maternal mortality did not decline
TBAs & “Clean Delivery”
In Matlab, Bangladesh, TBAs were trained to use clean delivery practices.
The did use these practices, but maternal deaths due to infection
did not decline.
Assumption
Through prenatal screening,
We can identify the women
who will need medical care
The Math of Prediction
It works for groups
but not for individuals.
Example: Matlab, Bangladesh1968-70
Maternal Age 10-14 20-29
MM Ratio 1770 450
Relative Risk 3.9 1
# Births 509 11,286
# Deaths 9 51
Example: United Kingdom1985-87
Maternal Age 20-24 45+
MM Ratio 37 188
Relative Risk 1 5.1
# Deaths 24 2
Risk and Prediction (cont.)
A big risk in a small population =
few deaths
A small risk in a big population =
many deaths
In Short ...
Once a woman is pregnant
most serious obstetric complications
cannot be predicted or prevented,
but they can be treated.
So
All pregnant women
need access to
emergency obstetric care
(EmOC)
Sri Lanka & Malaysia
How did they do it ? Expanding access to effective
maternity care by midwives and doctors Improving utilization and quality of care
with emphasis on making life-saving care free.
The World Bank, 2003
Assumption
EmOC is
“Hi-Tech”
Signal Functions of Basic EmOC :
Parenteral antibiotics, oxytocics, anticonvulsants
Manual removal of placenta Removal of retained products Assisted vaginal delivery Neonatal resuscitation (new)
Should be at health centers
Signal Functions of Comprehensive EmOC:
All Basic EmOC functions Blood transfusion Surgery (c-section)
Should be at District Hospitals
EmOC is not “Hi Tech”
It is mostly 1950s medicine !
EmOC is the foundation
Emergency Obstetric Care
SkilledAttendant Referral
Risk Screening
Social Mobilization
Waiting Homes
TBA Training
AntenatalCare
Assumption
EmOC is too expensive
Community-based workers
are more affordable
A cost-effectiveness exercise: unit cost
0 5000 10000 15000 20000 25000 30000 35000
Upgrading 1District Hospital
Upgrading 1Health Center
Training etc, 1MCHW
PROGRAM
Dollars
$350
$10,000
$30,000
Cost (cont.)
Suppose, per district, there are:
100 MCHW s 4 health centers 1 district hospital
Estimated program cost (in $000s)
0 10 20 30 40 50
Upgrading 1District Hospital
Upgrading 4Health Centers
Training etc, 100MCHWs
PROGRAM
Dollars
30
40
35
Estimated obstetric deaths prevented (%)
0 10 20 30 40 50 60
District hospital
Health centers
MCHW training
PROGRAM
50
Percent
25
15
Estimated cost per death averted ($000)
0 200 400 600 800 1000
Upgrading 1District Hospital
Upgrading 4Health Centers
Training etc, 100MCHWs
PROGRAM
$580
$845
$217
In short …
Something that is not effective
can never be cost-effective.
Measuring Progress:
Are we measuring the right things?
The Way It Should WorkEvidence
Interventions
Indicators
Strategy
But sometimes …
MDG Goal: Improve maternal health
Target: Reduce the MM Ratio by 3/4 by 2015
Indicators: Maternal mortality ratio Proportion of births attended by skilled
health personnel
Promoting SBAs
What is the evidence base
for this policy?
SBAs and MMR, 170 countries
R2 = 0.6124
0
500
1000
1500
2000
2500
0 20 40 60 80 100% Skilled Attendant at Delivery
Mat
ern
al M
ort
ality
Rat
io
(per
100
,000
live
bir
ths)
Country n=170
Source: Safe Motherhood Initiative website and Maternal Mortality in 1995: Estimates developed by WHO, UNICEF, UNFPA (2001)
SBAs and MMR, 50 Countries with MMR>400
R2 = 0.0818
0
500
1000
1500
2000
2500
0 20 40 60 80 100
% Skilled Attendant at Delivery
Ma
tern
al M
ort
alit
y R
ati
o
(pe
r 1
00
,00
0 li
ve
bir
ths
)
Country n=50
Source: Safe Motherhood Initiative website and Maternal Mortality in 1995: Estimates developed by WHO, UNICEF, UNFPA (2001)
This shows:
the relationship between
delivery by SBAs
and MMR
is not strong
for high-mortality countries
Source: Saving Lives: Skilled Attendance at Childbirth, W. Graham, 2000.
Source: Saving Lives: Skilled Attendance at Childbirth, W. Graham, 2000.
This shows:
the relationship between
delivery by midwives
and reduced MMR
is not clearcut –
probably due to regional variation in what midwives are trained and permitted to do.
Skilled Attendants need to be part of
a functioning health system
To Be Effective
Sri Lanka, 1970s >
HealthFacilities
SBAs
Many Proposed Programs
HealthFacilities
SBAs
In Reducing Maternal Deaths
There are really only 3 issues: COVERAGE OF SERVICES QUALITY OF CARE UTILIZATION OF SERVICES
The Road toMaternal Mortality Reduction:
Shortcuts or Detours ?
Pseudo-Interventions
“Safe Birth Kits”: No evidence of effectiveness in reducing maternal deaths, but consume effort, attention and funds.
Advocacy for Advocacy: If not linked to programs, advocacy can be a detour.
1-Complication MM Programs
Example: Home-based prevention of post-partum hemorrhage (PPH)
Hemorrhage = 25% of maternal deaths
Perhaps ½ preventable = 12.5%
Semi-Skilled Attendants
If you leave the skills out of
Skilled Birth Attendant
what do you get?
Institutional Delivery Targets
Easy to measure, but
no indication of quality of care
You can reach the target
But miss the goal !
“In the Meantime …”
If we don’t get started now
fixing health systems
in 20 years we will still be
in the meantime.
General Lesson:We must build health systems Need a strong evidence base Training and equipment are never
enough Management systems are crucial Even skilled personnel need support Learn from expensive failures