Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
Prepared for the 26th IUSSP Conference, Marrakech, September 2009
Session 037: Measuring Maternal Mortality through the 2010 round of population censusesg y g f p p
Maternal Mortality in South Africa:Maternal Mortality in South Africa:update from Community Survey 2007update from Community Survey 2007update from Community Survey, 2007update from Community Survey, 2007
Michel Garenne Michel Garenne (IRD/Institut Pasteur, Paris)(IRD/Institut Pasteur, Paris)
Robert McCaaRobert McCaa (MPC/IPUMS, Minneapolis(MPC/IPUMS, Minneapolis))Robert McCaa Robert McCaa (MPC/IPUMS, Minneapolis(MPC/IPUMS, Minneapolis))
Kourtoum NacroKourtoum Nacro ((UNPFA, AngolaUNPFA, Angola))
Updated: September 17, 2009
Maternal mortality in Maternal mortality in South AfricaSouth Africa
Controversial issueControversial issue Several sources:Several sources:
Vital registration Vital registration Census and surveys Census and surveys Demographic Surveillance SystemDemographic Surveillance System
Hospital dataHospital data Hospital data Hospital data Levels, Trends, DifferentialsLevels, Trends, Differentials
Major discrepancies on levelsMajor discrepancies on levelsMajor discrepancies on levels Major discrepancies on levels Agreement about recent increaseAgreement about recent increase Very large differences (regional, socioVery large differences (regional, socio--economic status …)economic status …)
Issues related with maternal Issues related with maternal mortality in developing countriesmortality in developing countries
Case definitionsCase definitions Demographic: “Pregnancy related death” (any death during Demographic: “Pregnancy related death” (any death during
maternal period) = direct + indirect + fortuitousmaternal period) = direct + indirect + fortuitous Medical: “Direct obstetric causes”Medical: “Direct obstetric causes”Medical: Direct obstetric causesMedical: Direct obstetric causes Major issue in case of high young adult mortality Major issue in case of high young adult mortality
(HIV/AIDS + external causes)(HIV/AIDS + external causes) Changing relationship with socioChanging relationship with socio economic statuseconomic status Changing relationship with socioChanging relationship with socio--economic status, economic status,
because of the HIV/AIDS epidemicbecause of the HIV/AIDS epidemic General issue for causes of deathGeneral issue for causes of death
i l l l f l lii l l l f l li Particularly relevant for maternal mortalityParticularly relevant for maternal mortality Earlier publication based on 2001 census data, in: Earlier publication based on 2001 census data, in:
Population Health MetricsPopulation Health Metrics, 2008 (4):1, 2008 (4):1--1313pp , ( ), ( )
Census data on maternal mortality Census data on maternal mortality in South Africain South Africa
Recent sources: Recent sources: Census, 2001 (1/10 sample available)Census, 2001 (1/10 sample available) Community Survey, 2007 (1/43 sample)Community Survey, 2007 (1/43 sample)
Both include:Both include: Births in past 12 monthsBirths in past 12 months
Deaths in past 12 monthsDeaths in past 12 months Deaths in past 12 months Deaths in past 12 months Maternal deaths, for women aged 12Maternal deaths, for women aged 12--50 / demographic 50 / demographic
definition = Deaths during pregnancy, delivery or within 6 definition = Deaths during pregnancy, delivery or within 6 k f d li / i f i di d i dik f d li / i f i di d i diweeks of delivery / No information on direct and indirect weeks of delivery / No information on direct and indirect
causescauses Deaths from external causes (accidents and violence)Deaths from external causes (accidents and violence)( )( )
Potential of census data in South Potential of census data in South Africa (2001 & 2007)Africa (2001 & 2007)
Allows to compute all maternal mortality indicatorsAllows to compute all maternal mortality indicatorsp yp y MMR: maternal mortality ratioMMR: maternal mortality ratio MDR: maternal death rateMDR: maternal death rate MMR: life time risk (maternal mortality quotient)MMR: life time risk (maternal mortality quotient)
Allows differential analysisAllows differential analysisVariety of socioVariety of socio economic factors in censuseconomic factors in census Variety of socioVariety of socio--economic factors in censuseconomic factors in census
Available only at household level for maternal deathsAvailable only at household level for maternal deaths Allows multivariate analysisAllows multivariate analysisAllows multivariate analysisAllows multivariate analysis
Compares cases (maternal deaths) with controls (women Compares cases (maternal deaths) with controls (women who delivered in past 12 months and survived)who delivered in past 12 months and survived)
Results: fertility and mortalityResults: fertility and mortalityResults: fertility and mortalityResults: fertility and mortality
Census, 2001Census, 2001 CS, 2007CS, 2007Census, 2001Census, 2001 CS, 2007CS, 2007
HouseholdsHouseholds 991,543991,543 246,618246,618
Fertility (CBR)Fertility (CBR) 22.822.8 24.024.0
Mortality (CDR)Mortality (CDR) 9.39.3 15.315.3
MaleMale qq 440440 606606Male, Male, 4545qq1515 440440 606606
Female, Female, 4545qq1515 320320 501501
Results: maternal mortalityResults: maternal mortalityResults: maternal mortalityResults: maternal mortality
Census, 2001Census, 2001 CS, 2007CS, 2007Census, 2001Census, 2001 CS, 2007CS, 2007
Nb maternal deathsNb maternal deaths 508508 168168
Percent time in Percent time in maternal periodmaternal period
7.1%7.1% 7.6%7.6%maternal period maternal period Proportion of Proportion of maternal deathsmaternal deaths
6.4%6.4% 4.3%4.3%
MMR / 100,000MMR / 100,000 542542 702702
95% CI95% CI 497497--591591 603603--816816
Levels and trends in Levels and trends in maternal mortalitymaternal mortality
LevelsLevels Abnormally high adult mortality (3 to 5 times that of model Abnormally high adult mortality (3 to 5 times that of model
life tables with same elife tables with same e°°(60)(60)D t h i i f ti HIV/AIDS d PTBD t h i i f ti HIV/AIDS d PTB Due to chronic infections: HIV/AIDS and PTBDue to chronic infections: HIV/AIDS and PTB
And to external causes (accidents, violence)And to external causes (accidents, violence) TrendsTrends TrendsTrends
Increasing trend in adult mortality from 2001 to 2007Increasing trend in adult mortality from 2001 to 2007 Increasing trend in maternal mortality (+29%; P<0.004)Increasing trend in maternal mortality (+29%; P<0.004) Decreasing proportion in maternal deathsDecreasing proportion in maternal deaths
Risk attributable to Risk attributable to the maternal periodthe maternal period
ComparisonComparisonpp Deaths expected from mortality among nonDeaths expected from mortality among non--pregnant pregnant
women women Deaths during maternal periodDeaths during maternal periodDeaths during maternal period Deaths during maternal period
Attributable riskAttributable risk Observed / Expected number of deathsObserved / Expected number of deaths
C 2001 AR 0 69C 2001 AR 0 69 Census, 2001: AR= 0.69Census, 2001: AR= 0.69 Community Survey, 2007: AR= 0.57Community Survey, 2007: AR= 0.57 No increased risk associated with pregnancy and delivery! No increased risk associated with pregnancy and delivery!
(pregnancy and delivery are protective)(pregnancy and delivery are protective) New situation, very different from preNew situation, very different from pre--AIDS era, where AIDS era, where
AR = 1.5 to 2.0AR = 1.5 to 2.0
Differential analysisDifferential analysisDifferential analysisDifferential analysis Household characteristics / SocioHousehold characteristics / Socio--economic statuseconomic status
Urban / RuralUrban / Rural ProvinceProvince RaceRace EducationEducation IncomeIncome IncomeIncome WealthWealth
Community variables / Provincial levelCommunity variables / Provincial levelyy HIV/AIDS prevalenceHIV/AIDS prevalence External causesExternal causes
H d li iH d li i Home deliveriesHome deliveries
MMR, by area of residenceMMR, by area of residenceMMR, by area of residenceMMR, by area of residence
1000888
800
1000
hs
20012007
604
505 539600
100,
000
birt
h
200
400
MM
R p
er
0
200
Rural UrbanRural Urban
MMR, by province of residenceMMR, by province of residenceMMR, by province of residenceMMR, by province of residence
245175Western Cape
415
347
245
509
445
Northern Cape
Limpopo
Western Cape 20072001
654
519
395
860
414
555
Free State
Mpumalanga
Gauteng
751
712
654
642
1236
North-West
Eastern Cape
ee State
772912
0 200 400 600 800 1000 1200 1400
Kwazulu Natal
MMR, per 100,000 live births, p ,
MMR, by population group (race)MMR, by population group (race)MMR, by population group (race)MMR, by population group (race)
1000
784800
1000
hs
20012007
614600
100,
000
birt
h
284 310
200200
400
MM
R, p
er
67
0
200
Black/African Coloured Indian/Asian White/EuropeanBlack/African Coloured Indian/Asian White/European
MMR, by level of educationMMR, by level of educationMMR, by level of educationMMR, by level of education9931000
899
753703
800
900
1000
ths
703
598 595
500
600
700
100,
000
birt
382
258
200
300
400
MM
R, p
er
0
100
200
0-4 years 5-8 years 9-11 years 12+ yearsy y y y
MMR, by income per capitaMMR, by income per capitaMMR, by income per capitaMMR, by income per capita1200
800
1000
1200
birt
hs
20012007
400
600
800
per
100,
000
200
400
MM
R,
0 None
declared < 2400 < 4800 < 9600 < 28800 >=28800
Income per capita (ZAR)Income per capita (ZAR)
MMR, by absolute wealth indexMMR, by absolute wealth indexMMR, by absolute wealth indexMMR, by absolute wealth index1200
1000
1200
birt
hs
20012007
600
800
per
100,
000
b
200
400
MM
R, p
0'0-1 '2-3 '4-5 '6-7 '8-9 '10-11 '12+
Wealth index (number of modern items in household)( )
MMR, by HIV prevalenceMMR, by HIV prevalenceprovinces of South Africaprovinces of South Africa
1000
1200
1400
CS-2001r= 0,73
1000
1200
1400
CS-2007r= 0,59
NWFSEC
KZN
600
800
1000
per
100,
000
birt
hs
FSKZN
600
800
1000
per
100,
000
birt
hs
LP
MP
GTNC
WC
200
400
600
MM
R, p
LP MP
GTNC
WC200
400
600
MM
R, p
0
200
0 10 20 30 40 50
Percent HIV+ (antenatal clinics)
WC
0
200
0 10 20 30 40 50
Percent HIV+ (antenatal clinics)
MMR, by external cause mortality, MMR, by external cause mortality, provinces of South Africaprovinces of South Africa
1000
1200
1400
rths
CS-2001r= 0,80
EC1200
1400
ths
CS-2007r= 0,59
MP
NWFS EC
KZN
600
800
1000
er10
0,00
0bi
r
NW
FSKZN
600
800
1000
r10
0,00
0bi
rt
LP
MPGT NC
WC200
400
600
MM
R,p
e
LPMP
GT NC
WC200
400
600
MM
R,p
e
0100 200 300 400 500
Mortality from external causes (per 10,000)
WC
0100 200 300 400
Mortality from external causes (per 10,000)y (p , ) y (p , )
MMR, by proportion of home MMR, by proportion of home deliveries, provinces of South Africadeliveries, provinces of South Africa
1000
NWFS EC
KZN
800
1000
birt
hs
CS-2001r= +0,29
LP
MPGT NC
400
600
er10
0,00
0
LPWC
200
400
MM
R,p
00% 5% 10% 15% 20% 25% 30%
Proportion of home deliveriesp
Conclusions (I)Conclusions (I)Conclusions (I)Conclusions (I) MMR much higher than previously thoughtMMR much higher than previously thoughtg p y gg p y g
Consistent with findings from Agincourt (same MMR)Consistent with findings from Agincourt (same MMR) Still within the range of other countries in AfricaStill within the range of other countries in Africa Unlikely to be overUnlikely to be over--estimated because of the low estimated because of the low
proportion of maternal deaths among women 15proportion of maternal deaths among women 15--4949 MMR has increased in recent years due toMMR has increased in recent years due to MMR has increased in recent years, due to MMR has increased in recent years, due to
HIV/AIDS & PTB, Accidents & ViolenceHIV/AIDS & PTB, Accidents & Violence Many hospital studies show a high proportion of maternal Many hospital studies show a high proportion of maternal
d h d HIV/AIDS PTBd h d HIV/AIDS PTBdeaths due to HIV/AIDS or PTBdeaths due to HIV/AIDS or PTB Evidence of increase in MMR in Cape Town peninsula Evidence of increase in MMR in Cape Town peninsula Evidence from the vital registration systemEvidence from the vital registration systemEvidence from the vital registration system Evidence from the vital registration system
Conclusions (II)Conclusions (II)Conclusions (II)Conclusions (II)
Very pronounced maternal mortalityVery pronounced maternal mortality Very pronounced maternal mortality Very pronounced maternal mortality differentials in South Africadifferentials in South Africa By race, ethnicity , provinceBy race, ethnicity , provincey , y , py , y , p Milder by level of educationMilder by level of education
Complex relationship with income and wealth:Complex relationship with income and wealth:p pp p middle income / average wealth have the highest middle income / average wealth have the highest
risksrisks Risks decline with distance from average, except Risks decline with distance from average, except
for the poorest for the poorest
Conclusions (III)Conclusions (III)Conclusions (III)Conclusions (III) Potential of population censuses for maternal Potential of population censuses for maternal p pp p
mortality:mortality: Exhaustive and unbiasedExhaustive and unbiased Large number of deaths and small confidence intervalsLarge number of deaths and small confidence intervals Large number of deaths, and small confidence intervalsLarge number of deaths, and small confidence intervals Potential for differential analysisPotential for differential analysis Potential for multivariate analysis at household levelPotential for multivariate analysis at household level
C b h i d b idi d ilC b h i d b idi d il Can be much improved by providing more details on Can be much improved by providing more details on causes of maternal deaths:causes of maternal deaths: Separate direct cause, indirect causes, external causesSeparate direct cause, indirect causes, external causesp , ,p , , Provide more details on timing of death (early pregnancy, Provide more details on timing of death (early pregnancy,
late pregnancy, delivery, postlate pregnancy, delivery, post--partum, postpartum, post--abortum)abortum) Question the actual “demographic” definition of MMRQuestion the actual “demographic” definition of MMRQuestion the actual demographic definition of MMRQuestion the actual demographic definition of MMR
South Africa, Rock Art