Date post: | 13-Jan-2016 |
Category: |
Documents |
Upload: | melissa-underwood |
View: | 216 times |
Download: | 1 times |
Lessons from Immpact: “Making sense of the
evidence”
Julia Hussein
Professor Wendy J Graham
www.immpact-international.org
Emerging evidence culture for policy decision-making
Evidence Values
Resources
Opinion-based decision making
Evidence-based decision making
Pressure on resources
0
100
200
300
400
500
600
700
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
Mat
ern
al d
eath
s p
er 1
0000
0 liv
e b
irth
s
Sri Lanka Thailand Malaysia
Honduras Egypt Matlab, Bangladesh
Bangladesh MM Survey 2001 India
“Success stories”
China
Immpact is the international research Initiative for Maternal
Mortality Programme Assessment
Sept 2002- to date
Goal: to improve the evidence-base for decision-makers on strategies to reduce maternal & newborn mortality
III. Communication: multi-faceted. e.g. high-profile networking,
12 policy briefs, >100 journal papers, 3 websites, toolkit, etc.
IV. Facilitating uptake of research outputs
II. Outputs:• New tools;• Evidence from major evaluations in 3 main countries + 5 more;• Strengthened research capacity• Launched technical advisory arm (Ipact)
Project cycle 2002- 2008
I. Aim: strengthen the evidence-base on mortality reduction strategies
Evaluations in three main focus countries at sub-national level
• Ghana: effects of free delivery care policy
• Indonesia: effects of village midwife programme
• Burkina Faso: effectiveness of skilled delivery care initiative
What can we reasonably expect of evidence?
0
20
40
60
80
100
2000 2020 2040 2060 2080 2100
% o
f m
ater
nal
mo
rtal
ity
in 2
000
?Trend from a “magic bullet”
?Secular trend with “development”
?Trend from evidence-informedstrategies
The good news - population coverage of “skilled attendants” increased at sub-national levels where:
• More personnel were provided at health centres: evidence from Burkina Faso
• Financial barriers were reduced: evidence from Ghana
• Midwives practised close to the community: evidence from Indonesia
……….BUT….
Immpact key message
Skilled Care at Delivery: the complete “package”
Skills to promote utilisation of
delivery care and to conduct normal
deliveries
Skills to provide basic emergency
obstetric care
Skills to provide comprehensive
emergency obstetric care
Referral
Skilled attendants
Community
Enabling environment
0
10
20
30
40
50
60
70
80
90
100
Poorest Poor Middle Rich Richest
Before fee exemption
After fee exemption
% of deliveries in health facilities by wealth quintiles
Utilisation of delivery care before & after fee exemption in Ghana
Serang and Pandeglang (2004-2006)
Wealth quartile
Poorest
Low middle
Upper middle
Richest
Source: Immpact data;
Relationship between maternal mortality & delivery with skilled attendants: Indonesia
Proportion attended by skilled
attendant
Maternal deaths per 100,000 live births (95% CI)
Yes 10% 2303 (1487-3292)
No 90% 541 (420-684)
Yes 17% 1218 (773-1830)
No 83% 278 (201-373)
Yes 33% 778 (541-1076)
No 67% 280 (195-388)
Yes 71% 257 (181-351)
No 29% 202 (107-334)
0
50
100
150
200
250
300
350
400
450
1990 1995 2000 2005 2010 2015
Ma
tern
al
de
ath
s p
er
10
0 0
00
liv
e b
irth
sMillennium Development Goal 5a* is off-track
MM Target
% deliveries with skilled attendants
100%
50%
0%
* 75% reduction in maternal mortality ratio from 1990-2005
“Quality facility deliveries”
“Ensuring that the care received is good quality will enable substantial additional benefits to be reaped in terms of fewer deaths to mothers& babies.” (p.50)
Importance of financial cover for catastrophic costs for maternal care: Indonesia
-1.0
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Cases (poorest to richest)
An
nu
al h
ou
seh
old
inco
me
(Rp
. mill
ion
)
Household income
Household income after catastrophic payments (without Askeskin)
Poverty level
Poverty level
Higher income
An example of a phased strategy for West JavaD
epth
Breadth
Banten - universalAttended delivery + catastrophic
care for all ($1.22 per capita plus investment)
Banten - possible transitionCatastrophic care + SBA in poorest 40% sub-districts +
small transport/incentive allowance ($0.47 or $0.92 per
capita with EmOC for all)
Public health spending $8 pc (Indonesia)Maternal Health $0.3 pc (Banten)
Potential of skilled birth attendants to reduce maternal and newborn mortality is contingent on effective coverage of a package of quality care at delivery (missed opportunity)
Phased options for overcoming implementation bottlenecks to effective coverage are context-specific (no one size fits all)
Implications
Context,
Context,
Context.
‘Science’ of quality improvement interventions
– adequate resources; – active engagement of health professionals; – sustained managerial focus; – multi-faceted interventions; – coordinated action throughout health system; – major investment in training; – availability of robust & timely monitoring.
Context-specific implementation package
Crucial translation step
Research priority-setting
Knowledge-generation & dissemination
Evidencetranslation
Policy-making processes
Source: Alliance for Health Policy and Systems Research. 2007.
Who should do translation of research evidence?
Knowledge brokers work at the interface between research organizations and their target audiences.
http://www.research-transfer.org
Knowledge brokers aim to provide evidence that is accessible, timely, credible and trusted, and packaged in user-friendly format, relevant to the local context.
Media
Advocates, civil society, NGOs, parliamentarians
Research institutions
Think tanks
Funding bodies
Government bodies
Demand for evidence is now more diverse
Researchers are from Venus.
Policy makers are from Mars.
Different worlds, different “communities”
Way forward for academies?
I. Robust descriptions of context to help share lessons
II. Evidence bases built around context
III. Stronger support for local-level data for local decision-makers
IV. More serious discussion of context-specific, phased & staged strategies
V. Strengthened translational skills to boost evidence uptake
0
100
200
300
400
500
600
700
800
900
1850 1870 1890 1910 1930 1950 1970 1990 2010
Year
Mat
erna
l Dea
ths
per 1
00,0
00 li
ve
birt
hsMaternal mortality trends: United Kingdom
“The great blot on public health
administration”Minister of Health, 1935
*