The critical role of pro-poor health systems in maternal

Health systems concepts and health
systems research
Barbara McPake,
Institute for International Health and
Queen Margaret University, Edinburgh
What is a health system?
• A health system consists of all the organizations,
institutions, resources and people whose
primary purpose is to improve health (WHO,
• Useful starting point but presents a static
framework, emphasises components
rather than relationships, and fails to
identify what is ‘systemic’ about this
system (the interconnectedness of its
• Has implications for how ‘health systems
strengthening’ (being used by global
health initiatives) has been conceptualised
and is being implemented.
• Marchal, Cavalli and Kegels (2009) find that these
GHIs are doing 3 kinds of things:
– Providing inputs or resources
– Reinforcing capacities that are directly related to
disease control programmes
– Integrating programme activities into general
health services
• ‘(These) HSS strategies are essentially a
means to deliver targeted interventions
more efficiently, rather than being
strategic and directed towards the root
causes of health system weaknesses.
• ‘Most current HSS strategies are in fact
selective, disease-specific interventions,
and their effects may undermine
progress towards the long-term goal of an
effective, high-quality, inclusive health
The effect of the global fund on the drug
distribution system in Uganda
Kyagonza, P. and McPake, B. (2007; unpublished)
• Increased investment in procurement of ART drugs
has not been followed by a proportionate increase in
investment in strengthening drug supply systems.
• Scaling up ART has affected the mainstream supply
chain for essential drugs and medical supplies
through the creation of a parallel supply chain for
ART drugs, and the precedence given to
management of ART drugs over that of other
essential drugs and medical supplies.
• So, not only do the building blocks seem
to lack something quite important in terms
of how they describe a health system, but
this seems to be reflected in how things
are actually being done and what
investments are being made, with negative
practical implications
• Health system = system whose output is
access to effective (preventative and
treatment) health interventions (or
activities whose primary purpose is to
promote, restore or maintain health)
Dynamic responses model of the health
Dynamic responses:
How people (‘users’ and
‘providers’) react and
interact in response to
formal structures and rules
De facto system:
De jure system:
Organisational structures
Intended incentives
Management procedures
Training courses
Services as
experienced by (poor)
For example: access;
The exclusion of the poor from health interventions
is systematic – some examples
• Worrall et al. (2005) reported differentials in the uptake
of malaria control interventions through a global
literature review
– preventive measures (coils, sprays etc.),
– ownership and use of bed nets
– choice of healthcare provider for treatment
• All used less by the poorest (measured through a
variety of indicators).
• The poorest were more likely to opt for self-treatment
and less likely to use private or higher level public
(Compiled by Riquelme and Thiede)
Socio-economic conditions impact on
access to TB care
• China’s 2003 National Household Health Survey
revealed that both breadth and depth of TB services
were comparatively low in poorer rural areas; both
receipt of care and affordability of TB services declined
with socio-economic position (Zhang et al., 2007).
• A study conducted in Malawi found that direct costs
associated with TB service (e.g. food and transport) as
well as indirect costs (e.g. work days lost), were
particularly significant for women and the poor (Kemp et
al., 2007).
• A study on access to India’s TB control programme
revealed that poor and socially marginalised patients
systematically received worse services (Singh et al.,
(Compiled by Riquelme and Thiede)
Systematic factors underpinning the
exclusion of the poor
• There are costs of accessing care, even when ‘free’
– the poor have more difficulty in covering these
• Non-financial resources including social capital can
also be involved in securing access to services –
the poor have more limited access to these
• Public health services constitute a significant
resource – people with power use it to capture an
unfair share
• As the de jure features of systems are changed to
try to counteract these forces, people use their
money, social capital and power to reassert the
status quo
Examples of systemic factors in maternal
The posting system
• Almost everywhere there is a serious imbalance of staffing
between rural and urban areas that excludes poor rural
dwellers from accessing services
• Failure of system that directs staff to posts is understood
• But alternatives are rarely developed and piecemeal
• In Kenya, ‘emergency hire’ programme recruited staff to
specific roles in rural areas – but intends to regularise
these new staff as public servants - the same strategies
that allow staff to relocate to urban areas will become
available to them
• The posting system entrenches interests that are more
powerful than the stated policy objective of equitable
access to health services
Constraints on migration
• In the early 2000s, staff were
haemorrhaging from low-income country
health systems, especially to the UK
• In Ghana nurse migration was constrained
by a bond (increased from C2m to C200m
in 2005), payable by anyone seeking a
qualification verification statement
• No similar measure taken to constrain
doctor migration which was higher in
proportionate terms
Divide between family planning and
MNC health services
• Common for these to be in separate Ministries or
branches of Ministries in South Asia
• Recognition that this causes overlapping
responsibilities for populations, duplication of effort
in some areas, gaps in service provision in others
• Attempts to integrate the functions founder on the
conflicts of interest between the different branches
• Family planning creates a stronger power base
because its outcomes are more measurable and it is
usually better funded externally, than maternal
An ethnography of two labour wards in
South Africa
• Policies implemented nominally but their
intentions ignored
– Clinical guidelines
– Name badges for nursing staff
– Suggestion boxes
• More complex or difficult changes,
particularly if they impacted on the culture
of the facility or challenged existing power
dynamics were ignored.
‘Rural allowances’ for staff in South
• Only professional nurses and not lower
grades of staff received the allowance
• High levels of tension in maternity wards
• Demotivated staff who did not get the
allowance but also demotivated staff that
did who felt guilty, embarrassed or
• ‘You get the allowance, you do the work’
• Attempts by bilaterals and multilaterals in last decade to
construct aid as a contract: system of rewards and
penalties for good and bad performance
• SWAp in Uganda – funding to be provided in response to
delivery of agreed undertakings
• The achievement of a satisfactory performance rating
was facilitated by the agreeing of undertakings that were
under-demanding, vaguely formulated and lacking
quantitative benchmarks against which progress could
be measured.
• However, even when poor performance was readily
observable, penalties failed to be applied by donors.
Dynamic responses model of the health
Dynamic responses:
How people (‘users’ and
‘providers’) react and
interact in response to
formal structures and rules
De facto system:
De jure system:
Organisational structures
Intended incentives
Management procedures
Training courses
Services as
experienced by (poor)
For example: access;
• It is also possible to use the model to
better understand how to develop
strategies to tackle health system
problems; and to better understand the
role of health systems research
• No ‘solutions’ – changes to the de jure system
have less than fully predictable impacts on the
de facto system – context is critical
• System interventions are normally better
understood as ‘tools’ than solutions in
• Tools have to be appropriate to the task in hand
and can be refined as the fit between tool and
task is better understood.
• They also have to be applied well, requiring
capacities of the user and benefiting from the
user learning as the task progresses.
Why performance-based contracting failed in Uganda
Freddie Ssengooba , Barbara McPake and Natasha Palmer
• World Bank and MoH implemented a bonus system
for PNFP providers and undertook a controlled trial
to evaluate it
The World Bank evaluation: a ‘black
Their conclusion:
• ‘..assignment to the performance-based
bonus scheme has not had a systematic
or discernible impact on the production of
health care services provided by PNFP
Our evaluation
What did it really
consist of?
Design features
Who came into contact with
the intervention?
How did they react?
How did they influence
What chains of effects were
initiated and how was
hospital performance
What has been
measured? What has
Key findings
• 2-3 members of the hospital management team
given a few hours in a one day meeting to choose
service targets
• Implementers changed the rules; refused to allow
managers to change the targets for the second year
• The reliability of measured output volumes
compromised by ad hoc adjustments to programme
• Staff in hospitals that received bonuses frustrated
when managers didn’t use them in an agreed,
transparent manner
• Furious control group: ‘YYY hospital got 10 million and
yet zzz hospital is doing far much better. It’s frustrating
and lost meaning – done better but no bonus'
Insight into the development of strategy
• Pay attention to process
• Monitor, manage, adapt
• Sweeping conclusions – PFP ‘good’ or
‘bad’ cannot be arrived at on the basis of
one or a few case studies. PFP is never
the same thing in two places
• See PFP as a tool, learn, improve, adapt,
and compare to alternative tools aiming at
the same outcomes
Insight into health systems research
• Avoid the black box – don’t know what the
intervention is or why outcomes
• Ask the right questions – avoid questions that
demand sweeping conclusions from a limited
evidence base
• Use research to support monitoring, managing,
adapting to context
• Over time, build sufficient case study evidence
for a picture of which tools most promising in
which kinds of context
• Current ‘health systems strengthening’ investments
may be misdirected – in part because they misconceptualise the health system
• Effective interventions will tackle the systemic
causes of the exclusion of the poor – by tackling
incentives and attempts to manipulate incentives;
challenging distribution of control over resources;
not just by making inputs available
• Effective health systems strategies and health
systems research will use and test interventions as
tools rather than implement or reject as solutions in
their own right.

similar documents