Health financing in post conflict settings

Funded by

Health financing in post conflict settings

Barbara McPake Nossal Institute for Global Health

School of Population and Global Health University of Melbourne

Funded by

Health financing debate in LMICs -summary of main issues

• Since late 1980s, debate about the mechanisms by which funds flow from individuals and households to health service providers

• taxation system - can fund use of public or private health providers

• insurance systems - public and private

• out-of-pocket payments - in public and private sectors

• Clear consensus from a body of research that out-of-pocket payments significantly deter use of health care where important

• Further conclusion that out of pocket payment (% in total health financing) and ‘catastrophic’ payment (incidence) strongly correlated

Funded by

• However, early evidence that ‘user fees plus quality improvement’ can maintain or increase use of health care

• Raises tension between effects of intervention on demand and supply sides and their inter-relationships

• Also raises issues of distribution of impacts - unlikely to be uniform change in use of care across population groups

• Measures to target groups likely to be most negatively impacted - for example exemption systems have largely worked poorly though some experience of funded exemption has been better.

Funded by

• Generating sufficient resources to provide effective services may not be feasible through the tax system -either for economic or political reasons

• Insurance mechanisms may better protect and stabilisehealth budgets

• Public insurance systems operate similarly to tax - limits to fiscal space and politics constrain them

• Private (voluntary) insurance systems exclude important population groups and increase inequities - although they may work well for large sections of the population

• Community based insurance systems focused on poorer populations require subsidy; still often exclude the poorest

Funded by

• More attention in last decade to the need to support the ‘supply side’ while removing constraints to the ‘demand side’

• Recognises constraints to subsidy levels by focusing subsidies on target services - for example maternal and child health

• Large numbers of countries have aimed to remove out of pocket payments while channelling additional resources to compensate for the loss of revenues

• Mixed results - difficult to channel resources in ways that generate incentives for effective care delivery, but some successes

Funded by

ReBuild program

• 6 year 6 partner research program funded by UK DFID

• Partner institutions in UK, Cambodia, Uganda, Sierra Leone and Zimbabwe

• ‘Path dependency’ idea at centre of design. What is possible with respect to policy development in post conflict period (short post conflict lens: N. Uganda and Zimbabwe)? What is the long term impact of changes made in the post conflict period (long post conflict lens: Cambodia and Sierra Leone)

Funded by

• Much stronger emphasis on the role of aid in these settings

• manage transition from humanitarian to development aid

• strengthening government stewardship and capacity

• coordination

• impact of Global Health Initiatives

Funded by

• Equity impacts of conflict create some unusual distributional contexts

• Conflict may have ‘levelled down’ the economic situation of the population

• May be strong regional patterns in distribution of impacts of conflict

• Pre-conflict inequities may differ from post-conflict inequities and distributional questions may need to be considered from both perspectives

• Free health care may be part of a post-conflict citizenship rights settlement

Funded by

• Context of health care provision - generally larger role for NGOs than in stable states

• Provision transition expected to accompany aid transition and financing transition - shift of service delivery to public sector

• But in practice, two main trends in health financing post-conflict are an increasing reliance on informal payments and donor funding

Funded by

• Strong parallels between user fee analysis in conflict affected settings and LMICs more generally

• Greater emphasis on maintaining whatever is working and has survived the extremes of conflict conditions - often fee-paying NGO facilities

• Afghanistan has provided a case study of successful exemption policy in a post-conflict setting

• Cambodia has demonstrated positive impacts of health equity funds - funded exemption system

• Rwanda has had most success with highly-subsidisedcommunity based insurance mechanism

Funded by

• Significant gaps in the literature

• Weak literature methodologically - few papers proceed on the basis of clear methodology; piecemeal and small scale studies

• Weak definitions of ‘fragile and conflict affected states’. Grouping highly diverse contexts facing very different challenges without clear basis of differentiation

• Focus overwhelmingly on role of donors - much less attention to role of government policies

• Emphasis on immediate post conflict period - few studies with longer perspective

Funded by

ReBuild work in this area aimed to use our four case studies to strengthen the understanding of the impacts of health financing policies on poor people in post conflict contexts

4 very different contexts

Different sets of health financing policies and changes

Different data opportunities

2 case studies (Sierra Leone and Uganda) illustrate these issues – both unfinished work in progress.

Funded by

• Current cost recovery scheme introduced 2006 - flat fee charged for all health services except medicines for which full cost recovery fee applies

• National guidelines to exempt children, adults over 60, pregnant/lactating women and disabled

• Poorly implemented - group too large for resources at facility level; few in fact receive waivers

Funded by

• Free Health Care Initiative April 2010

• Children <5; pregnant and lactating women - free care -funded by government and donors

• Range of health sector reforms - medicine supply management, human resources management

• In first few months, use of health care by target groups increased sharply, but then gradually declined

• Decline associated with shortages of medicines, informal charges

Funded by

• Study seeks to:

• quantify impact of FHCI on child and maternal health service use and out of pocket payment

• for children: a regression discontinuity design using 2011 Sierra Leone Integrated Household Survey (SLIHS)

• for mothers: a time-trend adjusted before-after estimation approach using 2013 Sierra Leone Demographic and Health Survey (DHS)

Funded by

• SLIHS - nationally representative household survey 6800 households. Study uses subsample of children 0-120 months. Data on out of pocket payment, utilisation (used outpatient care in two week period preceding interview) in public and private facilities but excluding NGO facilities.

• DHS - 16,658 women of reproductive age, most recent child birth over 5 year recall period and services received -information exists on births occurring before and after FHCI

Funded by

• Regression Discontinuity Design - exploits discontinuity in entitlement to free health care in relation to child age. If FHCI effective, a trend discontinuity at 60 months expected.

• However, not all children < 60 months succeed in receiving free health care and some non-eligible children will have done, so => ‘fuzzy RDD’

• Time trend adjusted before-after estimation approach - 4 binary outcome variables compared - 4+ ANC visits; delivery in public facility; vit A supplementation up to 2 months; DPT+ vaccination in first year.

Funded by

Descriptive results: children

Funded by

Funded by

Funded by

Funded byDescriptive results: women

Funded by

• Results statistically significant (though small) for simple comparison

• After time trends and interaction terms included, no longer significant for facility births, delivery with skilled health workers or 4 ANC visits for all facilities

• But significant for ANC, PNC, vit A and DPT+ significant for public facilities and fairly substantial for PNC, vit A and DPT+

• Effects larger and more significant in rural areas

Funded by

Before

After

Funded by

Funded by

• No clear impact for children - might relate to lack of clarity about which children were exempted

• DHS suggests increase in service use for children but may be longer term trend as there appears to be for women

• Statistically significant increases in service use for women, substantial for some indicators and for rural areas

• Overall disappointing impact may relate to continued costs, medicine shortages, targeting errors, insufficient supply side reforms

Funded by

Self-reported health, health utilisation, and food consumption in the post IDP camp period in Uganda

• Fu-Min Tseng, Tim Ensor, Ijeoma Edoka, Robert Bataringaya, Sarah Ssali and Barbara McPake

Funded by

• Armed conflict Northern Uganda from early 1990s

• By 2005, 2m internally displaced persons (IDPs) including 90-95% of the population of Acholiland

• Government declared it safe to leave camps in late 2006

• By 2009, IDP population had fallen to 450,000

Funded by

• As people return from camps…

• reduction in exposure to camp specific risks including infectious disease, stresses of displacement, lack of life choices

• access to health services may worsen as camp services inaccessible

• need to re-establish livelihoods, planting cycles, housing and land rights - basic services including health may be secondary

Funded by

• Study investigates changes in health indicators, healthcare utilisation and food consumption of people living in districts highly affected by internal displacement over the period in which most returned

• Analyses the Uganda National Household Surveys of 2005/6 and 2009/10 using difference in difference method

• ‘Treatment group’ = 3 districts most exposed to conflict; excluded = 9 districts partially exposed to conflict; ‘control group’ = remaining districts not exposed to conflict

• 5 outcomes - self-reported illness incidence in past 30 days; productive day loss caused by illness in last 30 days; visits to health facilities in the past 30 days; health expenditure in last 30 days; food consumption in the past 7 days

Funded by

• Descriptive statistics

Funded by

Funded by

• No significant evidence that self-reported health and frequency of healthcare utilisation changed after IDPs returned, but evidence of significant increase in food expenditure

• Insignificant change in self-reported health may balance counteracting effects of fewer camp related risks but more limited availability of infrastructure and services.

• Shift from formal private to informal care - probably reflects differing range of options.

Funded by

Overall conclusions

• Literature on health financing in post conflict contexts is limited

• Post conflict contexts are varied; policies diverse and data opportunities variable, so 4 case studies, even when fully complete will only add marginally

• Many of the issues appear similar to those in other LMICs

• Others specific to particular conflict related phenomena such as IDP return

• Any level of generalisation will have to wait.