The Experience of Ghana in Implementing a User Fee Exemption Policy to Provide Free Delivery Care

B

The Experience of Ghana in Implementing a User Fee

Exemption Policy to Provide Free Delivery Care

Sophie Witter,a Daniel Kojo Arhinful,b Anthony Kusi,c Sawudatu Zakariah-Akotod

a Research Fellow, IMMPACT, University of Aberdeen, Aberdeen, Scotland.E-mail: [email protected]

b Research Fellow, Noguchi Memorial Institute of Medical Research, College ofHealth Sciences, University of Ghana, Accra, Ghana

c Researcher, IMMPACT, Noguchi Memorial Institute of Medical Research, Universityof Ghana, Accra, Ghana

d Researcher, IMMPACT, Noguchi Memorial Institute of Medical Research, Universityof Ghana, Accra, Ghana

Abstract: In resource-poor countries, the high cost of user fees for deliveries limits access to skilledattendance, and contributes to maternal and neonatal mortality and the impoverishment ofvulnerable households. A growing number of countries are experimenting with different approachesto tackling financial barriers to maternal health care. This paper describes an innovative schemeintroduced in Ghana in 2003 to exempt all pregnant women from payments for delivery, in whichpublic, mission and private providers could claim back lost user fee revenues, according to an agreedtariff. The paper presents part of the findings of an evaluation of the policy based on interviews with65 key informants in the health system at national, regional, district and facility level, includingpolicymakers, managers and providers. The exemption mechanism was well accepted andappropriate, but there were important problems with disbursing and sustaining the funding, andwith budgeting and management. Staff workloads increased as more women attended, and levels ofcompensation for services and staff were important to the scheme’s acceptance. At the end of 2005,a national health insurance scheme, intended to include full maternal health care cover, wasstarting up in Ghana, and it was not yet clear how the exemptions scheme would fit into it.A2007 Reproductive Health Matters. All rights reserved.

Keywords: delivery care, user fees, health financing, Ghana

ARRIERS to women delivering in healthfacilities include poor quality services,negative attitudes of health workers, dis-

tance to facilities and cultural preferences.1,2

Financial barriers have also been important,but there have been relatively few studies ofwhether and how changing fees influenceswomen’s uptake of maternity services and atwhat cost.

Most developing countries currently rely onuser fees to fund part of their health care. Asurvey of 16 African countries found thatuser fees contributed an average of 5% (range

1–20%) of recurrent health service expendi-tures.3 Although small, this revenue is impor-tant, given the problems of low governmentbudgets, unreliable donor aid and poor resourceallocation. There is now widespread recognitionof the problems caused by user fees, both interms of inequitable access and inefficiency.4

However, the debate about whether to abolishfees totally and how to replace the lost revenueon a sustainable basis continues. Current dis-cussions focus on the complementary measureswhich will be needed to ensure that removal ofuser fees is carried out successfully.5,6

A 2007 Reproductive Health Matters.All rights reserved.

Reproductive Health Matters 2007;15(30):61–710968-8080/06 $ – see front matterPII: S0968-8080 (07 )30325-Xwww.rhm-elsevier.com www.rhmjournal.org.uk

61

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

If full removal of user fees is considered unten-able, there is a case for partial removal of fees forspecific services such as maternity care, whichhave high social priority. Maternity care exemp-tions would be expected to contribute to reducingmaternal mortality (by increasing supervised deliv-eries) and reducing the impoverishing effect onhouseholds of high and unpredictable paymentsfor deliveries (especially complicated deliveries).

This paper describes an innovative schemeintroduced recently in Ghana to exempt allwomen from delivery fees and the findings fromthe first stage of an evaluation of it, using keyinformant interviews.

BackgroundGhana has persistently high maternal mortalityratios, estimated to range from 214 to 800 per100,000 live births.7 Ghana also has growingsocial inequalities for this indicator, with rates ofskilled attendance either stagnant or decliningfor poorer women.8 While deliveries with healthprofessionals rose from 85% to 90% from 1993to 2003 for the richest quintile, according toDemographic and Health Survey data, deliverieswith health professionals for the poorest quin-tile dropped from 25% to 19%. Nationally, 45% ofbirths were attended by a medical practitioner(79% in urban areas, 33% in rural); 31% by tra-ditional birth attendants (TBAs) and 25% wereunsupervised. There were also significant regionalvariations. The three northern regions have thehighest levels of poverty and maternal mortalityand the lowest levels of supervised deliveries.9

In the Ghana health system, basic obstetric andantenatal care are provided by health centres,health posts, mission clinics and private mid-wifery homes. Each health centre or post servesa population of approximately 20,000. In therural areas, TBAs continue to carry out deliveries,though they are trained to refer more complexcases. Comprehensive emergency obstetric careis available from district hospitals and regionalhospitals, as well as national referral hospitals.Most are run by the Ghana Health Services,though the mission sector plays a significantrole, especially in more remote regions.7 All careis paid for, unless the service is exempt or theperson has private or public health insurance,though user fees are subsidised by public inputsinto the services.

62

Financial barriers are believed to be one of themost important constraints to seeking skilledcare during delivery in Ghana.7 A study costingmaternal health care in one district in 1999found cost recovery rates of between 152% fordeliveries and 211% for caesareans in missionhospitals, but did not shed light on affordabilityrelative to women’s income.10 Problems such asunder-funding of exemptions from user fees ingeneral have also been found,11–13 which havemeant that exemptions are available in theorybut not always in practice if the provider is notreimbursed for lost income.

The Government of Ghana introduced exemp-tions from delivery fees in September 2003 inthe four most deprived regions of the country,which in April 2005 was extended (withoutformal evaluation) to the remaining six regions.The aim was to reduce financial barriers tousing maternity services to help reduce maternaland perinatal mortality and contribute to pov-erty reduction.14

The policy was funded through HighlyIndebted Poor Country (HIPC) debt relief funds,which were channelled to the districts toreimburse public, mission and private facilitiesaccording to the number and type of deliveriesthey attended monthly. A tariff was approved bythe Ministry of Health which set reimbursementrates according to type of delivery (e.g. normal,assisted or caesarean section) and type of facil-ity. Mission and private facilities were reim-bursed at a higher rate, because they did notreceive public subsidies.14 Women would thenonly have to bear the costs of reaching facilities.

In 2005, an evaluation by IMMPACT of thefree delivery policy was initiated.15 The firststage was based on a series of key informantinterviews to establish the state of implementa-tion of the exemption policy and seek the viewsof stakeholders. These findings fed into theoverall evaluation, which included tracking offinance flows, a household survey, a survey ofhealth workers and TBAs, qualitative investiga-tions in communities and among providers, andquality of care assessments. These were com-pleted in 2006.16

The policy of user fee exemptions is now tosome extent being superseded by a new NationalHealth Insurance system, which has reachedeffective coverage of just under 20% of thepopulation.17 This provides protection against

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

user fee costs for a wide range of health services,including maternal health care, for formal sectorworkers, those in the informal sector who takeup voluntary coverage, children of members,pensioners and a small category of ‘‘indigents’’.However, the delivery fee exemption policyremains in place, officially, at this writing.

MethodsFor the evaluation two regions were chosen:one of the four regions to join the exemptionscheme first (Central) and the other from themore recent wave (Volta). Six focus districtsfrom each region (roughly half of the districts ineach region) were chosen, and matched for pop-ulation size, poverty status, urban profile andhealth infrastructure.15

The 65 key informant interviews were con-ducted from September–December 2005. Thekey informants included key stakeholders in theMinistry of Health, Ghana Health Services, mis-sion sector and development partners9; RegionalDirectors of Health Services, Senior MedicalOfficers, regional hospital directors and admin-istrators, and regional accountants7; and DistrictDirectors of Health Services and senior publichealth staff, District Assembly staff and accoun-tants.32 A sample of in-charges, matrons andsenior facility staff were also interviewed atfacility level.16 The facilities were selected by aprocess of stratified random sampling, to repre-sent each of the six focus districts and to covera range of facility types, including district hos-pitals, health centres, mission hospitals, privatematernity homes and mission clinics.

The authors carried out the interviews usinga semi-structured questionnaire. The questionscovered:

� at national level: perceptions of the programme,successes and failures, implementation, allo-cation of resources, disbursement mechanisms,sustainability and future funding options, pri-ority of the programme, impact on servicesand staff, interaction with health insurancein future and suggestions for future changes;� at regional level: similar questions but more

detail about the process of establishing theprogramme, current implementation, adequacyof funding, degree of dissemination in the com-munity and impact on quality of services;

� at district and facility levels: questions wereadded about dates of operation, whether patientswere making contributions of any kind, delaysin funding, whether funding can be vired toother uses, appropriateness of reimbursementtariffs, effect on defaulter and referral rates,whether reimbursement covers loss of user feeincome, whether revenue is shared with staffin facilities, impact on private midwives (whowere meant to be included in the scheme) andTBAs (who were not included but whose busi-ness was likely to be affected), whether thescheme was efficient to manage, and otherprogrammes in the district which might affectsupervised delivery rates.

The responses were analysed by topic, leveland region and triangulated, where possible,with figures available from facility records orannual reports. The findings were initially pre-sented in an internal report for IMMPACT18 andthen in a policy brief.19

For a qualitative process, 65 is a relativelylarge number and should give findings that arerepresentative for those regions (though not nec-essarily nationally). Discussions at national levelsuggest that the findings in these two regionswere not exceptional. Findings from such qual-itative research need to be set in the context ofother data gathering tools, and this is done tosome extent in the discussion, comparing find-ings from interviews with other evaluation com-ponent findings.

FindingsOverall perceptionsThere was generally a positive reception of thefree delivery policy as an effective approach toan important problem, which allowed for earlyreporting and better handling of complications.Stakeholders within the health system found theadministration of the scheme manageable andthere was reasonable consistency in the inter-pretation of the policy. These informants wereonly able to provide weak evidence on commu-nity perceptions, but their view was that commu-nities were both informed and enthusiastic. Thescheme was publicised through meetings withtraditional rulers, broadcasts and durbars (publicmeetings) at churches and other public places.The trends in utilisation – both up, when the

63

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

scheme was functioning, and down, when itstopped - suggested that women were cost-awareand informed about the charging regime.

Availability and adequacy of fundingThe main constraints, which were causing con-cern at all levels, were the shortfalls and unpre-dictability of funding. Funds were issued at thestart of the financial year, without guidance formanagers as to how they had been calculated,how long they should last or when they would bereplenished. The funds were not adequate for a fullyear and further instalments were expected butnot received until the next financial year.

Managers were used to unpredictable, erraticfunding – the norm with previous exemptions,such as for antenatal care, which is routinely inarrears. However, the financial implications offunding not arriving for deliveries were far moreserious because they were relatively expensiveprocedures that had brought in much of thefacility revenue.

The failure to reimburse adequately andpromptly had negative effects at all levels ofthe system. Patients, having been told theywould receive free services, were reported to beangry when they were asked to pay, and staffwere suspected of taking a cut of the funds.Facility staff wondered if the funds had beensiphoned off higher up the system. District andregional health managers were caught betweenfacilities accumulating debts and the need topersist with the policy.

‘‘It is difficult to charge now, as people will thinkyou are cheating them. But what do I do when Ihave no drugs left?’’

As a consequence of the funding shortfall, threeout of six districts visited in Central Region hadreverted to charging, and the others were closeto joining them. The regional hospital claimed tohave substantial amounts owing from the districtfunds for referral care provided free. Having pro-vided items on credit, the regional medical storefound it could not be reimbursed by districtswhose exemption funding had run out. Analysisof funds received in Central in 2005 comparedwith expected delivery numbers and unit costssuggested that the funds would be adequate onlyfor one-third of the year.

The situation in Volta Region was even worse.The scheme only began in May or June 2005 in

64

most districts, and many had run out of fundsby August or even earlier. Facilities were stillsupposed to be providing free services, but pay-ments were in arrears, it was not clear whetherfurther funds would arrive, and some facilitiesclaimed they had not yet received any reim-bursements at all. Four out of six districts visitedhad run out of funds by early October 2005.

Management problemsA number of management failures were describedby national level informants to explain the irreg-ular funding flows. The complexity of fundingchannels and multiple actors meant that respon-sibility for the policy was unclear and monitoringweak. In Central Region, for example, the firstfunding flows in late 2003–early 2004 passedthrough the District Assemblies (local govern-ment bodies). The second tranche in early 2005passed via the Ghana Health Services in bothregions, but the national level claimed not tohave been informed of the funding allocationsand hence had not issued instructions for theiruse. In Volta, there was confusion about thefunding channels stated in the national guide-lines and the actual channel through which fundsarrived. In addition, guidelines for monitoringwere not enforced. At national level, oversightinformation on numbers of deliveries carried outand delivery types and reimbursement amountswere not available. In Volta, concerns wereexpressed that there was no additional fundingfor the administration of the scheme.

Allocation of funds to districtsThe 2004 national guidelines applied a per capitaallocation of 61,510 per capita (US$0.17) for thepoorer regions (including Central), and a lowerrate for the relatively richer regions (includ-ing Volta) of 61,354 (US$0.15). However, in prac-tice, there seemed to be some local adjustment.In Central Region, districts with more facilitieshad received a higher allocation. Despite this,they had exhausted their funds more quickly.An agreed system for funding cross-borderpatient flows was also needed. This was par-ticularly relevant when the policy was limitedto certain regions; women were said to becoming from Accra to deliver in Central, forexample, when Accra was not yet included inthe policy.

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

In Volta, funds were allocated on a per capitabasis to the districts, with no variation for facil-ity number or size.

‘‘The whole design was poor from the beginning:they didn’t ask the regions how much theyneeded. We in the region developed our owncriteria, based on past utilisation, but we havenever received any funds since then.’’

Interpretation of the packageIn Central Region, the understanding was thatthe exemption covered deliveries only, but notcomplications during pregnancy or post-partum.Some suggested that the definition should bebroadened to include pregnancy-related compli-cations and also transport costs. In Volta, theregional director also wanted to provide a moreinclusive package, but at facility level, the inter-pretation was restricted to deliveries.

Reimbursement ratesDespite there being a national tariff setting outthe reimbursement rates by type of delivery andfacility, reimbursement rates were found notto be the same across the two regions. In oneregion, normal deliveries were paid at a rela-tively generous rate, but complications andcaesarean sections were paid at less than thenational rate. In the other, the opposite approachhad been taken, with discontent about the ratesfor normal deliveries, but top-end rates beingpaid for complicated cases. Satisfaction amongstproviders correlated with these differences.

Some mission in-charges in Central were dis-satisfied with the payments for more complexdeliveries, compared to what they had chargedwomen before. Others were receiving repaymentrates well above previously charged user fees.In addition, at least while funds were availablefor the policy, the facilities did not have toworry about chasing the 25–50% of women whowere struggling or unable to pay (particularlyfor caesareans).

Volta drew up its own reimbursement rates,based on the national guidelines, but also includeda provision to pay trained TBAs for deliveries ata lower rate. Of the six districts visited by theresearch team, only one was including the TBAs,however. Facilities were billing according tomaterials used, rather than at fixed rates. Therewas at least one report of inappropriate billing for

procedures carried out before the policy wasin effect, emphasising the need for vetting andauditing. Key informants in Volta cited lowerbut still substantial defaulter rates prior to thepolicy, higher for hospitals than health centres,which find it easier to enforce payment, beingcloser to households.

Impact on staffThe national guidelines made no mention ofincentive payments to staff, but Central Regionhad agreed a small payment of 20,000 cedis(US$2) for midwives and ancillary staff perdelivery. This was justified by the increasedworkload they faced, as well as lost income fromselling small items to women and getting dona-tions from them. Staff and managers appreci-ated the regularity of the monthly paymentsfrom the district when the funding was avail-able, and the fact that they did not have tocontend with women who could not pay, whowere often detained and whom health staff oftenhelped with their bills. In Volta (and in some ofthe mission facilities in Central), staff receivedno direct payments relating to the exemptionspolicy. They cited income lost from petty salesto women, though it is possible that this wascontinuing in places.

Quality of careChanges to quality of care were measured byother evaluation components, but the key infor-mant interviews probed perceptions of changes.The view in Central was that quality was im-proved by the more reliable funding flows forservices, while the policy was adequately funded,even though staff workloads but not staffinglevels were increased. In general, the brain drainand issues of retention were the biggest headachefaced by managers, though this was a widerissue, not directly related to the exemptionsscheme. Tiredness and overload were reported,but most in-charges felt they were still able tocope and had not reached breakdown levels.

Attitudes were less positive in Volta. The stafffelt that quality of care was more or less thesame, and that workloads were too heavy beforeand had not been improved. The scheme had notadded sufficient extra resources to have hadany beneficial effect. Some dissatisfaction wasexpressed with the quality of services, e.g. withpoor use of partograms, and negative attitudes

65

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

of midwives towards women, but these were notseen as affected by the scheme.

Informal paymentsInformal payments by women, an importantfeature of the health system prior to exemptions,were reported to be diminished or removed inCentral Region, where exemptions had been inplace for longer. In Volta, which was only juststarting the scheme and where financial prob-lems were already apparent, evidence on infor-mal payments was mixed.

ReferralsGiven the shortage of medical and nursing staffin Ghana, referrals are often the result of theabsence of a key member of staff or of key sup-plies, such as blood. Under the exemptions policy,women presenting at health centres and districthospitals were eligible for delivery fee exemp-tions, but attendance at regional hospitals requireda referral to qualify. However, there was no formalsystem for splitting the reimbursement paymentsbetween facilities when women were referred.This led to concerns that lower level facilitieswould fail to refer women at appropriate timesin order not to lose the income from their case, asthe facility which finally carried out the deliverygot the full reimbursement. This was denied byhealth staff, who pointed to maternal death auditsas one way of monitoring this tendency. In prac-tice, some degree of flexibility about reimburse-ment was found, such that direct costs such astransport, incurred by lower level facilities, couldbe claimed back, even if the delivery eventuallytook place elsewhere.

In Central, exemption funds were held bydistricts, and the regional hospital was experi-encing difficulties getting reimbursements fromthe districts. In Volta, the system operated dif-ferently, and 5% of funds were top-sliced to aseparate account in advance to pay for referrals.At the time of the research, this fund was stillin credit.

Impact on mission and private sectorsThe mission sector is increasingly being incor-porated into the public health system on a parwith public facilities. Thus, mission hospitalsand clinics were participating in the exemptionspolicy, though some smaller clinics complainedthey had not been well informed. Private midwives

66

were also participating, but their numbers werelimited. In some districts in Volta, mission facil-ities had reportedly opted not to join the schemeas they were dissatisfied with the repaymentrates offered.

Prior to the exemptions policy, many missionfacilities operated a ‘‘poor and sick fund’’, whichraised funds locally and internationally to payfor those deemed unable to pay. This used tocover about 15–20% of deliveries, according tonational level mission informants. If new schemesundermine the old and are not sustained, therewill be a net loss to society.

Changes to uptake of services and outcomesPerceptions of increased utilisation were triangu-lated with routine reports, where available. Theseshowed different patterns in different districts.For example, in one district in Central, skilledattendance rates, including deliveries by trained(but not untrained) TBAs, had remained con-stant, but with evidence of a switch to facility-based deliveries. In another district, a big increasepreceded the policy and was probably linked toan increase in TBA training. In others, the policyappeared to be linked to increases in facility-based deliveries.

The exemption scheme was felt to have hada very beneficial effect in terms of encourag-ing women to come in early, so that complica-tions were detected and better managed, savinglives. This was reflected in the increase in morecomplex interventions. In Gomoa district, forexample, there were 86 caesarean sections in2002; 138 in 2003 and 190 in 2004 (or 1.4%,2% and 2.6% of supervised deliveries in thoserespective years). The district hospital in Abura/Asebu Kwamankese also reported a doubling ofnumbers of caesareans. For Central region as awhole, 3.8% of supervised deliveries were cae-sareans in 2004. The regional reproductive healthreport showed a declining trend in facility-basedmaternal mortality: 450 per 100,000 in 2001,dropping to 206 in 2002, 159 in 2003 and 134in 2004.

Given the short period of implementation, itwas harder to assess changes in utilisation inVolta. Informants reported substantial increasesin utilisation (it had doubled, according to theregional director). For specific facilities though,the picture was more mixed. For a private mid-wife located near a district hospital, it meant the

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

halving of her business, as women could accessfree services at a better-equipped facility. In onedistrict where funds ran out at the end of June,utilisation, which had increased, was reported tobe on the decline again.

Recommendations by informantsSome informants suggested broadening theapproach to cover all health care in pregnancyto include, for example, malaria treatment forpregnant women and post-partum care. Otherssuggested that some transport funds should beadded to enable women in remote areas to bene-fit. However, the main concern was the sus-tainability of the scheme, and to that end,respondents suggested reducing overall costs bytargeting only poor households, targeting onlypoor regions, only reimbursing deliveries athealth centres or, conversely, only reimbursingcomplicated deliveries (which are least affordablefor households), and limiting the number ofeligible pregnancies for exemption to two.

‘‘It can end up like the other policies, wherebythe first two or three chunks of money come, andlater it takes years.’’

‘‘Though some will see it as discriminatory, thefact is that some can pay and they must pay tohelp the system.’’

A few informants felt that services should inprinciple not be free, that Ghana has beenmoving from a free health service throughvarious stages of cost recovery towards healthinsurance, that it gives the wrong message tooffer totally free services, and that maybe somesmall co-payment should have been included.It was observed that people pay large amountsfor other services, such as prayer camps or scansat private facilities in early pregnancy, and donot value anything they get free.

Relationship to the National HealthInsurance SchemeAt the end of 2005, a national health insurancescheme was starting up in Ghana.20 It was notclear to most stakeholders how the exemptionsschemewould fit into this new initiativewhich, intime, was intended to provide fullmaternal healthcare cover. Formal sector workers were auto-matically enrolled, but coverage for the informal

sector (the majority of Ghanaians) was voluntaryand at low levels. In principle, exemption fundscould be re-routed in future to provide free orsubsidised cards to pregnant women, but detailsof this were still under discussion.

Many of the issues highlighted by the reviewof the exemptions programme are relevant to thehealth insurance scheme. For example, the mis-sion sector was concerned that reimbursementrates under health insurance would be low andthat quality would be compromised, leading to apreference for fee-paying patients. Large increasesin attendance were already being reported by mis-sion facilities in some areas where the nationalhealth insurance was functional.

The development of national health insurancecomplicated the process of raising funds and bud-geting for exemptions, as there were unrealisticexpectations about how quickly health insurancecould take over this social protection mechanism.The build-up of formal sector health insurancefunds, based on contributions from employeesand a levy on VAT payments, also made it harderfor the Ministry of Health to argue for continuedHIPC funds to pay for delivery exemptions.

DiscussionGhana’s experience of this scheme is important,given that a growing number of countries areexperimenting with different approaches totackling financial barriers to maternal healthcare. Nepal, for example, is piloting a policy thatcombines subsidies for all deliveries with freedeliveries and assistance with costs such as trans-port for women in poorer regions.21,22 There isexperience in Yunnan, China, and Bangladeshin using vouchers that entitle poor women tofree delivery services.23 In Bolivia, a social insur-ance scheme, which provides free care for preg-nant women and under-fives, has increasedutilisation, especially by the poor.24 In Senegal,an evaluation is being finalised of a scheme toprovide free deliveries and caesareans in allregions outside the capital.25

The decision in Ghana to cover all deliveries,initially within poorer regions and later nation-ally, reflects awareness of the documented dif-ficulties of individual targeting,26 but it alsoincreased the cost of the scheme. It is debatablewhether the decision to extend the scheme to theremaining six regions so quickly was appropriate,

67

GIACOMOPIROZZI/PA

NOSPICTURES

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

given funding constraints and the fact thatthe first stage had not yet been evaluated. Theevaluation of financial flows found that thepolicy was under-funded by 34% in 2004, rising to73% in 2005 when all ten regions were covered.27

Analysis of utilisation changes from the house-hold survey, comparing the 18 months beforeintroduction with the 18 months afterward, foundan increase of around 12% in women deliveringin facilities in Central Region but not in Volta,presumably due to the short period of imple-mentation. It also found significant reductionsin mean out-of-pocket payment by patients fordelivery care (direct and indirect payments) athealth facilities, both for spontaneous vaginaldelivery and caesarean section. The total pay-ments for caesareans fell by 21.6% and normaldeliveries in health facilities by 18.9%, and therewas a reduction in the number of householdshaving to make catastrophic payments for deliv-eries.16 Other out-of-pocket payments for deliv-eries remained, but were reduced from 4.78% to4.15% of household income during the imple-mentation of the policy in these two regions for

68

Ghana, 2

the poorest quintile, while for the wealthiestquintile, they fell from 3.3% to 2.59%. Giventhe relatively high cost-recovery rates in Ghanaand reasonable geographic access to services,the emphasis on reducing fees was appropriate,whereas in Nepal, transport costs constitute thelargest cost element for households in accessingmaternal care.22

Ghana’s experience suggests that wherefacilities are understaffed, as in much of sub-Saharan Africa, they should be closely involvedin the development of policies such as theseand a package of incentives provided for staffto deliver quality care. The health worker surveyin this evaluation found that although very fewdirect financial incentives were provided as partof the delivery exemption policy, the overallincrease in pay as part of wider pay reforms inthe public sector compensated for increasedhours of work for different types of staff.28

Qualitative work with midwives and communi-ties suggests that comprehension of the policy wasnot always strong,29 highlighting the need forclear, simple designs and better communication.

003

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

The impact on maternal mortality has notbeen established, but case note extraction in asample of hospitals and routine data collected inhealth centres indicated that quality of care wasunchanged on the whole, and poor,16 e.g. rou-tine monitoring of labour in hospitals occurredin only 31% of cases.

International reviews of waivers and exemp-tions have already noted the importance of pro-viding funding to reimburse facilities for revenuelost so that they have a strong incentive to offerexemptions.30,31 Ghana’s experience corroboratesthat: user fees have been increasing in recentyears and amounted to 26% of Ghana HealthServices income in 2003.32 At the facility level,internally generated funds constitute the mainsource of flexible funding for drugs, suppliesand other minor purchases which are essentialto the running of services. Because maternalhealth care is an important income-generatingactivity for facilities.10 exemptions or a healthinsurance scheme will need to offer paymentsthat reflect those costs, without rewarding ineffi-ciency and taking into account the various sub-sidies which different facilities already receive.

Ghana has been bold in including the mis-sion sector, private midwives and even, in someareas, TBAs in its scheme. The issue of whichtype of providers should be included shouldbe a pragmatic issue, however, reflecting cost,quality and coverage. There are equity and effi-ciency arguments for including TBAs: they workin more remote communities, where women maynot be able to reach facilities, even if deliverieswere free; they are less costly to reimburse; andthere are often strong cultural preferences forhome deliveries. There is a strong case for includ-ing those whose quality of care is good and whosereferrals are appropriate and timely.

Anecdotal evidence suggests that the find-ings in Central and Volta regions were typical ofother regions. One informant said the schemehad more or less come to a halt nationally. Therewere also reports that delivery fees were beingexempted in some areas, but that this was nowinterpreted narrowly, so that items such as drugs,

References1. Amuti-Kagoona B, Nuwaha F.

Factors influencing choiceof delivery sites in Rakai district

of Uganda. SocMedicine 2000;

2. Wagle R, Sabroe

food, bed nets and admission costs were beingcharged again.

ConclusionUnderlying the specific budgeting and man-agement problems noted by the stakeholdersin Ghana are the more general issues of over-loaded systems and poor capacity, which anykind of additional vertical programme exacer-bates. The exemptions policy, while not separatein terms of service delivery, had a separate fund-ing channel and reporting requirements, and tothat extent added to the workload of strugglingofficials. The detailed design and management ofthis kind of scheme is critical to its success, andthere is a need for improved communication, bothvertically within the health system and acrossregions, so that the policy can be reviewed andadapted in light of successes and failures. Strongnational leadership is also a critical element insustainability. The experience of Ghana showsthe potential of schemes to increase skilled atten-dance by reducing the costs of services to users,though also some of their most common pitfallsin terms of sustaining commitment and funding.

AcknowledgementsWe would like to acknowledge the support ofthe research team in the Noguchi Institute forMedical Research, University of Ghana, and theUniversity of Aberdeen, in particular MargaretArmar-Klemesu, the Country Technical Partnerleader. Thanks also to the key informants for theirgenerous donation of time to share their viewswith us, and also to those who commented onthe paper, especially Tim Ensor. This work wasundertaken as part of an international researchprogramme, IMMPACT (Initiative for MaternalMortality Programme Assessment), funded by theBill & Melinda Gates Foundation, Departmentfor International Development, European Com-mission and USAID. The funders have no respon-sibility for the information provided or viewsexpressed in this paper. The views expressed hereinare solely those of the authors.

69

ial Science and50(2):203–13.S, Nielsen B.

Socioeconomic and physicaldistance to the maternityhospital as predictors for place of

ion

S Witter et al / Reproductive Health Matters 2007;15(30):61–71

delivery: an observation study the Volta reg

70

Childbirth 2004;4(8).3. Gilson L. The lessons of user fee

experience in Africa. HealthPolicy and Planning1997;12(4):273–85.

4. Witter S. An unnecessary evil?User fees for health care in lowincome countries. London7 Savethe Children, 2005.

5. Pearson M. Assessing the casefor abolishing user fees: lessonsfrom the Kenya experience.London7 HLSP, 2005.

6. Gilson L, McIntyre D. Removinguser fees for primary care inAfrica: the need for carefulaction. BMJ 2005;331:762–65.

7. UNFPA, Ministry of Health.Ghana health sector five yearprogramme of work(2002–2006) - an in-depthreview of the health sectorresponse to maternal mortalityin Ghana by 2003. Accra7UNFPA/MoH, 2004.

8. Graham W. Poverty andmaternal mortality: what’s thelink? Keynote address deliveredat Annual Research Meeting,Noguchi Memorial Institute forMedical Research, University ofGhana, Legon, 2004.

9. World Bank. Ghana LivingStandards Survey. WashingtonDC7 World Bank, 1999.

10. Levin A. Cost of maternal healthcare services in South Kwahudistrict, Ghana. Bethesda, MD7Partnerships for Health Reform/Abt Associates, 1999.

11. Waddington C, Enyimayew K.The impact of user charges inAshanti-Akim district, Ghana.International Journal of HealthPlanning and Management1989;4:17–47.

12. Garshong B, Ansah E, DakpallahG, et al. ‘‘We are still paying’’:a study on factors affectingthe implementation of theexemptions policy in Ghana.Accra7 Health Research Unit,Ministry of Health, 2001.

13. Nyonator F, Kutzin J. Health forsome? The effects of user fees in

y a1999;14:329–41

14. Ministry of Heafor implementinpolicy on materAccra7 MinistryReport No. MoHPlanning, MoniEvaluation-59.

15. Noguchi Institutevaluation of thuniversal fee exedelivery care: Ghprotocol. Accra7Institute/IMMPA

16. Armar-Klemesuevaluation of Guniversal fee exdelivery care: pfindings. AberdIMMPACT, 2006

17. Dubbledam R, WIndependent ReProgramme of WAccra7 Ministry

18. Witter S, Kusi T,Zakariah-AkotoEvaluation of thpolicy in Ghanakey informant inAberdeen7 IMMP

19. IMMPACT. Policimplementationdelivery prograAberdeen/AccraIMMPACT, 2005

20. Ministry of HeaHealth InsurancFramework forversion). Accra7Health, 2004.

21. Ensor T. Cost shalleviating finandelivery care: reproposed schemOptions, 2005.

22. Borghi J, Ensoret al. Financialskilled attendanNepal. TropicalInternational H2006;11(2):228–

23. Kelin D, KaininA draft report ochild health: PoFund study in C

of Ghana. Washington DC7

e, IMMPACT. Ane policy ofmption forana operationalNoguchiCT, 2005.M, et al. Anhana’s policy ofemption forreliminaryeen7.itter S, et al.

view ofork 2006.of Health, 2007.

S, et al.e free delivery: findings fromterviews. Accra/ACT, 2005.y brief:of the free

mme in Ghana.7

.lth. Nationale PolicyGhana (revisedMinistry of

aring system forcial barriers toview of thee. London7

T, Neupane B,implications ofce at delivery inMedicine andealth37.g Z, Songuan T.n mother andverty Actionhina.

24. Dmytraczenko T, Aitken I,Carrasco S, et al. Evaluation ofthe national security scheme formothers and children in Bolivia.Bethesda MD7 Abt Associates forPHR, 1998.

25. Ministry of Health, IMMPACT,UNFPA, Centre de Formationet de Recherche en Sante de laReproduction. An evaluation ofthe policy of fee exemption fordeliveries and caesareans inSenegal: operational protocol.Aberdeen7 IMMPACT, 2006.

26. Gilson L, Russell S, Buse K. Thepolitical economyofuser feeswithtargeting: developing equitablehealth financing policies. Journalof International Development1995;7(3).

27. Witter S, Aikins M, Kusi T.Funding and sustainability ofthe delivery exemptions schemein Ghana. Aberdeen7IMMPACT, 2006.

28. Witter S, Kusi A, Aikens M.Working practices and incomesof health workers: evidencefrom an evaluation of a deliveryfee exemption scheme in Ghana.Human Resources for Health2007;5(2).

29. Arhinful D, Zakariah-Akoto S,Madi B, et al. Effects of freedelivery policy on provisionand utilisation of skilled careat delivery: views fromproviders and communities inCentral and Volta regions ofGhana. Aberdeen/Accra7IMMPACT, 2006.

30. Bitran R, Giedion U. Waiversand exemptions for healthservices in developing countries.Washington DC7 World Bank,2003.

31. Tien M, Chee G. Literaturereview and findings:implementation of waiverpolicy. Bethesda, MD7 Abt forPartnerships for Health ReformPlus, 2002.

32. Ghana Health Service. GHS2003 Performance Report.Accra7 GHS, 2004.

71

ResumeDans les pays pauvres, le montant eleve descontributions demandees aux patientes pour lesaccouchements limite l’acces a des soins dequalite, tout en contribuant a la mortalitematernelle et neonatale et a l’appauvrissementdes menages vulnerables. Un nombre croissantde pays experimentent differentes methodespour lever les obstacles financiers aux soins desante maternelle. Un projet novateur, introduitau Ghana en 2003, exonerait toutes les femmesenceintes du paiement des accouchements etprevoyait que les praticiens publics, prives et desmissions pouvaient recuperer leurs honorairesperdus, conformement a un bareme convenu.L’article presente une partie des conclusionsd’une evaluation de cette politique, sur la based’entretiens avec 65 informateurs cles dans lesysteme de sante aux niveaux national, regional,des districts et des maternites, notamment desresponsables politiques, des gestionnaires et despraticiens. Lemecanisme d’exoneration a ete bienaccepte et etait satisfaisant, mais il connaissaitde graves problemes de decaissement des fondset de maintien du financement, ainsi que debudgetisation et de gestion. La charge de travaildu personnel a augmente car davantage defemmes ont demande des soins, et les niveauxde compensation pour les services et le personneletaient importants pour l’acceptation du projet.Fin 2005, un plan national d’assurance maladie,devant couvrir totalement les soins de santematernelle, demarrait au Ghana et la manieredont le projet d’exoneration s’y adapterait n’etaitpas encore claire.

ResumenEn los paıses con pocos recursos, el alto costo delas tarifas por partos limita el acceso a asistenciacalificada y contribuye a la mortalidad maternay neonatal y al empobrecimiento de hogaresvulnerables. En un creciente numero de paıses seesta experimentando con diferentes estrategiaspara afrontar los obstaculos financieros a losservicios de salud materna. En este artıculo sedescribe un plan innovador presentado enGhana en 2003 para eximir a todas las mujeresembarazadas de pagos por parto, medianteel cual los prestadores de servicios publicos,misioneros y privados pueden reclamar ingresosperdidos de tarifas de usuarias, de acuerdo conuna tarifa acordada. Se expone parte de losresultados de una evaluacion de la polıticabasada en entrevistas con 65 informantesclave (incluidos formuladores de polıticas,administradores y prestadores de servicios) delsistema de salud a nivel nacional, regional,distrital y local. El mecanismo de exencion fuebien aceptado y apropiado, pero hubo problemasimportantes con el desembolso y sustento delfinanciamiento, ası como con los presupuestos yla administracion. El volumen de trabajo delpersonal aumento a medida que se atendıan masmujeres, y los niveles de remuneracion porservicios y personal fueron esenciales para laaceptacion del plan. A fines de 2005, se estabainiciando en Ghana un plan nacional de seguromedico, con el objetivo de incluir coberturacompleta de servicios de salud materna, peroaun no era claro como el plan de exenciones seintegrarıa a este.

S Witter et al / Reproductive Health Matters 2007;15(30):61–71