X0 24276 April 2002 Vouchers for Health Peter Sandiford, Anna Using Voucher Schemes for Output-Based Aid Gorter, and Micol Salvetto This Note examines an innovative, donor-supported voucher scheme to treat and prevent sexually transmitted diseases in Nicaragua. Petei Sandiford ii 7wit/i the Instit,te for Health Voucher schemes enable donors to purchase outputs rather than Sector in iL hLon Anna inputs while also offering beneficiaries a choice of provider, a feature Gorter is withl tile Londlon School of Hygiene and that sets them apart from other output-based approaches, such as 7TopicalMedicine Anid supply-side subsidies to providers operating under performance-based Plicol Salvetto is withi Iinstituito Cettoarneiic(nno contracts. Choice creates incentives to lower prices or raise quality de la Saluid in Managua (or both). A donor-supported voucher scheme in In developed countries a range of special pro- Nicaragua provides treatment and prevention grams have been set up to address the health services for sexually transmitted infections needs of high-risk groups In developing coun- (STIs) to high-risk populations such as com- tries governments have been less successful in mercial sex workers and their partners and reaching these groups and meeting their special clients. Beyond detecting and treating STIs, needs. They often resort to counterproductive these health services can also raise awareness of coercive measures, such as having the police a. nrisks and promote safer behavior, leading to oblige sex workers to attend government-run widespread benefits. Because STIs increase the STI clinics. But sex workers shy away from dedi- transmiiission of HIV by a factor of three to five or cated STI clinics because the clinics are stigma- more, the population groups that are extremely uzing, tend to treat their patients discourteously, vulnerable to STIs and HIV infection facilitate and are perceived as having scant respect for the entry and spread of AIDS in the general pop- patient confidentiality. Sex workers tend to visit m tuilation. Thus small interventions targeted to general outpatient clinics instead, choosing not such high-risk groups can have a large impact on to reveal their occupational risks. Consequently, the spread of STIs and HIV. But these groups- they rarely receive adeqtiate diagnosis or treat- O) -sex workers, regular clients and partners of sex ment for their sexual health problems. workers, adolescent glue-sniffers, migrant labor- In the absence of special client-friendly pro- 7 1 ers, long-distance tranisport workers-tend to be grams, the use of sexual health services among difficult to reach and are often mobile. these vulnerable groups is low and usually inef- 4 VOUCHERS FOR HEALTH USING VOUCHER SCHEMES FOR OUTPUT-BASED AID fective. What these grouips need is subsidized vouchers non transferable, for two reasons. First, access to convenient, courteous, confidential preventing the vouchers from being used by services of high quality that identify their sexual someone other than the original recipient is health needs in a nonstigmatizing manner and costly. And second, the secondary recipienit of provide appropriate counseling, diagnostic the voucher might be at higher risk of having an information, and treatment. STI than the primary one, so the transfer could serve the program's objectives. Providers are How the scheme works unlikely to exchange vouchers for cash, because 2 The Central Amencan Health Institute (Instituto that would simply reduce their income and in Centroamericano de la Salud, or ICAS)l has any case each redeemed voucher must be attempted to provride such services in Nicaragua accompanied by a blood test and other samples. through a voucher scheme since 1995 At six- The government has no direct involvement month intervals ("rounds") the scheme distrib- in the scheme. ICAS acts as the voucher agency, utes 2,000 vouchers to the vuilnerable groups (in contracting and monitoring clinical and labo- later rounds, including clients and partners of ratory services, training clinic staff (doctors, sex workers), some directly and some (at no cost) nurses, and receptionists) in handling sex work- through community-based organizations in close ers in a nondiscriminatoiy manner, defining contactwiththegroups(figutire1) Thevoucliers, the service package covered by the voucher, which remain valid for two and a half months, analyzing data, and monitoring technical qual- entitle the bearer to a predefined package of ity and patient satisfaction, both generally and "best practice" sexual health services free of for each clinic In addition, its fieldworkers charge at any one of about 10 contracted clinics. maintain a map of the 50-60 prostittition sites Clinic staff need not ask patients embarrassing in and around the capital city ofManagua where questions about their membership in the tar- the vouchers are distributed. geted grotip, since this is declared uith the pres- Clinics compete for contracts on the basis of entation of a voucher. price, quality, and location. For the initial round To preventcounterfeiting, vouchers are indi- ofcontracting clinics were individually invited to vidually numbered, stamped with the ICAS seal, participate (since a newspaper advertisement and laminated. Their expiration date is printed generated little interest). Those responding on them. No meastires are taken to make the were assessed against predefined criteria. With 1 7 Voucher agency Vocher a payment J | Cashoa \ Cimmunty-based organizations Voucher men ~ ~ ~ ~~~ouhe Donor and government Sex workers ClinVcsu Source Diagram adapted from Harper and others (2000) each clinic that met the criteria ICAS negotiated zations (NGOs). Since rounld four, however, a price for the servce package, whilch consists of only private and NCO clinics have participated a medical consultation, follow-up visit, counsel- The program dropped the public sector clinics ing, and the taking of test samples (treatment is because they attracted few voucher redeemers dispensed separately). and had long waiting times and unfriendly For new clinics ICAS uses the range of prices "gatekeepers." The private for-profit clinics sur- paid under existing contracts as a benchmark, vive on their fees, but most of the NGOs also but it also considers their strategic importance receive subsidies and can therefore charge to the program. ICAS gives particular weight to lower prices. The public clinics charged nomi- clinics' location and the perceived quality of nal user fees. their service, both important determinants of Since 1995, in 12 rounds, the scheme has dis- use by sex workers. The contracts between ICAS tributed more than 15,000 vouchers, provided and the provider-s require staff to follow a speci- more than 6,000 consultationis, and treated fied treatmen1t protocol and to participate in countless cases of STIs. Of a dynamic popula- training sessions ICAS regularly invites new cdn- tion of about 1,150 female sex workers active at ics to join the scheme, but to keep the adminis- any one time, more than 40 percent have trative burden manageable it contracts no more redeemed their voucher. More than 2,500 sex than about 10 providers for each round. Clinics workers have participated, with the highest rates are paid according to the nuimber of voLtchers of redemption among the poorest women and they return (along with data collection sheets). among the groups with the higlhest initial rates ICAS revqews the performance of participat- of STIs ing clinics in each round and does not renew The program reduced the prevalence of gon- the contracts of those judged to have been orr-lea in the female sex worker population by unsatisfactory To monitor- quality, it conducts an average 5.25 percent a year, and the inci- semistructured interviewvs with 10 percent of the dence in repeat users by 11 5 percent a year. It patients redeeming vouchers at each clinic, ana- reduced the prevalence of syphilis by an average lyzes medical record-keepinig, and looks at 10.25 percent a year The poorest sex workers other indicators of quality, such as the number had the higlhest initial prevalenice rates for gon- of vouchers redeemed and the share of women orrhea and syphilis; these rates fell by an aver- attending their follow-up consultation. (The age 9.4 percent and 8.6 percent a year. interniews, conducted by a small, multidiscipli- Although prevalence at follow-up constiltations nary team, also explore ideas for increasing the was not zero, these women remained free of redemption of vouchers ) The voucher agency STIs longer, which considerably reduces the gives clinics feedback based on the results of the risk of being infected with HIV or infecting their quality monitoring. clients. HIV prevalence in sex workers in ICAS contracts a single laboratory to per- Managua was 0.8 percent in 1991,1.3 percent in form diagnostic tests, to simplify the logistics 1997, and 2.0 percent in 1999, a rate of increase and allow some economies of scale. The labora- well below that observed in the sex worker pop- tory provides swabs and transport media, col- ulations of other major cities. lects samples daily, reports results, and Today the voucher scheme IS sustained at a distributes treatment. A second laboratory pro- cost of US$60,000 a year, reflecting the relatively vides quality control. Drugs, condoms, medical low cost of patient visits. While a baseline stud)' supplies, and health education material are cen- estimated the cost of an outpatienit constiltation trally procured by competitive tender. in public facilities to be US$7.65, the average price the voucher scheme paid to clinics for a Results consultation and follow-up visit wvas initially only By 2001 the program had contracted 20 service US$6.70. And despite increasing slightly at first, providers at one time or another. These have the average price was only US$5.15 in the 12th been a mix of public and private providers, round. This cost redtiction was due mostly to the including nonprofit, nongovernmental organi- gradual devaluation of the currency, but also in VOUCHERS FOR HEALTH USING VOUCHER SCHEMES FOR OUTPUT-BASED AID part to competitive pressures and effective nego- ent. But this approach would be vulnerable to col- tiation. In fact, contracted prices may be quite lusion between clinics and probably would not close to the clinics' marginal costs Reducing achieve the broad geographic coverage necessary costs further through cost recovery fi-om the sex to ensure high redemption rates. Moreover, the workers is unrealistic and would exclude the current process allows the program to take advan- poorest, who also have the greatest health nieeds. tage of the subsidies received by some of the NGO p As it is, the costs to sex workers in transport and clinics by contracting the NGOs at a lower pnce lost income are significant, and for some a rea- than the private providers and thereby "leveling is an open forum to son not to use their voucher. the playing field" to some degree. encourage dissemination of The direct cash benefits to the clinics partici- Some of the potential problems of voucher g disseination of public policy innovations for pating In the scheme are not large. The largest schemes have not arisen. One is counterfeiting. private sector-led and provider has received only about US$1 0,000 iii In theory a counterfeiter could reproduce the market-based solutions for six years. But the program offers clinics steady, vouchers, selling them to users or sermce development. The views reliable income Moreover, some clinic directors providers But because of the measures taken to published are those of the report that the presence of voucher-bearing reduce this risk, a counterfeiter would need authors and should not be clients has helped to fill the clinics and attract some sophistication in producing vouchers as attributed to the World paying clients. And participation has improved well as confidence that the vouchers could be Bank or any other affiliated the technical quality of services, since the clinics sold (or, if sold to a provider, redeemed by the organizations. Nor do any of usually apply the lessons learned in the training voucher agency). A related risk is black market the conclusions represent to their services for all clients. Participation also sales ofvouchers. The market would probably be official policy of the World confers a certain status on the clinics, improving significant only if the vouchers covered a broad Bank or of its Executive their prospects when competing for other con- range of senices of value to a large part of the Directors or the countries tracts, such as with the nationial social insurance population or interventions that are particularly they represent. program. For these reasons providers were pre- costly. Moreover, black market sales are not nec- pared from the outset to offer services at prices essarily a bad thing in this case. The secondary To order additional copies well below their standard rates recipients of vouchers would probably be mem- contact Suzanne Smith, bers of the target population or at least poorer managing editor, Conclusion or at higher risk than the general population. Room 19-017, The World Bank, The voucher scheme has provided access to- 1 81I8 H Street. NW, and increased the use of-high-quality, tailored sexual health services in a nonistigmatizing man- Washington, DC 20433 ner for commercial sex workers and their regu- Note lar sexual contacts. It has used scarce publlc 1 ICAS is an indepenidenit, nonilofit, nonlgovern- 001 202 458 7281 resources cost-effectively to provide services to a mncital organization with no political or religious affilia- Fax group whose members have major health needs tioIn whose sole aim is to nrnprovc the healthi of Central 001 202 522 3181 and are generally unable to pay for care. And it Americans Its healthi programs are fundic( primilai-ly by Emal has reduLced the risk of STIs, incltuding HIV, donor agenicies ssmith7@worldbank org among the general population Few input-based programs of HIV/AIDS prevention can lay Reference claim to such stuccess. I-larper, Malcolm,Joige Ar-ioyo, Tusliar Bhattacharya, Printed on recycled paper But the voucher scheme also has some limita- and Tom Bulmain 2000 Pubhc Senvce5 thiough Puvate tions and risks. The scheme's admilistrative costs Enterpnse Milrcw-Pnvalisation fo? Impioved Dehveiy London account for a large share of total spending, Initermediate Technology Development Giotip although this share might decline if the pro- gram's geographic coverage expancled. In addi- tion, the scheme's price formation, through individual negotiations with each clinic, might raise questions abotit transparency. Fixing the number of clinics to be contracted and selecting just the lowest bidders might be more transpar- This Note is available online: www.worldbank.org/viewpoint/