Integration of immunization services with other health interventions in the developing world: what works and why? Systematic literature review

Authors


Corresponding Author Aaron Wallace, 1600 Clifton Road MS-E05, Atlanta, GA 30333, USA. Tel.: +1 404 639 6208; Fax: +1 404 639 8573; E-mail awallace@cdc.gov

Summary

Objective  To assess benefits, challenges and characteristics of integrating child and maternal health services with immunization programmes.

Methods  Literature review using journal databases and grey literature. Papers meeting the inclusion criteria were rated for the quality of methodology and relevant information was systematically abstracted.

Results  Integrated services were vitamin A supplementation, bednet distribution, deworming tablet distribution, Intermittent Preventive Therapy for infants and referrals for family planning services. Two key characteristics of success were compatibility between interventions and presence of a strong immunization service prior to integration. Overburdened staff, unequal resource allocation and logistical difficulties were mentioned as risks of integration, whereas rapid uptake of the linked intervention and less competition for resources were listed as two key benefits of integration.

Conclusion  The theoretical strengths of integrating other health services with immunization services remain to be rigorously proved in practice. When additional interventions are carefully selected for compatibility and when they receive adequate support, coverage of these interventions may improve, provided immunization coverage is already high. Evidence for the effectiveness of integration in increasing efficiency of resource use was insufficient and most benefits and challenges were not statistically quantified. More substantive information about the costs of integrated vs. vertical programmes and full documentation of the impacts of integration on immunization services should be published.

Abstract

Objectif:  Evaluer les bénéfices, les défis et les caractéristiques de l’intégration des services de santé maternelle et infantile dans les programmes de d’immunisation.

Méthodes:  Revue de littérature utilisant les bases de données de journaux et la littérature grise. Les articles répondant aux critères d’inclusion ont été classés selon la qualité de la méthodologie et les informations pertinentes ont été systématiquement prélevées.

Résultats:  Les services intégrés étaient la supplémentation en vitamine A, la distribution de moustiquaires, la distribution de comprimés de déparasitage, le traitement préventif intermittent pour les nourrissons et les transferts vers des services de planification familiale. Deux caractéristiques clés de succès étaient la compatibilité entre les interventions et la présence d’un service solide d’immunisation avant l’intégration. Le personnel surchargé, la répartition inégale des ressources et les difficultés logistiques étaient mentionnés comme des risques de l’intégration. Alors que l’absorption rapide de l’intervention intégrée et moins de concurrence pour les ressources étaient énumérés comme deux principaux avantages de l’intégration.

Conclusion:  Les atouts théoriques de l’intégration d’autres services de santé avec les services d’immunisation doivent encore être rigoureusement prouvés dans la pratique. Lorsque des interventions additionnelles sont soigneusement sélectionnées pour assurer la compatibilité et lorsqu’elles reçoivent un soutien adéquat, la couverture de ces interventions peut être améliorée à condition que la couverture vaccinale soit déjàélevée. La preuve de l’efficacité de l’intégration dans l’augmentation de l’efficacité de l’exploitation des ressources était insuffisante et la plupart des avantages et défis n’étaient pas statistiquement quantifiés. Des informations plus substantielles sur les coûts des programmes versus ceux des programmes verticaux ainsi que la documentation complète sur les impacts de l’intégration sur les services de vaccination devraient être publiées.

Abstract

Objetivo:  Evaluar los beneficios, retos y características de integrar los servicios de salud materno infantil con los programas de inmunización.

Métodos:  Revisión de la literatura realizada mediante bases de datos de revistas y literatura gris. Los artículos que cumplían con criterios de inclusión fueron categorizados según la calidad de la metodología, mientras que la información relevante fue extraída sistemáticamente.

Resultados:  Los servicios integrados fueron la suplementación con vitamina A, distribución de redes mosquiteras, distribución de medicación antiparasitaria, Tratamiento Preventivo Intermitente para niños y referencia a otros centros para servicios de planificación familiar. Dos características clave en el éxito de la estrategia eran la compatibilidad entre las intervenciones y la presencia de un servicio de inmunización sólido antes de la integración. La sobrecarga de trabajo del personal sanitario, una asignación de recursos no equitativa y las dificultades logísticas eran mencionadas como riesgos de la integración, mientras que el aceptar rápidamente la intervención integrada y una menor competencia por los recursos eran mencionadas como dos beneficios claves de la integración.

Conclusión:  Las fortalezas teóricas de integrar otros servicios sanitarios con los servicios de inmunización aún han de ser probados rigurosamente en la práctica. Cuando se seleccionan cuidadosamente intervenciones adicionales para compatibilidad y cuando reciben el soporte adecuado, la cobertura de estas intervenciones puede mejorar, siempre y cuando la cobertura de inmunización ya sea alta. La evidencia de la efectividad de la integración, al aumentar la eficiencia del uso de los recursos, era insuficiente, y la mayoría de los beneficios y retos no estaban estadísticamente cuantificados. Debería publicarse más información sobre los costes de programas integrados versus programas verticales y una documentación completa del impacto de la integración sobre los servicios de inmunización.

Introduction

The Expanded Program on Immunization, targeting primarily mothers and children, is recognized as one of the most widely implemented and well-established health programmes in the world (WHO 2005). Its characteristics led to the concept of ‘integrating’ the delivery of other maternal health and child-survival interventions with immunization services. Interest in this approach increased when the World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF) included such integration as one of four key strategic areas of the 2006–2015 Global Immunization and Vision Strategy (GIVS) framework (WHO 2005).

Although the theoretical strength of integration is recognized, we have been unable to find any comprehensive review of its feasibility, sustainability or impact (Unger 2003). Thus, we systematically reviewed the literature on interventions that have been integrated with immunization services at the delivery level to identify the programmes which have tried this approach, describe key benefits and risks of the integration strategy, highlight programme characteristics critical to successful integration and outline future research needs. This review may assist health programme managers considering integration of similar programmes, guide those designing studies of integrated programmes in selecting key variables to analyse and show both managers and researchers where the knowledge gaps lie in our understanding of the integration strategy.

Methods

Literature search

We used 48 keywords, singularly and in combination, to search grey and peer-reviewed online literature databases for literature published from 1970 to May 2005 in English or Spanish (Table 1). Grey literature was defined as material not published in a peer-reviewed journal; it was collected through contact with field experts, internet searches and requests to organizations known to conduct integrated projects. The bibliographies of identified literature were also used as a source for relevant articles. Grey and peer-reviewed papers identified were screened for primary research on the operational integration of routine or campaign immunization services with any other health intervention at the service delivery level in a low- or middle-income country as defined by the World Bank. All papers which met all these criteria were systematically reviewed.

Table 1.   Search strategy
  1. EPI, Expanded Program on Immunization.

Keywords used or in combination
 IntegrationDewormingIntermittent
 SynergyMalariaPolio
 SynergiesMaternalAsia
 Partnership$ReproductiveIndia
 Linkage$HealthJoint
 Immunization$MeaslesCombine
 Vaccination$Bednet$Combination
 Link$Net$ 
 CollaborateVaccination$ 
 Collaboration$EPI 
 IntegratingDisease$ 
 Partner$Vitamin A 
 AfricaFilariasis 
 Integrate$Family planning 
 Immunization plusPolio 
 RoutineTrachoma 
 Campaign$Onchocerciasis 
 MalariaIPTI 
 Lymphatic filariasisOnchocerciasis 
 FilariasisVitamin A 
 SchistosomiasisTrachoma 
Literature databases searched
 Access UNCSA-illumina databasesPAIS
 AccessScienceDissertation abstractsPOPLINE
 AGRICOLAEMBASEPopulation Index
 Bioline InternationalExpanded academic ASAPProquest Research Library
 BioMed CentralGlobal healthPubMed
 BIOSISIBSSSIGLE
 CAB AbstractsIndMedUNDP Project Reports
 CHID OnlineLexisNexis academicWeb of Science
 CINAHLLILACSWHOLIS
 Cochrane LibraryOvid MEDLINE 
Websites reviewed
 http://www.basics.orghttp://www.ifrc.orghttp://www.savethechildren.org
 http://www.bl.uk/services/document/greylit.htmlhttp://www.msh.orghttp://www.savethechildren.org.uk
 http://www.care.orghttp://www.nyam.org/library/greylitorgs.shtmlhttp://www.trachoma.org
 http://www.fhi.orghttp://www.paho.orghttp://www.un.org
 http://www.filariasis.orghttp://www.path.orghttp://www.undp.org
 http://www.gavialliance.orghttp://www.pathfind.orghttp://www.vaccinealliance.org
 http://www.greynet.orghttp://www.psi.orghttp://www.who.int/library/
 http://www.hki.orghttp://www.redcross.org 
Organizations contacted
 BASICSLiverpool School of Tropical MedicineWorld Bank
 International Federation of Red Cross & Red Crescent SocietiesCenters for Disease Control & PreventionRollins School of Public Health
 World Health OrganizationLondon School of Hygiene and Tropical Medicine 

Systematic review method

To assist in comparing research results, each paper was first sorted into one of three groups by study design: observational study, pre-/post-intervention study or controlled trial study. Next, information was systematically collected from each paper using a data extraction form based on one developed for similar literature reviews (Pegurri et al. 2005). Extracted data comprised project location, types of interventions, stated benefits and risks of integration, characteristics of successfully integrated programmes, extent of integration, costs and results. Lastly, to help us in interpretation of the paper’s results, we looked for the following critical scientific components: (1) presentation of target population data, (2) use of randomization in study design, (3) presence of well-defined research outcomes, (4) extent of data analysis, (5) discussion of study limitations and (6) comparison of results to published literature. Abstracted data were grouped by the type of intervention integrated with immunization.

Results

The initial keyword search yielded 150 papers, of which 27 met our criteria for inclusion in the review. Nineteen (70%) of the 27 papers were peer-reviewed and eight (30%) were grey literature. Of the 27 papers, 22 (81%) were published after 1998. These 27 papers covered 19 integration projects. In some instances, papers reported on the same integration project.

Of the 19 integration projects studied, 15 (79%) were set in Africa, 11 (58%) integrated delivery of other health interventions with routine immunization services and eight (42%) used immunization campaigns as the integration platform (Table 2).

Table 2.   Summary descriptions of the 19 integration projects
CountryLinked intervention(s)Immunization delivery typeMain integration strategyProject dateStudy typeLiterature typeReferences
ZambiaBednets, mebendazole, vitamin ACampaignDeliver all interventions together to child2003Pre-/post-evaluationGreyBwalya (2003), Red Cross & Red Crescent Societies (2004)
TogoBednets, mebendazoleCampaignDeliver all interventions together to child2004Pre-/post-evaluationGreyWolkon et al. (2005), Irby (2005)
GhanaBednetsCampaignDeliver all interventions together to child2001Pre-/post-evaluationPeer- reviewedGrabowsky et al. (2005)
MaliBednets, vitamin ARoutine w/Supplemental CampaignDeliver bednets when child fully immunized2003–2005Pre-/post-evaluationGreyUNICEF (2005)
SenegalBednets, vitamin ARoutine w/Supplemental CampaignDeliver bednets when child fully immunized2003–2005Pre-/post-GreyUNICEF (2005), WHO and UNICEF (2004)
NigerVitamin ACampaignDeliver all interventions together to child1990sObservationGreyAguayo et al. (2003)
IndiaVitamin ACampaignDeliver all interventions together to child2000ObservationPeer- reviewedSwami and Thakur  (2001)
EthiopiaFamily planningRoutineFamily planning referral message to mother at time of child’s vaccination. Integrated Data Management1991Pre-/post-evaluationPeer- reviewedWalley and McDonald (1991)
TogoFamily planningRoutineFamily planning referral message to mother at time of child’s vaccination1994TrialPeer-reviewedHuntington and Aplogan (1994)
ZaireFamily planningRoutineFamily planning referral message and counselling to mother at time of child’s vaccination1992Pre-/post-Peer-reviewedFisher (1992)
BurundiFamily planningRoutineFamily planning referral message and counselling to mother at time of child’s vaccination1993Pre-/post-evaluationPeer-reviewedFisher (1992), Birmingham et al. (1993)
IndiaFamily planningRoutineFamily planning referral message and counselling to mother at time of child’s vaccination2004Pre-/post-evaluationGreyNayak (2004)
MadagascarFamily planningRoutineFamily planning referral message and counselling to mother at time of child’s vaccination2004Pre-/post-evaluationGreyBASICS II (2004)
IndiaFamily planningRoutineFamily planning referral message to mother at time of child’s vaccination1983ObservationGreyTaylor et al. (1983)
TanzaniaMalaria treatmentRoutineSafety of giving Intermittent Protective Treatment of infants alongside immunizations1999–2001Clinical trialPeer-reviewedSchellenberg et al. (1999), Galindo et al. (2000), Acosta et al. (1999)
TanzaniaMalaria treatmentRoutineSafety of giving amodiquine with immunizations2003Clinical trialPeer-reviewedMassaga et al. (2003)
IndonesiaVitamin ACampaignImpact of giving vitamin A with immunizations1999–2001Clinical trialPeer-reviewedSemba et al. (1999, 2001)
Ghana, India, Peru (multi-location  study)Vitamin ACampaignImpact of giving vitamin A with immunizations1998–2002Clinical trialPeer-reviewedBahl et al. (2002a,b), WHO/CHD study group (1998)
Guinea-BissauVitamin ACampaignImpact of giving vitamin A with immunizations1997Clinical trialPeer-reviewedBenn et al. (1997)

Integrated programme descriptions and strategies

Three major integration strategies were identified within the 19 projects. The first, which we termed ‘linked referral’, used a routine immunization contact to inform the child’s mother of other health services available in the same facility. The second strategy, termed ‘integrated routine delivery’, used a routine immunization contact to deliver other health services to the child at the same encounter. The third strategy, termed ‘integrated campaign delivery’, used an immunization campaign to deliver additional health services to the child.

Seven (26%) of the 27 papers described the offering of family planning services to the child’s mother during routine immunization contacts; five (19%) explored the administration of anti-malarial medication to infants during routine immunization contacts (Intermittent Preventive Therapy for infants); eight (30%) described the distribution of vitamin A or deworming tablets in conjunction with campaigns or routine immunization visits; and six (22%) explored the distribution of insecticide-treated bednets (ITNs) through immunization campaigns. In some instances, more than one intervention was integrated with delivery of vaccines (Table 3).

Table 3.   Classification of search results for 27 papers describing integrated projects
Search results, = 27 papersStudy methodology
Pre-/post-evaluation, = 13 (48%)Observation, = 2 (7%)Trial, n = 12 (44%)
n = 5 (19%)n = 1 (4%)n = 12 (44%)Papers rated as ‘good’ or ‘fair’
n = 7 (26%)n = 1 (4%)n = 0 (0%)Grey literature papers
n = 5 (19%)n = 0 (0%)n = 0 (0%)Papers with financial costs included
n = 11 (41%)n = 2 (7%)n = 6 (22%)All integration projects
Programmes integrated with immunization
 n = 5 (19%)n = 1 (4%)n = 1 (4%)Integration of family planning education
 n = 1 (4%)n = 1 (4%)n = 6 (19%)Integration of vitamin A distribution
 n = 6 (22%)n = 0 (0%)n = 0 (0%)Integration of bednet distribution
 n = 0 (0%)n = 0 (0%)n = 5 (19%)Integration of intermittent protective treatment of infants

Quality of data collection

Although the literature demonstrates that integrated immunization projects exist, many of these papers did not include information that would have been useful in fully evaluating the impact of integration on cost, feasibility and coverage. For example, although all 12 peer-reviewed papers reporting a trial used control groups, 11 focused on comparisons of biological interactions of drugs administered rather than programmatic aspects of integrating health interventions (Benn et al. 1997; WHO/CHD Immunization-linked Vitamin A Supplementation Group 1998; Acosta et al. 1999; Schellenberg et al. 1999, 2001; Semba et al. 1999, 2001; Galindo et al. 2000; Bahl et al. 2002a,b; Massaga et al. 2003). Of the 15 non-trial methodology papers, six (40%) included no information on pre- or post-immunization coverage rates and seven (47%) reported either only pre- or only post-immunization coverage rates. However, 14 (93%) reported pre- and post-coverage rates for linked interventions; the remaining one recorded post-intervention coverage rates. Three (20%) of the 15 papers reported on use of control groups to measure the operational impact of integrating vs. not integrating programmes and three (20%) papers reported any information on financial costs (Red Cross & Red Crescent Societies 2004; Grabowsky et al. 2005; UNICEF 2005).

Reported impact on coverage rates

The five (19%) of 27 papers describing changes in vaccine coverage after integration of other services reported no negative impact on immunization service coverage (Table 4). Immunization coverage rate increases varied from 0% in measles vaccinations during a 5-month intervention to 50% in fully immunized children during a 36-month intervention.

Table 4.   Percentage point change in coverage rates documented by integration studies
LocationSettingImmunization coverage percentage point changeLinked intervention coverage percentage point changeIntervention lengthReferences
  1. ITN, insecticide-treated bednets.

ZambiaLinked campaign18%– fully immunized63%– ITN ownership1 weekBwalya (2003), Red Cross & Red Crescent Societies (2004)
TogoLinked campaign17%– measles61%– ITN ownership
54% child sleeping under net
1 weekWolkon et al. (2005), Irby (2005)
GhanaLinked campaign 75%– ITN ownership
56% child sleeping under net
1 weekGrabowsky et al. (2005)
MaliLinked routine51%– measles
44%– fully immunized
50%– vitamin A
80%– child sleeping under ITN
3 yearsUNICEF (2005)
SenegalLinked routine13%– measles
50%– fully immunized
26%– vitamin A
86%– child sleeping under ITN
3 yearsUNICEF (2005), WHO and UNICEF (2004)
EthiopiaLinked referral1%– measles4%– family planning acceptance1 yearWalley and McDonald (1991)
TogoLinked referral 27%– family planning acceptance2 monthsHuntington and Aplogan (1994)
ZaireLinked referral 19%– family planning acceptance7 monthsFisher (1992)
BurundiLinked referral0%– measles5%– family planning acceptance5 monthsBirmingham et al. (1993)
MadagascarLinked referral37%– fully immunized48%– breastfeeding
32%– child with 5 meals per day
2 yearsBASICS II (2004)

The length of the intervention varied by strategy: one project focusing on integrated routine delivery reported a 36-month intervention period whereas the six projects using linked referral averaged 11 months (range: 2–24 months).

Benefits of integration

Eighteen papers (66%) did not discuss benefits of integration, and those benefits that were discussed in other papers were rarely quantified. Nonetheless, it was possible to identify a number of benefits associated with programme integration.

Rapid uptake occurred when an intervention was linked to a robust immunization programme with high coverage rates leading to a rapid increase in coverage for the linked intervention to levels comparable to that of the immunization programme. In two papers, family planning services showed a very rapid uptake over a 2-month period; managers attributed this to linking the new intervention to robust immunization services with 90% pre-integration coverage rates (Birmingham et al. 1993; Huntington & Aplogan 1994). Competition fell after integration when two programmes had been competing for the same resources before being integrated (Fisher 1992; Bahl et al. 2002b; Diene et al. 2004; Grabowsky et al. 2005). For instance, one integration project merged ITN distribution with an immunization campaign and allowed the integrated programme to maximize use of human resources rather than competing for workers (Grabowsky et al. 2005). Programme managers also described integrated projects as more user-centric because they offered multiple services at once (Grabowsky et al. 2005). One paper cited how the community preferred distribution of ITNs and vitamin A tablets through immunization campaigns because this saved them trips to the regional health clinic (Grabowsky et al. 2005). Other benefits mentioned were: better staff morale because of a more diverse workload and new learning opportunities (Grabowsky et al. 2005); more equity between population sectors by offering multiple services to hard-to-reach communities which had previously been receiving only one service during outreach sessions (Wolkon et al. 2005); higher efficiency by reducing the need for the duplicate structures required when health programmes were not integrated (Walley & McDonald 1991); and less stigma attached to traditionally stigmatized services such as family planning (Prabhakara et al. 2002).

Challenges of integration

A number of factors that could hamper integration were cited. The most frequently mentioned was persistently unequal resource allocation – defined as one intervention continuing to receive disproportionately more resources than the other intervention after integration (Nayak 2004) – resulting in poor uptake of the less-funded service.

For example, in an intervention linking family planning to delivery of immunizations, donor financing could not be used to educate health workers regarding pre-natal care. This lack of education was considered responsible for relatively poor community acceptance of pre-natal care compared to other family planning services offered (Walley & McDonald 1991).

Papers mentioned that the non-linked components of semi-integrated programmes might be ignored or that linked interventions were not always accompanied by appropriate education. For example, in a linked bednet-deworming-measles campaign, helminth-control managers were concerned that a focus on drug distribution alone would jeopardise their comprehensive approach of improved sanitation, health education and drug distribution, all of which they felt were needed to reduce helminth carriage (Bwalya 2003). In a post-evaluation of an integrated child health week including ITN distribution and vaccinations, the percentage of children <5 years found to be sleeping under ITNs (44%) was 19% lower than the percentage of households reporting ITN ownership (63%) after the integrated campaign and 47% lower than the reported ITN coverage of eligible children in the campaign (91%), perhaps because no education had been provided at the time of net distribution; this finding prompted door-to-door training in appropriate ITN use (Wolkon et al. 2005).

Programme managers have raised concerns about overburdening staff (Swami & Thakur 2001; Wolkon et al. 2005). This risk was perceived to be particularly great when staff delivering services received inadequate training. For instance, when distribution of ITNs was integrated with a measles vaccination campaign, workers felt they were not given time to be educated about ITNs and consequently felt they could not educate the community about ITN use (Wolkon et al. 2005).

Another study cited the risk of poor integration of data management systems; in this instance, the integrated patient record form was oversimplified to such an extent that it lacked many necessary fields of information (Walley & McDonald 1991). Lastly, one author said that bringing together multiple programmes, ITN distribution, deworming tablet distribution and vaccinations, for an integrated campaign could present logistical challenges. The researchers felt that integrating these services could result in a large and unwieldy system with communication, coordination and planning complexities and would need further investigation (Wolkon et al. 2005).

Key characteristics of successful integration

Several characteristics of successful integration programmes were identified by the papers’ authors. When adequately taken into account by programme managers, these characteristics were noted to be facilitators of success; inadequately addressed, they were noted to be barriers.

The most frequently discussed characteristic of successful integration was programme compatibility, mentioned in 15 of the 27 papers, which is the appropriate matching of programmes based on health worker skill requirements, programme objectives, recommended timing of interventions, target populations and drug/treatment characteristics. Also important was the existence of a robust immunization service providing high quality training and good supervision to its staff which could translate easily into high coverage with multiple interventions in the target population (Walley & McDonald 1991; Huntington & Aplogan 1994; Diene et al. 2004; UNICEF 2005). Support from key stakeholders was critical; these are the community, staff, donors and upper management (Huntington & Aplogan 1994; Aguayo et al. 2003; Bwalya 2003; Diene et al. 2004). Decentralization of health services, which led to good facility access by providing village-based services in remote areas, was cited as a key to successful integration. These characteristics appeared to build upon each other; for instance the decentralization of health services to the village level was also seen as boosting community support, which in turn led to key stakeholder support for the integrated programme (Walley & McDonald 1991; Aguayo et al. 2003; Bwalya 2003; Diene et al. 2004).

Discussion

This review revealed that the theoretical strengths of the integration of other health services with immunization are yet to be rigorously proven. Nonetheless, our review shows that when these additional interventions are carefully selected and adequately supported, coverage of these interventions may improve, provided immunization coverage is already high. Integration may lead to strengthening of immunization services (WHO 2005); although this seems at odds with our finding that successful integration may depend in part upon the presence of an already robust immunization service. Where immunization coverage is low, cost and staffing efficiency, a possible benefit of integration, may be a key reason for using the integration strategy. Yet, this strategy may be impeded if one of the linked services lacks sufficiently high coverage before integration. Partnering with any high-coverage, compatible health service may yield positive results for a poorly performing intervention; but linking a low-coverage immunization service with another poorly performing intervention will not result in high coverage; additional research is necessary to provide these answers.

The critical question of whether integrated programmes use resources more efficiently than vertical programmes has yet to be answered.

Cost savings are seen as a major reason to integrate programmes and the basis for WHO’s recommendation in Strategy 14 of GIVS is that integration will improve efficiency in comparison to vertical programmes (WHO 2005). In measuring any efficiency gained from integrating systems, costs associated with moving from separate to integrated programmes should be considered, which can include worker re-training and re-education of community members.

Across all studies, programme managers and researchers commented on the critical importance of compatibility between programmes prior to integration to fully capture the potential advantages of integration. Compatibility characteristics included similarities in target population, logistical needs, worker training, stakeholder support, costs and supply chain requirements. Another aspect of integration to consider is the intensity of integration; in some projects, only a single component of a complex health intervention was integrated with immunization services. As such collaborations indicate, programme managers do not need to fully integrate all activities of both programmes but rather should only integrate where value is added.

The majority of the papers we reviewed described benefits of integration that are unique to the integration strategy. When integration is done carefully, there appears to be a decrease in competition for resources and duplication of health systems. The more patient-centred approach which is created can result in increased coverage rates of interventions partnered with a strong health service. The challenges to integration, on the other hand, are not unique to an integrated health system. Rather, they are challenges any health intervention faces: inappropriate resource allocation, poor logistics and data management and overburdened staff. We may then conclude that the risks inherent in integrating health services are not unique, whereas the potential benefits offered are.

When considering integration, programme managers should ensure compatibility of integrated programmes, strengthen proven operational strategies such as training, supervision, outreach and community education, and consider the optimal intensity of integration. Programme managers should build integrated programme strategies through developing a comprehensive plan that addresses re-training of workers so they have balanced knowledge of all linked interventions and are competent to deliver all interventions at once and promote intensive community education to ensure community support and understanding of the need for a transition to an integrated programme.

Conducting and publishing additional research on integration is critical to better understand the risks and benefits of this strategy. Our review was limited to literature databases that we could access by computer and to experts whom we were able to contact via email or phone. Our search discovered many integration programmes that have not been evaluated; studies in of these current programmes could provide useful data (Molyneux & Nantulya 2004). Searches of literature databases in other languages such as French, Chinese and Arabic also would likely yield additional articles of value. Further research is needed into ‘failed’ integration projects for a comprehensive view of risks and benefits; this future research would also benefit from control groups to permit comparison of integrated and non-integrated programmes. A standardized set of process and outcome indicators would facilitate evaluating the feasibility and effectiveness of interventions; possible indicators include efficiency of resource use at all levels, costs of switching from vertical to integrated services, extent of cost sharing between programmes, user satisfaction with the integrated programme and documentation that immunization services will not be harmed by integration.

Ancillary