SEARCH

SEARCH BY CITATION

Keywords:

  • HIV services;
  • tuberculosis;
  • integration;
  • low- and middle-income countries;
  • systematic review

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

Objectives

Given the imperative to scale up integrated tuberculosis (TB) and HIV services in settings where both are of major public health importance, we aimed to synthesise knowledge concerning implementation of TB/HIV service integration.

Methods

Systematic review of studies describing a strategy to facilitate TB and HIV service integration, searching 15 bibliographic databases including Medline, Embase and the Cochrane library; and relevant conference abstracts.

Results

Sixty-three of 1936 peer-reviewed articles and 70 of 170 abstracts met our inclusion criteria. We identified five models: entry via TB service, with referral for HIV testing and care; entry via TB service, on-site HIV testing, and referral for HIV care; entry via HIV service with referral for TB screening and treatment; entry via HIV service, on-site TB screening, and referral for TB diagnosis and treatment; and TB and HIV services provided at a single facility. Referral-based models are most easily implemented, but referral failure is a key risk. Closer integration requires more staff training and additional infrastructure (e.g. private space for HIV counselling; integrated records). Infection control is a major concern. More integrated models hold potential efficiencies from both provider and user perspective. Most papers report ‘outcomes’ (e.g. proportion of TB patients tested for HIV); few report downstream ‘impacts’ such as outcomes of TB treatment or antiretroviral therapy. Very few studies address the perspectives of service users or staff, or costs or cost-effectiveness.

Conclusions

While scaling up integrated services, robust comparisons of the impacts of different models are needed using standardised outcome measures.

Intégration des services de tuberculose et VIH dans les pays à faibles et moyens revenus: une revue systématique

Objectifs

Compte tenu de l'impératif du déploiement des services intégrés de la tuberculose (TB) et du VIH dans les contextes où les deux sont d'une importance majeure de santé publique, nous avons cherché à synthétiser les connaissances en matière d'intégration des services TB/VIH.

Méthodes

Revue systématique des études décrivant une stratégie visant à faciliter l'intégration des services TB et VIH, recherchant dans 15 bases de données bibliographiques dont Medline, Embase et la Bibliothèque Cochrane, et des résumés de conférences pertinents.

Résultats

63/1936 articles publiés et 70/170 résumés répondaient à nos critères d'inclusion. Nous avons identifié cinq modèles: entrée via le service de TB, avec aiguillage pour le dépistage et les soins du VIH; entrée via le service de TB, dépistage du VIH sur place et aiguillage pour les soins du VIH; entrée via le service VIH avec aiguillage pour dépistage et le traitement de la TB; entrée par le service VIH, dépistage de la TB sur place et aiguillage pour le diagnostic et le traitement de la TB; et services TB et VIH fournis dans le même établissement. Les modèles basés sur l'aiguillage sont plus facilement implémentés, mais la défaillance dans l'aiguillage est un risque majeur. Une plus étroite intégration nécessite plus de formation pour le personnel et d'infrastructure supplémentaire (ex. un espace privé pour les conseils sur le VIH, des dossiers intégrés). Le contrôle des infections est une préoccupation majeure. Plus de modèles intégrés ont une efficacité potentielle à la fois du point de vue du prestataire et de l'utilisateur. La plupart des articles font état de «résultats» (ex. la proportion de patients TB testés pour le VIH); peu de reports font état des «impacts» ultérieurs tels que les résultats du traitement antituberculeux ou de la thérapie antirétrovirale. Très peu d’études abordent les points de vue des utilisateurs du service ou du personnel, ou les coûts ou le rapport coût- efficacité.

Conclusions

Lors de l'extension des services intégrés, des comparaisons robustes de l'impact des différents modèles sont nécessaires en utilisant des mesures standardisées des résultats.

Integración de los servicios de tuberculosis y VIH en países con ingresos medios y bajos: una revisión sistemática

Objetivos

Dada la necesidad imperativa de llevar a gran escala un servicio integrado de tuberculosis (TB) y VIH en aquellos lugares en los que ambos tienen una gran importancia a nivel de salud pública, hemos intentado reunir lo que se conoce actualmente acerca de la implementación de la integración de los servicios de TB/VIH.

Métodos

Revisión sistemática de estudios que describen una estrategia para facilitar la integración de los servicios de TB y VIH, mediante una búsqueda en 15 bases de datos bibliográficas, incluyendo Medline, Embase y la Cochrane; así como resúmenes relevantes en congresos.

Resultados

63/1936 artículos con revisión por pares y 70/170 resúmenes cumplían con nuestros criterios de inclusión. Identificamos cinco modelos: vía de entrada al servicio de TB, con remisión a la unidad de prueba y cuidados para VIH; vía de entrada al servicio de TB, prueba para VIH in situ, y remisión para el tratamiento y cuidado del VIH; entrada vía servicios para VIH con remisión para pruebas y tratamiento de TB; entrada vía servicios VIH, prueba TB in situ y remisión para el diagnóstico y tratamiento de la TB; y prestación de servicios para TB y VIH en una misma instalación. Los modelos basados en la remisión de pacientes son los más fáciles de implementar, pero el fallo en la remisión es un riesgo clave. Una integración más cercana requiere de más personal entrenado y de infraestructuras adicionales (por ejemplo un lugar privado para realizar el aconsejamiento para VIH; historias clínicas integradas.) El control de la infección es una preocupación importante. Los modelos con una mayor integración tienen eficiencias potenciales tanto desde la perspectiva del proveedor como de la del usuario. La mayoría de las publicaciones reportan “resultados” (por ejemplo la proporción de pacientes con TB que se hicieron la prueba del VIH); pocos informan de “impactos” resultantes, tales como los resultados obtenidos en el tratamiento de la TB o sobre la terapia antirretroviral. Hay muy pocos estudios en los que se tengan en cuenta las perspectivas de los usuarios del servicio o del personal clínico; o costes o coste-efectividad.

Conclusiones

Mientras que se llevan a escala los servicios integrados, se requieren comparaciones robustas sobre el impacto de los diferentes modelos utilizando mediciones estándares para los resultados.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

Tuberculosis (TB) is a major cause of illness and death worldwide with an estimated 8.8 million incident cases and 1.3 million deaths in 2010 (WHO 2011a). The human immunodeficiency virus (HIV) pandemic has driven the resurgence in TB since the 1990s, with sub-Saharan Africa most severely affected by the confluence of these epidemics. In 2010, the African region accounted for 80% of the 1.1 million global HIV-positive TB cases (WHO 2011a). Globally, HIV prevalence among TB cases varies greatly; for example, in 2010, 1.7% of TB patients in China had HIV infection, whereas in South Africa 60% did (WHO 2011a).

In 2004, WHO published an interim policy on TB/HIV collaborative activities with three domains: establishing mechanisms for collaboration; reducing the burden of TB in people living with HIV; and reducing the burden of HIV in people living with TB (WHO 2004). Revised indicators for monitoring and evaluating collaborative TB/HIV activities were defined in 2009 (WHO 2009a). The Global Plan to Stop TB (2006–2015) included a series of milestones for implementation of collaborative TB/HIV activities (Gunneberg et al. 2008). Yet, in 2010, only 34% of TB patients underwent HIV testing, and only 46% of HIV-positive TB patients received antiretroviral therapy (ART; WHO 2011a). Thus, there is an urgent need for effective integration of TB and HIV services, but little information is available about how best to achieve this. To address this gap, our aims were to synthesise knowledge concerning experience with different models of integration of TB and HIV services in low- and middle-income countries (LMIC); to summarise barriers and enablers of integration; and to make recommendations for research and policy.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

Study design

We conducted a systematic review of literature describing the delivery of integrated TB and HIV services at health facility level in LMIC.

Data sources

We searched 15 bibliographic databases from inception to May 2010 including Medline, Embase and the Cochrane library; abstracts from the International AIDS Society Online Resource Library from 2006 to 2009 and the HIV/AIDS Implementers' Meetings from 2007 to 2009.

Study selection

Studies (published articles, abstracts and reports) were included if they were a description or evaluation of service delivery aiming to increase the integration between HIV and TB care. Studies had to describe practical experience of integration, not a theoretical account, in a low- or middle-income country as defined by the World Bank (The World Bank Group 2010), in English, Spanish, French, German, Portuguese or Russian.

To ensure a comprehensive descriptive review, we did not exclude studies based on their design or absence of outcome measures. We did not restrict by ‘quality measures’ but excluded ‘fatally flawed’ studies (Dixon-Woods et al. 2006).

Search strategy

The search strategy (Box 1) was consistent with other systematic reviews of health service integration (Briggs & Garner 2006; Atun et al. 2010). Briefly, we used the terms ‘tuberculosis’ AND ‘HIV’ (and their synonyms) AND (‘delivery of health care, integrated’ OR a range of other terms representing integration) AND a list of LMIC, as defined by the World Bank.

Box 1. Search strategy for Medline

  1. (exp Tuberculosis OR (tuberculo* or TB).mp. [mp = title, original title, abstract, name of substance word, subject heading word, unique identifier]).
  2. (exp HIV OR (Human immunodeficiency virus or HIV).mp. or HIV/AIDS [mp = title, original title, abstract, name of substance word, subject heading word, unique identifier] OR exp HIV Infections.
  3. ((vertical or horizontal or integrat* or coordinat* or link*) and (program* or care or service*)).mp. [mp = abstract, title, original title, broad terms, heading words].
  4. Exp ‘Delivery of Health Care, Integrated'.
  5. List of low- and middle-income countries as defined by the Word Bank including subject heading and mp.
  6. 1 and 2 and (3 or 4) and 5.

We hand-searched reference lists and consulted experts to check the completeness of the electronic searches and included additional papers as appropriate.

*Truncation term; mp: title, original title, abstract, name of substance word, subject heading word, unique identifier; exp: explode

Data extraction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

HLQ and CM independently screened titles and abstracts of the first 100 studies retrieved by the search. There was high concordance that inclusion criteria could be determined based on the title (92% agreement, kappa 0.76 on titles; 96% agreement, kappa 0.88 on abstracts). Subsequently, all titles were independently screened for inclusion by two researchers, results were compared and disagreement was resolved by consensus. If a study was potentially eligible for inclusion based on the title, its full text was assessed by two reviewers, and studies considered eligible by both reviewers were included in the final analysis. Any disagreement was resolved by consensus, involving the senior author if necessary. The screening process was described by using an adapted Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram (Moher et al. 2009).

Data were extracted using a standardised proforma (Appendix Standardised proforma for data extraction). Initially, HLQ and CM independently extracted data and compared the results, to assure a common data extraction strategy and for quality control (undertaken for 60% of reports). Subsequently, in view of time constraints, data were extracted by one researcher and verified against the source document by the second.

Data synthesis and analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

We organised the review findings around two conceptual frameworks: a framework of health system functions based on WHO's six operational ‘building blocks’ (WHO 2007), useful for identifying system constraints (Shakarishvilli 2009); and a framework – which emerged from the reviewed literature – of integration models based on the route by which the user enters the integrated service.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

Systematic search results

Our search yielded 1954 papers and 170 conference abstracts (Figure 1). The titles of 1516 non-duplicated records were screened for inclusion, of which 275 full-text records were further assessed for eligibility, and 133 included in the review. Only studies in English fulfilled the inclusion criteria.

image

Figure 1. Results of search strategy.

Download figure to PowerPoint

Models of integration of HIV and TB services: a framework

Models of integration range from providing referral between services with no shared service provision (least integrated), through to care provision for both HIV and TB by the same healthcare worker (most integrated). We identified five models of integrated care, based on how the service user enters the healthcare system (Figure 2).

image

Figure 2. Models of integration.

Download figure to PowerPoint

Where users enter the system through TB services, in the least integrated model (‘TB refers’), they are referred to HIV services for HIV testing and subsequent care (ART and co-trimoxazole preventive therapy). With closer integration, TB clinics may test for HIV on-site, but continue to refer to a separate service for HIV treatment (‘TB: tests and refers’). Where users enter through HIV services, in the least integrated model (‘HIV refers’), they are referred to a TB clinic for screening for active TB; in a more integrated model (‘HIV: screens and refers’), TB screening is undertaken within the HIV clinic, but users are referred to TB services for treatment. In the most integrated model (‘Single facility’), treatment for both HIV and TB is provided in one health facility.

Integrated care models: evolution and geographical distribution

The number of studies of HIV and TB service integration has increased over the last decade. Early studies generally involved referral, with or without on-site testing; the first ‘single facility’ study was published in 2001. Table 1 illustrates the geographical origin of reports, subdivided by the model of care. In settings where HIV prevalence among TB patients is low (e.g. India, former Soviet Union), most studies were of referral-based integration. Most ‘single facility’ studies were from sub-Saharan Africa.

Table 1. Models of integration according to geographical locationa
ModelTB refersTB: tests and refersSingle facilityHIV: tests and refersHIV refers
  1. a

    One report unless otherwise stated (N in brackets).

  2. The number of examples does not equal the number of reports because some reports describe more than one example (either an example in more than one category, or an example in more than one country). DRC, Democratic Republic of Congo.

Eastern EuropeRussiaRussia  Russia
Ukraine
AfricaMalawi (2)South Africa (5)South Africa (12)Zambia (3)Kenya (2)
DRCDRC (3)Uganda (7)South Africa (3)Mozambique
EthiopiaMalawi (2)Tanzania (5)Malawi (2)Rwanda
MozambiqueMozambique (2)Kenya (4)Mozambique (2)South Africa
RwandaKenya (2)DRC (3)Tanzania (2)Tanzania
South AfricaZambia (2)Rwanda (2)Uganda (2) 
TanzaniaCameroonZambia (2)Ethiopia 
KenyaCote d'IvoireBeninKenya 
 LesothoEthiopiaNigeria 
 TanzaniaLesothoRwanda 
 UgandaLiberia  
  Malawi  
  Nigeria  
AsiaIndia (4)Indonesia (2)Myanmar (2)Cambodia (2)India (4)
CambodiaVietnam (2)Thailand (2)Vietnam (2)Cambodia
KyrgyzstanThailandIndiaIndia 
TajikistanIndia Myanmar 
Uzbekistan  Thailand 
Latin America/Caribbean GuatemalaGuatemalaHaiti (3) 
  Haiti (3)  

‘TB refers’: entry via TB service, referral for HIV testing and care

In this model, described in 12 papers, three abstracts and one report, TB services refer patients to services providing HIV testing, with or without subsequent HIV care. For example, HIV testing was implemented among patients attending TB services in India by referral to the Integrated Counselling and Testing Centre (ICTC) in the same facility (if available) or at the nearest ICTC (Vijay et al. 2009). Implementing referral of TB patients for HIV counselling and testing requires minimal additional logistical and financial input from health services and has been achieved in India through relatively simple measures, such as joint training of health workers from both programmes, modification of existing record-keeping systems and referral forms and regular meetings of staff from both services to strengthen referral linkages (Shetty et al. 2008).

Strengths of this model include the relative simplicity of introducing the required measures and the low cost to health services. The key weakness is loss to follow up if referral fails. For example, in India, the source of most published reports, reasons for referral failure included lack of a standardised tool to determine which TB patients should be referred for HIV testing, and no log of referrals made (Shetty et al. 2008). In studies from Cambodia, Democratic Republic of Congo and India, concerns reported by patients about referral for HIV testing included travelling long distances, transport costs, lack of pre-HIV test counselling and stigma attached to attending an HIV testing centre (Laserson & Wells 2007; Van Rie et al. 2008; Thomas et al. 2009).

‘TB: tests and refers’: entry via TB service, HIV test on-site, referral for HIV care

In this model, described in 19 papers and 16 abstracts, TB clinics offer HIV testing on-site. In Malawi, a cross-sectional study in 2000 described an initiative to integrate HIV testing into a TB clinic. Trained HIV counsellors performed point of care HIV testing in a private space within the clinic; co-trimoxazole preventive therapy and management of opportunistic infections were provided (Zachariah et al. 2003). The main barriers to integration were the additional financial and human resources needed to implement the programme.

This model requires more investment in infrastructure than ‘TB refers’ (for both HIV testing and treatment). In a survey of TB clinics in Kinshasa, Democratic Republic of Congo (DRC), barriers to integrated care included lack of access to a counselling room; lack of standardised recording systems for HIV results; insufficient staff trained in HIV testing; inefficient use of skilled staff; and poor access to care and support for people with HIV (Martinot et al. 2008). In some cases, because TB clinic staff were already overburdened, an additional staff member was appointed to facilitate integration (Nguyen et al. 2009).

‘HIV refers’: entry via HIV service, referral for screening and treatment for TB

Like ‘TB refers’, this model, described in six papers, four abstracts and one report, is based on referral for screening and treatment and faces similar issues. Few studies described how patients were selected for referral: criteria included symptoms such as cough (Wambua et al. 2007; Ramachandran et al. 2009); or routine referral for HIV-positive patients (Micek 2005; USAID 2008).

As with ‘TB refers’, referrals were implemented by measures such as joint training of health workers from both programmes, use of existing record-keeping systems and referral forms and regular meetings between staff of the two services. Relatively little additional logistical or financial input was required.

‘HIV: screens and refers’: HIV service entry, screening, then referral for TB treatment

In this model, described in 11 papers, two reports and 17 abstracts, the infrastructure needed for integration of services varied considerably, depending on whether isoniazid preventive therapy (IPT) was offered by the HIV clinic, which was reported to require health workers trained to provide IPT; a system for procuring and storing tuberculin for skin testing; and medical officers to exclude active TB among sputum-negative clients (Ssemafumu et al. 2007). In clinics not providing IPT, requirements for HIV centres to manage TB included tools for screening, prevention, treatment and quality control (Jacquet et al. 2006). This model (if not providing IPT) may require less infrastructural investment than ‘TB: tests and refers’, as there is no need to provide private space for HIV testing. In the ‘HIV: screens and refers’ model, TB screening is provided, which requires screening guidelines and consideration of infection control issues if sputum samples are collected on-site.

In programmes providing IPT in HIV care facilities, authors recommend ensuring adequate isoniazid supplies (Ssemafumu et al. 2007), effective referrals between TB and HIV services, accurate recording and reporting of TB/HIV data and human resources to supervise and monitor the programme (Vandebriel et al. 2007).

‘Single facility’: TB and HIV services provided at a single facility

In single facility models, described in 16 papers, 26 abstracts and two reports, encompassing services in one or adjacent buildings, the risk of nosocomial spread of TB was the commonest concern (Dong et al. 2007; Gandhi et al. 2009), requiring in all cases infection control measures such as open-air environments and adequate ventilation of the premises. The greatest risk of nosocomial TB transmission to HIV-infected individuals is from patients with undiagnosed infectious TB (Gandhi et al. 2009). Service integration should improve early diagnosis and treatment of TB, which would reduce the risk of TB transmission, not increase it. Active TB case finding amongst all HIV-infected individuals, waiting areas that maximise natural ventilation (preferably outdoors) and designated ‘cough monitors’ (staff members who quickly identify coughing patients in waiting areas and isolate them) are recommended to improve infection control (Gandhi et al. 2009).

One of the biggest challenges of this model relates to the different cultures in the management of HIV and TB programmes and differences in nurse roles in the two programmes (Musa et al. 2009). Programmes can learn from each other to improve patient care, for example, the more patient-centred approach of ART programmes could be replicated in TB programmes (Dong et al. 2007).

Reported barriers to the equitable expansion of access to services in the ‘single facility’ model of integration usually relate to ART provision. These include over-rigorous adherence to ART programme guidelines, for example, the requirement for patients to attend adherence training sessions, even if very sick; or the requirement that a ‘treatment buddy’ accompany the patient (which may result in postponed treatment; Dong et al. 2007).

Box 2. Barriers to integrated HIV and TB services

Service delivery

  • Service users unaware, or unconvinced, of the importance of testing for HIV or tuberculosis.
  • Users reluctant to undergo testing because of stigma, particularly tuberculosis patients unwilling to undergo HIV testing (especially if this involves attending a facility identified as serving people with HIV).
  • In models based on referral, barriers to access of the relevant service: distance between facilities, cost of travel, being too ill to travel, inconvenient opening hours.
  • Staff using inappropriate criteria to decide which patients need testing.
  • In models based on referral, poor communication between services, and failure to share information concerning the care of individual patients.
  • Lack of access to care and support for people diagnosed with HIV.
  • Facilities not designed to facilitate infection control for tuberculosis, both in general areas and specifically for areas where sputum samples are produced.
  • Lack of private space for HIV counselling and testing.
  • Data recording systems poorly designed for coordinated care.
  • Ineffective referral systems.

Human resources

  • Lack of staff trained to manage both HIV and tuberculosis.
  • High staff turnover, requiring continuing training activities.
  • In models where additional activities are introduced, staff already overburdened.
  • Staff attitudes: not motivated to carry out co-ordinating activities, particularly if this is perceived as extra work; may be reluctant to implement HIV testing early in tuberculosis treatment.

Supply of medicines and products

  • Lack of IPT, suitably formulated (300 mg tablet) and packaged.
  • Unreliable supplies, including drugs and point of care HIV tests.
  • Shortage of ART in general, and specifically shortage of regimens suitable for people taking tuberculosis treatment.
  • Where tuberculin skin tests are used to determine IPT eligibility, lack of tuberculin and space for its cold storage.

Several studies suggest that this model of care requires multiple components to work effectively, including training, linking providers within a facility and throughout the down-referral network and maximising access to existing support services (Dong et al. 2007). Nonetheless, this model appears to be preferable for the patient (Levin et al. 2006), particularly the less well off, whose travel costs should be reduced. The most integrated models also overcome the problem of losing patients when referrals between services fail (Wandwalo et al. 2004).

There are few studies describing patients' attitudes towards integrated services providing HIV and TB treatment. The vast majority of patients receiving care at an urban South African TB clinic reported willingness to attend an integrated TB/HIV service (Levin et al. 2006), although the main limitation of this study was that it involved a small sample of individuals who were already accessing TB services (Levin et al. 2006).

Barriers to integration

Reported barriers to integrated services are summarised in Box 2. For service users, in models based on referral, the cost of travel between separate facilities was a major obstacle. For service providers, key issues were training and retaining skilled staff and, in integrated models, provision of space for specialised activities such as HIV testing. In all models, infection control emerged as an important issue requiring managerial, administrative and environmental control measures (WHO 2009b). The few studies describing enablers of integration mentioned, in referral models, activities such as training, which brought staff from both services together; and having a staff member with specific responsibility for integration activities. No studies specifically mentioned experience of improving retention in referral models.

Measures of effectiveness of integration

Adopting terminology from WHO's guide to monitoring and evaluation of collaborative TB/HIV activities (WHO 2009a), we defined ‘outcomes’ as ‘coverage’ of integration activities, such as the number of patients with TB tested for HIV, and ‘impacts’ as longer-term biological measures, such as mortality among patients in TB or HIV treatment programmes.

Ninety-four studies described at least one measure of ‘effectiveness’, primarily outcomes representing a measure of coverage of integration, such as undergoing HIV testing, starting co-trimoxazole preventive therapy or being referred to the other service. Fewer studies described relevant impacts such as outcomes of TB treatment (proportion cured, died, etc.) or outcomes relevant to HIV care such as mortality at a specified period after ART initiation (Table 2).

Table 2. Measures of effectiveness of integration, by the model of integration
Model of integrationTypes of outcome indicator reported (N reporting outcome)Types of impact indicator reported (N reporting impact)
  1. TB, tuberculosis; CPT, co-trimoxazole preventive therapy; ART, antiretroviral therapy; IPT, isoniazid preventive therapy.

  2. Only the most commonly reported indicators are shown, and there may be more than one indicator shown in each report; hence, the number of examples does not always equal the total reported in the first column. TB treatment outcomes include death on TB treatment.

TB refers (N = 16, 14 report a measure of effectiveness)% TB patients referred for HIV testing (10)TB treatment outcomes (2)
% TB patients with an HIV test result (8)Death (1)
% HIV-positive TB patients starting CPT (7) 
% HIV-positive TB patients starting ART (4) 
TB: tests and refers (N = 35, 25 report a measure of effectiveness)% TB patients with an HIV test result (23)TB treatment outcomes (1)
% HIV-positive starting ART (11) 
% HIV-positive TB patients starting CPT (10) 
Single facility (N = 44, 29 report a measure of effectiveness)% HIV-positive starting/continuing ART (14)TB treatment outcomes (11)
% TB patients with an HIV test result (14)Death (4)
% HIV-positive screened for TB (7) 
% HIV-positive starting/continuing CPT (7) 
% HIV-positive starting IPT (2) 
HIV: screens and refers (N = 30, 19 report a measure of effectiveness)% HIV-positive screened for TB (15)TB treatment outcomes (2)
% HIV-positive diagnosed with TB (5) 
% HIV-positive starting IPT (5) 
% HIV-positive starting/continuing CPT (4) 
HIV refers (N = 11, 7 report a measure of effectiveness)% referred to TB clinics (4)Death (1)
% HIV-positive screened for TB (4) 
% HIV-positive starting IPT (1) 

Three studies compared outcomes before and after integration activities were introduced (Burua et al. 2008; Harris et al. 2008; Chan et al. 2009). For example, in Malawi, introduction of a daily TB/HIV clinic where all TB patients were tested for HIV and referred for HIV care in the same physical area increased ART uptake from 10% to 30% (Chan et al. 2009).

Two studies compared the outcomes or impacts of different models of care (Miti et al. 2003; Van Rie et al. 2008). For example, a cross-sectional study in rural DRC compared three models of HIV testing for TB patients (TB clinic vs. primary health centre where the TB clinic was based vs. referral to a free-standing HIV testing centre) and found that testing uptake was higher in test on-site models (97.7% in TB clinic, 94.8% in primary health centre) than the referral model (68.5%; Van Rie et al. 2008).

Five studies included some economic evaluation (Hausler et al. 2006; Shrestha et al. 2006; Terris-Prestholt et al. 2008; Sutton et al. 2009; Vassall et al. 2010), but no studies measured cost-effectiveness directly, or compared costs between different models of care.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

Globally, the epidemic of HIV-associated tuberculosis remains poorly addressed, with limited scale-up of integrated TB/HIV interventions. Limited evidence concerning effective delivery models hinders service development, while increasing ART availability makes integration ever more urgent. A recent review described individual interventions in sub-Saharan Africa, but not models of integrated care (Uyei et al. 2011).

The review reveals a diverse range of integrated TB and HIV service delivery activities, evolving from early descriptions of implementation of HIV testing in TB services to models of integration within a single facility, ranging from two separate services under one roof to a single provider delivering both TB and HIV care.

The five-model framework facilitates comparison of the advantages and disadvantages of different integration models (Table 3). Models based on referral only, common where HIV prevalence among TB patients is low, are relatively simple to implement and require minimal changes to existing services, primarily staff training, but depend critically on a robust referral system.

Table 3. Summary of advantages and disadvantages of integration models
Model of integration     
 TB refersTB: tests and refersSingle facilityHIV: screens and refersHIV refers
  1. ART, antiretroviral therapy; IPT, isoniazid preventive therapy.

Potential advantagesEasy to implement, requires minimal additional resources.

Increases uptake of HIV testing.

Reduces potential stigma associated with attending HIV testing centre.

More efficient than referral-only model, as reduces unnecessary referrals.

Cross-fertilisation of expertise & best practice approaches between services.

Rapid diagnosis & treatment of infectious TB should reduce risk of transmission.

Potential to increase ART uptake among HIV+ TB patients.

Users benefit from lower transport costs.

Minimises lost referrals.

Increases TB case detection.

Reduces nosocomial TB transmission to patients and staff.

Facilitates IPT provision.

More efficient than referral-only model, as reduces unnecessary referrals.

Easy to implement, requires minimal additional resources.

May increase accessibility/availability of TB services to marginalised populations.

Special considerations if high HIV prevalence Routine testing should maximise access to ART.

Efficient for providers & patients; where most TB patients have HIV, particular gains by avoiding need for referral.

Better coordinated care for most patients.

  
Potential disadvantages

Disadvantages common to all referral-based models:

Referral may fail, especially if ineffective referral criteria and/or referral pathway complex.

Patient costs increased by attending multiple clinics. Potential risks to patient if communication inadequate between services.

Difficult for referring service to monitor outcomes at receiving service.

Requires financial and human resources to implement and adequate space.

System vulnerable to stock-outs of HIV test kits.

Requires standardised reporting system for HIV results.

Plus all disadvantages of referral-based models

Potential high risk of TB transmission if infection control inadequate.

May require significant investment to implement.

Requires staff training and suitably ventilated space where sputum specimens can be produced.

IPT implementation may need additional trained staff.

Plus all disadvantages of referral-based models.

All disadvantages of referral-based models.
Special considerations if high HIV prevalenceNosocomial spread of TB. HIV service staff may be over-burdened by high referral load.Selective criteria for HIV testing may result in inequitable access to services.

More patients with HIV- associated immunosuppression at risk of acquiring TB if infection control inadequate.

TB-only patients may experience slower service.

 TB service staff may be overburdened by high referral load.

Implementing HIV testing in TB services and vice versa requires more organisational changes, particularly in terms of staff training and time, but may also require changes to physical infrastructure to make private space for HIV counselling and testing and suitably ventilated space where sputum specimens can be produced. While these models are theoretically more efficient – in that they identify more accurately individuals who need referral to the other service and thus should reduce the number of unnecessary referrals – they share the same critical dependence on a robust referral system for individuals who require further assessment and/or treatment. The biggest challenge specific to the ‘HIV: screens and refers’ model is to systematise TB screening as part of routine HIV care. The new WHO recommendations on intensified TB case finding, based on any of cough, weight loss, fever or night sweats (WHO 2011b), have helped promote coverage of screening (WHO 2011a).

Models where care is integrated within one facility are likely to require additional resources, because few services have been designed with integration in mind. Referrals may still be necessary if the two services are delivered in separate rooms within one facility, but close proximity should facilitate effective referral. Users should benefit by having lower transport costs if visits are on the same day or, better still, if one provider manages both TB and HIV at the same visit. Data documenting such benefits are, however, lacking. TB transmission is a key risk: integrated facilities have the potential for harm if individuals with HIV-associated immunosuppression have increased exposure to infectious TB cases (as compared with physically separated services), and measures to minimise this risk, including cough screening, are essential. However, the greatest risk of transmission is from those with undiagnosed infectious TB. Thus, if integrated care promotes rapid diagnosis and treatment of TB, the net effect may be a reduction in the risk compared with separate services.

Limitations of this review include publication bias. Inclusion of recent abstracts as well as peer-reviewed articles limits this bias to some degree; however, our review cannot encompass the wealth of unpublished practical experience. In addition, most studies reflect small projects rather than national initiatives. Our review does not touch on the broader question concerning the advantages and disadvantages of integration of TB and HIV services into primary health care (Maher 2010).

Prompt commencement of ART among individuals with HIV-associated TB and rapid identification, treatment and prevention of TB among individuals with HIV infection are key to saving lives and addressing the HIV/TB epidemic. Much effort has gone into developing an array of better integrated services, despite multiple obstacles, including separate programme management and geographical separation of services.

Integrated service delivery models should focus on the one patient with two life-threatening diseases who often has to navigate complex health systems to secure access. Intuitively, in settings with the greatest confluence of TB and HIV, care integrated within one facility should serve patients best (with caveats around TB infection control) and promote efficient use of resources. The balance of costs vs. benefits of closer integration is likely to differ where the prevalence of HIV among TB patients is lower (Table 3). However, the advantages of, at minimum, testing for HIV in TB services (and vice versa) seem to outweigh the disadvantages at any HIV prevalence and offer a relatively simple first step towards closer integration.

There are few robust data comparing models of TB and HIV service integration to aid policy makers and service providers. Few studies have measured patient-relevant impacts, made direct comparisons of effectiveness or cost-effectiveness between models, or reported user or provider perspectives. To support policy making, programmes should report standardised measures of coverage of integrated services. Scaling up of TB/HIV services should be underpinned by operational research to investigate potential efficiencies of integrated care from provider and service-user perspectives. Robust comparisons of the impacts of different models should compare facilities as the unit of observation, ideally using phased implementation or cluster-randomised designs.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information

HL-Q, CM, PK, AG had financial support from the Global Fund to Fight AIDS, Tuberculosis and Malaria and the Stop TB Department of the World Health Organization, who commissioned this work as a background paper for the First Global Symposium on Health Systems Research to develop evidence-based policies on service delivery. HL-Q, PK and AG were also supported by the Evidence for Action research consortium, a project funded by the Department for International Development (DfID), for the benefit of developing countries. The views expressed here are not necessarily those of DfID.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information
  • Atun R, De Jongh T & Secci F (2010) A systematic review of the evidence on integration of targeted health interventions into health systems. Health Policy and Planning 25, 114.
  • Briggs C & Garner P (2006) Strategies for integrating primary health services in middle- and low-income countries at the point of delivery. The Cochrane Database of Systematic Reviews 2, CD003318.
  • Burua A, Musisi A, Tumwesigye B et al. (2008) The integration of tuberculosis-HIV/AIDS Care in 89 health facilities in Uganda [Abstract no. 554]. HIV/AIDS Implementers' Meeting. Kampala, Uganda.
  • Chan A, Njala J, Kanyerere H & Van Lettow M (2009) Improved uptake of antiretroviral therapy (ART) following integration of TB and HIV services in a district in southern Malawi [Abstract no. CDD062]. 5th IAS Conference on HIV Pathogenesis Treatment and Prevention. Cape Town, South Africa.
  • Dixon-Woods M, Cavers D, Agarwal S & Annandale E (2006) Conducting a critical interpretive synthesis of the literature on access to healthcare by vulnerable groups. BMC Medical Research Methodology 6, 35.
  • Dong K, Thabethe Z, Hurtado R et al. (2007) Challenges to the success of HIV and tuberculosis care and treatment in the public health sector in South Africa. Journal of Infectious Diseases 196, S491S496.
  • Gandhi NR, Moll AP, Lalloo U et al. (2009) Successful integration of tuberculosis and HIV treatment in rural South Africa: the sizonq'oba study. Journal of Acquired Immune Deficiency Syndromes 50, 3743.
  • Gunneberg C, Reid A, Williams BG, Floyd K & Nunn P (2008) Global monitoring of collaborative TB-HIV activities. The International Journal of Tuberculosis and Lung Disease 12, 27.
  • Harris JB, Hatwiinda SM, Randels KM et al. (2008) Early lessons from the integration of tuberculosis and HIV services in primary care centers in Lusaka, Zambia. The International Journal of Tuberculosis and Lung Disease 12, 773779.
  • Hausler HP, Sinanovic E, Kumaranayake L et al. (2006) Costs of measures to control tuberculosis/HIV in public primary care facilities in Cape Town, South Africa. Bulletin of the World Health Organization 84, 528536.
  • Jacquet MS, Auguste M & Grand'pierre R (2006) Integration of TB services at HIV centers in Haiti [Abstract no. TUPE0830]. XVI International AIDS Conference. Toronto, Canada.
  • Laserson K & Wells C (2007) Reaching the targets for tuberculosis control: the impact of HIV. Bulletin of the World Health Organization 85, 377386.
  • Levin L, Irving K, Dikgang M, Punwasi J, Isaacs M & Myer L (2006) TB patients' perspectives on integrated TB/HIV services in South Africa. Tropical Doctor 36, 173175.
  • Maher D (2010) Re-thinking global health sector efforts for HIV and tuberculosis epidemic control: promoting integration of programme activities within a strengthened health system. BMC Public Health 10, 394.
  • Martinot A, Van Rie A, Mulangu S, Mbulula M & Jarrett N (2008) Baseline assessment of collaborative tuberculosis/HIV activities in Kinshasa, the Democratic Republic of Congo. Tropical Doctor 38, 137141.
  • Micek M (2005) Integrating TB and HIV Care in Mozambique: Lessons From an HIV Clinic in Beira [Online]. Health Alliance International, Seattle. http://pdf.usaid.gov/pdf_docs/PNADG189.pdf (accessed 26 October 2010).
  • Miti S, Mfungwe V, Reijer P & Maher D (2003) Integration of tuberculosis treatment in a community-based home care programme for persons living with HIV/AIDS in Ndola, Zambia. The International Journal of Tuberculosis and Lung Disease 7, S92S98.
  • Moher D, Liberati A, Tetzlaff J & Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 339, b2535.
  • Musa BM, Gebi U, Etiebet M et al. (2009) TB/HIV collaboration in Nigerian teaching hospitals: challenges and opportunities [Abstract no 1442]. HIV/AIDS Implementers' Meeting. Windhoek, Namibia.
  • Nguyen NV, Le TH, Trinh TT et al. (2009) Expansion of provider-initiated HIV testing and counseling for TB patients to 14 provinces in Vietnam 2007–2008 [Abstract no. 706]. HIV/AIDS Implementers' Meeting. Windhoek, Namibia.
  • Ramachandran R, Chandrasekaran V, Muniyandi M, Jaggarajamma K, Bagchi A & Sahu S (2009) Cross-referral between HIV counselling and testing centres and smear microscopy centres in Tamil Nadu. The International Journal of Tuberculosis and Lung Disease 13, 221225.
  • Shakarishvilli G (2009) Building on Health Systems Frameworks for Developing a Common Approach to Health Systems Strengthening [Technical Workshop on Health Systems Strengthening]. World Bank, Global Fund to Fight AIDS, Tuberculosis and Malaria, and GAVI Alliance, Washington DC.
  • Shetty PV, Granich RM, Patil AB et al. (2008) Cross-referral between voluntary HIV counselling and testing centres and TB services, Maharashtra, India 2003–2004. The International Journal of Tuberculosis and Lung Disease 12, 2631.
  • Shrestha RK, Mugisha B, Bunnell R et al. (2006) Cost-effectiveness of including tuberculin skin testing in an IPT program for HIV-infected persons in Uganda. The International Journal of Tuberculosis and Lung Disease 10, 656662.
  • Ssemafumu E, Mabirizi D, Kasasa S, Luzze H & Muswani I (2007) Evaluation of the Integration of Isoniazid Prophylactic Therapy (IPT) in HIV/AIDS Prevention and Control Programs at the AIDS Information Center (AIC) Uganda. Final report. Management Systems International, Washington, D.C.
  • Sutton BS, Arias MS, Chheng P, Eang MT & Kimerling ME (2009) The cost of intensified case finding and isoniazid preventive therapy for HIV-infected patients in Battambang, Cambodia. The International Journal of Tuberculosis and Lung Disease 13, 713718.
  • Terris-Prestholt F, Kumaranayake L, Ginwalla R et al. (2008) Integrating tuberculosis and HIV services for people living with HIV: costs of the Zambian ProTEST initiative. Cost Effectiveness and Resource Allocation 6, 2.
  • The World Bank Group (2010) Country and Lending Groups [Online]. The World Bank. http://data.worldbank.org/about/country-classifications/country-and-lending-groups#Lower_middle_income (accessed 3 April 2010).
  • Thomas BE, Dewan PK, Vijay S et al. (2009) Perceptions of tuberculosis patients on provider-initiated HIV testing and counseling-a study from south India. PLoS ONE 4, e8389.
  • USAID (2008) Integrating TB and HIV Services in the General Health Care System in Russia. USAID, Bethesda, MD.
  • Uyei J, Coetzee D, Macinko J & Guttmacher S (2011) Integrated delivery of HIV and tuberculosis services in sub-Saharan Africa: a systematic review. The Lancet Infectious Diseases 11, 855867.
  • Van Rie A, Sabue M, Jarrett N et al. (2008) Counseling and testing TB patients for HIV: evaluation of three implementation models in Kinshasa, Congo. The International Journal of Tuberculosis and Lung Disease 12, S73S78.
  • Vandebriel G, Kabanda G, Turinawe K, Sahabo R, Mugabo J & Gasana M (2007) Early results of implementation of a national policy on TB screening in people living with HIV attending ART clinics in Rwanda [Abstract no. 596]. HIV/AIDS Implementers' Meeting. Kigali, Rwanda.
  • Vassall A, Seme A, Compernolle P & Meheus F (2010) Patient costs of accessing collaborative tuberculosis and human immunodeficiency virus interventions in Ethiopia. The International Journal of Tuberculosis and Lung Disease 14, 604610.
  • Vijay S, Swaminathan S, Vaidyanathan P et al. (2009) Feasibility of provider-initiated HIV testing and counselling of tuberculosis patients under the TB control programme in two districts of South India. PLoS ONE 4, e7899.
  • Wambua N, Odhiambo J, Njoroge A, Phillips E, Juma A & Ombeka V (2007) Implementing TB screening in HIV testing sites in an urban slum in Nairobi, Kenya [Abstract no. 1783]. HIV/AIDS Implementers' Meeting. Kigali, Rwanda.
  • Wandwalo E, Kapalata N, Tarimo E, Corrigan CB & Morkve O (2004) Collaboration between the national tuberculosis programme and a non Governmental organisation in TB/HIV care at a district level: experience from Tanzania. African Health Sciences 4, 109114.
  • WHO (2004) Interim Policy on Collaborative TB/HIV Activities [Online]. World Health Organisation, Geneva. http://www.who.int/hiv/pub/tb/en/Printed_version_interim-policy_2004.pdf (accessed 25 October 2010).
  • WHO (2007) Everybody's Business: Strengthening Health Systems to Improve Health Outcomes: WHO's Framework for Action [Online]. World Health Organization, Geneva. http://www.who.int/healthsystems/strategy/everybodys_business.pdf (accessed 25 October 2010).
  • WHO (2009a) A Guide to Monitoring and Evaluation for Collaborative TB/HIV Activities [Online]. World Health Organization, Geneva. http://www.who.int/hiv/pub/tb/en/guidetomonitoringevaluationtb_hiv.pdf (accessed 25 October 2010).
  • WHO (2009b) WHO Policy on TB Infection Control in Health Care Facilities, Congregate Settings and Households [Online]. World Health Organization, Geneva. http://whqlibdoc.who.int/publications/2009/9789241598323_eng.pdf (accessed 1 August 2010).
  • WHO (2011a) Global Tuberculosis Control 2011 [Online]. World Health Organization, Geneva. http://www.who.int/tb/publications/global_report/2011/gtbr11_full.pdf (accessed 8 November 2011).
  • WHO (2011b) Guidelines for Intensified Tuberculosis Case-finding and Isoniazid Preventive Therapy for People Living with HIV in Resource-constrained Settings [Online]. World Health Organization, Geneva. http://whqlibdoc.who.int/publications/2011/9789241500708_eng.pdf (accessed 30 November 2011).
  • Zachariah R., Spielmann M. P., Harries A. D. & Salaniponi F. M. L. (2003) Voluntary counselling, HIV testing and sexual behaviour among patients with tuberculosis in a rural district of Malawi. The International Journal of Tuberculosis and Lung Disease 7, 6571.

Appendix 1:: Standardised proforma for data extraction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information
Author, year 
Country 
Model of integration 
Integration activities 
Study design 
Author's definition of integration 
Measure of integration assessed 
Suggested measures of integration 
Patient-level measures of effectiveness of integration 
Suggested measures of effectiveness of integration 
Other outcomes measured (e.g. patient preferences) 
Other outcomes suggested 
Strengths of integration model 
Weaknesses of integration model 
Barriers to integration 
Research priorities identified by authors 
Impact on other health services 
Comments 

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Data extraction
  6. Data synthesis and analysis
  7. Results
  8. Discussion
  9. Funding
  10. References
  11. Appendix 1:: Standardised proforma for data extraction
  12. Supporting Information
FilenameFormatSizeDescription
tmi12029-sup-0001-TableS1.docWord document649KTable S1. Studies included in the review.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.