2001-2005: The vertical scale up of HIV/AIDS services
An HIV prevention structure was initiated in 2001. VCT constituted the main element of the approach in Manica and Sofala provinces, and was established as a separate programme with its own freestanding sites and data gathering system. PMTCT services were initiated in 2001 in selected health centres, with parallel data collection and activity duplication for maternal & child health nurses.
The initial approach to ART scale up in 2004 focused on a vertical, donor-initiated, day hospital model in which new freestanding HIV treatment hospitals were constructed in large population centres alongside existing hospital compounds. Day hospitals included their own pharmacies, data systems, health workforce, waiting areas and receptions. Using this separate infrastructure, patients identified as HIV positive from other sectors of the health system (VCT, PMTCT, blood bank and laboratory) were referred to day hospitals to register for HIV care, and to follow a sequence of visits for clinical staging, CD4 testing, social worker visits, treatment for opportunistic infections, and initiation and follow up of ART.
The day hospitals included specifically allocated staff (often expatriate) and better working conditions than other sectors. This vertical approach may have contributed to high loss-to-follow-up rates and missed opportunities that limited the uptake of patients initiating ART. All the data presented in this case description are derived from routine health system data systems. Paper registries are used to collect facility-level data that are later computerized at district level into the MOH health information system. Health Alliance International technical advisors supported data collection, compilation and analysis for programme evaluation to produce findings presented here.
ART scale up
In Manica and Sofala, the first day hospitals were completed by 2004 in the cities of Chimoio and Beira, respectively. In the first two years, the day hospitals successfully placed nearly 4000 patients on ART. However, providers and planners soon realized that the vertical model had major limitations:
• Day hospitals were only accessible to local urban populations.
• Major loss to follow up (LTFU) at a number of steps in the treatment cascade limited patient uptake. In 2005, only 78% of HIV-positive patients referred to day hospitals returned for CD4 testing, and only 46% of those who returned for results and were found to be ART eligible succeeded in starting treatment.
• Poor linkages with other specific services contributed to LTFU, missed opportunities for testing, and low referral rates.
• Greater human and material resources for HIV-related activities, including salary top-ups, created resentment and limited support from other sectors of the health system.
• Day hospital carrying capacities limited new patient registration.
• Allocation of HIV resources did not strengthen the wider system.
Voluntary counselling and testing
VCT sites began referrals to vertical day hospitals when ART became available, but distance between facilities contributed to high loss-to-follow-up rates. In 2005, only 59% of those testing positive at VCT sites managed to enrol in HIV care at the day hospitals. CT was not offered in general outpatient and inpatient wards. Doctors could only refer suspected HIV cases to separate freestanding VCT sites. After nearly two years of ART scale up, only 5% of TB patients had been tested before integration efforts began in late 2005. HIV patients were also not being routinely tested and referred to TB care.
Opportunistic infection identification and management
The initial vertical approach prevented the referral of patients presenting with HIV-related opportunistic infections (OIs) in other service areas because clinicians outside day hospitals were not trained to recognize OIs or make referrals. The missed opportunities were compounded by the lack of provider-initiated CT in outpatient or inpatient services. Even if a clinician did recognize an OI, she had to refer the patient to an off-site VCT centre, where HIV-positive patients would be referred back to the day hospital, creating additional visits and greater loss to follow up.
Prevention of mother to child transmission
PMTCT was initially established as a prevention activity that focused on single-dose nevirapine distribution coordinated through antenatal care services. When ART became available, HIV-positive mothers were referred to day hospitals for treatment, but in practice they suffered high loss to follow up as few women managed to register. In 2005, only 30% of pregnant women who tested positive in PMTCT programmes enrolled at the day hospitals. By the end of 2005, only 20% of eligible mothers had initiated ART.
By 2005, separate data, pharmacy systems, supervision and infrastructure had been set up in two provinces to support the ART scale up. Dozens of new staff, including doctors, nurses, physicians' assistants and administrative staff, had been placed in the day hospitals to focus on ART.
2005-2008: Integration and decentralization of HIV/AIDS services
By mid-2005, the MOH recognized these challenges and determined that both decentralization and integration of HIV-related services would be necessary to increase coverage, maximize efficiency and improve quality. Decentralization was necessary to expand coverage to widely distributed populations. Integration would make decentralization possible by maximizing utilization of limited space, infrastructure and health workforce, while improving system efficiency and quality through better service linkages to reduce LTFU and missed opportunities.
Decentralization and integration of ART into PHC
ART initially expanded into rural hospitals and health centres in areas with higher population densities, while sites were established at smaller centres that would receive regular visits by teams of ART providers. Sites were chosen based on geographic coverage needs, patient volume and assessments of sufficient infrastructure. By December 2008, ART was being provided in 67 sites (of a total 222 health facilities), distributed throughout all 23 districts in the two provinces, where an estimated 360,000 were HIV positive and an estimated 75,000 were ART eligible. Nearly 180,000 people living with HIV/AIDS (PLHIV) registered in the system, and nearly 30,000 had initiated ART out of an estimated 60,000 determined to be ART eligible through CD4 counts.
The actual model of ART provision varies somewhat by site, depending on space, available workforce, and patient volume. In some sites, there are still staff members dedicated for ART, but they are co-located in the same facilities where other key services are provided. In other smaller sites, health centre clinicians were trained in ART provision, which was then integrated with other routine activities, including the regular outpatient consults, inpatient treatment, and ANC. Health Alliance International helped cover "gap year" funding (while the MOH prepared to absorb the new hires) for newly trained nurses and physician's assistants to speed new personnel into the PHC system to help mitigate workforce shortages and support overburdened staff.
Impact on patient flow and loss to follow up
Analysis of routine data comparing the previous day hospital vertical sites with newer and smaller integrated sites shows that a greater percentage of local PLHIV managed to start ART, and that the percentage of patients initiating ART in less than 90 days (from registration at a health unit) was significantly greater in integrated sites than in vertical day hospitals (see Figure ). Co-location of ART services with other PHC services in the same facilities reduced LTFU from testing to registration for HIV care, as well as the time from registration to initiation of treatment.
The percentage of eligible patients starting ART by health facility type and province.
Monitoring and evaluation, laboratory, pharmacy
Supervision visits for ART are now integrated and conducted with other programme heads. As a result of a national initiative, monitoring and evaluation and other routine data collection tools now integrate ART-related data into the national health information database. Most routine data are still collected through paper-based registries at facility level, but are now computerized at district level and include key ART indicators.
The day hospital model did not include separate laboratories, but the MOH-managed provincial lab network has been strengthened via rehabilitation of laboratory facilities, provision of additional equipment, and training to support dispersed sites providing ART services (see Figure ). National policy mandated that parallel pharmacies in day hospitals be phased out and tasks integrated into existing pharmacies. Provincial and district pharmacists received additional training to integrate the distribution of antiretrovirals (ARVs) into the national system while basic logistics systems were improved to support new ART sites.
Province-level PHC laboratory system strengthened using HIV-specific resources.
Counselling and testing
Provider-initiated HIV testing was progressively introduced into other PHC services as health workers were trained in CT. CT is provided routinely in ANC and TB programmes, and offered to patients who present with OIs in other outpatient and inpatient services. Counselling is provided through individual, group or video sessions. In 2004, 20,000 people had been tested in 19 separate VCT sites in both provinces. With integrated CT, more than 100,000 were tested in 103 sites in 2007 alone, including TB services (Figure ).
The number of testing sites and patients tested for HIV (2002-2007).
Before these efforts, less than 5% of TB patients were being tested for HIV. With integration efforts, TB staff was trained in CT and new protocols for proper referrals and follow up to HIV care. Items were added to existing patient registries to ensure patient follow up. Clinicians were trained to ensure that HIV patients were also routinely tested for TB and referred, using new protocols and modified paper registries. (See Figure for testing increases in one facility.) By December 2008, more than 90% of TB patients had been tested for HIV at 28 facilities with TB services, and 65% of eligible TB patients had initiated ART.
The number of TB patients tested for HIV per month before/after integration.
By December 2008, 727 doctors, physicians' assistants (called "tecnicos" in Mozambique) and nurses had received specific OI training, covering virtually all clinical health workers in the 67 integrated sites. The OI courses integrated closely with CT and referral training. New protocols and modified patients registries were introduced to facilitate referrals and follow up.
With integration, PMTCT services were included as an integral component of antenatal care (ANC) services within the integrated sites that also initiated ART services. (Routine syphilis testing and treatment, and intermittent preventive therapy for malaria were included in the integrated package.) Maternal & child health nurses and staff were trained in the integrated protocol. Figure compares sites in 2006 offering ART but not ANC (two of the first day hospitals), and sites where both ART and ANC are offered in the same health facility.
Number of HIV-positive women referred from PMTCT/ANC and registered for HIV care <30 days post-test. RR (relative risk) 2.53 for health facility with ANC vs health facility w/o ANC (1.88, 3.40); p < 0.001.
Data were gathered for the same time period for both sets of sites. In 2006, the first two larger sites initially remained vertical, while smaller surrounding sites were integrated. The two integrated sites referenced in Figure are similar to the many other integrated sites in the two provinces. Data were gathered for both sets of sites over the same time period. At the large vertical sites, HIV-positive women were referred from other health units with ANC services, while in integrated sites, patients were referred from ANC services within the same health unit.
Loss to follow up from referrals of HIV-positive women from PMTCT services to ART services was reduced from 70% in vertical sites to 25% at integrated sites, based on analysis of routine data. Nearly all integrated sites across both provinces demonstrate similar results. Larger patient volume at the vertical sites may have contributed to higher LTFU, but the dramatic improvement at integrated sites suggests that integrating ANC and ART in the same health units helps reduce LTFU.
By the end of 2008, HIV-related resources had contributed to the rehabilitation or construction of 40 staff houses, 22 laboratories, 11 pharmacies and warehouses, and dozens of maternities and ANC service areas. Haematology and biochemistry equipment has been provided for 14 laboratories, and 100 dual system (gas and electricity) refrigerators were purchased for cold chain improvements. Fifty-five motorcycles and 19 cars have been provided to the health system over three years and used for integrated activities. Had a vertical approach been maintained, these funds would have been channelled only to HIV-specific facilities and systems. Table lists the types of infrastructure support provided by HIV-related funding.
PHC infrastructure improvements supported by HIV-related resources