Design, implementation and evaluation of a national campaign to distribute nine million free LLINs to children under five years of age in Tanzania

In March 2007, the GFATM invited Tanzania's Country Coordinating Mechanism (CCM) to submit a Rolling Continuation Channel (RCC) application to extend its Round 1 grant to increase ITN ownership and use among vulnerable groups. After extensive stakeholder discussions, NMCP proposed the continuation of the voucher programme plus the launch of a free, one-time LLIN distribution campaign for U5s (under five catch-up campaign - U5CC). The World Bank under its Booster Programme for Malaria Control in Africa and PMI simultaneously contributed funds to expand the scope of the distribution. In addition, Malaria No More/UNICEF, World Vision Switzerland, the UK Department for International Development (DfID), and the Swiss Agency for Development and Cooperation (SDC) contributed additional funding. Contributions raised during the Davos Economic Forum in 2005 and unobligated MoHSW funds closed the final budgetary gaps to complete the national campaign.

In compliance with Tanzania's tendering rules and required product specifications, a single LLIN tender was issued by MoHSW funded by GFATM and the World Bank. LLINs financed by PMI were the subject of a separate tender. The only polyethylene net with the required qualifications at the time (2008), specifically a full WHO Pesticide Evaluation Scheme recommendation, was the Olyset^® net produced by Sumitomo Chemical and A-Z Textiles Ltd. This requirement was exceptionally agreed to by the three donors in order to ensure that the same LLIN was delivered throughout the country following initial delays negotiating this issue with the different donors lasting several months. Because the local manufacturer (A-Z Textiles) also won the contract for distribution to village level, the management of the logistics was greatly facilitated.

The five grant sub-recipients had already been identified through a competitive procurement mechanism conducted by the Country Coordinating Mechanism prior to the development of the RCC grant proposal. The MoHSW, through its Procurement Management Unit, contracted the grant sub-recipients for the five key components of the campaign: (1) Logistics - MEDA Economic Development Associates; (2) Training - World Vision Tanzania (WVT) (3) Social mobilization - Population Services International; (4) Monitoring and Evaluation - Ifakara Health Institute (IHI) who in turn sub-contracted technical support to the London School of Hygiene and Tropical Medicine; (5) Financial and procedural audit - KPMG. The proposal submission, evaluation and contracting process for the five sub-recipients took more than seven months to complete. The delivery contractor, (A to Z Textiles Ltd), was identified by the logistics contractor MEDA Economic Development Associates through a separate competitive tender. The LLIN Hang-up Campaign (conducted by Tanzania Red Cross) was separately contracted by USAID and co-funded by DfID.

Since 2003, the ITN Cell, a unit within the NMCP, has coordinated the National Insecticide-Treated Nets Program (NATNETS) programme, with technical and financial support from the Swiss Agency for Development and Cooperation through its executing agency, the Swiss Tropical and Public Health Institute. During the U5CC, the ITN Cell and other NMCP staff coordinated planning in Dar es Salaam and in the field where they facilitated contacts between the local government and the government contractors. The U5CC proceeded on a rolling basis, entering a new zone (comprising 2-3 regions each) every five weeks, while simultaneously completing later phases of the campaign in other zones. The U5CC covered Tanzania's eight zones in order of their malaria prevalence, with the zones having the highest malaria prevalence visited first. Tanzania's mainland population in 2009/10 exceeded 41 million, with over 80% living in areas with stable perennial to stable seasonal malaria transmission [4]. The U5CC started with a pilot programme in Mpanda District (highlighted in Figure 1) in October 2008 to test the planned methodology. The lessons from the pilot programme were incorporated into the main campaign, which is detailed below and diagrammed in Figure 2.

Prior to departure of the field team (NMCP, ITN Cell and contractors' staff) from Dar es Salaam, the MoHSW sent a letter to Regional and District government officials, alerting them to the upcoming U5CC in their respective regions and districts. Upon arrival in a region, the field team-consisting of contractors and an ITN Cell representative-jointly briefed the regional government authorities on the U5CC. The training contractor organized regional sensitization meetings for the regional authorities including the regional health management teams who had a role in regional supervision.

Approximately one month after LLIN issuance, an organized effort was made to encourage household members to hang-up and use their new LLINs. The effort was implemented by the Tanzania Red Cross volunteer network and local government officials (including WEOs and VEOs) in each zone. In areas lacking a Red Cross volunteer structure, division-level extension officers worked as supervisors and ward health workers served as volunteers. With one volunteer per village, each volunteer had 12 days to visit every household in the village to ensure the new LLIN was hung and used and to share malaria messages with residents by distributing illustrated leaflets. Volunteer reports were collected by local supervisors and entered into a national database.

The monitoring and evaluation contractor, Ifakara Health Institute, assisted by the London School of Hygiene and Tropical Medicine, conducted a series of post-distribution district household surveys on ITN coverage and other U5CC related activities one to three months following net issuing in a given zone. Districts were selected based upon the availability of baseline data from a 2008 nationally representative malaria survey, including five Districts, (Figure 1 - Ifakara Health Institute and the London School of Hygiene and Tropical Medicine 2009, unpublished data). From each district, 30 clusters (villages) were selected with probability proportional to the size of the village. Within each selected cluster, one sub-village was chosen randomly and 30 households were chosen from that sub-village, based upon a modified Expanded Immunization Programme sampling procedure [15]. Questions about household ownership and use were asked using the standardized Malaria Indicator Survey format

Nine different donors funded the U5CC (Table 1), and contributions ranged from $25,258,382 (GFATM) to $317,817 (Davos World Economic Forum Fund). Table 1 represents the direct financial costs for the U5CC, not including local government contributions. All of these funds were provided in unadjusted USD, with the exception of the Swiss Agency for Development and Cooperation and MoHSW contributions, which were calculated in USD based upon the exchange rate at the time of money transfer. The primary donors (GFATM, World Bank, and PMI) committed funds in 2008. The overall cost of the U5CC totaled USD $63,831,113, of which USD $47,340,943 (74.2%) was used to purchase 9,034,677 LLINs. The financial cost per LLINs distributed was USD $7.07 per LLIN, of which USD $5.24 were used to purchased the net, and USD $1.83 were for LLIN transport, training, logistics, management, social mobilization/BCC, and M&E.

Towards the end of the campaign, the U5CC faced a 1,822,954 LLIN shortfall due to the difference between the estimated number of LLINs needed for the U5s (original budget) and the actual numbers of LLINs ordered based upon the household registration process. The original budget relied upon National Bureau of Statistics (NBS) 2008 population projections from the 2002 Tanzania National Census, which projected a need for 7,220,083 LLINs. However, the actual order for LLINs in the first two zones (Southern and Lake) surpassed the original projections by 42%. An analysis detailed in Table 2 showed six main contributions to the discrepancy between the NBS estimates and actual LLIN needs. A total of 20.2% could be attributed to various buffers: 5% as village buffer, 5% at district level, 5% due to bale-rounding because only full bales of 40 LLINs were delivered to the village level, and a village level estimate of unregistered children capped at 5% of the total village LLIN order. A comparison between the NBS projections and 4,000 entries in a sample of U5CC register books from 12 villages indicated that another 7% of the discrepancy could be attributed to registration of overage children (>59 months), particularly five-year old children. The source of the remaining 14.8% discrepancy was most likely due to an NBS projection error or inaccurate census data in 2002. Table 3 illustrates the discrepancy between NBS estimates and LLIN orders on a zone basis. The total number of buffer LLINs provided for each zone was simply the difference between U5s registered and total LLINs sent.

Following the Table 2 analysis (conducted after LLIN issuing in the first two zones), policies on buffer stocks became more restrictive to reduce the anticipated future shortfall of nets. Overall, the vast majority (96.1%) of children that attended issuing posts received LLINs, as illustrated in Table 3. This included the children registered by volunteers as well as the children missed in the registration process who received LLINs from buffer stocks. Village reports documented that a total of 222,712 LLINs (the difference between LLINs delivered to a village and LLINs issued to U5s) were delivered to villages in excess of the village's U5 registered population. These LLINs were reallocated to 1) unregistered U5s in the village; 2) unregistered U5s in the ward; and (if any surplus LLINs remained) to needy members of the community. While the village and ward authorities were tasked with reallocating the LLIN surpluses, this occurred after the reporting period ended, so there is no data on this re-allocation.

Table 4 details the line item costs associated with the U5CC implementation in one of mainland Tanzania's 121 districts. The U5CC budget for Rural Kyela District was selected as an illustrative example of components of U5CC district budgets. This district is smaller in U5 population (38,007) than the average Tanzanian district (68,444), but the contractor budgets were comparable to other Tanzanian districts [16]. These costs exclude the LLIN procurement and delivery to village.