ApproachesforUniversal LLIN Distribution in Four Eastern Districts UsingGlobal Fund LLIN Distribution Phase 1 RegistrationData
CONCEPT NOTE

BACKGROUND

According to the Uganda Malaria Indicator Survey of 2009, 47% of households own an insecticide treated net (ITN).This is a significant increase from the 17% measured in the Uganda Demographic and Health Survey (UDHS) in 2006. In 2010, another 7.2 million LLINs obtained from the Global Fund to Fight AIDS, TB and Malaria (GFATM) Round 7, Phase 1, were distributed, leading to an estimated household ITN ownership coverage of 74% (UDHS 2011). The same year,the Uganda National Malaria Control Program (NMCP) declared universal coverage of the entire population at risk of malaria as the strategic approach to community LLINs distribution. However, the 2010 distribution was carried out as a targeted distribution to children under five and pregnant women on the adviceof the GFATM. Nonetheless, household registration during these distributions included all families, irrespective of whether or not they included any members of the target group. This was done in 685 out of the then 964 sub counties in Uganda[1]with the aim of recording the required number of long-lasting insecticide treated nets (LLINs) for subsequent distribution during Round 7, Phase II. Unfortunately, the Phase IILLIN distribution did not take place as had been planned (within one year of Phase I). This has resulted in threevariations of LLIN coverage. There are sub counties where registration was done and received LLINs; sub counties that were registered but did not receive LLINs; and sub counties that were not registered and did not receive LLINs. Moreover, the LLIN coverage in the sub counties that received LLINs is uneven. At this time, the Phase I registration data is no longer valid.It is now approximately two years since the Phase I registration was done and natural shifts in populationhave led to significant changes in number of households in each village as well as the number of household occupants.

Uganda is now preparing for Round 7, Phase II LLIN distribution, which is likely to start in the second half of 2012.There is a need to validate Phase I registration data, and to determine the best approach to implement the Phase II LLIN distribution using Phase I household registration data. Under GFATM Phase II, there are no funds to conduct another household registration in sub counties which conducted registration under Phase I.

Based on the different LLIN coverage situations, the Ministry of Health, NMCP requested partners to bridge the LLIN gap in order to reach universal coverage. Support was also requested to determine the best approach to implement Phase II LLIN distribution using Phase I registration data. The United States President’s Malaria Initiative (PMI)responded to this request by offering to procure 650,000 LLIN and to provide technical support to the distribution exercise, through the PMI funded Stop Malaria Project (SMP). SMP will provide technical and financial assistance to distribute LLINs in fourdistricts in Eastern Uganda which did not receive nets but were registered under Phase I. As part of the distribution process, SMP will also assist the NMCP to evaluatevarious distribution approaches to determine which one is most -effective. SMP will also provide ongoing technical support to NMCP at the central and district level for the purpose of Phase II LLIN distribution.

PURPOSE

The purpose of this concept note is to twofold:

  1. To support the achievement of universal LLIN coverage in four selected districts in Eastern Uganda and;
  2. To testpotential approaches to household registration and Phase I data validation and draw lessons that would inform Phase II LLIN distribution across the country.

UNIVERSAL DISTRIBUTION OF LLINS IN FOUR DISTRICTS OF EASTERN UGANDA

Criteria for Selecting the FourDistricts

The criteria below were used to select the fourdistricts that will be targeted under PMI support:

  1. Districts that missed LLIN distribution in Phase I, yet had been targeted
  2. Proximity of the districts to the lake(water bodies)
  3. Number of available LLINs
  4. Duration of time since the district had any LLIN distribution (an indication of the existence or absence of valid LLINs within the targeted areas)

Based on these criteria, the fourdistricts selected are: Bugiri, Kaliro, Serere and Mayuge.

APPROACHES TO HOUSEHOLD REGISTRATION AND PHASE I DATA VALIDATION

Phase II will cover 953 Sub Counties in 104 districts country wide,considering the available 10,371,134 nets. In 627 of those sub counties, PhaseIregistration was conducted, covering over 80% of households. As registration was conducted two years ago in 2010, this registration data needs to be validated in all (627) sub counties. In the remaining 326 sub counties that did not participate in Phase 1, household registration will need to be carried out. We propose to test two different approaches to data validation and household registration in the fourdistricts, as described below. The results from this evaluation will inform Phase II data validation and household registration.

Validation of Phase I Household Registration Data

Before beginning the validation process, the existence and status of the original Phase I registration lists will be confirmed. The lists will be located, checked to ensure they remain in good condition (to ensure no damage was done during storage), and traced to ensure that lists exist for all villages.

The approaches below will then be tested to validate Phase I household registration data. The new lists generated by each approach will also be verified, through a process described below (p.4).

A. Simplified Approach – Village Leaders (LC I)

In this approach, village leaders (LC Is) will be asked to provide an updated (current) village list with information needed to validate the Phase I registration list (household heads, total persons and their status:migration in, migration out, missed out Phase I registration). This village register will be compared with Phase I registration listfor the village by the sub county District Health Management Team (DHMT) or CSO supervisors, in order to assess the validity of the Phase I registration list. A new, updated PhaseII registration list will then be generated based on the updated list.

Updating of the Phase I village register will involve:

  1. Removing those households that left the village
  2. Including new members to the village
  3. Including those who were left out in the Phase I registration

The validation exercise will be done at the parish level. This activity will last at least twodays per parish, as per the guidelines[2]. This approach will be applied in two sub counties in Kaliro (out of five sub counties), two sub counties in Bugiri(out of eleven) and two sub counties in MayugeDistrict (out of sixteen)where registration was done during Phase I LLINsdistribution.

B. Modified Simplified Approach – Village Leaders (LC I) with Another Opinion leader

The LC I, or his/her representative,together with an opinion leader in the village (e.g. “faith based” organizations representatives, cultural leaders, Red Cross, Rotarians or health club leader), shall develop a village register which will be compared to Phase I register. The appropriate opinion leader in each village will be nominated by the VHTs. The nomination of the opinion leader will be guided by the following criteria: most respected, knowledgeable about the village, willingness to serve the community, and be permanent resident in the village. The sub county DHMT or CSO supervisors will validate Phase I registration village list using the village list developed by the LC Itogether with the opinion leader. The validation exercise will be done at the parish level and will last two days, using the same methods described above. This approach will be tested in the remaining three sub counties in Kaliro (out of five sub counties) and five sub counties in Bugiri (out of eleven sub counties), where registration was done during Phase I LLINs distribution.

C. Testing both Approaches for the Purpose of Fill-in Distribution

Both approaches A and B will be tested in Kayunga district where Phase I registration and LLIN distributionwas done. Each approach will be conducted in 50% of randomly selected sub counties in the district. This validation exercise will be done to test the same validation approaches to calculate the fill-in LLIN needed in sub counties in which nets were distributed under Phase I. Similar methodology of validation will done as describe above.

KayungaDistrict was purposefully selected because it is representative of many other districts that benefited from Phase I LLIN distribution, and it is near Kampala, making the verification process easier.

The estimated degree of variance from each approach will be compared to determine the best approach to calculate the fill-in LLIN for Phase II. The resultant outcome (Phase I registration and distribution done) will be contrasted with the results from those districts where Phase I registration was done but didn’t receive LLIN.

Household Registration

Two household registration approaches will be tested in sub counties where Phase I household registration was not done. This will apply to all sub counties in Serere and seven sub counties in Mayuge Districts. The new registration lists generated by each approach will also be verified through a process described below (p.4).

A. House to House Registration by VHTs

Two VHTs per village will create a village register through a house to house registration. This exercise will last at least 3 days. As per the Toolkit for mass LLIN campaigns 2012, four categories of household size exist in every community. The table below indicates the expected household categories and expected proportions as per the AMP tool kit.[3]

Category / Household size / Expected proportion
1 / 1 – 3 / 20%
2 / 4 – 7 / 60%
3 / 8 – 11 / 15%
4 / 12 and above / 5%

The sub county supervisors will summarize the total number of households for each village list and categorize these according to expected household size. The supervisors shall then validate this data against the expected proportion for each household category to determine any variance.

This approach will be tested in threesub counties (out of sixteen) in Mayuge District and five sub counties(out of ten) in SerereDistrict.

B. Compilation of List of Households in the Village by the Village Leader

The LC I chairperson, or his/her representative, will compile or update the current list of all households in his/her village. The sub county supervisors will summarize the total number of households in each village and categorize these according to expected household size. The supervisors shall then validate this data against the expected proportion for each household category to determine any variance.

This approach will be applied in foursub counties in Mayuge(out of sixteen) and Serere(out of ten) Districts respectively.

For future LLIN mass distribution exercises, the registration lists generated from the best approach selected for each scenario above shall be considered in comparisonto the Tool Kit for Mass Distribution Campaigns.

Verification Process

All village registration lists will be verified to ensure that the lists have accurately captured all households, the appropriate number of household members, their status, etc. by the sub county DHMT or CSO supervisors.An appropriate sampling methodology will be identified, which may include the Lot Quality Assurance Sampling (LQAS) methodology. If LQAS is to be used, a total of 19 households will be systematically selected from each village register, proportionate to household size composition. A decision rule of 80% will be applied. If 80% or more households pass this decision rule, then the village register will be taken as true representation of village. For whatever sampling methodis selected, if the decision rule is not met, the whole exercise for the selected area will repeated and the LC I, opinion leader, or VHT will cease to be involved in the exercise.

PREPARATORY ACTIVITIES AHEAD

  1. Engaging twoTechnical Assistants to provide technical support at the central and district levels, to be approved by PMI
  2. Production of implementation tools
  3. Field implementation (field implementation will include; validation of data, re/registration, allocation, distribution, data compilation and report writing)
  4. Dissemination workshop of the findings

1

[1] The number of sub counties in Uganda has since changed as a result of the subdivision of districts and subsequent changes in administrative classifications and boundaries.

[2]National Guideline on LLIN distribution 2010, Ministry of Health

[3]Tool Kit for Mass Distribution Campaign, 2nd Edition 2012, The Alliance for Malaria Prevention