Malaria High-Risk Populations in Namibia

This is defined as the proportion of eligible HRPs who report receiving a full course of AL (presumptive treatment) at any time over the study period. The difference in intervention coverage between arms at end-line will be assessed using generalized linear models adjusted for potential confounders and a fixed effect to capture clustering at the health facility level.

This is defined as the proportion of eligible HRPs who report sleeping at a worksite in a structure sprayed with insecticide during the past 6 months (IRS), at any time over the study period. The difference in intervention coverage between arms will be assessed using a difference-in-difference approach using generalized linear models adjusted for potential confounders and a fixed effect to capture clustering at the health facility level.

This is defined as the proportion of eligible HRPs who report sleeping under a bednet (LLIN) the last night staying at a worksite, at any time over the study period. The difference in intervention coverage between arms will be assessed using a difference-in-difference approach using generalized linear models adjusted for potential confounders and a fixed effect to capture clustering at the health facility level.

Species-specific prevalence of malaria infection will be calculated as the proportion of people testing positive for each species malaria by PCR out of all tested HRPs. The reduction in malaria prevalence will be assessed using a difference-in-difference approach, comparing change (pre- versus post-intervention) in all-species infection between intervention and control groups.

The odds of clinical malaria associated with receiving each intervention will be measured by comparing the risk of exposure in cases (RDT-positive) to the risk in controls (RDT-negative) in HRPs within study areas.

The total confirmed outpatient malaria case incidence will be estimated for each of the study health catchment areas and stratified by HRP status. RDTs used for routine malaria diagnosis will be collected and a short questionnaire affixed to the back to allow collection of key indicators to stratify incidence estimates. Denominators will be based on health catchment populations and population size estimates obtained through this study. A reduction in malaria incidence in HRPs and the overall catchment population will be evaluated using a difference-in-difference approach, comparing change (pre- versus post-intervention) between intervention and control groups.

Malaria exposure in HRPs will be measured as the proportion of people with antimalarial antibodies present in their blood serum. ELISA assays will be used to detect biomarkers of Pf exposure, using collected dried blood spots during baseline and endline cross-sectional surveys. The reduction in malaria exposure will be assessed using a difference-in-difference approach, comparing change (pre- versus post-intervention) between intervention and control groups.

Test positivity rates will be calculated as the proportion of individuals screened during routine RACD who test positive for malaria by RDT and PCR. Rates will be stratified by HRP status and location of event (either within villages or at an HRP worksite).Differences in prevalence measures between HRP and non-HRP populations will be assessed using a χ2 test, as well as logistic regression models to account for potential confounding factors.

Entomological measurements including descriptions of vector occurrence, estimates of human biting rates, measures of indoor and outdoor densities, and insecticide susceptibility status and frequency. Measures will be compared between intervention and control study areas, for each type of intervention.

The acceptability of a targeted delivery of presumptive treatment and vector control interventions to HRPs will be assessed as the proportion of targeted individuals refusing each intervention, among those eligible for inclusion. Acceptability of IRS will be assessed as the proportion of targeted farms refusing LLIN, among those with at least one sprayable structure missed during the spray campaign.

Qualitative data collection methods such as focus groups and key informant interviews with HRPs, health extension workers, employers and other health sector staff will be implemented to assess intervention acceptability.

The costs and cost-effectiveness of the intervention package will be assessed as an incremental cost effectiveness ratio (ICER), as well as cost per population and case averted as measured through the difference in infections identified subsequent to the intervention. Program data on costs/expenditure for each intervention will be collected and combined with estimates of program effectiveness to estimate these measures.

Participant adherence to presumptive treatment will be assessed within a sub-sample of people distributed the drug and measured as the proportion of people who complete all pills in the pill pack after receiving the first dose of AL by DOT. Changes in proportions over time will be quantified between the two distribution rounds.

Self-reported user compliance to LLINs and topical repellents will be assessed through the quantitative baseline and endline cross-sectional surveys, as well as through a use tracking log tracking compliance over a 30-day period (Appendix 16 and 17) in the cohort included in the pill count. LLIN use and condition will be empirically assessed during the endline cross-sectional survey and at 30 days, as well as repellent containers weighed after 30 in the cohort. Qualitative information on use of these interventions will be collected during key informant interviews and focus group discussions.

The population size of target HRPs will be estimated in intervention areas using a combination of multiplier and multiple capture-recapture (CR) methods. In Ohangwena Region, the population is assumed fixed, although accessibility of cattle herders in Namibia may vary throughout the year. In Zambezi Region, the population is estimated at two distinct time periods in order to capture turnover between agricultural seasons.

The proportion of individual's time spent in different locations will be measured from aggregating GPS readings (located within pre-defined space-time windows) which are spatially located within a given type of area (farm, cattle post, grazing location, village etc). Where possible, common locations will be tagged and classified by field workers to distinguish individual farms, fields and other points of interests (such as bars, homes, etc). Aggregated GPS readings (located within pre-defined space-time windows) will be plotted to distinguish individual trips and measure specific characteristics of each trip, including distance of travel and frequency of trips.