Effectiveness of Seasonal Malaria Chemoprevention in Nampula Province, Mozambique: Type Two Hybrid Implementation Study

Overview

We describe a Type II hybrid effectiveness-implementation study design which evaluates the effects of a clinical intervention on relevant outcomes whilst collecting information on implementation. It is designed to determine feasibility and effectiveness of an innovative intervention, as well as the protective efficacy of the drugs used. The study consists of three components: 1) Conducting a cluster randomized controlled trial (cRCT) through household surveys establishing confirmed malaria cases in children; 2) Conducting a prospective cohort study to determine the chemoprevention efficacy of sulfadoxine-pyrimethamine and amodiaquine (SPAQ) and whether drug concentrations or parasite resistance influence the duration of protection; and 3) Conducting a resistance markers study in children 3-59 months to measure changes in resistance marker prevalence over time.

This implementation research study uses a convergent mixed-methods approach. SMC will be implemented in four monthly cycles between December 2021 and February 2022. The research will involve the following components: 1. Conducting a cluster randomized controlled trial (cRCT) through passive surveillance to establish confirmed malaria cases in children 2. Conducting a prospective chemoprevention efficacy cohort study to determine if SPAQ provides 28 days of protection from infection and whether drug concentrations and/or resistance influence the duration of protection 3. Conducting a resistance markers study in children 3-59 months in the two research districts plus the two standard intervention districts to measure changes in resistance marker prevalence over time. Four districts will be involved in the study: the two intervention districts from phase 1 (Mecubúri and Malema) and two districts where SMC has not previously been implemented (Lalaua and Muecate). The cRCT will only be implemented in the two new districts. The chemoprevention efficacy cohort study will only be conducted in one of the two new districts, which will be selected based on operational feasibility. The estimated target population for SMC in the study districts is 114,000 children aged 3-59 months. Methods 1. Cluster randomized controlled trial (cRCT) To assess the effectiveness of SMC using a cluster randomized control trial (cRCT) in two districts in Nampula province. Study design and sample selection The effectiveness of SPAQ (on an intention to treat basis) for prevention of clinically-significant cases of malaria among children aged 3-59 months in this setting will be investigated using a cRCT. This trial will comprise of two arms: one control arm and one intervention arm. Informed consent for childrens' participation in the trial will be obtained from caregivers. Clusters will be selected for the intervention arm at the level of comunidades or communities. The study will be powered to have an 80% chance of detecting, as significant at the 5% level using a two-tailed test, a 40% reduction in incidence (based on that assumption that efficacy of SPAQ in Northern Mozambique, where prevalence of resistance alleles in circulating parasites is high, is half that found in studies in West African settings). It will assume that study participants, in the absence of SMC, would experience 0.2 clinical episodes per child per high-transmission season (corresponding to the SMC round) of sufficient severity to present to a health facility. Sample size calculation for chi square test comparing two independent proportions in a cluster randomized design with the proportion of sample with the expected outcome: 0.80; Statistical power: 80 %; Confidence level: 95% (α=0.05); Assumed proportions of children developing confirmed case of malaria during follow-up: Intervention arm: 0.2; Control arm: 0.12 (Corresponding to protective effect of 40%); Ratio of children in control arm: intervention arm (K-ratio): 1.5; Assumed intra-cluster correlation (ICC) of malaria outcomes: 0.23; Selected cluster size: 15; Estimated total required sample size:n = 2,850, of which 1,710 in the control and 1,140 in the intervention arms; Estimated total number of clusters: n = 190, of which 114 in the control and 76 in the intervention arms. A total of 76 communities will be selected at random from a list of all communities with probability proportional to population size for inclusion in the intervention arm. Afterwards, 114 communities will be selected at random from among the remaining for inclusion in the control arm. All eligible children in the intervention areas will receive SMC as per the standard protocol. In all areas selected for inclusion in the control and intervention arms, 15 households will be selected at random from a household list in each community. In each household, one eligible child will be selected at random from among the eligible children and included into the study. The participant sampling frame was designed to ensure overall efficiency of the study in terms of number of clusters, overall sample size and number of communities receiving SMC. Attack rates in the intervention and control districts will be compared to ensure these are similar across the selected intervention and control areas. In the control arm, households will be randomly sampled by researchers from selected communities with one eligible child recruited at random in each household. After caregivers provide consent, children will be recruited and a short baseline questionnaire will be administered to collect individual data on each child and to confirm their eligibility. In the intervention arm, a researcher will follow a pair of community distributors as they administer SMC. Households in selected communities will be randomly sampled from those visited by the distributors. Before SMC is administered, consent sought, one eligible child will be recruited and a short baseline questionnaire will be administered to collect individual data and to confirm their eligibility. Blood samples will be taken and subsequent hemoglobin concentration measured using a Hemocue machine. Children recruited into the study presenting at clinics will be identified using information on their SMC record cards, and data on clinic visits including suspected malaria cases and results of rapid diagnostic tests (RDTs) will be matched to baseline questionnaire data to build a database for analysis. Date of clinic attendance and confirmation of malaria cases using rapid diagnostic tests will be recorded to allow for calculation of follow-up time and time to malaria events for fitting of Cox proportional hazards regression model. 2. Chemoprevention efficacy cohort study The aims of this study component is to ascertain whether current dosing regimens of sulfadoxine, pyrimethamine and amodiaquine given during SMC sufficiently protect against Plasmodium falciparum infection in children of 3-59 months for 28 days. The study design is similar to what has been proposed as a standard chemoprevention efficacy study (CPES) protocol. The aim is not only to explore if chemoprevention failure takes place, but if this is as a result of recrudescent infections, new infections or under-dosing. As proposed by the standard CPES protocol, both malaria thick smears and dry blood spots (DBS) will be taken at day 0, day 2, day 7, day 14, day 21 and day 28 after administration of the first three-day dose of SPAQ, which will be administered through directly observed therapy (DOT). In addition to what is proposed by the standard CPES protocol, all DBS samples on day 0 and day 28 will be processed in order to ascertain low density parasitemia estimates through qualitative polymerase chain reaction (qPCR). If sufficient DNA exists, day 0 and day 28 qPCR genotypes will be compared. If children are found to have malaria infections, presence of the same haplotype would indicate a recrudescent infection. If different haplotypes are present, this would indicate a new infection after SPAQ administration. The qPCR methodology is designed to amplify only parasites that have freshly emerged from the liver. DBSs will also be processed for parasitemia estimations and genotyping on days 2, 7, 14 and 21 if a child's thick smear on those days was positive or, retrospectively, if subsequent DBSs are Plasmodium falciparum positive. A DBS will also be collected whenever a child has fever and is malaria positive on a rapid diagnostic test (RDT). Drug level, qPCR parasite estimation and genotyping will also take place. Assuming a 3% breakthrough infection rate sampling 494 children will provide a 95% confidence interval of 1.7% to 4.9%. Assuming a 99% response rate and 1% loss to follow up a sample size of 500 children is considered appropriate. Finger pricks will be conducted by trained nurses or health personnel both for thick blood smears and for DBS. Dose, weight, age, length of mid-upper arm circumference, tympanic temperature, location, time and date will be recorded for each child on day 0 and all subsequent time points on day 2, 7, 14, 21 and day 28 when the follow up slides and DBS will be taken. Thick blood smears and DBS will be prepared for processing by a trained laboratory technician. All thick smears and DBSs will be sent to a suitable laboratory with high-sensitivity pharmacology capacity for microscopy and qPCR analyses. A questionnaire exploring if the child received other treatment or experienced any disease since SPAQ administration will be co-administered every time a sample is taken. Caregivers will be reminded at each visit to the household to seek care at a health facility or from a community health worker if the child becomes febrile. A system will be put in place to record RDT results and any treatments received. On day 28, data collectors will also administer a questionnaire to each caregiver to explore the child's illness history over the duration of the study. The laboratory will send the processed data back to Malaria Consortium, where the relevant mutations distributions and proportions will be analyzed comparing parasitological efficacy between groups of mutations. The drug levels will be analyzed as a cohort by mean, median and standard deviation and determine any correlations with treatment outcomes, in particular drug concentrations on day 7. Day 28 positivity will be correlated to antimalarial drug resistance genotype. 3. Resistance markers study The aim of this study component is to continue to monitor prevalences of SP and AQ resistance and any increase in resistance prevalence after one annual round of SMC (in the two new phase 2 study districts) or two annual rounds of SMC (in the two phase 1 districts) as assessed via molecular markers in the population. A specific objective is to detect prevalence over time in the proportion of symptomatic children with an RDT residing in the districts where SMC is implemented who carry parasites with mutations compared to children living in similar districts without SMC. Trends of SP and AQ resistance will be monitored in the two original intervention districts, as well as in the two research districts. A health facility-based, cross-sectional survey will be conducted before SMC project implementation (base line) and after one complete round of SMC distribution (end line). Monitoring the prevalence of alleles associated with resistance to drugs is, by standard protocol, done by collecting samples from symptomatic children with evidence of infection. The sample collection will be performed in four selected first level of care health facilities in each study district. The key markers to be monitored are: dihydrofolate reductase (dhfr): codons 108, 51, 59 and 164; dihydropteorate synthetase (dhps): codons 431, 437, 540, 581 and 613; Pf chloroquine resistance transporter gene (pfcrt): codons 72-76; Pf multidrug resistance gene 1 (pfmdr1): codons 86, 184 and 1246 and Plasmepsin 2 (plm2). The sample size for the survey was determined using a World Health Organization protocol for drug efficacy testing with the intention to estimate changes in prevalence of SP and AQ resistance markers with enough precision to provide reassurance that there has been no important increase, while having adequate power to detect changes if they occur. A sample of 300 children per district (150 each at base line and end line) is expected to result in 90% power to detect a difference at 5% level between baseline and end line. The total sample size will therefore be 1,500, with 750 children to be sampled at both baseline and end line.