Study area
This study will be conducted in a selected villages in Tanzania, Côte d’Ivoire and India (Table 1). Study villages will be selected based on the availability of a large number of households required for the study, their accessibility throughout the year, the high abundance of malaria vectors that promotes mosquito net usage [37] and proximity to the bioefficacy testing laboratory in each study country.
Table 1. Description of study location in Tanzania, Côte d’Ivoire and India.
Features
|
Tanzania
|
Côte d’Ivoire
|
India
|
Location of the study
|
Bagamoyo district in the Coastal region located 70 km North of Dar Es Salaam
|
Tiassalé district in the Agnéby-Tiassa region located 135km North of Abidjan, in Southern Côte d’Ivoire.
|
Kurnool District in the Andhra Pradesh state. Study villages are on the embankment of the River Tungabhadra.
|
Global positioning system
|
Latitudes 6° 37ʹ South
Longitudes 38° 58ʹ East.
|
Latitude 5° 54’ North
Longitude 4° 50’ West.
|
Latitude 15° 81’ North
Longitude 77° 96’ East.
|
Climate
|
Tropical climate with an average annual rainfall of 900 mm and an average temperature of 28°C.
|
Tropical climate with an average annual rainfall of 1739 mm and an average temperature of 26.6°C.
|
Tropical with an average annual rainfall of 705 mm and temperatures ranging from 26°C to 46°C in the summer and 12°C to 31°C in the winter.
|
Economic activities of inhabitants
|
Most of the inhabitants are small-scale farmers
|
Most of the inhabitants are subsistence farmers and worker
|
Most of the inhabitants are small-scale subsistence farmers and worker
|
Study design
This will be a prospective household-randomized controlled trial conducted in Tanzania, Côte d’Ivoire and India simultaneously. The study will follow guidelines [8] for monitoring the durability of ITNs with slight modifications to the sample size and study procedures. A minimum of 6,500 households from the selected villages in the study area will be enrolled in the study in each country. In this study, a household is defined as a group of people who share living accommodation and who are eating from one pot. The households will be the unit of ITNs randomization and the individual nets as a unit of observation. At baseline, a household survey will be conducted to collect household information including demographic and socioeconomic characteristics and the location of the household. At the same time, mosquito nets will be distributed by a study team in each household to cover all sleeping areas, to ensure maximum coverage of ITNs for all members of the household. Follow-up surveys will be conducted to assess the rate of net loss in the receiving households (attrition rates due to all causes, as well as for wear and tear), physical integrity, bioefficacy and chemical residual content of ITNs collected from households at the time intervals up to 36 months (Table 2). The study flow diagram is shown below (Fig. 1). All participating households will be blinded to which net product they receive, and study investigators will also be blinded to intervention allocation.
Table 2: Schedule of activities
Activities
|
Month
|
Conducted at baseline
|
0
|
1
|
6
|
12
|
18
|
24
|
30
|
36
|
Community sensitization
|
x
|
|
|
|
|
|
|
|
Household recruitment
|
x
|
|
|
|
|
|
|
|
Baseline questionnaire
|
x
|
|
|
|
|
|
|
|
Baseline bioefficacy and chemical analysis
|
x
|
|
|
|
|
|
|
|
ITNs distribution
|
x
|
|
|
|
|
|
|
|
Conducted after baseline
|
|
|
|
|
|
|
|
|
Adverse effects survey
|
|
x
|
|
|
|
|
|
|
Monitoring the attrition
|
|
|
x
|
x
|
|
x
|
|
x
|
Monitoring fabric integrity for damage
|
|
|
x
|
x
|
|
x
|
|
x
|
Monitoring bioefficacy using WHO bioassays
|
|
|
x
|
x
|
x
|
x
|
x
|
x
|
Chemical content analysis
|
|
|
|
x
|
|
x
|
|
x
|
Monitoring bioefficacy using the I-ACT
|
|
|
|
x
|
|
x
|
|
|
The durability monitoring components that will be evaluated are: attrition, bioefficacy, chemical residual content, and damage to fabric. Also, the Ifakara ambient chamber testing (I-ACT) will be conducted as an additional bioassay using nets sampled from the field in Tanzania, 20 times laboratory washed and unwashed nets
Study ITNs
Four different products of ITNs will be evaluated: Olyset® Plus, Veeralin®, Tsara® Boost and MAGNet® (Table 3). All test items have been prequalified by WHO [39]. To ensure households and field workers are blinded to intervention allocation during the study, all ITNs will be rectangular in shape, white in colour, and of standard dimensions (190 cm x 180 cm x 150 cm) and will be labelled with a water-resistant six digits numeric self-laminating code attached to one of the six hanging loops. This code will uniquely identify each ITN and contain numbers representing the country, product, and individual net. Each product of ITN will be assessed at baseline in each country to ensure that they meet WHO bio-efficacy criteria before distribution (Table 4) using well-characterized laboratory-reared Anopheles mosquitoes in the WHO cone bioassay and/or WHO tunnel tests. Insecticide content at baseline will be measured using appropriate Collaborative International Pesticides Analytical Council (CIPAC) methods, at an accredited laboratory.
Table 3: Characteristics of study ITNs
Test Item
|
Material type
|
Denier
|
Bursting strength
|
Mesh size
|
Active ingredient
|
Estimated
AI Dose
|
Treatment method
|
Manufacturer
|
WHO Status
|
Tsara® Boost
|
Polyethylene
|
130
|
≥400kPa
|
21 holes/cm2
|
Deltamethrin and PBO
|
120 mg/m2 of deltamethrin and 440 mg/m2 of PBO
|
Incorporated
|
NRS Moon Netting FZE, United Arab Emirates
|
[55]).
|
Olyset® Plus
|
Polyethylene
|
150
|
≥250kPa
|
6.45 holes/cm2
|
Permethrin
and PBO
|
800 mg/m² of Permethrin and 400 mg/m2 of PBO
|
Incorporated
|
Sumitomo Chemical, Japan
|
[58]
|
Veeralin®
|
High-Density Polyethylene
|
130
|
≥350kPa
|
13 holes/cm2
|
Alpha-cypermethrin and PBO
|
216 mg/m2 of alpha-cypermethrin and
79.2 mg/m2 of PBO
|
Incorporated
|
V.K.A Polymers, India
|
[56]
|
MAGNet®
|
High-Density Polyethylene
|
150
|
≥450kPa
|
20 holes/cm2
|
Alpha-cypermethrin
|
261 mg/m² of alpha-cypermethrin
|
Incorporated
|
V.K.A Polymers, India
|
[59]
|
Table 4: ITNs durability components and outcome measured
Component
|
Definition
|
Test conducted
|
Outcome Indicators
|
WHO criteria or industry standard
|
Analysis Method*
|
Attrition
|
Net loss from household through discarding or use for alternative purpose‡
|
Households Survey
|
Net presence
|
-
|
Binary logistic regression
|
Physical integrity
|
The physical state of the net to estimate bite protection
|
Count the number, location and size of the hole(s) of max. 3 nets per household
|
The holed surface area is measured by the proportionate Hole Index (pHI)† [43]
Median hole surface area (cm2)
|
pHI 0-64: good
pHI 65-642: damaged
pHI≤642: serviceable
pHI≥643: too torn or unserviceable
|
Negative binomial regression
|
Functional survival** [43]
|
Estimation of nets still in households in serviceable condition
|
Presence or absence of net
|
(Number of nets present and serviceable) / (Number of nets originally received and not given away or lost to follow up)
|
Median net survival in years = time point at which the estimate of functional survival crosses 50%
|
Logistic regression
|
Biological efficacy
|
The ability of the net to incapacitate or kill susceptible anopheline mosquitoes after contact with the insecticide
|
WHO cone/tunnel test using 25x25cm pieces[8]
|
The proportion of net samples meeting the optimal WHO bioefficacy criteria
|
1hr Knock-Down ≥ 95%
Or 24hr Mortality ≥ 80% or
Blood feeding Inhibition ≥ 90%
|
Logistic regression
|
Ifakara Ambient Chamber Test (I-ACT) using whole nets
|
The proportion of mosquitoes dead at 24 hours
The proportion of mosquitoes not blood-fed
|
-
|
Logistic regression
|
Insecticide content
|
Amount of active ingredients in the net
|
Permethrin, Alpha-cypermethrin and piperonyl butoxide: Gas Chromatography with Flame Ionisation Detection (GC-FID)
Deltamethrin: High-Performance Liquid Chromatography with UV Diode Array Detection (HPLC-DAD)
|
The proportion of ITNs meeting with WHO specifications at baseline
|
MAGNet: 6 [4.5-7.5] g/kg alpha-cypermethrin
Olyset Plus: 20 [15-25] g/kg permethrin; 10 [7.5-12.5] g/kg PBO
Tsara Boost: 3 [2.25-3.75] g/kg deltamethrin; 11 [8.25-13.75] g/kg PBO
Veeralin 6[4.5-7.5] g/kg alpha-cypermethrin; 2.20 [1.65-2.75] g/kg PBO
|
Binary logistic regression
|
Community sensitization
Community sensitization will be conducted to inform community leaders and community members about the study objectives, study rationale and to request their co-operation during study implementation. All attendees of the community sensitization meetings will be encouraged to ask questions of the study investigators to enable the community to understand the study objectives and the risks and benefits of study participation.
Baseline surveys and distribution of ITNs
Upon obtaining written informed consent from the head of household or adult resident, field workers that have been trained on the study protocol and procedures will conduct the baseline questionnaire as outlined in [40] and will distribute the nets required for each household with one net distributed per sleeping space. Non-study nets found in the households will be withdrawn from households to ensure only study ITNs will be present and used in the households. In Tanzania, the withdrawn nets will be stored in bags labelled with the name of the household head and the study village to be returned to their respective households after the study, while the withdrawn nets will be discarded in Côte d’Ivoire and India following specific country procedure.
The questionnaire forms will be written in local languages, Kiswahili for Tanzania, in French for Côte d’Ivoire and in Telugu for India and will be pre-loaded in Open Data Kit (ODK) Collect software installed on a hand-held Samsung Tab A tablet computer. Data collected during the baseline surveys are the household roster (names, ages, sex, education, occupation and the relationship to the household head), household wealth indicators [41], house characteristics, number of sleeping areas, ownership of mosquito nets, use of mosquito nets, net attitude indicators [42] and household coordinates in the global positioning system (GPS).
Randomization to study arms
ITNs distribution
After the baseline survey questionnaire has been completed, on the same day, one out of the four products of ITNs will be assigned to the household based on block randomization using a lottery method. All net types are packaged in identical cloth bags. Each field interviewer is given a large bag containing all four products that they select at random and once the net is selected, additional nets of the same product (starting with the same digit code) will be provided to cover all sleeping areas in the household.
Generating master lists for follow-up: attrition, fabric integrity and bioefficacy
The unique identifiers (UID) consist of six numbers generated by the study statistician – the first number identifies the country, the second identifies the treatment arm and the remaining four numbers identify nets distributed per arm. At baseline, the UID of each ITN given to the household will be recorded in the baseline questionnaire and in the ITNs master list with each unique net identification code linked to the household identifying codes and GPS coordinates. Two lists will be generated using a random number generator from the household masterlist: 1) attrition and fabric integrity monitoring masterlist (A-list) and 2) destructive sampling for bioefficacy and chemical retention monitoring masterlist (B-list). These are separate because bio-efficacy and chemical retention monitoring will require ITNs to be removed from the cohort and replaced with new nets. For attrition and fabric integrity monitoring, the same ITNs will be followed up to the end of the study period unless the household is lost to follow-up or withdraw consent.
Follow up surveys
Study ITNs per household visited during attrition monitoring at 6, 12, 24 and 36 months will be inspected for fabric integrity. All study ITNs listed in the master list provided will be inspected while non-study nets found in the household will only be recorded. In the households selected for bioefficacy testing, the first ITN listed will be withdrawn, and if that ITN is not present in the household, the second ITN on the list provided will be sampled. If no study ITNs are present in the household, another household with the same net product will be visited as a replacement.
Adverse effects and ITN usage
One month after distribution, 250 households for each ITNs product will be selected randomly from the household master list for the assessment of ITN usage and adverse events. The head of the selected households will be interviewed on perceived adverse effects from all members using the adverse effect questionnaire adapted from the WHO guideline [8].
Monitoring attrition
Attrition (the proportion of ITNs no longer found in their respective households) will be assessed at 6, 12, 24, and at 36 months after distribution. After 6 months, 2712 individual households (678 per product per location) will be selected randomly from the A-list prepared for each country separately. The same households will be followed up longitudinally at 12, 24 and 36 months. The presence or absence of all the ITNs that were distributed will be recorded and if absent the reasons for absence in the household will be provided by the head of the household. ITNs missing due to wear and tear will be considered as attrited and those missing because they were stolen, sold or given away to others, will be considered as lost to follow-up.
Monitoring fabric integrity
Each ITN found during the attrition monitoring will be draped over a collapsible net frame [40] and the number of holes of four different sizes categories [8] will be counted and recorded in a hole tally sheet. The four size categories of the holes are 0.5-2 cm, 2-10 cm, 10-25 cm, and above 25 cm for sizes 1, 2, 3 and 4, respectively approximated using a thumb, fist, head and larger than a head [43]. The overall physical condition of the net will be obtained by weighting the number of holes of each size by 1, 23, 196 and 576 for the four size categories of holes (based on the average assumed surface area for each size category) to obtain a proportionate hole index (pHI) following WHO guidelines [8]. The pHI value obtained will be used to classify the net as serviceable if the pHI value is less or equal to 642 and too torn if the pHI value obtained is greater than 642. Additionally, three new nets per product from the same production batch used in this study will be assessed for the Resistance to Damage Score [31] at CITEVE (Centro Tecnológico das Indústrias Têxtil e do Vestuário de Portugal).
Monitoring bio-efficacy
Mosquito species
Both pyrethroid susceptible and pyrethroid-resistant strains will be used for bioefficacy testing in each country. Mosquitoes are maintained following MR4 guidance [44] by feeding larvae on Tetramin fish food and adults on blood meal between 3 and 6 days after emergence and 10% glucose solution ad libitum. Temperature and humidity within insectaries are between 27ºC ± 5 ºC and 70% ± 20% relative humidity.
In Tanzania, metabolic pyrethroid-resistant Anopheles arabiensis (Kingani strain) and a full pyrethroid susceptible Anopheles gambiae sensu stricto (Ifakara strain) will be used. In Côte d’Ivoire, metabolically resistant An. gambiae s.s. (Tiassalé strain) and fully pyrethroid susceptible An. gambiae s.s. (Kisumu strain) will be used. In India, a pyrethroid-resistant An. culicifacies strains and fully pyrethroid susceptible An. stephensi strains will be used.
Mosquito net sampling and preparation
Cone bioassays will be conducted at baseline using 30 nets per product from B-list to ensure nets are of sufficient quality before distribution. At baseline, five net pieces (25 cm x 25 cm) will be cut from position 1 to position 5 of the net [8] and all net pieces will be tested in cone bioassay [38]. For each sampled net from the field, four net pieces (25 cm x 25 cm) will be cut from position 2 to position 5 of the net and tested in cone bioassay. A piece from position 1 (the bottom of the net) will not be cut, as it may be exposed to excessive abrasion due to tucking under the bed [38].
Bioefficacy testing will be conducted on a total of 30 nets (one from each of 30 households) sampled at random from the B-list for each product at 6, 12, 18, 24, and 30 months post-distribution survey. Only at 36 months, 50 nets per product will be sampled for bioefficacy and chemical testing estimates. However, 60 (30 selected for bioefficacy testing and a further 30) nets per product will be sampled at 12 and 24 months for additional testing in the I-ACT [45]. If the listed ITN is not present, another net (of the same product) within the same household will be selected. After a net is withdrawn for bioefficacy testing, the household will not be eligible for a next round of bioefficacy sample collection from the B-master list. The withdrawn net samples will be replaced. The head of the household will be offered any of the ITN products in the trial and allowed to select their replacement based on preference, and this choice will be recorded.
WHO Cone test procedure
The WHO cone bioassays will be conducted at the Vector Control Product Testing Unit (VCPTU) in Bagamoyo, Tanzania, at the Swiss Centre for Scientific Research (CSRS) in Tiassalé district, Côte d’Ivoire, and at the ICMR-National Institute of Malaria Research (NIMR) field unit in Bengaluru, Karnataka state, India. Four cones will be attached to each net piece and 5 mosquitoes exposed per cone, 80 mosquitoes per strain in total exposed to each ITN at each post-distribution survey and 100 mosquitoes at baseline [8]. Non-blood-fed, 2-5 days old, female mosquitoes each of the susceptible and resistant strains for the PBO nets and susceptible strain for the pyrethroid only net will be exposed to the ITN for 3 minutes [8]. After the exposure, mosquitoes will be removed and kept in holding cups provided with 10% glucose or sucrose solution. Mosquito knockdown will be recorded after 60 minutes (KD60) and mortality after 24-hours (M24). ITNs that do not meet WHO bioefficacy thresholds of KD60 ≥95% or M24 ≥ 80% with the susceptible mosquito strains will be tested in the WHO tunnel test [8].
WHO Tunnel Test
From each ITN that did not meet WHO efficacy criteria (i.e., ≥95% KD60 or ≥80% M24) in the cone test against the susceptible strains of mosquitoes, only one out of four net pieces will be selected for the tunnel test. This is the piece that gave mortality closest to the average mortality in the cone test of the four pieces (i.e., average for that net). The selected piece will be fixed in the tunnel for testing. One tunnel with untreated netting will be used as a negative control. Fifty non-blood fed female susceptible Anopheles mosquitoes aged 5–8 days sugar starved for 6 hours will be released in a tunnel (square section 25 cm x 25 cm) made of glass, 60 cm in length [8]. For pyrethroid only nets, susceptible mosquitoes will be used, while for PBO nets, both susceptible and resistant strains will be used.
At one-third of the length, the netting sample is fixed, with the surface of netting available to mosquitoes of 400 cm2 (20 cm x 20 cm) with nine holes each 1 cm in diameter: one hole is located at the centre of the square; the other eight are equidistant and located at 5 cm from the border. In the shorter section of the tunnel, a small rabbit, which will be restrained and unable to move but available to mosquitoes, will be placed as bait. In the cage at the end of the longer section of the tunnel, 50 female mosquitoes will be introduced at 18:00. The following morning from 09:00, the mosquitoes will be removed using a mouth aspirator and counted separately from each section of the tunnel and mortality and blood-feeding rates will be recorded. During the test, ambient conditions will be maintained at 27°C ± 5°C and 60% - 100% relative humidity. Acceptable feeding success and mortality in controls will be ≥50% and ≤10%, respectively.
Chemical residual analysis
The residual chemical concentration will be estimated for all products at baseline and after every 12 months using nets allocated in B-list. At baseline, five net pieces (25 cm x 25 cm) will be cut from 30 different mosquito nets (position 1 to 5 of each net) per product [8] and at 12, 24 and 36 months, four net pieces will be cut from position 2 to 5 from the nets sampled for bioefficacy testing of each product. These pieces will be rolled up and placed in a labelled aluminium foil and stored at 4oC until they are shipped for chemical analysis following CIPAC procedures to check if the chemical concentrations are within the tolerance limit as per the manufacturer’s specifications (Tsara Boost 333/LN/(M)/3, 33/LN/(M)/3, Veeralin 454/LN/(M)/3 33/LN/(M)/3, Olyset Plus 331/LN/(M)/3, 33/LN/(M)/3, MAGNet 454/LN/(M)/3).
Mortality and blood-feeding for PBO ITNs in Ifakara Ambient Chamber Test
Whole ITNs returned from the field at 12 and 24 months for bioefficacy testing in Tanzania and unwashed and 20 times laboratory washed nets of each product will be evaluated in the 18 chambers of the Ifakara Ambient Chamber Test (I-ACT) in Tanzania [40, 46]. Overall, the 18-arm study will include 4 arms of field aged nets, 4 arms of unwashed and 4 arms of 20 times laboratory washed nets. Also included are two negative controls and positive control nets (unwashed and 20 times washed); PermaNet® 3.0 and standard Olyset® nets. Unwashed, 20 times washed nets as well as positive and negative controls are deliberately holed with six holes (4x4 cm) as per guidance [8]. Eighteen volunteers will rotate sequentially among the 18 chambers of the I-ACT nightly based on a prepared roster to allow equal sleeping under nets per product among volunteers. Each chamber of the I-ACT will have a bed net frame over which the ITN will be draped and a foam mattress upon which one volunteer will sleep. The study will run for seventy-two experimental nights per time point (12 months and 24 months). After three rounds of 18 nights (54 experimental nights) each unwashed, 20 times washed nets and positive controls are replaced with field aged nets for the last round (18 nights) from the I-ACT to allow four field aged nets of each product tested for each of the remaining 18 experimental nights, leading to 126 replicates per field-aged product in total. This is done to allow a sufficient sample size for the non-inferiority comparison of field aged nets against Olyset Plus field-aged net
Fifteen nulliparous, sugar starved female An. gambiae s.s and An. arabiensis mosquitoes will be released at 21.00 hours in each of the compartments occupied by one volunteer sleeping under one of the test ITNs or a negative control net. These species are morphologically identical, thus one strain will be marked with a fluorescent dye that does not affect their survival or behaviour [47]. At 06.00 hours, each volunteer will collect mosquitoes inside the compartment using a mouth aspirator in paper cups. Recaptured mosquitoes will be sorted by species and recorded as fed alive, fed dead, unfed alive and unfed dead, then will be provided with 10% sucrose solution, and held under standard laboratory conditions to assess delayed mortality at 24-hours.
Sample size
The sample size for the ITN surveys
The sample size was calculated based on the primary outcome measure of net attrition together with additional nets for the bio-efficacy components. Assuming an average of 2.7 nets per household and a coefficient of variation of 0.25, the formula on page 110 of Hayes & Moulton [48] gives a sample size of 678 households per arm to detect the difference in attrition between two products assuming three-year attrition rates of 47.5% and 52.5% with 80% power. An additional 260 households for bio-efficacy, chemical analysis and Ifakara Ambient Chamber test (I-ACT) have been added to 678 making a total number of 938 households. Since the median lifespan of ITNs products from durability studies in Tanzania is less than 3 years [29, 49, 50], a 41% loss to follow-up of households is assumed making addition of 657 households to the 938 households. The final total number of households that will be recruited at baseline per product per country is 1,595. With an average of 2.7 nets per household [51], the total number of ITNs required per product per country will be 4,521 including 5% for error due to poor quality or nets with open seam prior to distribution, Table 5.
Table 5: Description of the numbers of ITNs distributed in the study area
Households for longitudinal attrition and fabric integrity monitoring
|
678
|
Households for I-ACT, cone bioassay and chemical testing
|
260
|
Unadjusted total households per product per country
|
938
|
Loss to follow up households
|
657
|
Adjusted total households per product per country
|
1,595
|
The average number of ITNs per household per product
|
2.7
|
Total ITNs per product
|
4,305
|
5% error and nets for replacement
|
215
|
Total ITNS required for each product per country
|
4,521
|
NOTE: The same net used in the I-ACT will be used for cone bioassay and chemical testing
The sample size for the entomological outcomes in the I-ACT
Simulation is used to determine the sample size for the bioefficacy testing of ITNs in the I-ACT. Each PBO net product will be assessed for non-inferiority against Olyset Plus on mortality at 24 hours. Each product consisting of sixty replicate nets will be tested 126 times (three rounds with 18 replicates each and the fourth round with 72 replicates) with 15 mosquitoes per strain per chamber per night. The variation between individual field-aged nets per product was assumed at a standard deviation of 0.1 and between chamber-nights at 0.15, both on the log-odds scale. Assuming mortality of 70% for the PBO nets and Olyset Plus, 1000 simulation trials were conducted to assess non-inferiority for each trial and the power, estimated by the proportion of trials that showed non-inferiority, was over 80%.
To assess the equivalence of field nets to 20-times washed nets on mortality, each product consisting of sixty replicates of field-aged nets as well as four replicates of twenty times laboratory washed and unwashed nets will be tested 54 times (three rounds with 18 replicates each). Assuming 70% mortality for each product of field and washed nets, 0.1 standard deviation between individual replicate nets per product and a chamber-night standard deviation of 0.15 on the log-odds scale, and 15 mosquitoes per strain per chamber per night the study is powered at 80%.
Samples size for chemical analysis
The baseline quality assurance of the ITNs used in the durability trial is conducted by chemical analysis and biological efficacy testing. A problem with the batch is detected if one or more of the sampled nets fails to meet the acceptance criteria. We propose that 10 ITNs are evaluated per site to give a total sample size of 30 ITNs tested at baseline. We base our sample size calculation on the power to detect a problem with the batch. With 30 nets, there would be at least 80% power to detect a problem with the batch if there was a probability of failure for each net individually of 0.008 or greater. The 30 net samples for chemical analysis will be sampled during follow up study period as per standard WHO procedures[8]
Data management
Field data will be collected using electronic data capture format in Open Data Kit (ODK) Collect. The data will be sent to the secure server located at Ifakara Health Institute in Tanzania. Cleaned data sets will be returned to CSRS and NIMR India. The information from the baseline questionnaire will be linked to the follow-up surveys using the unique household and country identification number as well as unique net identification numbers. Data collected on paper forms in the laboratory and in the I-ACT will be entered in excel using double-entry to facilitate cross-referencing and validation. The cleaned excel file of the data will be uploaded to the IHI server. Access to the data in the server will be limited to the data manager.
Data analysis
The durability outcomes attrition, fabric integrity, bioefficacy and insecticide content will be estimated. These outcomes will be estimated for each net product by survey period and country. A pooled estimate from three countries will also be presented, if appropriate.
The non-inferiority of Veeralin® and Tsara® Boost nets to Olyset® Plus will be carried out using logistic regression for the binary outcome of mosquito mortality with ITN product, compartment, volunteer and day as covariates. A random effect for chamber-night will be included. The 95% confidence interval for the estimated odds ratio for the effect of Tsara Boost® and Veeralin® compared to Olyset® Plus will be presented. Non-inferiority is shown if the confidence interval excludes an unacceptably worse performance. The bound of the margin of non-inferiority for mortality will be set at 0.7. If the lower bound of the confidence interval for the effect of the candidate net compared to Olyset® Plus is greater than 0.7, then we will conclude that the net product is non-inferior [52].
The utility of washed nets as a proxy for field nets will be assessed on the mortality and blood-feeding inhibition endpoints. The estimated odds ratios for the effect of washed nets compared to the negative control and field-aged nets compared to the negative control will be compared. The ratio of these odds ratios will be estimated along with the 95% confidence interval using interaction terms in the logistic regression model.