Comparative functional survival and equivalent annual cost of 3 long-lasting insecticidal net (LLIN) products in Tanzania: A randomised trial with 3-year follow up.

BACKGROUND: Two billion long-lasting insecticidal nets (LLINs) have been procured for malaria control. A functional LLIN is one that is present, is in good physical condition, and remains insecticidal, thereby providing protection against vector-borne diseases through preventing bites and killing disease vectors. The World Health Organization (WHO) prequalifies LLINs that remain adequately insecticidal 3 years after deployment. Therefore, institutional buyers often assume that prequalified LLINs are functionally identical with a 3-year lifespan. We measured the lifespans of 3 LLIN products, and calculated their cost per year of functional life, to demonstrate the economic and public health importance of procuring the most cost-effective LLIN product based on its lifespan. METHODS AND
FINDINGS: A randomised double-blinded trial of 3 pyrethroid LLIN products (10,571 nets in total) was conducted at 3 follow-up points: 10 months (August-October 2014), 22 months (August-October 2015), and 36 months (October-December 2016) among 3,393 households in Tanzania using WHO-recommended methods. Primary outcome was LLIN functional survival (LLIN present and in serviceable condition). Secondary outcomes were (1) bioefficacy and chemical content (residual insecticidal activity) and (2) protective efficacy for volunteers sleeping under the LLINs (bite reduction and mosquitoes killed). Median LLIN functional survival was significantly different between the 3 net products (p = 0.001): 2.0 years (95% CI 1.7-2.3) for Olyset, 2.5 years (95% CI 2.2-2.8) for PermaNet 2.0 (hazard ratio [HR] 0.73 [95% CI 0.64-0.85], p = 0.001), and 2.6 years (95% CI 2.3-2.8) for NetProtect (HR = 0.70 [95% CI 0.62-0.77], p < 0.001). Functional survival was affected by accumulation of holes, leading to users discarding nets. Protective efficacy also significantly differed between products as they aged. Equivalent annual cost varied between US$1.2 (95% CI $1.1-$1.4) and US$1.5 (95% CI $1.3-$1.7), assuming that each net was priced identically at US$3. The 2 longer-lived nets (PermaNet and NetProtect) were 20% cheaper than the shorter-lived product (Olyset). The trial was limited to only the most widely sold LLINs in Tanzania. Functional survival varies by country, so the single country setting is a limitation.
CONCLUSIONS: These results suggest that LLIN functional survival is less than 3 years and differs substantially between products, and these differences strongly influence LLIN value for money. LLIN tendering processes should consider local expectations of cost per year of functional life and not unit price. As new LLIN products come on the market, especially those with new insecticides, it will be imperative to monitor their comparative durability to ensure that the most cost-effective products are procured for malaria control.

RCT Entities:   Population Interventions Outcomes

Introduction

The use of long-lasting insecticidal nets (LLINs) remains the most cost-effective way to control malaria and reduce mortality [1], notwithstanding insecticide resistance [2]. However, despite the procurement of 254 million LLINs in 2017 alone, global LLIN coverage remains inadequate, with only 56% of the population in endemic areas estimated to have access to a LLIN [3]. LLINs are mostly distributed through periodic mass distribution campaigns, and as a result, population access to LLINs fluctuates over time. Access is typically high directly after a mass campaign and then declines as nets wear out, often to 50% or less, until the next campaign. This fluctuating pattern of coverage, caused by nets wearing out, is seen across the African region [4], where gains in malaria control have stalled, and fewer than 50% of endemic countries remain on track to reach critical malaria reduction targets [3]. Investment in malaria control has stagnated and was US$2.3 billion (50%) below the resources required to meet the World Health Organization (WHO) targets of 40% reductions in malaria case incidence and mortality rates by 2020 [5]. These gaps in funding and coverage emphasise the need to deploy products that present the best value for money.

A report to the Malaria Policy Advisory Committee (MPAC) advised that increasing the functional life of LLINs by 1 or 2 years would reduce the cost of malaria control by between US$500 million and US$700 million over a period of 5 years [6]. A functional LLIN is one that is present, is in good physical condition, and remains insecticidal, thereby providing protection against vector-borne diseases through preventing bites and killing disease vectors [6]. Durability, or functional survival, of LLINs varies between geographical regions [7] and environments [8,9] and remains an undervalued but critical determinant of the success and efficiency of malaria control programmes [10,11]. How long LLINs remain protective under user conditions will dictate how frequently they should be replaced, which has both public health and economic implications [12]. In 2011, it was calculated that in Tanzania, for mean LLIN lifespans of 2, 3, and 4 years, 89, 63, and 51 million LLINs, respectively, would be needed over 10 years to achieve national access targets [10].

Currently, WHO prequalifies products that demonstrate adequate insecticidal activity 3 years after deployment, but does not appraise the physical deterioration of nets over time as part of the LLIN prequalification assessment [13]. Historically, pyrethroid-treated LLINs were assessed in multi-country studies for physical and chemical durability over an anticipated lifespan of 3 years and 20 washes. In the mid-2000s, when these procedures were designed, we did not yet know the relative importance of attrition—the disappearance of nets from study households—as one of the main factors limiting the duration of protection from LLINs. Unfortunately, even after the importance of attrition had become very clear, the evaluation criteria were never changed to take account of it. Thus, of the nets tested in the current study, PermaNet 2.0 received WHO recommendation (now prequalification) based on pooled prospective data from 6 countries, where 80% of remaining nets met bioefficacy and net fabric integrity criteria [14]; Olyset received recommendation based on pooled retrospective data from 7 countries, where 77% of nets passed bioefficacy criteria, although net loss and damage could not be accurately assessed [15]; and NetProtect did not receive full recommendation due to inconsistencies in data between WHO-sponsored studies [16,17], and was withdrawn from the market after the trial reported here had started.

The WHO prequalification website lists a number of newly prequalified products as long-lasting (LLINs) [18], including some with active ingredients other than pyrethroids. The listing of these products was based on experimental hut data from 2 or 3 sites. Fabric integrity, residual chemical content, and bioefficacy data for products after operational household use through longitudinal studies or post-marketing surveillance are requested, but are not a requirement for prequalification. This has resulted in a tendering process where donors assume LLINs are identical, and procurement is weighted by the unit price of the commodity without regard to actual product lifespan [19]. However, all the available data suggest that the assumption of a uniform 3-year lifespan for all LLIN products is unrealistic [4]. There is a clear need for a more integrative economic approach, with purchasing decisions based on value for money and cost per effective unit of LLIN coverage [6,19]. New product classes of LLINs with novel active ingredients for insecticide resistance management are becoming available [20], but they remain susceptible to the same forces of physical disintegration, being discarded, and losing insecticidal activity. Moreover, in most cases, they are more expensive. This emphasises the need to consider the price of LLINs in terms of cost per year of functional life [12].

Here we report results from a large randomised trial of 3 LLIN products (PermaNet 2.0, Olyset, and NetProtect), conducted in 8 epidemiologically and ecologically distinct districts in Tanzania. The proportion of LLINs remaining in use and still protective against malaria mosquitoes was measured over 3 years of follow-up after deployment. We calculated relative LLIN cost-effectiveness in terms of the equivalent annual cost (EAC), which is a conventional financial indicator used to compare products with different effective lifetimes. The median functional survival of each product and its EAC were calculated to inform optimal procurement of cost-effective LLINs.

Methods

The trial has been described in detail previously [21]. It took place in 8 districts in Tanzania, selected to be representative of national environmental, ecological, and epidemiological settings (Fig 1). Within each district, 10 villages were randomly selected (except for Kinondoni [Dar es Salaam], where only 6 areas were available), and within each village, 45 households were recruited to participate in the trial. All households were randomised to receive 1 of 3 LLIN brands on a 1:1:1 ratio, stratified by village. The 3 brands were Olyset (manufactured with an enhanced knitting pattern that was introduced in 2013; permethrin incorporated in 150 denier polyethylene; Sumitomo Chemicals, Japan), PermaNet 2.0 (deltamethrin coated on 100 denier polyester; Vestergaard Frandsen, Switzerland), and NetProtect (deltamethrin incorporated in 110 denier polyethylene; BestNet, Denmark). Distribution of trial nets took place between October and December 2013. All nets owned by the participating households were collected and replaced with enough nets to cover all sleeping spaces. Before distribution, a sample of 10 nets per product was quality tested. Nets were the same size and colour and labelled by a 5-digit serial number so that participants and investigators remained blinded to the LLIN product until data collection was complete. In total, 3,393 households were randomised (1,132 to Olyset, 1,127 to PermaNet 2.0, and 1,134 to NetProtect), to which 10,571 nets were distributed.

Fig 1

Map of trial districts with 2015 malaria prevalence data (percent of children aged 6–59 months diagnosed with malaria by rapid diagnostic test and microscopy) [22].

Open-access shapefiles from https://www.nbs.go.tz/index.php/en/census-surveys/gis.

Surveys were conducted among all consenting trial households when the LLINs were distributed and at 3 follow-up points: 10 months (August–October 2014), 22 months (August–October 2015), and 36 months (October–December 2016) (S1 Table). The serial numbers of the nets, linked to household-identifying codes in a master list, enabled follow-up of each net at each time point. At each follow-up visit, information on each LLIN was collected, including whether the net was present in the house, whether the net was in use, and, if the net was not present, reasons why it was not present. Physical integrity of LLINs was measured on a random sample of 3 nets per household by counting the number, location and size of holes [13,23]. Socioeconomic variables and a household member roster were also recorded. Electronic data capture was used for all surveys.

In addition to the data collected as part of the household surveys, at each time point 48 LLINs from each brand were randomly sampled from the master list and returned to a laboratory in Bagamoyo, Tanzania, for bioefficacy and chemical analysis using standard WHO methods [13,23] and, additionally, the Ifakara Ambient Chamber Test (IACT) [24]. Households received new nets to replace those removed for destructive sampling. Once a house had been sampled, it was eliminated from the master list to prevent confounding of results. Table 1 describes the different components of LLIN durability, the tests conducted to obtain the data, the outcome indicators for statistical analysis, and the corresponding WHO threshold criteria [6,13,23]. The numbers of LLINs tested for each of the components of LLIN durability are listed in S1 Table.

Table 1

LLIN durability components.

Component	Definition	Test conducted	Outcome indicators	WHO criteria or industry standard
Attrition	Net loss from household through discarding or alternative use	Household survey	Net presence
Physical integrity	Physical state of the net to estimate bite protection	Count number, location, and size of hole(s) of a maximum 3 nets per household	Holed surface area measured by pHI [6] or MHSA (cm2)	pHI 0–64, MHSA ≤ 79 cm2: good
				pHI 65–642, MHSA 80–789 cm2: damaged
				pHI ≤ 642, MHSA ≤ 789 cm2: serviceable
				pHI ≥ 643, MHSA ≥ 790 cm2: too torn/unserviceable
Functional survival [6]	Estimation of nets still in households in serviceable condition	Median survival analysis	(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%
Biological efficacy	Ability of net to incapacitate or kill susceptible anopheline mosquitoes after contact with insecticide	IACT: whole nets [24]	Proportion of mosquitoes dead at 24 hours
			Proportion of mosquitoes not blood fed
		WHO cone/tunnel test: 25 × 25 cm pieces [13]	Net samples meeting optimal bioefficacy criteria	1-hour knock-down ≥ 95%or24-hour mortality ≥ 80%orblood feeding inhibition ≥ 90%
Insecticide content	Amount of active ingredient in the net	Permethrin: GC-FIDDeltamethrin: HPLC-DAD	Compliance of nets with WHO specifications at baseline; loss of active ingredient over time	Olyset: 20 g/kg ± 25% [15–25 g/kg]PermaNet: 1.4 g/kg ± 25% [1.05–1.75 g/kg]NetProtect: 1.8 g/kg ± 25% [1.35–2.25 g/kg]

GC-FID, gas chromatography with flame ionisation detection; HPLC-DAD, high-performance liquid chromatography with diode array detection; IACT, Ifakara Ambient Chamber Test; LLIN, long-lasting insecticidal net; MHSA, median hole surface area; pHI, proportionate hole index; WHO, World Health Organization.

First, the protective efficacy of whole nets returned from the field was evaluated using IACT [24]. Each night, 10 male volunteers slept underneath 1 of the nets (or an untreated control net to monitor the quality of the bioassay) between 9 PM and 6 AM in a small chamber similar in size to a bedroom, within a screened compartment. At 9 PM, 30 laboratory-reared mosquitoes were released into the chamber. The next morning, all mosquitoes within the compartment were recaptured, and scored for 24-hour mortality and blood feeding inhibition. Each LLIN was tested twice on 2 consecutive nights. Subsequently, net pieces (25 × 25 cm2) were cut following the WHO sampling pattern and standard WHO cone bioassays were carried out [13]. If nets did not meet WHO optimal bioefficacy criteria for cone tests (Table 1), WHO tunnel tests were conducted [13]. All mosquito assays were conducted with fully pyrethroid-susceptible 2- to 8-day-old nulliparous female Anopheles gambiae sensu stricto (Ifakara strain). Insecticide content analyses were performed at Walloon Agricultural Research Centre (CRA-W) using standard Collaborative International Pesticides Analytical Council Limited (CIPAC) methods for determining LLIN insecticide content (Olyset, 331/LN/M/3; PermaNet 2.0, 333/LN/(M)/3; NetProtect, 333/LN/(M2)/3).

Statistical analysis

All statistical analyses were conducted using Stata release 13 (StataCorp, College Station, TX). Data from the surveys at 10, 22, and 36 months were used to calculate attrition and functional survival (Table 1) using Kaplan–Meier estimators. For both attrition and functional survival, nets reported as given away, sold, or stolen were treated as lost to follow-up. Hazard ratios (HRs) for the difference in attrition and functional survival were calculated using discrete time survival analysis using a complementary log-log model [25]. Robust standard errors were used to account for the highest level of clustering (district) [26]. Of nets that were present, net condition was defined, following WHO recommendations, as ‘good’ or ‘damaged’ (combined as ‘serviceable’) or ‘too torn/unserviceable’ (Table 1). Negative binomial regression was used to compare hole surface area between net products. Data on WHO bioassays and the IACT test came from the 48 nets sampled at each time point. For WHO bioassays and the IACT test, if control mortality for an assay of a section of net was over 10%, the data from that section were not included in the analysis. A chi-squared test assessed the proportion of nets of each product passing the WHO bioefficacy criteria based on combined cone and tunnel tests. Logistic regression was used to analyse mortality and blood feeding inhibition from the IACT test; results were adjusted for chamber and experimental night, and robust standard errors were used to take account of nets being tested multiple times. A further analysis was conducted to test for differences in mortality between net brands in the IACT test based on net condition, in which net condition (defined above) was adjusted for as a fixed effect.

Economic analysis

The EAC of an LLIN was calculated according to the standard formula [27]. To assess the value of longer functional survival, we used Eq 1, where b is the ratio of the lifespan of the more durable product to the lifespan of reference net n. The variable r is the discount rate. This relationship shows, for any change in net lifespan from n to bn, the relative increase in price, a, that would yield an identical EAC for the 2 products. Other factors being equal, a relative price increase less than a would favour the new, longer-lasting LLIN, while relative price increases greater than a would favour the reference net.

Simulation of EACs for products tested in the trial was conducted using Monte Carlo methods, assuming a 3% discount rate, as is standard in health economic analysis. The baseline survival function for LLINs was estimated by regressing the survival proportions of Olyset nets derived from Kaplan–Meier analysis against time. The survival function was converted into a baseline hazard, and net failure lifetimes were simulated for a cohort of 500 LLINs assuming a Weibull distribution of time to failure (in terms of functional survival). The results of the cohort were summarised by estimating the median lifetime, and this process was repeated 10,000 times for each net type, yielding an estimate of the expected median lifetime and quantiles of its expected distribution. Results were converted into EACs with 95% quantiles. Distributional assumptions for the baseline hazard and the parameters of the Weibull distribution were fitted to the results. The baseline hazard and proportional hazard were simulated with log normal distributions (S2 Table).

Ethics

Ethical approval was granted by ethical review committees at the London School of Hygiene & Tropical Medicine (6333/A443), Ifakara Health Institute (IHI/IRB/AMM/No: 07–2014), and the Tanzanian National Institute for Medical Research (NIMR/HQ/R.8c/Vol. I/285). Community sensitisation meetings were held prior to trial inception, and written informed consent was obtained from the head of the household or another adult household member of participating households before each survey. Volunteers for the IACT experiment were all Ifakara Health Institute staff members with appropriate training who gave written informed consent.

Results

A total of 3,393 households were randomised, to which 10,571 nets were distributed (3,520 Olyset [33%], 3,513 PermaNet 2.0 [33%], and 3,538 NetProtect [33%]). The 3 trial arms were similar in number of participants, number of nets allocated, household characteristics, house design, and socioeconomic characteristics (Table 2). The proportion of households lost to follow-up was 20% over the 3 years of the trial.

Table 2

Household and socioeconomic characteristics of participating households in each trial arm.

Characteristic	Olyset	PermaNet 2.0	NetProtect
Number of nets distributed	3,520	3,513	3,538
Number of participants	6,061	6,024	6,200
Number of households	1,132	1,127	1,134
Average household size	5.8	5.8	6.5
Mean sleeping spaces per household	3.65	3.55	3.55
Mean nets per household	2.92	2.96	3.04
Male household members (%)	49	48	49
Female household members (%)	51	52	51
Age distribution of household members (%)
≤5 years	16.64	17.21	17.56
6–17 years	33.16	33.27	34.19
18–50 years	37.61	39.16	37.73
≥51 years	12.60	10.36	10.52
Highest level of education of household head (%)
No education	21.62	19.99	20.69
Some primary education	30.23	29.26	20.69
Completed primary school	32.60	33.54	39.66
Secondary education	6.45	6.75	5.17
Housing materials (%)
Roof: thatch	19.88	17.11	17.08
Roof: tin	79.89	82.60	82.56
Walls: mud and sticks	17.30	14.96	14.65
Walls: mud brick	24.15	21.81	22.18
Walls: burned brick	40.32	43.54	43.98
Walls: cement brick	18.23	19.69	19.19
Floor: mud	52.97	48.42	49.89
Floor: cement	43.17	46.13	44.48
Socioeconomic quintile (%)
1 (least wealthy)	21.90	18.99	19.23
2	20.59	19.06	20.60
3	19.85	20.12	20.29
4	19.70	20.65	19.52
5 (most wealthy)	17.96	21.18	20.37

Functional survival

There were significant differences in functional survival (defined as presence of serviceable net) of the 3 products (Table 3). Estimated median functional survival was 2.0 years (95% CI 1.7–2.3) for Olyset, 2.5 years (95% CI 2.2–2.8) for PermaNet, and 2.6 years (95% CI 2.3–2.8) for NetProtect (p < 0.001). There was no significant difference in net use by net product (S3 Table).

Table 3

Percentage net functional survival (defined as presence of the net in the house and in serviceable condition) and simulated equivalent annual cost (assuming S$3.0 purchase price) by net product and time point.

Net product	Percent functional survival (95% CI)			Median survival in years (95% CI)†	Hazard ratio (95% CI), p-value	Simulated equivalent annual cost in US dollars (95% CI)
	10 months	22 months	36 months
Olyset	82 (79, 85)	54 (47, 62)	27 (20, 34)	2.0 (1.7, 2.3)	1	1.5 (1.3, 1.7)
PermaNet	88 (85, 90)	65 (57, 72)	38 (31, 46)	2.5 (2.2, 2.8)	0.73 (0.64, 0.85),p = 0.001	1.2 (1.1, 1.4)
NetProtect	88 (84, 91)	67 (61, 72)	40 (34, 45)	2.6 (2.3, 2.8)	0.70 (0.62, 0.77),p < 0.001	1.2 (1.1, 1.4)
					p = 0.001*

†Details of the survival analysis are provided in S4 Table.

*p-Value for the comparison between the 3 nets. For the difference between PermaNet and Netprotect, p = 0.199.

Simulation results show that the expected EAC in US dollars of the 3 LLINs in the trial varied between $1.2 (95% CI $1.1–$1.4) for PermaNet and NetProtect and $1.5 (95% CI $1.3–$1.7) for Olyset, assuming that each net was priced identically at $3.0 (Table 3). The longer-lived net products (PermaNet and NetProtect) were approximately 20% lower in EAC than the shorter-lived Olyset product.

Components of functional survival and secondary outcomes

Attrition

There were significant differences in attrition between net products. Olyset nets were lost at a faster rate than PermaNet 2.0 and NetProtect nets (Table 4). After 3 years, 55% of Olyset nets were no longer present in households, compared to 42% of PermaNet 2.0 and 46% of NetProtect nets (p < 0.001; Table 4). Of the 10,571 nets distributed, 4,964 (46%) were lost over the whole trial period (S5 Table).

Table 4

Percentage attrition (defined as net loss due to discarding or alternative use of nets) and hazard ratios after 36 months by net product and time point.

Net product	Percent attrition (95% CI)			Hazard ratio (95% CI), p-value
	10 months	22 months	36 months
Olyset	7 (5, 8)	25 (21, 29)	55 (49, 61)	1
PermaNet	5 (3, 6)	20 (17, 24)	42 (38, 46)	0.71 (0.64, 0.79), p < 0.001
NetProtect	6 (4, 8)	22 (18, 26)	46 (43, 50)	0.81 (0.71, 0.93), p = 0.008
				p < 0.001*

Details of the analysis are provided in S5 Table. Number of nets remaining in households by time point: 10 months, 8,269 nets; 22 months, 6,324 nets; 36 months, 3,942 nets.

*p-Value for the comparison between the 3 nets. For the difference between PermaNet and NetProtect, p = 0.006.

Physical integrity

The condition of nets that remained in households deteriorated over the course of the trial. At each time point, Olyset had the largest proportion and NetProtect had the smallest proportion of ‘too torn’ nets (Fig 2). The median hole surface area in Olyset nets increased from 38 cm2 at 10 months to 459 cm2 after 36 months, compared to 6 cm2 to 295 cm2 for PermaNet 2.0 and 8 cm2 and 152 cm2 for NetProtect (S6 Table). Questionnaire data showed that at 3 years, 70% of nets no longer in use had been discarded when they were perceived as too damaged to be useful. Others were given away (17%), stolen (3%), or repurposed (3%).

Fig 2

Physical condition of long-lasting insecticidal nets remaining in households at time of survey according to WHO categorisation using proportionate hole index (pHI) [5] for the 3 net products and time points.

Green shows percent of nets in good condition (pHI 0–64), orange shows percent nets in damaged condition (pHI 65–642), and red shows percent of nets defined as ‘too torn’ (pHI ≥ 643). The sample sizes at 10 months were as follows: Olyset, 3,520; PermaNet, 3,513; NetProtect, 3,538. The sample sizes at 22 months were as follows: Olyset, 2,592; PermaNet, 2,622; NetProtect, 2,617. The sample sizes at 36 months were as follows: Olyset, 1,687; PermaNet, 1,827; NetProtect, 1,746.

Bioefficacy

At baseline, all products met optimal WHO bioefficacy criteria. After field use, there were significant differences in the bioefficacy of the net products measured using standard WHO cone and tunnel tests over time (Table 5). At 10 months, 100% of NetProtect and PermaNet 2.0 nets met WHO optimal bioefficacy criteria, compared to 73% of Olyset nets (p < 0.001). Nets decreased in bioefficacy through time, but even after 3 years, 96% of NetProtect, 85% of PermaNet 2.0, and 75% of Olyset nets met WHO criteria for bioefficacy (p = 0.017; Table 5).

Table 5

Percentages of net products meeting optimal WHO bioefficacy criteria by time point.

Net product	WHO cone test			WHO tunnel test			Overall (cone + tunnel)
	10 months	22 months	36 months	10 months	22 months	36 months	10 months	22 months	36 months
Olyset	4	8	14	72	78	71	73	79	75
	(1, 14)	(2, 20)	(5, 27)	(57, 84)	(62, 89)	(54, 85)	(58, 85)	(65, 90)	(60, 87)
	[2/48]	[4/48]	[6/44]	[33/46]	[34/44]	[27/38]	[35/48]	[38/48]	[33/44]
PermaNet	98	92	73	100	50	46	100	96	85
	(89, 100)	(80, 98)	(58, 85)	(3, 100)	(7, 93)	(19, 75)	(92, 100)	(85, 99)	(72, 94)
	[46/47]	[44/48]	[35/48]	[1/1]	[2/4]	[6/13]	[47/47]	[46/48]	[41/48]
NetProtect	100	100	73	n/a	n/a	85	100	100	96
	(92, 100)	(93, 100)	(58, 85)			(55, 98)	(92, 100)	(93, 100)	(86, 99)
	[47/47]	[48/48]	[35/48]			[11/13]	[47/47]	[48/48]	[46/48]
							<0.001*	<0.001*	0.017*

95% confidence intervals in parentheses. Numbers passing/numbers tested in square brackets [n/N]. Nets are tested by cone test, and those that fail WHO optimal insecticide effectiveness criteria of ≥95% knock-down after 60 minutes or ≥80% 24-hour mortality are then further tested by tunnel test. Optimal criteria for the tunnel test are ≥80% 24-hour mortality or ≥90% blood feeding inhibition. Overall pass (cone and tunnel) is based on a net achieving 1 or more of these 4 criteria.

*p-Value for the comparison between the 3 nets. For the differences between Olyset and PermaNet, the p-values were <0.001, 0.014, and 0.208 at 10, 22, and 36 months, respectively. For the differences between Olyset and NetProtect, the p-values were <0.001, <0.001, and 0.004 at 10, 22, and 36 months, respectively. For the differences between PermaNet and NetProtect, the p-values were 1.0, 0.153, and 0.080 at 10, 22, and 36 months, respectively.

When whole nets were tested after 3 years using IACT, 88% of Olyset, 96% of PermaNet 2.0, and 92% of NetProtect nets passed WHO optimal criteria of ≥80% mortality and ≥90% blood feeding inhibition. There were differences between products in 24-hour mortality. Olyset showed lower mortality (p < 0.001), but all 3 products showed similar levels of feeding inhibition (Fig 3; S7 Table). Mosquito mortality was higher for nets defined as ‘too torn’ (odds ratio = 0.65 [95% CI 0.49–0.88], p = 0.005), and the differences in mosquito mortality between the net products remained significant after adjusting for physical condition. Similarly, protection from mosquito bites (feeding inhibition) was considerably lower among nets that were ‘too torn’ (OR = 0.12 [95% CI 0.08–0.18], p < 0.001), but the differences between the net products remained non-significant after adjusting for physical condition.

Fig 3

Ifakara Ambient Chamber Test (IACT) results for mosquito mortality and blood feeding inhibition by net product and time point.

Mosquito mortality (top panel) and blood feeding inhibition (bottom panel). Orange, Olyset; blue, PermaNet; maroon, NetProtect. Optimal WHO criteria (80% mortality; 90% blood feeding inhibition) are indicated by dashed lines. The number of mosquitoes used at 10 months was as follows: Olyset, 2,700; PermaNet, 2,730; NetProtect, 2,730. The number of mosquitoes used at 22 months was as follows: Olyset, 2,880; PermaNet, 2,880; NetProtect, 2,880. The number of mosquitoes used at 36 months was as follows: Olyset, 2,880; PermaNet, 2,880; NetProtect, 2,880.

Active ingredient content

At baseline, 100% (10) of Olyset and PermaNet 2.0 and 50% (5) of NetProtect samples complied with their target doses of active ingredient (S8 Table).

At 10 months, 22 months, and 36 months, mean permethrin content in Olyset nets decreased to 16.2 g/kg, 14.8 g/kg, and 13.0 g/kg, corresponding to a loss of 20%, 27%, and 36% of the original dose, respectively. Mean deltamethrin content of PermaNet 2.0 nets decreased to 0.75 g/kg, 0.47 g/kg, and 0.40 g/kg, corresponding to a loss of 48%, 68%, and 72% of the original dose, respectively. Mean deltamethrin content of NetProtect nets decreased to 0.91 g/kg, 0.52 g/kg, and 0.40 g/kg, corresponding to a loss of 33%, 61%, and 70% of the original dose, respectively (S8 Table). While this loss of insecticide did not negatively impact the bioefficacy of the nets against a pyrethroid-susceptible strain of mosquito, it is plausible that it would impact the efficacy of the nets against more resistant mosquitoes.

Discussion

We conducted a randomised trial with 10,571 new LLINs of 3 brands (3,520 Olyset, 3,513PermaNet, and 3,538 NetProtect) distributed among 3,393 households in 76 villages in 8 districts in Tanzania and followed up annually for 3 years. This was done to measure the rate at which the 3 net brands became damaged, lost bioefficacy, and were discarded by households. The findings of this trial demonstrate that there is considerable variability in the lifespan of pyrethroid-treated LLIN products. Our data also confirm that the median functional life of the LLINs in our study was less than 3 years in Tanzania, as also suggested by a systematic review of LLIN retention data in 39 sub-Saharan African countries [4]. A WHO-sponsored evaluation of NetProtect and PermaNet 2.0 conducted in Kenya showed very similar results to those found here, with a median time to failure of 2.5 years for PermaNet 2.0 and 2.5 years for NetProtect [16]. A full literature review of durability data available for the products evaluated in this trial is included in S1 Text. Summary net durability data available from peer-reviewed publications and WHO reports agree with the data in our trial for estimates of bioefficacy and fabric integrity after 3 years of operational use. The proportions of nets passing WHO bioefficacy criteria were above 80% for NetProtect and PermaNet 2.0 and slightly below 80% for Olyset. NetProtect and PermaNet had similar fabric integrity after 3 years of domestic use, with a higher proportion of serviceable nets relative to Olyset.

While there have been substantial economic investments to find new active ingredients, insecticide combinations, and synergists to combat the negative effects of insecticide resistance [28], the importance of durability for LLIN effectiveness has been side-lined. Consideration of its importance in vector control by key stakeholders such as the WHO may re-awaken the LLIN market to reward more durable products. This should, in turn, create incentives for investments in technological advances, research, and development by LLIN manufacturers [19]. There are indications that LLINs can be made substantially more durable for a small increase in unit price [29], and rapid technological evolution may be possible if there are appropriate market incentives.

The WHO’s Guidelines for Procuring Public Health Pesticides [30] recommends that procurement decisions consider ‘operational cost’ rather than unit price, and an appropriate measure to compare value for money of LLINs would be ‘cost per median year of net life under local conditions’. We measured the relative durability of nets using functional survival estimates, in terms of the EAC, and demonstrated that this approach outlined by WHO would indeed be useful. The cost analysis showed approximately 20% lower EAC when a longer-lasting LLIN (PermaNet 2.0 or NetProtect) was chosen over a shorter-lasting LLIN (Olyset), assuming prices for the products were identical. The economic modelling showed that the relative increase in price that is acceptable for a new product coming to market is also much smaller when the lifetime of the standard product increases (S1 Fig). Thus, the extension of the life of an innovator product is much more valuable if the standard product is relatively short-lived, as was seen in this study.

WHO requests LLIN manufacturers to provide data from 3 longitudinal field evaluations in different ecologies (e.g., West Africa, East Africa, and Asia) to retain prequalification listing. While it is recognised that durability is context-specific, we argue that it is possible to routinely generate median functional survival estimates and EACs for at least 3 locations using the WHO methodology outlined [13,23], albeit with a more limited sample size than the present study. The EAC may be a useful metric to compare cost-effectiveness of products, rather than the current practice of assessing products based simply on a minimum threshold of insecticidal activity after 3 years.

The limitation of the EAC metric is that it only captures the relative weighting of price and effective lifetime, while full cost-effectiveness and cost (including non-commodity costs) will result from a complex interaction of net durability, distribution modality, cost, and effectiveness. A limitation of the simplified approach here is that it does not fully consider these interactions, but it presents a straightforward and easily applicable approach to judging the relative cost and lifetime of a product.

Attrition and fabric integrity, the 2 factors that define physical survival of LLINs [6,31], differed significantly between the 3 net products. Olyset demonstrated more rapid accumulation of damage and faster attrition. In the current study and in previous work, we demonstrated that most LLINs were discarded because they were perceived by users as too damaged to offer protection against mosquito bites or malaria [32]. Attrition and fabric integrity are highly variable between contexts, and information on these factors is simpler to collect than bioefficacy or chemical content data. Further consideration should be given to developing simple tools to allow countries to assess attrition and fabric integrity during routine surveys (e.g., Malaria Indicator Surveys or Demographic and Health Surveys) to inform planning of intervals between mass distribution campaigns.

Of those nets still present after 3 years, 25%–40% were categorised as no longer physically serviceable, depending on the brand. However, even after 3 years, nets remained highly insecticidal when tested by bioassays against insecticide-susceptible malaria vectors. Damage actually increased the mortality of mosquitoes that entered nets through holes and became trapped, as also observed in other studies [33]. Indeed, torn LLINs continue to provide a degree of individual and community protection from malaria [34,35]. Our IACT experiments demonstrated that the 3 brands were all highly protective, although Olyset killed significantly fewer mosquitoes than PermaNet 2.0 and NetProtect. It is of note that the most common location for damage to the nets is on the bottom section of the nets at the point where they are tucked under a mat or mattress (S2 Fig). The act of tucking makes these holes inaccessible to mosquitoes even though the net appears to be badly damaged to the user, which may motivate them to discard the net.

A limitation of the trial is that only susceptible mosquitoes were used for bioefficacy testing. Pyrethroid resistance is widespread and increases feeding success and reduces mortality of mosquitoes [33]. Another limitation is the fact that the trial was only conducted in Tanzania (albeit in a wide range of epidemiological settings). Functional survival varies by country (S1 Text), so the single country setting is a limitation. However, the setting is more likely to affect absolute net survival rates than the comparison between LLIN products. Furthermore, the trial only included 3 brands of LLINs, all of which are treated with pyrethroids. As new LLIN products come on the market treated with different insecticides, insecticide combinations, or synergists, such as piperonyl butoxide (PBO), it will be imperative to monitor their comparative durability to ensure that the most cost-effective products are procured for malaria control. Functional life will have important implications for the selection of new products for resistance management that have higher unit costs. New pyrethroid plus PBO nets may not be as durable as standard pyrethroid nets because PBO is lost rapidly from nets during washing, which reduces their efficacy [36]. However, in Tanzania, PBO nets continued to have superior public health benefits 2 years after distribution [20]. If the median functional survival of pyrethroid LLINs is 2 years, then PBO nets may remain cost-competitive.

Our findings confirm that even after 3 years, nets that are still in households, despite holes, still give partial protection against mosquito bites and continue to kill mosquitoes, providing some personal and community protection. However, if nets are discarded, or no longer used because they are perceived as too damaged, then they have no public health benefit at all. While it is possible to encourage users to retain their damaged, but still insecticidal, nets through behavioural change communication, a more effective and safer strategy would be to distribute more physically durable LLINs [29]. LLINs are the largest single cost item in the global malaria control budget. If LLIN effective lifespans became longer, net replacement needs would be substantially reduced, aiding in improving population access to this life-saving intervention despite the current stagnation in financial support for malaria control. It is technically feasible to manufacture more durable LLINs. However, this will happen only if institutional buyers consider cost-effectiveness for coverage [30] and give greater market share to longer-lasting and better value-for-money products.

Relationship between increased net lifetimes in years and the acceptable increase in price.

(TIF)

Click here for additional data file.

The location of damage on nets by year after distribution and net brand measured by proportionate hole index.

(PDF)

Click here for additional data file.

(PDF)

Click here for additional data file.

Study flow.

The number of interviews completed each year, loss to follow-up, and the number of study nets evaluated for each durability component is shown.

(PDF)

Click here for additional data file.

Parameters used in simulation of lifetimes for equivalent annual cost simulation analysis.

(PDF)

Click here for additional data file.

Reported net use the previous night by net product and time point.

Data represent numbers of respondents (percent) reporting use of nets.

(PDF)

Click here for additional data file.

Number at risk (functional survival).

(PDF)

Click here for additional data file.

Number at risk (attrition).

(PDF)

Click here for additional data file.

Median hole surface area (in cm2) and interquartile range (IQR) by net product and time point.

(PDF)

Click here for additional data file.

Ifakara Ambient Chamber Test (IACT) results for mosquito mortality and blood feeding inhibition by net product and time point (in months).

(PDF)

Click here for additional data file.

Number of nets, mean active ingredient (AI) content (g/kg), range (g/kg), and between net variation (%RSD); percentage of active ingredient lost over time; mean R-alpha isomer content (g/kg); and percentage of deltamethrin (only for PermaNet 2.0 and NetProtect) in net samples at baseline and 3 follow-up time points.

(PDF)

Click here for additional data file.

Literature review on durability of PermaNet 2.0, Olyset, and NetProtect nets.

(PDF)

Click here for additional data file.

30 Mar 2020

Dear Dr. Moore,

Thank you very much for submitting your manuscript "Comparative functional survival and equivalent annual cost of three long lasting insecticidal net (LLIN) products in Tanzania: a three-year prospective cohort study of LLIN attrition, physical integrity, and insecticidal activity" (PMEDICINE-D-19-02821) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to three independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Apr 20 2020 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1- I think the title can be shortened: “Comparative functional survival and equivalent annual cost of three long lasting insecticidal net products in Tanzania: a three-year prospective cohort study”

2- Thank you for listing the authors’ competing interests. Please list, either in the CI statement or in a supplemental file if the list is very long, the manufacturers that the authors work with.

3- Thank you for agreeing to make the data available and providing the link for data access. Please also provide any accession number(s) or search terms a researcher will need to obtain access to this specific dataset.

4- The Abstract Background is a bit long, please shorten by about half.

5- Please amend “We conducted the largest prospective study of LLIN…” with “to our knowledge” or similar.

6- Please include full date ranges for the study in the Abstract Methods and Findings section.

7- In the Abstract Methods and Findings section, please include both 95% CIs and p values for the study outcomes, including the cost estimates.

8- Please delete the final sentence of the Abstract Methods and Findings section (“However, the setting is more likely to...”).

9- Please preface the Abstract Conclusions with: “These results suggest…” or similar.

10- At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary

11- Thank you for providing the STROBE checklist. Please replace the line numbers with paragraph numbers per section (e.g. "Methods, paragraph 1"), since the line numbers of the final published paper may be different from the page numbers in the current manuscript.

12- Please remove the registered trademark symbols throughout the manuscript.

13- S1 Table could be moved to the main text.

14- Please begin the Discussion with a brief summary of what was done, before getting to the findings and interpretations.

15- The current first sentence of the Discussion would be better stated: “The findings of this study demonstrate that there is considerable variability in the lifespan of pyrethroid-treated LLIN products.”

16- On line 317, perhaps “Our findings demonstrate that nets that are still in households…”

Comments from the reviewers:

Reviewer #1: Although at first glance, this study may not appear particularly novel, and the results do not yield any major surprises it is an important paper that raises some fundamental points that urgently need to be addressed by all involved with bednet manufacturer, procurement and distribution.

As the authors point out, long lasting insecticidal nets (LLINs) represent a huge market with billions having been distributed across Africa in the 21st century. To be classified as 'long lasting' these nets must last three years under operational use and the World Health Organisation have issued guidelines on how these nets should be evaluated. But such prospective studies are rarely conducted in practice and surprisingly little has been published on the effective life of LLINS, particularly considering the billions of dollars spent on these. This paper looks at this specific issue in a study across multiple districts in Tanzania. In this particular study they find that one brand of net in inferior compared to two further nets tested. This is an important result for Tanzania and arguably for other countries too; although some might argue that the results will be context specific, as the authors point out it seems reasonable to expect the relative performance to remain.

The paper is weakened by not testing any of the net samples on resistant mosquitoes; this point is acknowledged by the authors but it cannot be overlooked. The vast majority of Anopheles mosquitoes in Tanzania, and across Africa, will have some level of resistance to pyrethroid insecticides. When nets are initially distributed they may still have sufficient insecticide to kill both susceptible and resistant mosquitoes but as the amount of active ingredient (AI) decreases (as observed in this study where two of the products have lost a fifth or a third of the AI within 10 months of use) it is possible that resistant mosquitoes will survive a short exposure to these nets and hence the operational efficacy of the nets will be compromised. The failure to measure the bioefficacy against resistant mosquitoes is a major limitation, particularly given it may have altered the key conclusions of the paper.

I believe the paper could be strengthened in the following ways

1. It is essential to provide greater clarity how these nets were originally classified as LLINs. In the introduction (lines 27-29) the process is briefly described but, to evaluate the results of this manuscript, it is important that readers are aware of the specific evaluations that these three net products have undergone previously. This is important to understand if the current results are an anomaly or agree with previous studies. In the discussion this is mentioned again (line 275-277) but again no information on the previous evidence warranting the WHO classification of LLIN is provided. I would like to see a supplementary file, providing a summary of the durability data used by WHO to evaluate the three net types included in the current study, plus any subsequent published data on the durability of these nets (the authors don't even mention whether these results agree with their own retroactive study in Tanzania (Ref 26)). Furthermore, a fuller discussion of the extent to which WHO PQ are following these guidelines in their listings of new nets (particularly the new LLINs which contain more than one insecticide) must be included in the discussion. I agree with the authors conclusions that durability is too often overlooked but the extent to which this is happening, and the potential consequences of this could be more clearly stated.

2. If I remember correctly the weave design of Olyset changed at some point in the last six years; given the poor fabric integrity for Olyset reported it is critical to clarify if nets used here are those currently on the market.

3. Methods: The sampling strategy requires clarification

a. Were the three nets sampled per household pre-selected before entering (as opposed to convenience sampling on arrival)

b. When nets were sampled for bioeffiacy assays, was the net replaced? If so was the replacement clearly labelled to avoid being sampled in subsequent rounds?

4. Figure 2 legend should make clear what the denominator is (nets remaining in household at time of survey?) and indicate sample size for each row. This figure gives a somewhat misleading picture as it ignores nets that have been removed from households, many of which might be too torn. An additional (or replacement) figure including nets lost to follow up would be informative.

5. Figure 3 requires sample sizes. Please also change y scale from 0-100 as the current graph presents a misleading picture on the difference between the nets.

6. The results section 'Active ingredient content' is confusing for non-chemists. What is the biological significance of the R-alpha content? I assume this is not bioactive? The very high rate of loss of AI in two of the nets is not discussed further. This does not appear to relate to bioefficacy against susceptible mosquitoes but as the study did not test resistant strains, it is not possible to state whether this loss of insecticide would affect their operational efficacy.

7. Discussion, line 282, states that attrition and physical integrity are the two factors that define functional survival of LLINs - again I would query whether this is the case if you considered pyrethroid resistance. Churcher et al, e-Life 2018 have shown how net durability is influenced by resistance. This need acknowledging

8. There was a national distribution of LLINs in Tanzania from 2015-2017, was the current study area included and, if so, how did this impact on the current study?

9. Discussion, line 253, states that median functional lifespan was closer to 2 years. However Table 2 states the median life span to be 2.0, 2.5 and 2.6 for the three nets - is this really closer to 2 years? A more accurate statement could be obtained if the lifespan of the pooled analysis was presented. If it is closer to 2.5 years rather than 2 years, state this. Six months is important in this context!

10. The analysis of the IACT tests mentions adjustment for physical condition. Please include further explanation of how this was performed in the data analysis section.

11. Discussion, line 299 'most damage to the nets is accumulated on the bottom section of the nets….'. Is this from this study? I couldn't find any information on the location of holes from the supplementary data tables. If not from this study a reference must be given.

Reviewer #2: Statistical review

This paper reports a randomised study comparing the longevity and laboratory-assessed efficacy of mosquito nets. The authors show there are significant differences between the different types in both of these characteristics as assessed by different outcomes.

I had some comments on the statistical methods and reporting, which I have provided below.

1. Title: I was not sure why the title refers to this as a prospective cohort study, but the rest of the paper refers to it as a RCT - I would recommend this is kept consistent.

2. Abstract and results - I think it would be useful if confidence interval in differences of net functional survival was given between groups together with CIs and p-values; currently unclear (at least in the abstract) if the differences here are significant or not.

3. I did not find it completely clear from the methods section which outcomes were assessed by which data source (i.e. the inspection of the house or the lab tests on the restricted number). From what I understand, attrition, physical integrity and functional survival come from the household inspections and biological efficacy/insecticide content from the lab tests - please clarify if this is wrong.

4. Line 110 "adjusted for control mortality" - I didn't follow what exactly is being adjusted for here.

5. Methods - how was missing data handled in the analysis?

6. Statistical/economic analysis methods - was a pre-specified analysis plan developed for this analysis? It should be provided if so.

7. Page 10 line 159: is the p-value here testing if there's any difference between the three types of net? If not, could it be clarified what the p-value is testing?

8. Table 3 - is there any significant difference between permanet and netprotect here?

9. Table 4 - similarly here, it doesn't appear that permanet or netprotect would be significantly different?

James Wason

Reviewer #3: The manuscript presented by Lorenz et al. on the comparative functional survival and equivalent annual cost of three long-lasting insecticidal nets (LLIN) is clearly the best systematic assessment of LLIN durability to-date. The issue of how long LLINs last in the field is rapidly becoming the top priority topic for LLIN programmess, allongside insecticide resistance. Given that public entities spend in excess of USD 250 millions per year on LLIN deployment, and given that this represents our primary malaria prevention measure in most malaria-endemic settings, the importance of any major factor affecting the performance of LLINs can not be over-emphasized.

The authors present here a well designed and well powered study comparing directly and in a randomized design the three brands of LLINs which largely dominate the market. These results are eagerly awaited by the malaria community and it is important to publish them rapidly. In addition to these important results, this manuscript also presents a number of new approaches to quantifying the issue of uselife of nets (or validates their use on a large scale), as well as a methodologically appropriate way to estimate the cost versus the expected effects of a product on the basis of its field characteristics. As a result of the above, I consider this a manuscript of high public health importance.

The methodology of this work is largely sound and well implemented, and the results are well analyzed and described. The article is also well written and nearly all required information is there. I have therefore only one issue for which clarification is required.

1. Economic analysis (p 9 and 11). The authors have done the economic analysis on the basis of a uniform net price of USD 3. They should better justify their decision to take this price value for two reasons: (1) most factory prices of nets are lower than that amount these days, and (2) This price does not include the distribution costs as well as all the non-commodity costs associated with net distribution, such as Behaviour Change components, etc. The distribution costs (planning, census, transport, etc.) are particularly important to consider in this work because a net lasting longer would allow to lengthen the time interval between distributions, and that part of the total LLIN-in-the-field cost would be reduced proportionally. I appreciate that the authors have probably not taken this into account in their calculations, but they should at least discuss this point and outline what the wider implications are of an expended durability of LLINs in the field.

Finally, the limitations of the study are well described, once the points above are included.

Any attachments provided with reviews can be seen via the following link:

[LINK]

22 Apr 2020

Submitted filename: PMEDICINE-D-19-02821 response to reviewers_.docx

Click here for additional data file.

18 Jun 2020

Dear Dr. Moore,

Thank you very much for re-submitting your manuscript "Comparative functional survival and equivalent annual cost of three Long Lasting Insecticidal Net (LLIN) products in Tanzania: a three-year prospective cohort study" (PMEDICINE-D-19-02821R1) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by xxx reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

Our publications team (plosmedicine@plos.org) will be in touch shortly about the production requirements for your paper, and the link and deadline for resubmission. DO NOT RESUBMIT BEFORE YOU'VE RECEIVED THE PRODUCTION REQUIREMENTS.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Jun 25 2020 11:59PM.

Sincerely,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

Comments from Reviewers:

Reviewer #1: The authors have made the requested additions/clarifications to the manuscript.

Reviewer #2: Thank you to the authors for addressing my previous points.

My only remaining minor point is related to my previous point 1: there are still references that make the study appear to be a randomised trial: the first line of the methods and findings of the abstract, together with other references to randomised allocation of the nets, together with descriptions of the study as a 'trial' in several places. I'm not sure why it is described to be an observational study when assignment is randomised. Perhaps a line somewhere in the methods would be useful to explain.

Any attachments provided with reviews can be seen via the following link:

[LINK]

14 Aug 2020

Submitted filename: PMEDICINE-D-19-02821 response to reviewers_.docx

Click here for additional data file.

17 Aug 2020

Dear Dr. Moore,

On behalf of my colleagues and the academic editor, Dr. Elizabeth Ashley, I am delighted to inform you that your manuscript entitled "Comparative functional survival and equivalent annual cost of three long lasting insecticidal net (LLIN) products in Tanzania:  a randomised trial with three-year follow up" (PMEDICINE-D-19-02821R2) has been accepted for publication in PLOS Medicine.

PRODUCTION PROCESS

Before publication you will see the copyedited word document (in around 1-2 weeks from now) and a PDF galley proof shortly after that. The copyeditor will be in touch shortly before sending you the copyedited Word document. We will make some revisions at the copyediting stage to conform to our general style, and for clarification. When you receive this version you should check and revise it very carefully, including figures, tables, references, and supporting information, because corrections at the next stage (proofs) will be strictly limited to (1) errors in author names or affiliations, (2) errors of scientific fact that would cause misunderstandings to readers, and (3) printer's (introduced) errors.

If you are likely to be away when either this document or the proof is sent, please ensure we have contact information of a second person, as we will need you to respond quickly at each point.

PRESS

A selection of our articles each week are press released by the journal. You will be contacted nearer the time if we are press releasing your article in order to approve the content and check the contact information for journalists is correct. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact.

PROFILE INFORMATION

Now that your manuscript has been accepted, please log into EM and update your profile. Go to https://www.editorialmanager.com/pmedicine, log in, and click on the "Update My Information" link at the top of the page. Please update your user information to ensure an efficient production and billing process.

Thank you again for submitting the manuscript to PLOS Medicine. We look forward to publishing it.

Best wishes,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org