Implementing a community vector collection strategy for monitoring vector-borne diseases in Ghana

Background: Monitoring vector-borne diseases requires sampling of very large numbers of disease vectors in order to corroborate infections in the human population. This can be challenging, as current vector collection tools are either inefficient, or expensive to implement from a public health perspective. To circumvent this challenge, this study compared a community vector collection strategy using a double-netted mosquito collection method (a tent trap (TT)) to the traditional human landing collection (HLC) method in three communities in lymphatic filariasis-endemic districts in Ghana. Methods: Following community entry and sensitization, community volunteers appointed by the community leaders were trained in the mosquito collection and storage methods and provided with supplies for mosquito collection over a 7-month period. They were visited occasionally by the study team to retrieve the mosquito samples for identification. The collectors were also assessed to evaluate their perspectives on using community vector collectors for monitoring vector-borne diseases. Results: The results of the study indicated that the TT method collected significantly more mosquitoes (63%) over the collection period than HLC (37%). Thus, the TTs were observed to be performing relatively better than the HLC (P<0.001). The collectors knew the importance of mosquitoes in transmitting diseases, could identify the main diseases that were locally transmitted within their communities. They appreciated the involvement of the community in the collection as this enhanced community ownership of the programme as well as providing some financial incentives to those directly involved in the collection. Conclusions: The study revealed that use of community volunteers for the collection of mosquitoes for xenomonitoring purposes can be a viable strategy in the monitoring of vector-borne diseases. However, further development of the strategies and assessments of the costs involved will be required to make this a sustainable approach to monitoring vector-borne disease interventions and enhance community ownership of the programmes.


Introduction
Vector-borne diseases remain an important threat to the health of the human population.These are diseases caused by pathogens in humans, and transmitted by mosquitoes and other arthropod organisms.It is estimated that more than one billion people get infected with vector-borne diseases every year, with more than one million deaths 1 , and many more suffering from permanent disability and morbidity.Among the most important vector-borne diseases in the world are malaria, onchocerciasis, lymphatic filariasis, dengue, yellow fever, leishmaniasis, chikungunya, West Nile virus, Chagas and more recently emerging infections such as Zika virus 2 .
Of all arthropods that transmit diseases, mosquitoes remain the best-known vectors, transmitting diseases such as malaria and lymphatic filariasis, two of the best-known diseases in Africa.Malaria for instance leads to an estimated 210 million cases and 430,000 deaths annually 3 , while lymphatic filariasis is responsible for an estimated 36.45 million infections, 19.43 million hydrocele cases and 16.68 million lymphedema cases 4 .While significant efforts and achievements have been made in the control of such diseases, through vector control such as indoor residual sprays (IRS), use of long-lasting insecticide nets (LLINs) 5,6 and treatment of endemic populations 7,8 , there is a need to assess the effectiveness of these interventions.
The most direct timely measure of transmission (or the lack thereof) is through the examination of vectors for the presence of infective stages of the parasites responsible for the infection.To this effect, xenomonitoring (the examination of disease vectors for pathogens) serves as an important assessment tool 9,10 .However, in order to document that transmission has been interrupted, it is necessary to screen large numbers of insects to ensure that the prevalence of infective stages of disease pathogens has reached a relatively low threshold, below which the parasite population cannot be sustained.While pool-screen PCR methods have been developed that can efficiently screen large numbers of vectors 11 , the challenge of collecting large numbers of vectors remains.The most common method of collecting large number of mosquitoes is the human landing method, which is ethically questionable due to the exposure of the collectors to infections 12,13 .As a result, modifications of this approach to involve a doublenetted human baiting method have been designed.
A further challenge to the collection of large number of mosquitoes is the use of trained entomologists, which significantly increases the cost of the surveys 14 .To circumvent this challenge, this study opted for the use of community volunteers in the collection of the mosquitoes [14][15][16] .Operation of the traps by community members themselves has many advantages.First, this would be much less expensive than if trained individuals were required to operate them, and second, this would permit the traps to be more widely distributed and to be operated for longer periods in each community than would be possible if trap operation were confined to traveling teams of trained individuals.Such a community-based operation of the traps has the potential of providing a much more comprehensive estimate of the intensity of transmission, both over time and space, than can be obtained using teams of trained individuals.The usefulness of community vector collection approaches has been demonstrated in Mexico 17 , and Togo 18 .This study therefore sought to evaluate a community-based vector collection approach and to assess the performance of the double-netted human baiting method, the tent traps (TT) for the Anopheles vectors of lymphatic filariasis in Ghana.

Selection and sensitization of communities
The study communities were Akonu, Asemko and Aketenkyi in the Nzema East, Ellembele and Ahanta West Districts, respectively.All study communities are located in the Western region of Ghana, and the selected districts are lymphatic filariasis endemic.The Districts are approximately 300 km west of Accra, the capital of Ghana.The chiefs, opinion leaders and communities were informed on the purpose of the study, using the information document developed for the study 19 .

Recruitment and training of community vector collectors
To enhance community ownership, the chiefs and elders were asked to identify individuals within the community who will serve as vector collectors.Thus, no strict inclusion or exclusion criteria were applied, so long as the collectors were deemed capable by the community.The vector collectors were either male or female, with formal or informal education.No prior knowledge in mosquitoes or mosquito collection was required from the collectors.The vector collectors were trained on setting up the TT (Figure 1), and collecting the mosquitoes using aspirators.Two traps were provided for each community, with two vector collectors per trap.They were also trained on mosquito collection using the HLC method.

Mosquito collection
The mosquitoes were collected by the volunteers, without supervision from the study team.Collections were done from June to December 2017.Two TT and two HLC collections were carried out every month at a time suitable to the collectors, but at two weeks interval.On each collection day, the mosquitoes collected for each method were stored in a 50-ml falcon tube, provided to the collectors and labelled with all the information including the collection method and month of collection.
Once every two months, a member of the study team visited the communities to collect the mosquitoes to the laboratory at the Noguchi Memorial Institute for Medical Research, Accra where they were identified, counted and stored.

Evaluation of community xenomonitoring study: perspectives of collectors
The study also assessed the views of the community vector collectors, using a simple questionnaire designed for this purpose 19 .This was done to enhance future operational activities.
Following community sensitization, verbal approval was obtained from the chiefs and elders of the communities.Written consent was obtained from the vector collectors.

Statistical analysis
The number of mosquitoes collected was represented through graphs drawn in Microsoft Excel.The difference between the number of mosquitoes collected using the TT and HLC methods was assessed using the chi square test in IBM SPSS Statistics 20.0.Statistical significance was assessed at a p value ≤0.05.

Results of mosquito collection
A total of 3363 mosquitoes (Table 1) were collected from June to December 2017 comprising; 1581 (47%) An. gambiae, 1266 (37.6%)Culex spp., 6 (0.2%) Aedes sp. and 510 (15.2%) Mansonia sp.A significantly (P<0.001) higher proportion of mosquitoes were collected using the TT (63%) compared to the HLC (37%).The average monthly collections also revealed higher collections by the TT compared to the HCL, with significant differences observed during the dry season months of August, September, November and December (Figure 2).Generally, the variation (range) in mosquito numbers collected each month was relatively smaller for the HLC (apart from June and October) compared with the TT (Figure 2).
The general trend in the total number of mosquitoes collected reflects the influence of seasons on the behavior and the population of mosquito species, with the Anophelines and Culicines revealing opposing trends.The main collection peak for An.gambiae was from June to August, which coincides with the major rains and the beginning of the dry season.On the other hand, the main peak for Culex was during the minor rains and dry season months of October to December (Figure 3).A similar trend was observed in mosquitoes collected using the TT and HLC (Figure 4).
Among the three communities where collection was done, the highest number of mosquitoes was collected from Aketenkyi and the least from Akonu (Figure 5).In Akonu though the number of mosquitoes collected from TT was generally higher than those from the HLC between June and September, the reverse was the case between October and December.Asemko showed higher numbers in TT than HLC in almost all the months apart from June and July, whiles Aketenkyi showed similar pattern except from June where HLC numbers were higher.

Evaluation of community xenomonitoring study: perspectives of collectors
A summary of the demographic information of mosquito collectors and responses to questionnaire is shown in Table 2.There were 12 collectors involved in the study, three of whom were females.The age of the collectors ranged from 16 to 58 years with a median age of 34 years.Seven of the collectors had received primary/JHS education, 3 secondary education and the remaining 2 post-secondary education.Four of the collectors were farmers, 3 unemployed, 2 seamstresses, 2 students and 1 teacher.
The assessment showed that the mosquito collectors had a fairly good idea about the importance of mosquitoes in disease transmission especially malaria and lymphatic filariasis.This is exemplified by some of the responses below:

"They bite and bring diseases like malaria & elephantiasis, kill pregnant women and children cause us to spend a lot" "To investigate the disease they transmit" "They suck blood and contain parasites in them to give diseases" "They are strange animals that bite and cause diseases such as high fever, rashes & elephantiasis"
Generally, the volunteers were aware of the main diseases transmitted by mosquitoes in their community.They all identified malaria as a disease transmitted by mosquitoes.In addition, others mentioned elephantiasis, with two individuals mentioning cholera and tuberculosis.On the importance/benefits of using community     members for mosquito collection, the main views expressed were with the speed of collection, the willingness of the community and individuals to be involved in the study and feel a part of the disease elimination process, and the opportunity costs to the funders and collectors.These are reflected in the views below: "When we are involved, we believe the results, it also helps to get some income and we feel part of eliminating the disease" "We will help the GHS reduce cost of travelling to community and also get some income" "We are able to collect fast.The team from Accra sometimes come late and we cannot do much….The people also know us so they allow us into their houses" "So, we can help ourselves to drive away mosquitoes to prevent diseases"

"Saves time and cost, no tension in sample collection, shows commitment level of community"
When asked which of the mosquito collection methods they preferred, 10/12 volunteers stated that they preferred the HLC.
The reasons given for their preference were; the number of mosquitoes collected using the HLC, and the speed and ease of collection.The preference to the TT was due to the absence of mosquito bites.All the volunteers responded in the affirmative to the question of using the community to collect mosquitoes for the GHS as part of disease monitoring activities.However, 11/12 volunteers indicated that payment for the collection must come from the government or the Ghana Health Service.When asked whether they will consider collecting mosquitoes for the community without payment, 10/12 volunteers said "yes".The main reasons given were to help protect the community, themselves and their families from infection.However, two of the collectors said they will not collect the mosquitoes without payment.Below are some of the views expressed:

"Disease is destroying our community & I need to help; I have been doing volunteer work already for NGOs"
"To help the community eliminate the disease and protect my children" "I could get the disease and the cost will be more to me, also I will receive blessings" "It is volunteer work and that is my contribution to the community" "Already involved in a lot of volunteer work without pay" The volunteers mentioned the need for more logistics such as consumables and better torch lights, malaria prevention using anti-malaria prophylaxis, improved methods for mosquito collection and enhance cohesion between the collection teams, as some of the ways of improving the mosquito collection and the work of the collectors.Complete answers to the questionnaire are available on OSF 19 .

Discussion
The surveillance of vector-borne diseases requires an assessment of the infection in the human population, whiles at the same time undertaking entomological surveillance to detect the infection in the vectors of the disease.Such is the plan for the post-intervention surveillance phase of lymphatic filariasis 20 .However, conducting entomological surveys especially in the context of elimination activities is considered expensive due to the limited resources 21 and the high number of samples that may be required to assess transmission of the disease after interventions as well as the cost involved in the collection and processing of the mosquitoes 14,16 .Thus, there is the need for more cost-effective strategies for entomological assessments of infections within vectors of diseases.
In this study, the TTs were observed to be performing relatively better than the HLC (P<0.001).This study showed that the TTs collected higher number of mosquitoes compared to HLC and may be a better tool for the collection of a larger number of mosquitoes for xenomonitoring surveys, even though a majority of the volunteers thought the HLC was better.
Mosquitoes are often observed to be more abundant during the rainy season compared to the dry season 22 .This is due to the presence of more breeding sites created during the rainy season.In this study this was not the case from the collections as there was no significant differences in the numbers collected in both seasons.On the contrary, the data showed more mosquitoes collected during the dry season.This observation may be attributed to disruptions during rainy days.In fact, the collectors complained about their inability to set up the traps for long periods during the rainy season.
The seasonal changes observed in the distribution pattern of the mosquito species can be explained by their habitat preferences 23 .The numbers of An. gambiae during the rainy season was higher than the other species but low during the dry season.This is not surprising because An. gambiae prefer breeding in shallow and fresh water which are often abundant during the raining season 24,25 .Culex spp. on the other hand was higher during the dry season when fresh water is less abundant (Figure 5).This is due to its ability to breed in polluted breeding sources 26,27 , which are prevalent in Akonu and Asemko.
The assessment of the volunteers' perspectives served to assess the level of understanding and importance of the task given to them as well as the feasibility of using such approach for future national programme implementation activities.The assessment revealed that generally, the collectors knew the importance of mosquitoes in transmitting diseases and were able to identify the main diseases that were locally transmitted within their communities.They appreciated the involvement of the community in the activities as this enhanced community ownership of the programme as well as providing some financial incentives to those directly involved in the collection (Table 2).For example, the community selection of ivermectin drug distributors as part of the Community Directed Treatment with ivermectin, improved access to the drugs and enhanced the community ownership of the activities 28,29 .The use of the community volunteers also provides opportunity costs to both the programme and the communities involved.To the national programmes, there are fewer expenses on the use of trained entomologists for vector collection and subsequent reduction in transportation and logistics costs.To the community volunteers, this provides some opportunities for them to earn some allowances, even though the work is entirely voluntary.However, the motivation of community volunteers is an important factor that determines the success of programme activities 30,31 .Recent use of community volunteers for large-scale monitoring activities have suggested that it is a useful strategy for programme implementation 18,32 .However, this needs to be tailored to country-specific situations.
In conclusion, the study revealed that use of community volunteers for the collection of mosquitoes as part of xenomonitoring purposes, can be an important strategy in the undertaking of monitoring of vector-borne diseases.However, further development of the strategies and assessments of the costs involved will be required in order to make this a public health approach to monitoring vector-borne disease interventions, and enhance community ownership and sustainability of the programmes.Though the use of community members for mosquito collection is promising, there is the need for further community education and training of community volunteers for xenomonitoring purposes.
The project contains the following extended data: • Collectors' questionnaire.docx(the questionnaire administered to the mosquito collectors).
• Information document for mosquito collection using the tent trap.docx.
The data is available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Introduction:
The authors speak about vector-borne diseases in general, but it seems they are thinking mostly about lymphatic filariasis.For example, the statement about a prevalence of the pathogen which "has reached a relatively low threshold, below which the parasite population cannot be sustained", is hard to imagine for a disease such as malaria.
The authors state that the human landing catch is ethically questionable and I think there are lots of arguments for this position, however, to support this statement, they cite a study that shows significantly less malaria detected in HLC collectors than in the general population.I would think there would be better references to cite unless I have missed the intention of the authors.

Methods:
The mosquito collection methods are inadequately described, making replication of this work impossible.What were the starting and ending times of the collections?Were they done indoors or outdoors?Were the HLC collections made in the same locations as the tent trap collections?Did the same collectors conduct both HLC and TT? Results: In the "Results of mosquito collection" section, second sentence: "proportion" is mentioned, but then percentages are given.Please be consistent between the text and the values provided.
In the following paragraph, please italicize Culex.When people stated their preference for HLC, one of the reasons was "the number of mosquitoes collected using the HLC" -can you explain this further?Did people prefer collecting fewer mosquitoes?Discussion: The authors are discussing mosquitoes in general, but disease transmission depends on the species collected.For me, one of the big questions I would have for this paper is whether the two methods were significantly different in the number of collected.

An. gambiae
The rain was discussed, but it is not clear how the rain delays were dealt with.Was trapping cancelled every time there was rain, or was there a certain amount of delay that was acceptable?
The authors make a statement about the ability of all Culex species to breed in polluted breeding sources.I assume they are really talking about , and if this is the case, they Culex quinquefasciatus should state it clearly.

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: Medical entomology I have read this submission.I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
In the following paragraph, please italicize Culex.

Comment:
This has been addressed accordingly.

Response:
Table 1: I think the last "subtotal" column should be bolded.Also, the numbers Comment: collected by the two methods don't look sufficient for the collection of large numbers needed for xenomonitoring.Finally, was the only Anopheles species collected?If not, Anopheles gambiae please note that the other Anopheles collected are not presented.
The last "subtotal" column has been bolded.This study looks at whether communities Response: can implement vector collection properly by themselves, and without supervision.Given that only two collectors were used, at each collection time for each method, we consider the numbers collected large enough.Thus, if the number of collectors and frequency of collection was increased, much more mosquitoes could be collected for xenomonitoring purposes, based on the same costs that would otherwise be used in paying experienced entomologists.
was the only species identified in this study.This has been stated An. gambiae s.l.Anopheles clearly in the results.See line 168.When people stated their preference for HLC, one of the reasons was "the number of Comment: mosquitoes collected using the HLC" -can you explain this further?Did people prefer collecting fewer mosquitoes?
When people stated their preference for HLC, one of the reasons was "the number of Response: mosquitoes collected using the HLC".This could be linked to the high human-mosquito contact, and thus the perception of high mosquito abundance.However, the results show that the TT collected more mosquitoes than the HLC.We have addressed this in the discussion.

Discussion:
The authors are discussing mosquitoes in general, but disease transmission depends Comment: on the species collected.For me, one of the big questions I would have for this paper is whether the two methods were significantly different in the number of collected.

An. gambiae
The number of collected from TT was higher than HLC but this was not Response: An. gambiae statistically significant ( = 0.213).However, this study shows that the TT can be an alternative to P the HLC, while helping reduce the likelihood of infection presented using the HLC.We have added this point to the text.
The rain was discussed, but it is not clear how the rain delays were dealt with.Was Comment: trapping cancelled every time there was rain, or was there a certain amount of delay that was acceptable?
Trapping for both HLC and TT were cancelled anytime there was rain.These were not Response: recorded too many times during the study.Therefore numbers obtained from the two methods can be compared.This has been addressed in the methods.
The authors make a statement about the ability of all Culex species to breed in polluted Comment:

Figure 1 .
Figure 1.Setting the tent traps.(A) The inner sleeping tent.(B) The outer tent.

Figure 2 .
Figure 2. Average number of mosquitoes collected per month.Error bars represent the minimum and maximum number collected.HLC, human landing collection.

Figure 3 .
Figure 3.Total number of mosquitoes and the species abundance per month.

Figure 4 .
Figure 4. Trends showing species composition of mosquitoes collected using (A) tent trap and (B) human landing collection.
This work received financial support from the Coalition for Operational Research on Neglected Tropical Diseases (COR-NTD), which is funded at The Task Force for Global Health primarily by the Bill & Melinda Gates Foundation (OPP1053230), by the United Kingdom Department for International Development, and by the United States Agency for International Development through its Neglected Tropical Diseases Program.

Figure 3 :
Figure 3: Please include in the figure legend what collection method was used.

Figure 3 :
Figure 3: Please include in the figure legend what collection method was used.Comment: Figure 3: This figure was to show the number of mosquito species collected out of the Response: total per month.Including the collection method may be a bit challenging using the bar chart and line graph combination we have in the figure.As mentioned, this figure was to show the number of mosquito species collected out of the total per month.The Figure title has been changed to reflect this.The breakdown of captures by collection method is shown in Figure 5. Additionally, Figures 2 and 5 all give information about the collection method.Figure 3 has now been renumbered as Figure 4 in the new version.

Table 1 :
I think the last "subtotal" column should be bolded.Also, the numbers collected by the two

Table 1 :
I think the last "subtotal" column should be bolded.Also, the numbers collected by the two methods don't look sufficient for the collection of large numbers needed for xenomonitoring.Finally, was the only Anopheles species collected?If not, Anopheles gambiae please note that the other Anopheles collected are not presented.