The methodology to develop the atlas of tsetse and AAT in Burkina Faso is broadly based on the FAO continental atlas [16,17,18], and a similar approach has already been used at the national level in Sudan [19], Mali [20], Kenya [21] and Zimbabwe [22]. The main difference between the continental and the national atlases is that the former is solely based on publicly available data from scientific journals, while the latter include all published or unpublished data collected in a given country.
The atlas in Burkina Faso covers a period of 30 years (1990–2019). Data were provided by all the institutions involved in tsetse and trypanosomosis research and control at the national level. In addition to IBD-CETT, these are the School of tsetse control (Ecole de Lutte Anti-Tsé-tsé—ELAT), the International Research and Development Centre on Livestock Farming in the Subhumid Zone (Centre International de Recherche-Développement sur l'Elevage en zone Subhumide—CIRDES), the French Agricultural Research Centre for International Development (Centre de coopération Internationale en Recherche Agronomique pour le Développement—CIRAD), the French Research Institute for Development (Institut de Recherche pour le Développement—IRD), the International Atomic Energy Agency (IAEA) and FAO.
Input data
Input data were collected either in the framework of tsetse and trypanosomosis control operations or in the context of research activities. Control operations were mostly carried out by PATTEC Burkina Faso by using impregnated targets and traps and animal treatment. Data included both baseline cross-sectional surveys (i.e. pre-intervention) and longitudinal surveys (i.e. during or post-intervention), and provided a large amount of unpublished information. The raw data from research activities were obtained from the authors of published articles, using as a basis the systematic review of the literature carried out by FAO for the continental atlas. The full list of publications that contributed to the national atlas in Burkina Faso is provided in Additional file 1: Text S1. When the raw data underpinning publications could not be retrieved, information on AAT and tsetse occurrence was extracted directly from the papers, as done for the continental atlas.
Tsetse data
In Burkina Faso, entomological surveys are mainly carried out with biconical traps [23], which may or may not be boosted with odour attractants [24]. However, other traps such as monoconical Vavoua trap [25] [26, 27], N’Zi traps, adhesive targets and electric targets were also occasionally used in experimental settings [28]. The trapping duration was usually 3 days for the surveys carried out in the framework of control activities, and 1 to 3 days for research activities. All traps were geo-referenced with GPS.
Entomological data were recorded at the level of individual traps. Data were initially captured in hard-copy recording sheets, which were subsequently converted into digital format (e.g. Microsoft Excel or Access). Field recording sheets include information such as the name of the surveyed location, its coordinates and administrative units (which in Burkina Faso are called regions, provinces and departments), the date of the survey, the time of trap deployment and removal (and the corresponding duration of trapping), and the number, species and sex of trapped tsetse flies. Complementary information (type of trap, use of odour attractants, etc.) is normally available in mission or project reports. For research activities linked to scientific publications, raw datasets at the trap level were normally obtained. When these were not available, information was mapped at the site level.
African animal trypanosomosis data
As for tsetse, AAT data were mainly collected in the framework of control activities, thus including both baseline and monitoring/longitudinal investigations. Supplementary data were generated in the context of research.
AAT surveys in Burkina Faso target the main susceptible animal species such as cattle, sheep, goats and donkeys, but data on horses and camels are also occasionally collected. In most surveys, the sampling locations and the animals are selected randomly. However, a non-random sampling was sometimes performed, for example in studies of trypanocidal drug resistance [29, 30], or for the isolation of trypanosomes. Cross-sectional surveys provided the majority of datasets included in the atlas, but longitudinal investigations were also included [14].
Field data recording sheets for AAT include the data source and the institution in charge, the name, geographical coordinates and administrative units pertaining to the survey village or site, the date of the survey and the sample size (i.e. the number of animals tested). The sheets also include parasitological results such as the number of AAT-positive animals by trypanosome species, the haematocrit (packed cell volume—PCV) and the possible use of trypanocidal drugs (including the type of drug).
AAT data are recorded at the level of individual animal in the parasitological recording sheets, and information at the animal level was retained in the atlas. When the animal level data were not available, aggregated information at the herd/site level was recorded.
Structure of the atlas
The atlas is composed of a structured data repository and two databases (tsetse distribution and AAT distribution). Data on tsetse infections are not yet included. Open-source PostgreSQL is used as the main database management system, together with its spatial extension PostGIS. Microsoft Access is used to facilitate a broader data utilization.
Data repository
The data repository includes digital copies of all the input files used to build the atlas, i.e. spreadsheets, databases, reports, thesis and scientific articles. At a first level, the repository is organized by type of data (i.e. tsetse or AAT), at a second level by geographical area (i.e. province), and last by individual data source or publication. File names for scientific publications are standardized as author name and year of publication (i.e. Surname_YEAR.pdf), and for unpublished works as title of activity and year.
Tsetse database
Entomological data recorded in the tsetse database are organized in six tables: 1) data sources; 2) geographical data; 3) entomological surveys (trap level); 4) tsetse catches (trap level); 5) entomological surveys (site level); and 6) tsetse catches (site level). As in all relational databases, records in each table are given a unique identifier, the ‘primary key’, which allows them to be linked to records in other tables through ‘foreign keys’.
The table on ‘data sources’ summarizes information on the input files stored in the repository. In particular, it records the following information: institution, author’s name, title, year of production/publication, presence of information on tsetse or AAT, availability of raw data and whether the source is published or not. The ‘geographical data’ table includes the location or village name, its geographical coordinates and the corresponding administrative units. The table on ‘entomological surveys’ (trap level) includes the code of the trap (as per survey recording sheet), the type of trap, the attractant used (if any), the geographical coordinates of the trapping site (latitude and longitude in decimal degrees on WGS84 datum), the trapping period and its duration. Information on the possible presence of tsetse control activities in the surveyed area is also recorded. The table on ‘tsetse catches’ (trap level) summarizes the results of the trapping and includes the tsetse species, sex and the number of flies caught.
The tables on ‘entomological surveys’ and ‘tsetse catches’ (site level) record data that were directly extracted from scientific publications when raw data at the trap level could not be obtained. These site-level tables of the national atlas for Burkina Faso have a similar structure to those of the continental atlas [17].
The detailed structure of the tsetse database, including the relationships between the tables, is provided in Additional file 2: Text S2.
African animal trypanosomosis database
The database for AAT contains seven tables: (1) data sources; (2) geographical data; (3) animal species; (4) diagnostic method; (5) AAT data (animal level); (6) AAT data (herd level); and (7) chemotherapy.
The tables ‘data sources’ and ‘geographical data’ are similar to those of the tsetse database. The table ‘animal species’ records the species and breed of tested animals, and the table ‘diagnostic method’ provides details on the diagnostic test used to detect AAT. The table ‘AAT data’ (animal level) includes the code of the animal (as per survey recording sheet), its age, sex, haematocrit/PCV, health status, treatment with trypanocidal drugs, date of last treatment, husbandry system, date of survey, presence or absence of trypanosomes, species of trypanosomes (including subspecies/groups), type of animal sampling (random or purposeful), type of survey (longitudinal or cross-sectional) and possible presence of vector control operations during the survey. The table ‘AAT data’ (herd level) is used to capture data extracted from scientific publications when the raw data at the animal level could not be obtained; as such, it follows the structure of the continental atlas [16]. The last table (chemotherapy) provides details on the possible treatment of animals with anti-trypanosomal drugs before the survey. It includes the date and cause of treatment, the drug used, the doses and the person who administered the treatment.
The detailed structure of the AAT database, including the relationships between the tables, is provided in Additional file 3: Text S3.
Development of the atlas
The first step in the development of the atlas was to collect input data from all partner institutions. Workshops and meetings were organized to facilitate the process. When input data could only be retrieved as hard copy, Microsoft Excel was used to enter data in digital format.
Merging data from a variety of sources into the atlas databases required a systematic process of verification and harmonization. Geographical coordinates were standardized as latitude and longitude (decimal degrees on WGS84 datum). When Universal Transverse Mercator (UTM) coordinates were reported in the input files, coordinate transformation was performed. Extensive harmonization was also needed for data on animal breeds, husbandry systems, names of geographical locations and the related administrative units. For tsetse data, the names of locations were often missing. For AAT data, a high proportion of epidemiological records lacked geographic coordinates; in these cases, the names of the survey sites were used to extract geographic coordinates from other sources (i.e. Geographical Institute of Burkina Faso, Google Earth and Google Maps). Also, input data sheets often lacked information on whether interventions against tsetse were taking place in the study area at the time of the survey, or whether trypanocidal drugs were being used. In these cases, reports and publications were often useful to fill the gaps. When raw data could not be retrieved, information was extracted directly from the papers; for example, the maps included in the publication were imported in the GIS and georeferenced to estimate the geographical coordinates of the study sites.