At the time of writing, approximately 610 epidemiological reports and files dating from 1985 onwards were collated and included in the data repository. These reports contain information for over 20,000 HAT cases and approximately 7,000 geographical entities. Data from seventeen countries have been included in the repository so far: Angola, Benin, Burkina Faso, Cameroon, Central Africa Republic (CAR), Congo, Côte d'Ivoire, DRC, Equatorial Guinea, Gabon, Ghana, Guinea, Malawi, Mali, Sudan, Togo and Uganda.
For eleven countries, data processing has been initiated and data are being imported into the HAT database. High priority is presently given to the most recent datasets (i.e. reports dating from the year 2000 onwards).
For six central African countries, namely Cameroon, CAR, Chad, Congo, Equatorial Guinea, and Gabon, all data available for the period 2000–2008 were processed, thus leading to the first preliminary regional component of the HAT Atlas initiative. Figure shows the spatial distribution of 1,580 locations, corresponding to 92 percent of all locations of epidemiological interest included in the reports. For 829 of these locations (red dots), HAT cases were reported either by passive surveillance of through active case-finding surveys. Epidemiological data used as input for this map were collected by the NSSCPs of the six countries; for CAR and Congo, data collected by the NGO Médecins Sans Frontières (MSF) were also included.
Figure 5 Atlas of HAT: an example of regional-level map. Red dots represent locations from which HAT case were reported (either through active screening or by passive surveillance); white dots indicate screened locations where no HAT case was detected. Study countries: (more ...)
We note that a significant number of affected locations can be found in border areas, where trade and population displacements may contribute to creating conditions conducive to disease transmission.
Table provides country-level statistics on the results of the geo-referencing activity for this first regional output of the HAT Atlas initiative. Overall, the geo-referencing exercise was very successful. The locations of epidemiological interest that could be mapped as point entities (e.g. camps, hamlets, villages, towns, etc.) allowed to geo-reference 98 percent of HAT cases reported from the six study countries. In addition to villages reporting HAT cases, 751 screened locations that did not result in any positive case were also included in the DB (white dots). The importance of including these sites in the DB can not be overstated, as they play a crucial role in drawing and updating the overall epidemiological picture. In actual fact, knowing how many villages have been screened and their location is of critical importance in providing information on where there is or there is not active transmission. Thus, a database of screened but negative villages is as important as a database of screened and positive ones.
Table 2 Results of the geo-referencing activity in 6 central-African countries. The table shows how the geographic locations and HAT cases contained in epidemiological reports were geo-referenced with either reported coordinates, gazetteers, or other means (e.g. (more ...)
For most geographical entities (approximately 61 percent), geographic coordinates were either available in the epidemiological reports or provided through consultation with WHO partners; 22 percent of the locations were geo-referenced using the combination of reported names and gazetteers, and for 9 percent position was estimated from other sources (digital or paper maps, out-of-scale maps enclosed with the reports, etc.). The remaining geographical entities (8 percent, associated with 2 percent of the HAT cases) have not been mapped yet.
As an example, Figure shows the cumulative number of sleeping sickness cases recorded in the area of Nkayi, Bouenza Region, Congo, between 2002 and 2007. Locations where no cases were detected, as well as the boundaries of HAT transmission area as depicted by WHO in 1998 [17
] are also included.
Figure 6 Atlas of HAT: an example of local-level map. Cumulative number of sleeping sickness cases recorded in the area of Nkayi, Bouenza Region (Congo), between 2002 and 2007. Data sources: Programme National de Lutte contre la Trypanosomiase Humaine Africaine (more ...)
As opposed to Figure , this zoomed-in image allows to fully appreciate the unprecedented spatial detail of the HAT Atlas. We note that the picture drawn in this area by the HAT Atlas appears substantially different from the previous map of HAT transmission areas. This is true also for the study region as a whole. Less than a third of the endemic locations we mapped in the six central African countries are situated within the boundaries of previously described transmission areas. Even though the spatial distribution of transmission areas may have undergone some changes in the last decade, it is believed that most of the differences between the past and present representation are to be ascribed to improved methodology for mapping control activities, rather than to a substantial evolution of the epidemiological conditions on the ground. It is important to stress that these preliminary results will be systematically verified in collaboration with NSSCPs, with a view to consolidating and sharing the outcomes of the HAT Atlas initiative.