SUPPLEMENTARY INFORMATION - Nature · 1.2 Background to malaria parasite prevalence 1.3 Assembling malaria survey data 1.3.1 Archive searches ... states of Sao Tome and Principe,

RW Snow, B Sartorius, D Kyalo, J Maina, P Amratia, CW Mundia, P Bejon, AM Noor

The prevalence of Plasmodium falciparum in sub Saharan Africa since 1900

1 Parasite prevalence survey data assembly

1.1 Geographic scope 1.2 Background to malaria parasite prevalence 1.3 Assembling malaria survey data

1.3.1 Archive searches 1.3.2 Electronic data searches 1.3.3 National household surveys post 2005 1.3.4 Other data sources

1.4 Survey data abstraction 1.4.1 Minimum data requirements 1.4.2 Age standardization 1.4.3 Survey location geo-coding

1.5 Data summaries

2 Administrative boundaries and the changing margins of Plasmodium falciparum

2.1 Defining national and sub-national boundaries 2.2 The natural maximal extent of malaria 2.3 Changing malaria risk extents and margins malaria since 1950

3: Statistical methods

3.1 Hierarchical space–time model for P. falciparum infection prevalence in children aged 2-10 years (PfPR2-10) from 1900-2015

3.2 Validation of model outputs

4. References

5. Acknowledgements

5.1 General archive assistance, data extraction and geo-coding 5.2 Regional research institutes and national malaria control programmes 5.3 Country level support

WWW.NATURE.COM/NATURE | 1

SUPPLEMENTARY INFORMATIONdoi:10.1038/nature24059

1 Parasite prevalence survey data assembly

1.1 Geographic scope

The analysis focuses on Africa south of the Sahara and excludes the small offshore Island states of Sao Tome and Principe, Cape Verde, Annobin and Bioko islands of Equatorial Guinea, Zanzibar and Pemba of the United Republic of Tanzania and the archipelagos of Comoros, Mauritius, Reunion and Mayotte. The large island of Madagascar has been included.

1.2 Background to malaria parasite prevalence

The parasite rate has been used to document the intensity of P. falciparum transmission for over 100 years in Africa1-4. This vast survey effort has not been previously completely assembled and geo-coded. Here we present an assembly of these data, summarised spatially and temporally for every country in Africa where P. falciparum has been endemic, and released in the public domain as a legacy dataset.

1.3 Assembling malaria survey data

Identifying infection prevalence survey data began in 1996 and ended on 31st December 2016. The basic principles of searching for information on malaria infection prevalence in Africa were established under the first Pan-African collaboration to assemble and archive malariometric data, Mapping Malaria Risk in Africa/Atlas du Risqué de la Malaria en Afrique (MARA/ARMA)5-7. In 2005, the Malaria Atlas Project (MAP) was established in Nairobi, Kenya to update information started by MARA and extend searches to provide a global data repository8,9.

Methods to identify sources of information have been opportunistic, cascaded approaches and did not adhere to strict methods proposed for systematic reviews or meta-analysis. Search strategies have included personal contacts, references to surveys in ministry of health reports, searches of archives and more traditional peer-reviewed publication searches.

1.3.1 Archive searches: Since 1996, manual searches have been undertaken at the archives and libraries of pre-independence tropical medicine institutes to locate unpublished reports from malariologists working in Africa. These included archives at the Institute of Tropical Medicine, Antwerp; Institute Pasteur, Paris; Department of History of Medicine, Sapienza - Università di Roma, Rome; Archivio Italiano di Scienze Mediche Coloniali, Rome; Instituto Higiene Medicina Tropical, Lisbon; The Wellcome Trust Library, London; The National Archives, Kew, UK; the London School of Hygiene and Tropical Medicine, UK and the Liverpool School of Tropical Medicine, UK. In addition, the archives of the World Health Organization (WHO) in Geneva, Brazzaville and Cairo were visited. Of particular note have been consultants’ trip reports and quarterly reports from malariologists working on behalf of the WHO from the 1950s through to the 1970s. These provide rich narratives and survey


SUPPLEMENTARY INFORMATIONRESEARCHdoi:10.1038/nature24059

data from pre-elimination campaigns, elimination progress reports and national malaria situation analyses.

National archives of the Ministry of Health offices were visited at Nairobi, Kisumu, Eldoret, Mombasa and Meru (Kenya), Entebbe and Jinja (Uganda), Sennar, Khartoum and Kassala (Sudan), Thiès (Senegal), Dar es Salaam and Amani (Tanzania), Accra (Ghana), Tzaneen (South Africa), Bobo-Dioulasso (Burkina Faso), Lagos (Nigeria), Freetown (Sierra Leone), Institute Pasteur, Antananarivo (Madagascar) and the personal archives of the ex-director of Tzaneen Malaria Research centre (covering South Africa, Namibia and Botswana). Records and reports held at the National Institute of Medical Research in Amani covered the period when the centre was the sub-regional, East African Institute of Malaria and Vector Borne Diseases (1954-1977). The Uganda eradication headquarters at Jinja, housed national household survey records from the 1960s, some had been re-located to Entebbe and were reviewed, those left at Jinja were unfortunately destroyed in 2011.

Annual medical and sanitation department reports from 1919 produced by the Colonial governments of The Gambia, Nigeria, Gold Coast (Ghana), Sierra Leone, Somaliland, Kenya, Tanganyika (Tanzania), Uganda, Belgian Congo (Democratic Republic of Congo), Bechuanaland (Botswana), Nyasaland (Malawi), Rhodesia (Zambia and Zimbabwe), Mauritius and Sudan were available in variable states of preservation, in the library founded by the Wellcome Trust in 1963 and now part of the National Public Health Laboratories of the Ministry of Health, Nairobi, Kenya.

1.3.2 Electronic data searches: Online electronic databases were used as one means of identifying peer-reviewed, published data on malaria infection prevalence, most notably those published since the 1980s, including: PubMed, Google Scholar, the Armed Forces Pest Management Board – Literature Retrieval System, the World Health Organization Library Database, and the Institute de Recherché pour le Development on-line digital library service. Regional journals, including the large number of national medical, public health and parasitological journals, were searched using African Journals Online. In all digital electronic database searches for published work the free text keywords "malaria" and "country-name" were used. We avoided using specialised Medical Subject Headings (MeSH) terms in digital archive searches to ensure as wide as possible search inclusion.

Titles and abstracts were used to identify possible parasite cross-sectional survey data in a variety of forms: either as community surveys, school surveys, other parasite screening methods or intervention trials. We also investigated studies of the prevalence of conditions associated with malaria when presented as part of investigations of anaemia, haemoglobinopathies, blood transfusion or nutritional status to identify coincidental reporting of malaria prevalence. In addition, it was common practice during early anti-malarial drug sensitivity protocols to screen community members or school attendees to recruit infected individuals into post-treatment follow-up surveys. Surveys of febrile populations, those attending clinics or within special groups, for example HIV positive individuals, were excluded.



Publications with titles or abstracts suggestive of possible parasite data were either downloaded from journal archives where these have been made Open Access (OA) or sourced from HINARI10. If publications were not available OA from HINARI we visited UK library archives at the London School of Hygiene and Tropical Medicine, the Liverpool School of Tropical Medicine, the Bodleian library at the University of Oxford or the Library at the Institute Pasteur, Paris to obtain copies.

All publications from which data were extracted were cross-referenced using the bibliographies for additional sources that may have been missed or that may correspond to unpublished or ‘grey’ literature, not controlled by commercial publishers. Authors of peer-reviewed papers were often contacted to seek additional information and directions to other possible unpublished work in their geographic area or from their institution.

1.3.3 National household surveys post 2005: In 2005, there was recognition that it was important to resurrect national household surveys of infection prevalence to monitor country-level progress11. The three principal survey vehicles for contemporary national parasite prevalence data have been the Demographic and Health Survey (DHS) malaria modules, managed by the US based DHS programme12, Multiple Indicator Cluster Surveys, managed by UNICEF13 and standalone Malaria Indicator Surveys (MIS) managed by national malaria control programmes and their survey partners. Data from national household sample surveys since 2005 were either available in downloadable formats from the ICF Measure programme12 or have been provided by national malaria control programmes.

1.3.4 Other data sources: For other possible unpublished, site-specific data on malaria prevalence web-sites of Non-Governmental Organizations working across Africa who may have undertaken health assessments were used as initial searches that led to contacts for unpublished data in Guinea, Sierra Leone, Côte d’Ivoire, Democratic Republic of Congo, Tanzania, Liberia, Somalia, Kenya, Uganda, Nigeria, Ethiopia, Mali and South Sudan. Tropical Medicine and malaria meeting abstract books were identified from as many sources as possible, produced as part of national, regional and international conferences and congresses. These were used to signal possible data that were followed up through correspondence with abstract authors. Post-graduate theses from faculties related to zoology, medicine and public health were searched periodically at universities in Africa and Europe. Our regional presence and connections to the wider African malaria research community has created an awareness of the purpose and ambition of malaria mapping research first started in 1996 under the MARA/ARMA collaboration. This regional connectivity of research scientists was used to directly contact colleagues working on the epidemiology of malaria to seek disaggregated site-specific and often unpublished data, all individually acknowledged in Section 5.

1.4 Survey data abstraction

1.4.1 Minimum data requirements: From each of the survey reports the minimum required data fields for each record were: description of the study area (name, administrative divisions and geographical coordinates, if available), start and end of survey



dates (month and year) and information about blood examination (number of individuals tested, number positive for Plasmodium infections by species), the methods used to detect infection (microscopy, Rapid Diagnostic Tests (RDTs), Polymerase Chain Reaction (PCR) or combinations) and the lowest and highest age in the surveyed population (decimal years). Given its ubiquity, and history, as a means for malaria diagnosis, the preferred parasite detection method was microscopy where more than one method was used and where slide preparation and reading were quality assured. Where microscopy was not available results from RDTs were used. We have not been able to define the quality of slide preparation, slide reading or RDT sensitivity across the entire data series, and have assumed each survey has equal parasite detection precision. For data derived from randomized controlled intervention trials, data were only selected when described for baseline, pre-intervention and subsequent follow-up cross-sectional surveys among control populations. When cohorts of individuals were surveyed repeatedly the first survey was included and subsequent surveys included if they were separated by at least three months from the initial survey. Occasionally, reports presented the total numbers of people examined across several villages and only the percentage positive per village; the denominator per village was assumed to be equivalent to the total examined divided by the total number of villages. The month of survey was occasionally not possible to define from the survey report. Descriptions of "wet" and "dry" season, first or second school term or other information was used to make an approximation of the month of survey. Some survey results were reported as an aggregate in space (e.g. a single PfPR for a group of villages) or time (e.g. a mean PfPR estimated from four different surveys conducted over time). In such cases, additional reports of the same surveys with higher spatial or temporal resolution were sought from authors.

1.4.2 Age standardization: Where age was not specified in the report but a statement was made that the entire village or primary school children examined the age ranges to be 0-99 years or 5-14 years were assumed respectively. There was a large diversity in the age ranges of sampled populations between studies. Correction to a standard age for P. falciparum was undertaken using an adapted catalytic conversion Muench model that used the lower and upper range of the sample and the overall prevalence to transform into a predicted estimate in children aged 2-10 years, PfPR2-1014.

1.4.3 Survey location geo-coding: During data extraction, each data point was recorded with as much geographic information from the source as possible and this was used during the geo-positioning, for example checking the geo-coding placed the survey location in the administrative units described in the report or corresponded to other details in the report on distance to rivers or towns when displayed on Google Earth. According to their spatial representation, data were classified as individual villages, communities or schools or a collection of communities within an area covering a 5-km grid or approximately 0.05 decimal degrees at the equator (point). Preference was given to point data, however, areas more than 5 km2 were classified as “wide-areas”, and those where data was only available across larger administrative units included as “polygons”. In practice this was a difficult criterion to audit as early survey reports did not provide enough detail on the size of the area surveyed. More recent use of Global Positioning Systems (GPS) during survey work enabled a re-aggregation of household survey data, to increase the sampling precision by



combining clusters of small sample sizes in space, while maintaining the 5 km grid criteria. While in theory GPS coordinates should represent an unambiguous spatial location, these required careful re-checking to ensure that the survey location matched the GPS coordinates and all coordinates located on populated communities. To position each survey location where GPS coordinates were not available, a variety of digital resources were used, amongst which the most useful were Microsoft Encarta Encyclopedia (Microsoft, 2004) and Google Earth (Google, 2009). Other sources of digital place name archives routinely used, importantly, across Africa several national digital, GPS confirmed, place-name gazetteers exist for populated places, health facilities or schools. These are increasing in number, precision and coverage. These were obtained on request from national census bureau’s, ministries of education and health and NGO partners and proved to be valuable locating communities in Angola, Burkina Faso, Chad, Kenya, The Gambia, Madagascar, Malawi, Mauritania, Mozambique, Namibia, Niger, Senegal, Somalia, South Africa, South Sudan, Tanzania, Uganda and Zambia.

1.5 Data summaries

Despite repeated efforts using multiple on-line digital gazetteers, national resources and personal communications it was not possible to identify with sufficient precision the geo-coordinates for 578 (1.1%) survey data points at 546 locations; including 48 locations from national household surveys where the survey agency had withheld coordinates. It was not possible to obtain higher spatial resolution data from 66 survey location reports where only the administrative unit (polygon) summary data were available. The final, complete data set included 50,424 surveys at 36,966 locations, reflecting the repeat sampling of many communities over short surveillance periods and longer term congruence of sampling sites over 115 years (Extended Data 1 and Extended Data 2). The data captured malaria parasite examinations of over 7.8 million blood samples. The estimates of PfPR2-10 varied significantly at any given locality during the data survey period (Extended Data 1). Data volumes varied between countries and between time intervals, reflecting largely the emphasis placed on the needs for empirical survey data to support changing elimination and control agendas (Extended Data 2).

The full dataset assembled for this study is available at http://dx.doi.org/10.7910/DVN/Z29FR0

2 Administrative boundaries and the changing margins of Plasmodium falciparum

2.1 Defining national and sub-national boundaries

We have used contemporary national borders, rather than those in existence at the turn of the last century or independence, thus treating South Sudan separate from The Republic of Sudan and Eritrea separate from Ethiopia. Throughout we have used as a basis for delineating national and sub-national boundaries the Global Administrative Unit Layer (GAUL) 2008 edition15 that assembles UN approved cartographies. We have adapted these at national boundary levels in some disputed areas of Africa that remain unresolved16. For the purposes of our work we have treated the Ilemi Triangle as part of Kenya's Turkana



http://dx.doi.org/10.7910/DVN/Z29FR0

region. We have regarded Walvis Bay in Namibia as always being part of Namibia, despite being a territory of South Africa for many years since Namibia's independence in 1990. Without making any political judgments we regard the region Abeyi as part of the Republic of Sudan. The Hala'ib Triangle on the Sudan-Egyptian border we treat as part of Sudan, although that since 2010 Egypt controls this area as part of its Red Sea State. The Bakassi region, a disputed border territory on the Atlantic coast between Nigeria and Cameroon we treat as part of Nigeria, despite final implementation of agreements for Cameroonian administration in 2014.

To create an adjacency (neighborhood) matrix for spatial analysis, administrative units for malaria predictions through time have been adapted from the first level administrative units developed by GAUL15. Where these covered vast spaces, we have adopted second level administrative units (Sierra Leone, Senegal). The Gambia, Rwanda, Burundi, Djibouti and Swaziland are treated as one administrative unit to the country boundary. We replaced the 11 GAUL regions that cover the vast expanse of the Democratic Republic of Congo (DRC) with the currently nationally approved 26 regions. In some countries, we have used a hybrid approach to retain first level administrative units, while separating large administrative units along second level administrative boundaries to provide a more equal spatial representation compared to other first level administrative units within the country or compared to neighboring countries (Niger, Mali, Senegal districts around Dakar, one admin unit in Benin, Togo, Sudan). Cabinda Province, belonging to Angola, is surrounded by administrative polygons belonging to DRC.

Where administrative boundaries straddle areas free of malaria (Section 2.2) we have separated the administrative unit into malaria free and malaria endemic polygons to allow predictions only to the latter (Angola, Namibia, Botswana, South Africa, Mali, Niger, Mauritania, Chad, Sudan). In Ethiopia (11) and Kenya (8) first level GAUL administrative regions included very large and very small spatial extents within their national borders and in both cases, administrative polygons were uneven and disconnected shapes. Here we have joined up areas based on continuity of neighboring administrative areas and used, where possible second level boundaries to achieve this using ArcGIS 10.1 (ESRI, USA). In both countries, therefore, these are not representative of defined administrative units, rather they represent more contiguous, even spatial units for the purposes of malaria prediction.

Small administrative first level areas that represent federal or sub-national capitals have been dissolved into surrounding administrative units as these vary in their distinction between countries and are very small (Somalia (Mogadishu), Mozambique (Maputo), Central African Republic (Bangui), Niger (Niamey), Benin (Littoral), Mali (Bamako), Mauritania (Nouakchott), Uganda (Kampala), Kenya (Nairobi), Guinea (Conakry) and Guinea Bissau (Autonomo De Bissau)). The exception to this rule was Bulawayo and Harare in Zimbabwe and Addis Ababa in Ethiopia that represent malaria free areas (Section 2.2).

In summary, the final spatial-temporal construction covered 498 polygons across mainland sub-Saharan Africa, and 22 polygons on Madagascar i.e. 520 polygons in total (Extended Data 3; Source Data 1).



2.2 The natural maximal extent of malaria in sub Saharan Africa and Madagascar

Plasmodium falciparum transmission probably reached its natural extent in Africa around 190017. It is difficult to empirically define the natural range of transmission, but the detailed reports of the extents of malaria risk generated by malariologists during the first half of the last century allow an approximation of the natural, pre-elimination margins of risk. These were synthesized during the 1960s18 and we have re-aligned these crude margins of risk with more detailed national reports from the 1920s to 1960s from archived material.

In Mauritania, the regions of Adrar, Inchiri, Dakhlet-Nouadhibou and Tiris Zemmour, north of 20°N have been described as malaria free with a small area of receptive risk at Bay du Levrier around Port Etienne19-22. The margins of stable transmission in Mali, Niger and Chad have been defined from mapped extents and narratives provided by previous authors and WHO consultants18,23-25. In Sudan, the geographical range of malaria transmission was described in reports and maps provided by the Sudan Medical Service (1930-1959) and other authors18, 26-27. In broad terms malaria transmission in Sudan has always been regarded absent above the 18th parallel except along the River Nile and its congruence; one exception is an area on North Eastern Red Sea border encompassing Port Sudan where local transmission has been recorded28. Malaria probably existed in Djibouti at the turn of the last century29, however, it was felt it was largely absent from the entire country between 1910 until circa 1973 when infected vectors may have been re-introduced by train from Ethiopia30,31. We have treated Djibouti as a country with the potential for stable transmission since 1970 only. Addis Ababa, Ethiopia, has been regarded as an area free from malaria since the Second World War22,32.

In Namibia, the southern areas from Grootfontein and Franzfontein to the Orange River have always been regarded as malaria free33 and Hardap, Karas and southern reaches of Kunene and Omaheke remain malaria free34. In Botswana, malaria transmission is constrained by latitude and the Kalahari Desert that makes up 70% of the land mass. Areas regarded as historically free from malaria include the southern parts of Kgalagadi province35 and we have used confirmatory parasite prevalence data to delineate areas where no infections have been detected. In Zimbabwe, autochthonous transmission of malaria in Harare and Bulawayo has not been reported18, 36,37. In South Africa, malaria risks historically extended to within Durban’s city limits, along the Indian Ocean and included Pretoria in the north, reaching the railway crossing point at Ramatlabama on the Botswana border. Malaria in South Africa, has been absent from large parts of the western region including all Northern, Eastern and Western Cape Provinces38-43. We have used maps assembled by Sharp and Le Sueur43 and a photographed copy of the 1936 risk map available in the Malaria Control Programme Offices at Tzaneen to define the natural extent in South Africa. Malaria in Swaziland is governed by altitude but has been reported at all levels, except a small high altitude area along the South Africa border44.

Temperature plays a key role in determining the transmission of human malaria based on its relationship with the duration of sporogony of the parasite in the mosquito vector. To provide a plausible mask to eliminate the possibility of transmission across Africa, we have



used a temperature suitability index (TSI)45. The TSI model uses a biological framework based on survival of vectors and the fluctuating monthly ambient temperature effects on the duration of sporogony that must be completed within the lifetime of a single generation of Anophelines. This was used to generate, at each 1x1 km pixel, periods of an average year when a vector’s lifespan would exceed the time required for sporogony, and hence when transmission was not precluded by temperature. If this time exceeded the maximum feasible vector lifespan, then the cohort was deemed unable to support transmission and the area classified as being at zero risk45. Here, we have used a TSI value of zero for P. falciparum to represent no transmission and TSI values above zero as areas able to sustain some parasite transmission. The P. falciparum temperature mask highlights the highland areas Ruwenzori mountains in Uganda, and mountains of East Africa, the Usambara, Udzungwa and Kilimajaro mountains of Tanzania, Mount Kenya, Elgon and Kinagop in Kenya, the mountains at the junction of Democratic Republic of Congo, Rwanda and Burundi, the highlands in Ethiopia and Eritrea, Mount Cameroun, Mount Kupe, Mount Manengouba and Mount Mekoua in Cameroun, the Imatong Mountains in South Sudan, the Shimbiris mountains in Somaliland, the Nyika Plateau in Malawi, Mount Nyangani in Eastern Zimbabwe and the cold Atlantic coastal areas of Namibia and Angola.

It seems plausible that the natural extents described at the turn of the last century persisted through to the period of the Second World War, despite a reduction in transmission intensity in urban areas and more wide-spread control46. The final delineation of the likely maximal extent of stable P. falciparum transmission, likely to have prevailed through to 1950 is shown in Extended Data 3 and provided in Source Data 1.

2.3 Changing malaria risk extents and margins malaria since 1950

The natural extents of malaria transmission have changed with increasing efforts to eliminate transmission across territories, mostly in Southern Africa. We have reviewed data on sub-national case-incidence, revised maps of malaria risk and narrative descriptions of areas where malaria was eliminated to re-define the margins of malaria free versus unstable/stable transmission by 1960, 1970, 1980, 1990, 2000 and 2015 (shown in main text Figure 1). We have not made any epidemiological distinction between stable and unstable endemicity, nor operational distinctions between pre-elimination, attack or consultation phases.

1960: In South Africa, phases of elimination attack and consolidation in Western and Central Transvaal rendered increasing areas malaria free41,47. Control efforts along the borders with South Africa, constrained the malaria extents in Botswana48, the southern districts of Tsabong through to Gaborone were malaria free, parasitological surveys found no infected infants in Tsha, Loda, Gaborone, Kanye, Moduchi and Ghanzi areas in the 1960s48. In Zimbabwe, areas under aggressive control since 1949 significantly reduced infection and case-incidence risks but failed to interrupt transmission49-52. There were significant reductions in malaria prevalence and case-incidence in other African countries in specific WHO-UNICEF “eradication” project areas during the late from the late 1940s to early 1960s46,53-56 but none of these were sustained to interrupt transmission beyond one or two seasons. The Malagasy highland elimination efforts using DDT IRS and



niviquinization57,58, were notably effective but residual transmission foci continued all the way through to 1979, when epidemics re-emerged54,59.

1970: Continued efforts to eliminate malaria in the Transvaal and KwaZulu-Natal provinces, in South Africa reduced the spatial extents of risk43,47,60,61. The elimination attempts in Ethiopia (1966-1977)62,63 and Somalia (1959-1966)64 significantly reduced parasite incidence but never interrupted transmission. In 1973, malaria was re-introduced for the first time in over 70 years to Djibouti30,31. Evidence from clinical records shows that the urban extent of Nairobi, Kenya was rendered free of malaria, probably through a process of urbanization, by 196965, the extent of built-up Nairobi in 1970 has been digitized from ariel photographs.

1980: In Botswana, malaria was reportedly confined to Ngamiland, Chobe and Francistown by 197966,67. In Namibia, combined medical intelligence based on case data generated by the Ministry of Health and Social Services showed that the regions of Khomas and Erango supported conditions that were malaria free68. In Zimbabwe, evidence suggests that the areas under control since 1959, P. falciparum transmission persisted in the Highveld69. In Swaziland, case incidence data were mapped in 1983 to show that stable risks were constrained to only the areas located on the east of the country70.

1990: In South Africa, epidemics re-emerged but were, by 1995, constrained to areas located along the Kruger national park and borders with Zimbabwe in the Limpopo and Mpumalanga Provinces [Philip Kruger and Aaron Mbuza, personal communication] and the two northerly districts of Ingwavuma and Ubombo in KwaZulu-Natal Province61,71,72.

2000: In Swaziland, clinical cases were largely constrained to Lubombo and Hhohho regions by 199973 [Simon Kunene, personal communication]; the highveld was regarded by this time as malaria free74. In South Africa, the emergence of drug and insecticide resistance led to substantial increases in malaria incidence toward the end of the 1990s46,75,76; the district-by-district extent of endemic versus pre-elimination risk in Kwa-Zulu Natal, Limpopo and Mpumalanga Provinces have been digitized from77.

2015: The Wadi Halfa area of Sudan, repeatedly showed an absence of transmission through repeat bi-annual community-based surveillance of malaria infection and vectors, as part of the Gambiae Project co-funded by Egypt, we have therefore regarded the area from Kimnar to Wadi Halfa on the Egyptian border as free of malaria by 2015 [FMoH Sudan, personal communication]. A national PCR survey in the Kingdom of Swaziland identified only two infected hosts (P. falciparum and P. malariae) in 201078, in 2012, malaria became a notifiable disease using combinations of active and passive surveillance79, with receptive risks concentrated by 2014 in a constrained area to the east80. Swaziland has embarked on a strategy of elimination and the remaining areas of locally acquired risk, Hhohho and Lubombo regions, while subject to imported malaria, is considered malaria free81 [Simon Kunene, personal communication]. The margins of stable endemic risk in Namibia remain congruent with those of the preceding decade, despite ambitions for elimination by 202034,82,83. In Botswana in 2012, malaria was made a notifiable disease84,85. By 2013, 456 cases were reported from Botswana and most were detected in Bowirwa district, which



remains the one stable endemic area of the country [Godira Segoea, personal communication]. The districts of Chobe, Okavango and Ngami continue to support unstable transmission, and are most likely to be suited to the country’s elimination strategy84 while the remaining areas of the country were probably malaria free by 2015 but subject to imported and epidemic risks85,86. By 2010/11, case incidence in South Africa had dropped dramatically. Information provided in publications and personal communications with provincial malaria programmes covering Limopopo, Mpumalanga and Kwa Zulu Natal have been used to delineate areas regarded as unstable and under pre-elimination phases by circa 201576,77,87-89 [Philip Kruger, Aaron Mbuza, Marlies Craig and Rajendra Mahara, personal communication]. The E8 consortia of countries now plan to interrupt the transmission of malaria within the boundaries of in Botswana, Namibia, South Africa, Swaziland, and southern Zimbabwe90 and southern regions of Mozambique (Maputo, Gaza and Inhambane)91.

Changing extents of P. falciparum transmission provided in Source Data 2.

3: Statistical methods

3.1 Hierarchical space–time model for P. falciparum infection prevalence in children aged 2-10 years (PfPR2-10) from 1900-2015

We used a Bayesian hierarchical model that simultaneously characterizes and estimates stable spatial and temporal patterns and departures from these stable components (due to sparse data) through predictable global space–time structures and specific space–time interactions92 with the aim of predicting the prevalence of infection in out of sample subnational regions. This approach extends traditional spatial conditional autoregressive (CAR) models93 by inclusion of the space-time effect to account for departures from the stable effects and through inclusion of a CAR temporal random effect based on the two adjacent period points (preceding and post).

Point location survey data and subnational regional (areal) data were aggregated by subnational area and period to produce a space-time data cube structured as follows: observed number of P. falciparum positive children aged 2-10, PfPR2-10 it , and total number of tested children aged 2-10 ,n it, for subnational region i = 1, … , 520 (total number of subnational administrative areas in sub-Saharan Africa, 498 on mainland and 22 in Madagascar, within malaria endemic limits) and period t = 1, … , 16. The periods were divided as follows: 1900-1929, 1930-1944, and five year intervals between 1945-1949 and 2010-2015.

A binomial model was used, rather than the usual Poisson approach, for P. falciparum count data92 with a logit link function in our space-time model. The binomial formulation was chosen over the Poisson approach as it provided a significantly better overall goodness of fit based on the Deviance Information Criterion (DIC)94. Furthermore, the Poisson approach did not perform any better with regards to the validation exercise described at the end of this section. The full Bayesian binomial space-time model hierarchical formulation was as follows:



Level 1 [within-area variability of the counts conditional on unknown risk parameters]:

PfPR2-10 it ~ binomial (nit, πit) [1]

Level 2 [risks on logit scale conditional on unknown risk parameters]:

Logit (π it) = α+ λ i, + ξ t, + ν it [2]

where π it is the risk (prevalence) of P. falciparum infection in subnational region i in period t, α is the overall baseline risk (intercept), λ i is the main spatial effect, ξ t is the main temporal effect, and ν it is the space–time interaction term. Random variables (effects) were assigned prior distributions that could borrow information across the space-time cube to better capture the underlying structure of the P. falciparum infection risk – these were spatial (λ I) and temporal (ξ t,).

The spatial dependence is represented by means of a neighbourhood matrix that defines for each subnational region i its set of adjacent neighbours and modelled through a conditional autoregressive process. The value of a parameter in one subnational region is influenced by the average value of its neighbouring regions. The notation to denote the conditional autoregressive process specified in [2], where μ is the vector (μ1, μ2, . . . , μ I)′ specified as follows:

λi ~ N (μi, σ2λ), i=1,…,522; μi ~ CAR(W, σ2μ) [3]

The CAR approach can also be employed to model the temporal random effect. Similarly, we define period neighbours simply by the two adjacent period points (namely preceding and post), and implemented using a period adjacency matrix. The specification is as follows:

ξt ~ N (γt, σ2ξ), t= 1,…,16; γt ~ CAR(Q, σ2γ) [3]

Apart from the spatial and temporal effects in [2], we included a space–time interaction parameter ν it,. This parameter is meant to account for any departure from predictable patterns based the temporal and spatial effects and for characterizing the stability of the underlying spatial patterns, with large fluctuations indicating instability of risk in subnational region i. The primary purpose of this additional effect is to help distinguish more stable predictable patterns from unusual instances and provide the necessary flexibility in the model to allow certain subnational regions to have “true” departures from the aforementioned “predictable” spatial (λi) and temporal (ξt) effects. For full details of the specification of this parameter see [92].

We have avoided using climate, land use, ecology and intervention covariates during the spatial-temporal modelling, because we regard parasite prevalence in each locality at a specific time to represent the product of these covariates, furthermore their inclusion



would introduce a circularity precluding the examination of the plausible role of either climate or intervention.

The WinBUGS code for our models are provided at http://dx.doi.org/10.7910/DVN/Z29FR0

The variance parameters for the random effects were treated as unknown (i.e. non-informative priors) and given hyper prior distributions. We chose inverse gamma distributions with parameter values of 0.5 and 0.0005 as suggested by Wakefield and colleagues95. The model was fitted using Markov Chain Monte Carlo simulation. Posterior distributions of parameters were obtained using WinBUGS software96.

Model convergence: A two-chain Markov Chain Monte Carlo simulation was used for parameter estimation to assess convergence. Model convergence was assessed by visual inspection of the series plot of each parameter, and using Gelman-Rubin statistics97

(Extended Data 4). The final posterior samples obtained after convergence were run until the Monte Carlo error for each parameter was less than 5% of the sample standard deviation.

Model validation: A random sample of 100 data points were drawn from the space-time cube. The data with these points removed were then re-inputted into WinBUGS. The posterior distributions for the predicted PfPR2-10 at these 100 removed data points were then compared against the observed values to ascertain the predictive power of the model. The percentage of observed PfPR2-10 that were contained within the credibility interval of the posterior distribution for the predicted PfPR2-10 were calculated.

The posterior predicted estimate of PfPR2-10 for each temporal interval for each of the 520 prediction polygons shown in main text Figure 1

3.2 Validation of model outputs

An assessment of model convergence using Gelman-Rubin statistics/plots (Extended Data 4) and MC error/SD less than 5% suggested convergence/stabilisation of the model afterapproximately 75,000 iterations (Table SI 1).

Table SI 1 Parameter posterior distributions with Monte Carlo (MC) error divided by the standard deviation (SD)

node mean SD MC error 2.50% median 97.50% start sample

MC error/SD

beta0 -0.9673 0.02547 0.00101 -1.016 -0.9685 -0.9161 5000 75000 3.97%

kappa 1.391 0.1482 0.003926 1.144 1.377 1.722 5000 75000 2.65%

sigma.nu[1] 0.8207 0.05235 0.001673 0.709 0.8241 0.913 5000 75000 3.20%

sigma.nu[2] 2.212 0.1851 0.005405 1.886 2.2 2.609 5000 75000 2.92%

sigma.t 0.3873 0.09134 0.002057 0.2458 0.375 0.601 5000 75000 2.25%

sigma.w 1.32 0.06961 0.001876 1.189 1.319 1.462 5000 75000 2.70%



http://dx.doi.org/10.7910/DVN/Z29FR0

A comparison of observed versus fitted PfPR2-10 from the full model suggested a very high degree of correlation (0.99, P<0.001) with a few notable differences e.g. many data points with observed prevalence of zero were corrected upwards by the model and/or smoothed towards local regional mean based on contiguity matrix in space and time. Of note is that 99/100 of the observed PfPR2-10 were contained in the 95% credible interval (CI) for its posterior distribution (Correlation coefficient 0.483, P <0.001) (Extended Data 5).

4. References

1. World Health Organization. Report on the malaria conference in equatorial Africa.Held under the joint auspices of the World Health Organization and of thecommission for technical co-operation in Africa south of the Sahara. Kampala,Uganda, 27 November–9 December, 1950

2. Metselaar, D. & van Thiel, P.H. Classification of malaria. Trop. Geogr. Med. 11, 157–161 (1959)

3. Cohen, J.M., Moonen, B., Snow, R.W. & Smith, D.L. How absolute is zero? Anevaluation of historical and current definitions of malaria elimination. Malar J 9, 213(2010)

4. Noor AM et al. The changing risk of Plasmodium falciparum malaria infection inAfrica: 2000–10: a spatial and temporal analysis of transmission intensity. Lancet383, 1739-1747 (2014)

5. Snow, R.W., Marsh, K. & le Sueur, D. The need for maps of transmission intensity toguide malaria control in Africa. Parasit Today 12: 455–457 (1996)

6. Le Sueur, D. et al. (1997). An atlas of malaria in Africa. Africa Health 19, 23-24 (1997)

7. MARA/ARMA (1998). Towards an atlas of malaria risk in Africa. First technicalreport of the MARA/ARMA collaboration. Durban, South Africa; https://idl-bnc.idrc.ca/dspace/bitstream/10625/31644/1/114833.pdf (accessed 5th March2017)

8. Hay, S.I. & Snow R.W. The Malaria Atlas Project (MAP): developing global maps ofmalaria risk. PLoS Med 3, e473 (2006)

9. Guerra, C.A. et al. Assembling a global database of malaria parasite prevalence forthe Malaria Atlas Project. Malar. J. 6, 17 (2007)

10. HINARI http://hinarilogin.research4life.org/



https://idl-bnc.idrc.ca/dspace/bitstream/10625/31644/1/114833.pdf

https://idl-bnc.idrc.ca/dspace/bitstream/10625/31644/1/114833.pdf

http://hinarilogin.research4life.org/

11. Roll Back Malaria Monitoring and Evaluation Reference Group, World HealthOrganization, United Nations Children’s Fund, MEASURE DHS, MEASURE Evaluation,and U.S. Centers for Disease Control and Prevention (2005). Malaria IndicatorSurvey: Basic documentation for survey design and implementation. Calverton,Maryland: MEASURE Evaluationhttp://apps.who.int/iris/bitstream/10665/43324/1/9241593571_eng.pdf(accessed 5th May 2017)

12. DHS programme http://dhsprogram.com/

13. UNICEF http://www.unicef.org/statistics/index_24302.html

14. Smith, D.L., Guerra, C.A., Snow, R.W. & Hay, S.I. Standardizing estimates of malariaprevalence. Malar. J. 6, 131 (2007)

15. GAUL http://www.fao.org/geonetwork/srv/en/metadata.show?id=12691

16. African Union Border Programme. Delimitation and demarcation of boundaries inAfrica: The User’s Guide. The Commission of the African Union / Department of Peaceand Security, African Union, Addis Ababa, Ethiopia (2014)

17. Carter, R. & Mendis, K.N. Evolutionary and historical aspects of the burden ofmalaria. Clin. Microbiol. Rev. 15, 564-594 (2002)

18. Lysenko, A.J. & Semashko, I.N. Geography of malaria. A medico-geographic profile ofan ancient disease. In: Lebedew AW, ed. Itogi Nauki: Medicinskaja Geografija.Moscow, USSR: Academy of Sciences, pp 25–146 (1968)

19. Sautet, J., Ranque, J., Vuillet, F. & Vuillet, J. Quelques notes parasitologiques sur lepaludisme et l’anophelisme en Mauritanie. Med. Trop. (Mars). 8, 32–39 (1948)

20. Hudleston, J. Programme de pré-éradication du paludisme: Kaédi, Mauritanie - 9.Report on mission 1st January 1963-9th March 1966. AFR/MAL/74; World HealthOrganisation Archives, Geneva (1966)

21. Khromov, A.S. Programme de pre-eradication du paludisme. World HealthOrganization, AFR/MAL/99; World Health Organization archive, Geneva (1969)

22. International Association for Medical Assistance to Travellers (IAMAT), 2015:https://www.iamat.org/ (accessed December 2016)

23. Ochrymowicz, J.W., Bakri, G.E.D. & Hudleston, J.A. Rapport sur la prospection faite envue d'une section antipaludique au Niger par l'équipe consultative régionale dupaludisme juin-novembre 1968. World Health Organisation Project AFRO 204,Niamey, December (1968)



http://apps.who.int/iris/bitstream/10665/43324/1/9241593571_eng.pdf

http://dhsprogram.com/

http://www.unicef.org/statistics/index_24302.html

http://www.fao.org/geonetwork/srv/en/metadata.show?id=12691

https://www.iamat.org/

24. Holstein, M. Etudes sur l’anophelisme en A.O.F. 1. Soudan français. A Bamako. Bull.Soc. Pathol. Exot. 42, 374-378 (1949)

25. Saugrain, J. & Taufflieb, R. Anophelisme sans paludisme au Nord Tchad. Bull. Soc.Pathol. Exot. 53, 150-152 (1960)

26. Nasr, A.H. Preparations for future malaria eradication programme in the Republic ofthe Sudan. Al Hakeem, University of Khartoum, Faculty of Medicine, Journal of theMedical Students Association 7, 178-190 (1968)

27. El Gaddal, A.A. Malaria in the Sudan. In Proceedings of the conference on malaria inAfrica: Practical considerations on malaria vaccines and clinical trials. Ed. AA Buckfor USAID and AIBS, Washington DC, USA, December 1-4, 1986. Page 157 (1986)

28. Hashim, A. Five-year plan of action in malaria control in Red Sea State 1998-2002.Blue Nile Research and Training Institute, Giezera (1998)

29. Bouffard, G. Ge ographie Me dicale: Djibouti. Ann. Hyg. Med. Colon. 8, 342-344 (1905)

30. Carteron, B., Morvan, D. & Rodhain, F. The question of endemic malaria in Republicof Djibouti. Med. Trop. (Mars). 38, 299-304 (1978)

31. Fox, E. et al. Plasmodium falciparum voyage en train d’Ethiopie a Djibouti. Med. Trop.(Mars). 51, 185–189 (1991)

32. Ovazza, M. & Neri, P. Vecteurs de paludisme en altitude (région d'Addis Abeba,Ethiopie). Bull. Soc. Pathol. Exot. 48, 679-686 (1955)

33. De Meillon, B. Malaria survey of South-West Africa. Bull. World Health Organ. 4, 333–417 (1951)

34. Noor, A.M. et al. The receptive versus current risks of Plasmodium falciparumtransmission in Northern Namibia: implications for elimination. BMC Infect. Dis. 13,184 (2013)

35. Bechuanaland Protectorate. Annual Medical and Sanitary Reports for theProtectorate for the years 1958-1963. Government Printers, Gaborone (1958-1963)

36. Thomson, J.G. Section of epidemiology and state medicine: endemic and epidemicmalaria in Southern Rhodesia. Proc. R. Soc. Med. 22, 1051-1058 (1929)

37. Ministry of Health [Southern Rhodesia]. Annual Medical Reports Southern Rhodesia,Federation of Rhodesia and Nyasaland, and Rhodesia (1928-1954)



38. Pratt-Johnson, J. The distribution of malaria in South Africa and a mosquito surveyof military hospital areas. J. Hyg. 19, 344-349 (1921)

39. Ingram, A. & de Meillion, B. A mosquito survey of certain parts of South Africa, withspecial reference to the carriers of malaria and their control. The South Africaninstitute for Medical Research, Johannesburg, January (1929)

40. Swellengrebel, N.H. Report on investigation into malaria in the Union of South Africa1930-1931. Annexure to Department of Public Health Report 1932, GovernmentPrinters, Pretoria (1932)

41. Brink, C.J.H. Malaria control in the northern Transvaal. S. Afr. Med. J. 32, 800-808(1958)

42. Le Sueur, D., Sharp, B.L. & Appleton, C.C. Historical perspective of the malariaproblem in Natal with Emphasis on the period 1928–1932. S. Afr. J. Sci. 89, 232–239(1993)

43. Sharp, B.L. & Le Sueur, D. Malaria in South Africa- the past, the present and selectedimplications for the future. S. Afr. Med. J. 86, 83–89 (1996)

44. Mastbaum, O. Past and present of malaria position of malaria in Swaziland. J. Trop.Med. Hyg. 60, 119-127 (1957)

45. Gething PW et al. Modelling the global constraints of temperature on transmission ofPlasmodium falciparum and P. vivax. Parasit. Vectors 4, 92 (2011)

46. Snow, R.W. et al. The changing limits and incidence of malaria in Africa: 1939-2009.Adv. Parasitol. 78, 169-262 (2012)

47. Hansford, C.F. Malaria control in the Northern Transvaal. S. Afr. Med. J. 48, 1265-1269 (1974)

48. World Health Organization. Report of WHO Advisory team on Malaria Eradication no.2, Bechuanaland Protectorate July 1961 - August 1962 (1963)

49. Alves, W. Preliminary notes on a Southern Rhodesian experiment in malaria control.S. Afr. J. Sci 47, 289-292 (1951)

50. Alves, W. & Blair, D.M. An experiment in the control of malaria and bilharzias. Trans.R. Soc. Trop. Med. Hyg. 47, 299-308 (1953)

51. Alves, W. & Blair, D.M. Malaria control in Southern Rhodesia. J. Trop. Med. Hyg. 58,273-280 (1955)



52. Wolfe, H.L. Epidemiological data concerning one year of a malaria surveillance pilotproject in Southern Rhodesia. Bull. World Health Organ. 31, 707-720 (1964)

53. Russell, P.F. Some epidemiological aspects of malaria control with reference to DDT.J. Natl. Malar. Soc. 10, 257-265 (1951)

54. Laing, A.B. The impact of malaria chemoprophylaxis in Africa with special referenceto Madagascar, Cameroon and Senegal. Bull. World Health Organ. 62, 41-48 (1984)

55. Kouznetsov, R.L. Malaria control by application of residual insecticides in tropicalAfrica and its impact on community health. Trop. Doct. 7, 81-91 (1977)

56. Najera. J.A., Gonzalez-Silva, M. & Alonso, P.L. Some lessons for the future from theGlobal Malaria Eradication Programme (1955–1969). PLoS Med. 8, 1000412 (2011)

57. Bernard, M.P. Trois ans de lutte antipaludique à Madagascar 1950-1951-1952.Bulletin de Madagascar, Publication Mensuelle du Service General de l’informationdu Haut-Commissariat, no. 96 (1954)

58. Joncour, G. La lutte contre le paludisme à Madagascar. Bull. World Health Organ. 15:711-723 (1956)

59. Mouchet, J. et al. Stratification épidémiologique du paludisme à Madagascar.Archives d’Institute Pasteur, Madagascar 60, 50-59 (1993)

60. Nethercott, A.S. Forty years of malaria control in Natal and Zululand. S. Afr. Med. J 48,1168–1170 (1974)

61. Craig, M.H., Kleinschmidt, I., Le Seuer, D. & Sharp, B.L. Exploring 30 years of malariacase data in KwaZulu-Natal, South Africa: Part II. The impact of non-climatic factors.Trop. Med. Int. Health 9, 1258-1266 (2004)

62. Delfini, L. & Shidrawi, G. A report on a visit to Ethiopia 23rd March to 9th April 1976.World Health Organization, unpublished report EM/MAL/144 dated June 1976;World Health Organization archives, Geneva (1976)

63. Gish, O. Malaria eradication and the selective approach to health care: Some lessonsfrom Ethiopia. Int. J. Health Serv. 22, 179-192 (1992)

64. Nouger, A. Assignment report Malaria pre-eradication programme in Somalia.EM/MAL/58, Somalia 0002/R, UNDP, UNICEF August (1967)

65. Mudhune, S.A. The clinical burden of malaria in Nairobi: a historical review andcontemporary audit. Malar. J. 10, 138 (2011)



66. Chayabejara, S., Sobti, S.K., Payne, D. & Braga, F. Malaria situation in Botswana.Report on a visit December 1973 - October 1974. AFR/MAL/144. World HealthOrganization Archives, Geneva (1975)

67. Franco de, A.L.T., Roche. S.M., Ariaratnam, V., Joia, H.S. & Chinien, V. Malaria situationin Botswana. Report on evaluation mission WHO consultant team, September-November 1984. Unpublished report, World Health Organization archives, Geneva(1984)

68. Snow, R.W. et al. Estimating the distribution of malaria in Namibia in 2009:assembling the evidence and modelling risk. Ministry of Health and Social Services,Republic of Namibia and Malaria Atlas Project, May (2010)

69. National Malaria Control Programme [Zimbabwe]. Strategy for Zimbabwe 2008 –2013. Ministry of Health and Child Welfare, Zimbabwe, October (2008)

70. Hansford, C.F. Malaria control programme in Swaziland: report on evaluation mission,WHO consultant 5th-16th, unpublished report September (1994)

71. Strebel, P.M., Hansford, C.F. & Kustner, H.G.V. The geographic distribution of malariain South Africa in 1986. South Afr. J. Epidemiol. Infect. 3, 4-8 (1988)

72. Kleinschmidt, I., Sharp, B.L., Clarke, C.P.Y., Curtis, B. & Fraser, C. Use of GeneralizedLinear Mixed Models in the spatial analysis of small-area malaria incidence rates inKwaZulu Natal, South Africa. Am. J. Epidemiol. 153, 1213-1222 (2001)

73. Ministry of Health & Social Welfare [Swaziland]. Malaria control in Swaziland.National Malaria Control Programme, Mbabane, Kingdom of Swaziland (1999)

74. Fontaine, R.E. Review of the Swaziland malaria control program conducted betweenMarch 30 – April 25, 1987. Vector Biology and Control Project, Medical ServiceConsultants Inc., AR-052 (1987)

75. Department of Health [South Africa]. Prevalence and distribution of malaria in SouthAfrica 2007. Directorate of Epidemiology, HIER Cluster, Department of Health,Republic of South Africa, March (2008)

76. Department of Health [South Africa] & Roll Back Malaria. Progress and Impact SeriesCountry Reports: Focus on South Africa. World Health Organization, Geneva, October(2013)

77. Maharaj, R. et al. The feasibility of malaria elimination in South Africa. Malar. J. 11,423–428 (2012)



78. National Malaria Control Programme [Swaziland]. Swaziland National MalariaIndicator Survey 2010. National Malaria Control Programme, Ministry of Health,Kingdom of Swaziland, July (2011)

79. Churcher, T.S. et al. Measuring the path toward malaria elimination: Developing newtargets and milestones from standard surveillance data. Science 344, 1230-1232(2014)

80. Reiner, R.C. et al. Mapping residual transmission for malaria elimination. eLife 4,e09520 (2015)

81. Kunene, S., Phillips, A.A., Gosling, R.D., Kandula, D. & Novotny, J.M. A national policyfor malaria elimination in Swaziland: a first for sub-Saharan Africa. Malar. J. 10, 313(2011)

82. Alegana, V.A. et al. Estimation of malaria incidence in northern Namibia in 2009using Bayesian Conditional-Autoregressive Spatial-Temporal Models. Spat.Spatiotemporal Epidemiol. 7, 25-36 (2013)

83. Smith-Gueye, C. et al. Namibia's path toward malaria elimination: a case study ofmalaria strategies and costs along the northern border. BMC Public Health 14, 1190(2014)

84. Ministry of Health [Botswana]. Malaria Strategic Plan 2010–2015: towards malariaelimination. Botswana National Malaria Control Programme, Department of PublicHealth, Ministry of Health (2010)

85. Simon C et al. Malaria control in Botswana, 2008–2012: the path towardselimination. Malar. J. 12, 458 (2013)

86. Chihanga, S. et al. Malaria elimination in Botswana, 2012–2014: achievements andchallenges. Parasit. Vectors 9, 99 (2016)

87. Gerritsen, A.A.M. et al. Malaria incidence in Limpopo Province, South Africa, 1998–2007. Malar. J. 7,162 (2008)

88. Ngomane, L. & de Jager, C. Changes in malaria morbidity and mortality inMpumalanga Province, South Africa (2001- 2009): a retrospective study. Malar. J.11, 19 (2012)

89. Silal, S.P., Barnes, K.I., Kok, G., Mabuza, A. & Little, F. Exploring the seasonality ofreported treated malaria cases in Mpumalanga, South Africa. PLoS One 8, e76640(2013)

90. Elimination 8 (2015). Elimination 8 Strategic Plan: 2015–2020. Elimination 8:Windhoek.



http://www.shrinkingthemalariamap.org/files/content/resource/attachment/E8%20Strategic%20Plan%20(2015-2020).pdf (accessed 6th March 2017)

91. Moonasar, D. et al. Towards malaria elimination in the MOSASWA (Mozambique,South Africa and Swaziland Region. Malar. J. 15, 419 (2016)

92. Abellan, J.J., Richardson, S. & Best, N. Use of space-time models to investigate thestability of patterns of disease. Environ. Health Perspect. 116, 1111 (2008)

93. Besag, J., York, J. & Molliè, A. Bayesian image restoration, with two applications inspatial statistics. Ann. Inst. Stat. Math. 43, 1-59 (1991)

94. Spiegelhalter, D., Best, N., Carlin, B., van der Linde, A. Bayesian measures of modelcomplexity and fit. J. R. Stat. Soc. B. 64, 583-639 (2002)

95. Wakefield, J.C., Best, N.G. & Waller, L. Bayesian approaches to disease mapping In:Spatial Epidemiology (Elliott P, Wakefield JC, Best NG, Briggs DJ, eds.). Oxford:Oxford University Press, 104–127 (2000)

96. Lunn, D.J., Thomas, A., Best, N. & Spiegelhalter, D. (2000) WinBUGS — a Bayesianmodelling framework: concepts, structure, and extensibility. Stat Comp 10, 325–337(2000)

97. Gelman, A. & Rubin, D. Inference from iterative simulations using multiplesequences. Stat Sci. 7, 457-472 (1992)



http://www.shrinkingthemalariamap.org/files/content/resource/attachment/E8%20Strategic%20Plan%20(2015-2020).pdf

http://www.shrinkingthemalariamap.org/files/content/resource/attachment/E8%20Strategic%20Plan%20(2015-2020).pdf

5. Acknowledgements

5.1 General archive assistance, data extraction and geo-coding

Camila Aros-Perez, Hoda Atta, Muriel Bastien, Mauro Capocci, Catherine Cecilio, Gilberto Corbellini, Joaquim Da Silva, Daniel Demellier, Virgílio do Rosário, Dominique Dupenne, Reynald Erard, Gehane Al Garraya, Zamani Ghassem, Priscilla Gikandi, Carlos Guerra, Nahla Ibrahim, Caroline Kabaria, Amir Kamal, William Kisinza, Damaris Kinyoki, Donna Kynaston, Anne-Marie Lallemand, Lorena Lucioparedes, Peter Macharia, Ruth Macharia, Claire MackIntosh, Alex Maina, Betsy Makena, Charles Mayika-Louvouezo, Anne Mbeche, Bernard Mitto, Sebastian Morrelo, Hatem Nour El-Din H. Mohamed, Winne Musivo, Jonesmus Mutua, Lydiah Mwangi, Stephen Oloo, Judy Omumbo, Viola Otieno, Paul Ouma, Marie Sarah Villemin Partow, Christian Pethas-Magilad, João Pires, Agnes Raymond-denise, Gilbert Sang, Christian Sany, Dirk Schoonbaert, Naomi Snow, Rebecca Snow, Patrick Tungu

5.2 Regional research institutes and national malaria control programmes who have provided archive assistance, access to national household survey data not in public domain, data extraction and geo-coding

The Medical Research Council Laboratories (The Gambia), the Kenyan Medical Research Institute collaborative partnerships with the Wellcome Trust/University of Oxford, US Centers for Disease Control, Nagasaki University and Walter Reed (Kenya), the Malaria Research Training Centre (Mali), Uganda Malaria Surveillance Project (Uganda), Ifakara Health Institute's collaborative partnerships with the Swiss Tropical Institute, London School of Hygiene and Tropical Medicine and US Centers for Disease Control (Tanzania), Dar es Salaam Urban Malaria Control Project (Tanzania), National Medical Research Institute at Amani (Tanzania), Institut Pasteur (Madagascar), Institut de Recherché pour le développement (Senegal), Centre de Recherché en Santé de Nouna (Burkina Faso), Institute for Endemic Diseases, University of Khartoum (Sudan), Blue Nile Health Project-Giezera State University (Sudan), Centro Investigaco Saude Angola (Angola), Liverpool School of Tropical Medicine, Wellcome Trust and College of Medicine, University of Malawi collaborative programme (Malawi), South African Medical Research Council (South Africa, Mozambique and Swaziland), Swiss Tropical and Public Health Institute's collaborative programme (Côte D'Ivoire), US Centers for Disease Control collaborative programme (Togo), Prince Leopold Institute's country level collaborative partnerships (Burundi, Benin and Rwanda), Malaria Institute at Macha (Zambia), Medical Research Centre/Uganda Virus Research Institute (Uganda), the School of Public Health Kinshasa (DRC), National Centre for Tropical Medicine, Health Institute Carlos III, Madrid and the Spanish Tropical Diseases Research Network (RICET) (Equatorial Guinea), Médecins sans Frontières (Guinea), MERLIN (Kenya), The Carter Center (Nigeria), The Food Security and Nutrition Analysis Unit (Somalia), Shape Consulting (DRC, Mozambique, Kenya, Sierra Leone, Tanzania), Save the Children (Mozambique), the Malaria Consortium (Ethiopia, Nigeria, South Sudan and Uganda) and MENTOR (Somalia) and the Southern African Malaria Consortium (Zimbabwe). In addition, we are enormously grateful to the generous support of the National Malaria Control Agencies of the following countries who have provided data as part of national mapping exercises and used here: Chad, Equatorial Guinea, Eritrea, Kenya, Namibia, Mozambique, South Sudan, Republic of Sudan, Republic of Somaliland, Republic of Puntland, Somalia, Swaziland, Tanzania, The Gambia and Zambia



5.3 Country level support

Angola: Ana Paula Arez, Assumpta Bou-Monclus, Fernando David, Dina Gamboa, António Langa, Akiko Matsumoto, John Mendelsohn, Cristina Mendes, Susana Nery, João Mário Pedro, José Sousa-Figueiredo, Fern Teodoro, Claudia Videira

Benin: Césaire Damien Ahanhanzo, Alexandre Biaou, Emile Bongo, Vincent Corbel, Umberto D'Alessandro, Georgia B Damien, Yves Eric Denon, Virgile Gnanguenon, Marie-Claire Henry, Florence Kouadio, Florence Migot-Nabias, Christian Lengeler, Alain Nahum, Mariam Oke, Razack Osse, Christophe Rogier, Peter Thomas

Botswana: Colleen Fraser, Musa Mabaso, Marlies Craig, Frank Hansford, Graham Root

Burkina Faso: Heiko Becher, Claudia Beiersmann, Clarisse Bougouma, Mamoudou Cisse, Diadier Diallo, Adama Gansane, Jean-Olivier Guintran, Sidonie Gouem, Amadou Konate, Bocar Kouyaté, Christian Lengeler, Laurent Moyenga, Olaf Müller, Issa Nebie, Jean-Bosco Ouedraogo, André Lin Ouédraogo, Hermann Ouédraogo, Christophe Rogier, August Stitch, Yazoume Yé, Issa Zongo

Burundi: Lidwine Barahadana, Marc Coosemans, Baza Dismas, Natacha Protopopoff

Cameroon: Eric Achidi, Toby Apinjoh, Rolland Bantar, Chi Hanesh, Wilfred Mbacham, Kenneth Ndamukong, Maria Rebollo, Innocent Takougang, Peter Uzoegwu, Samuel Wanji

Chad: Kerah Clement, Ephraim Djoumbe, Ibrahim Socé Fall, Etienne Magloire Minkoulou, Jose Nkuni, Nadjitolnan Othingué, Kaspar Wyass

Cote d’Ivoire: Serge-Brice Assi, Bassirou Bonfoh, Mark Divall, N'Goran K Eliezer, Marie-Claire Henry, Clarisse Houngbedji, Astrid Knoblauch, Benjamin Koudou, Barbara Matthys, Giovanna Raso, Fabian Rohner, Kigbafori Silue, Marcel Tanner, Andres Tschannen, Thomas Tuescher, Juerg Utzinger, Rita Wegmüller, Richard Yapi

Democratic Republic of Congo: Mike Bangs, Civeon Bazebosso, Morrison Bethea, Thierry Lengu Bobanga, Mark Divall, Ellen Dotson, Giovanfrancesco Ferrari, Louis Ilunga, Seth Irish, Jean-Emmanuelle Julo-Réminiac, Sakkie Hattingh, Yakim Kabvangu, Didier Kabing, Didier Kalemwa, Lydie Kalindula-Azama, Hyacinthe Kaseya, Onyimbo Kerama, Astrid Knoblauch, Phillipe Lafour, Christian Lengeler, Joris L. Likwela, Crispin Lumbala Wa Mbuyi, Eric Mafuta, Landing Mane, Emile Manzambi, André Mazinga, Jolyon Medlock, Steve Meshnick, Janey Messina, Flavien Mulumba, Ambroise Nanema, Marius Ngoyi, Henry Ntuku, Laura O'Reilly, Milka Owuor, Cédric Singa, Salumu Solomon, Eric Mukomena Sompwe, Edouard Swana, Katie Tripp, Antoinette Tshéfu, Francis Watsenga, Mirko Winkler

Djibouti: Mouna Osman Aden, Ifrah Ali Ahmed, Abdisalan Mohamed Noor, Christophe Rogier

Equatorial Guinea: Ana Paula Arez, Estefanía Custodio, Immo Kliendsmidt, Zaida Herrador Ortiz

Eritrea: Araia Berhane, Joe Keating, Azmera Gebreslassie, Tewolde Ghebremeskel, Samuel Goitom, Amanuel Kifle, Selam Mihreteab, David Sintasath, Abdulmumini Usman, Assefash Zehaie

Ethiopia: Ruth Ashton, Meshesha Balkew, Estifanos Biru, Simon Brooker, Peter Byass, Karre Chawicha, Wakgari Deressa, Tufa Dinku, Yeshewamebrat Ejigsemahu, Henok Kebede Ejigu, Paul



Emerson, Tekola Endeshaw, Teshome Gebre, Asrat Genet, Asefaw Getachew, Melaku Gimma, Patricia Graves, Afework Hailemariam, Sharon Hill, Don Hopkins, Daddi Jima, Abdissa Kurkie Kabeto, Bernt Lindtjørn, Addisu Mekasha, Endale Mengesha, Ayenew Messele, Aryc Mosher, Jeremiah Ngondi, Frank Richards, Teshale Seboxa, Bob Snow, Niko Speybroeck, Kassahun Tadesse, Zerihun Tadesse, Hiwot Solomon Taffese, Bekele Worku, Adugna Woyessa, Delenasaw Yewhalaw, Mulat Zerihun

Gabon: Steffen Bormann, Dieudonné Nkoghe, Francine Ntoumi, Odile Oukem

Gambia: Jane Achan, Onome Akpogheneta, Steve Allen, Sarah Atkinson, Kalifa Bojang, Sering Ceesay, Sian Clarke, David Conway, Baboucarr Daffeh, Umberto D'Alessandro, Chris Drakeley, Sam Dunyo, Fatty Fatoumatta, Malang Fofana, Brian Greenwood, Adam Jagne-Sonko, Cherno Jallow, Ebrima Jarjou, Musa Jawara, Momodou Kalleh, Balla Kandeh, Sharmila Lareef-Jah, Steve Lindsay, Karafa Manneh, Ousman Nyan, Margaret Pinder, Eleanor Riley, Judith Satoguina, Lorenz von Seidlein, Bob Snow, Sheriffo Sonko, Thomas Sukwa, Ebako Takem, Michael Walther, Emily Webb

Ghana: Benjamin Abuaku, Michael Acquah, Collins Ahorlu, Felicia Amo-Sakyi, Frank Amoyaw, Anthony Amuzu, Vivian Aubyn, Irene Ayi, Aba Baffoe-Willmot, Frank Baiden, Constance Bart-Plange, Fred Binka, Michael Cappello, Daniel Chandramohan, Amanua Chinbuah, Benjamin Crookston, Ina Danquah, Stephan Ehrhardt, Johnny Gyapong, Franca Hartgers, Debbie Humphries, Shamwill Issah, Kwadwo Koram, Margaret Kweku, Keziah Malm, Juergen May, Frank Mockenhaupt, Wahjib Mohammed, Philomena Efua Nyarko, Kofi Osae, Abena Asamoabea Osei-Akoto, Seth Owusu-Agyei, Felicia Owusu-Antwi, Philip Ricks, Sylvester Segbaya, Fredericka Sey, Harry Tagbor, David van Bodegom, Mitchell Weiss

Guinea: Jane Achan, Annick Antierens, Umberto D'Alessandro, Amadou Baïlo Diallo, Mark Divall, Timothé Guilavogui , Christine Jamet , Astrid Knoblauch, Cheick Tidiane Sidibe, Amanda Tiffany, Mirko Winkler

Guinea-Bissau: Poul-Erik Kofoed, Amabelia Rodrigues, Michael Walther

Kenya: Timothy Abuya, Kubaje Adazu, Willis Akhwale, Pauline Andang'o, Ken Awuondo, Fred Baliraine, Nabie Bayoh, Philip Bejon, Simon Brooker, Maria Pia Chaparro, Jon Cox, Meghna Desai, Mark Divall, Ulrike Fillinger, Lia Smith Florey, Andrew Githeko, Carol Gitonga, Joana Greenfield, Helen Guyatt, Katherine Halliday, Mary Hamel, Laura Hammitt, Allen Hightower, Tobias Homan, Susan Imbahale, Rachel Jenkins, Chandy John, Elizabeth Juma, Lydia Kaduka, Jimmy Kahara, Akira Kaneko, Simon Kariuki, Christine Kerubo, Charles King, Chris King, Rebecca Kiptui, Astrid Knoblauch, Yeri Kombe, Feiko ter Kuile, Kayla Laserson, Tjalling Leenstra, Eugiena Lo, Brett Lowe, Claire MacIntosh, Hortance Manda, Charles Mbogo, Margaret McKinnon, Noboru Minakawa, Sue Montgomery, Eric Muchiri, Richard Mukabana, John Muriuki, Charles Mwandawiro, Joseph Mwangangi, Tabitha Mwangi, Miriam Mwjame, Charlotte Neumann, Emmily Ngetich, Patricia Njuguna, Abdisalan Mohamed Noor, Oscar Nyangari, George Nyangweso, Christopher Nyundo, Christopher Odero, Edna Ogada, Bernards Ogutu, Bernard Okech, George Okello, Maurice Ombok, Raymond Omollo, Simon Omollo, Monica Omondi, Milka Owuor, Beth Rapuoda, Evan Secor, Dennis Shanks, Larry Slutsker, Bob Snow, David Soti, Jennifer Stevenson, Willem Takken, Feiko Ter Kuile, Jacobien Veenemans, Juliana Wambua, Vincent Were, Tom Williams, Shona Wilson, Guiyun Yan, Guofa Zhou, Dejan Zurovac

Liberia: Richard Allan, Kristin Banek, Joel Jones, Tolbert Nyenswah



Madagascar: Vololomboahangy Andrianaja, Alyssa Finlay, Ronan Jambou, Thomas Kesteman, Patrice Piola, Fanjasoa Rakotomanana, Benjamin Ramarosandratana, Louise Ranaivo, Milijaona Randrianarivelojosia, Heryinsalna Honore Rasamimanana, Arsene Ratsimbasoa, Jemima Andrianina Ravelonarivo, Vincent Robert, Christophe Rogier

Malawi: Doreen Ali , Adam Bennett, Cameron Bowie, Bernard Brabin, Simon Brooker, Marian Bruce, Job Calis, Tiyese Chimuna, John Chiphwanya, P. Chirambo, James Chirombo, Maureen Coetzee, Michael Coleman, Wilfred Dodoli, Thomas Eisele, Oliver Gadabu, Paul Prinsen Geerligs, Sarah Gibson, Timothy Holtz, Gertrude Kalanda, Lawrence Kazembe, Peter Kazembe, Immo Kleinschmidt, David Lalloo, Miriam Laufer, Misheck Luhanga, Alan Macheso, Ganizani Malata, Kingsley Manda, Don Mathanga, Malcolm Molyneux, Kelias Msyamboza, Piyali Mustaphi, Themba Mzilahowa, Monica Olewe, Kamija Phiri, Arantxa Roca-Feltrer, John Sande, Andrea Sharma, Bertha Simwaka, Jacek Skarbinski , Rick Steketee, Kevin Sullivan, Terrie Taylor, Anja Terlouw, Lindsay Townes, Peter Troell, Mark Wilson, Charles Yuma, John Zoya

Mali: Modibo Bamadio, Sian Clarke, Amadou Baïlo Diallo, Diadier Diallo, Seybou Diarra, Alassane Dicko, Abdoulaye Djimde, Ogobara Doumbo, Soce Fall, Boubacar Maiga, Natalie Roschnik, Saba Rouhani, Massambou Sacko, Issaka Sagara, Mahamadou Sissoko, Ousmane Toure, Manijeh Vafa

Mauritania: Ba Mamadou dit Dialaw, Sidi Ould Zahaf

Mozambique: Pedro Aide, Pedro Alonso, Kate Brownlow, James Colborn, Michael Coleman, Alexandre Macedo De Oliveira, Mark Divall, Celine Gustavson, Albert Kilian, Immo Kleinschmidt, Astrid Knoblauch, Samuel Mabunda, Eusébio Macete, Rajendra Maharaj, Susana Nery, Emilia Virginia Noormahomed, Milka Owuor, Natalie Roschnik, Allan Schapira, Ricardo Thomson

Namibia: Colleen Fraser, Frank Hansford, John Irish, Richard Kamwi, Stark Katokele, Musa Mabaso, Kudzai Makomva, John Mendelsohn, Benson Ntomwa, Andreas Reich, Christine Theron, Petrina Uusiku

Niger: Elisa Bosque´-Oliva, Jean-Bernard Duchemin, Isabelle Jeane, Zilahatou Tohon

Nigeria: Emmanuel Adegbe, Grace Adeoye, Philip Agomo, Oluwagbemiga Aina, Bolatito Aiyenigba, Idowu Akanmu, Oladele Akogun, Chiaka Anumudu, Ebere Anyachukwu, Matthew Ashikeni, Samson Awolola, Ebenezer Baba, Oluseye Babatunde, William Brieger, Marian Bruce, Paul Emerson, Emmanuel Miri, Emmanuel Emukah, Nnenna Ezeigwe, Bayo Fatunmbi, Patricia Graves, Celia Holland, Nnaemeka Iriemenam, Albert Kilian, Paddy Kirwan, Ibrahim Maikore, Addusalami Yayo Manu, Kolawole Maxwell, Mark Miare, Audu Bala Mohammed, Gregory Noland, Stephen Oguche, Olusola Ojurongbe, Patricia Okorie, Ladipo Taiwo Olabode, Folake Olayinka, Olanpeleke Olufunke, Peter Olumese, Ogu Omede , Yusuf Omosun, Sola Oresanya, Wellington Oyibo, Lynda Ozor, Frank Richards, Abdullahi Saddiq, Adamu Sallau, Arowolo Tolu, Yemi Tosofola, Aminu Mahmoud Umar, Peter Uzoegwu

Rwanda: Caterina Fanello, Jean-Bosco Gahutu, Alain Kabayiza, Corrine Karema, Michael Loevinsohn, Frank Mockenhaupt, Alphonse Mutabazi, Alphonse Rukondo, Kevin Sifft

Senegal: Mady Ba, Christian Boudin, Badara Cisse, Sian Clarke, Ndiacé Dangoura, Mame Birame Diouf, Ibrahim Socé Fall, Florie Fillol, Oumar Gaye, Ibou Guisse, Sylvia Males, Florence Migot-Nabias, Paul Milligan, Rick Paul, Moussa Thior, Jean Francois Trape



Sierra Leone: Matthew Burns, Mark Divall, Astrid Knoblauch, Edward Magbity, Musa Sillah-Kanu, Samuel Smith

Somalia: Abdi Abdillahi, Ali Abdulrahmann, Jamal Amran, Imanol Berakoetxea, Mohammed Borle, Matthew Burns, Waqar Butt, Craig Von Hagen, Abdikarim Hussein Hassan, Abdi Hersi, Grainne Moloney, Abdisalaam Mohamed Noor, Abdi Noor, Ismail Rage, Abdulhamid Osman Salah, Tanya Shewchuk, Miriam Warsame, Randa Youssef, Fahmi Essa Yusuf

South Africa: Karen Barnes, Marlies Craig, Frank Hansford, Gerdalize Kok, Philip Kruger, Dave Le Sueur, Aaron Mabuza, Rajendra Maharaj, Brian Sharp

South Sudan: Robert Azairwe, Steve Barya, Emmanuel Chanda, Bill Gueth Kueil, Richard Laku, Margaret Lejukole, Olivia Lomoro, Charles Agono Mona, Harriet Pasquale, Samuel Patti, Heidi Reid

Sudan: Mohamed Abbas, Alnazear Abdalla, Tareg Abdelgader, Nasruddin Abdul-Hadi, Abdalla Ahmed, Mubashar Ahmed, Dalin Abdelkareem Altahir, Sahar Bakhite, Mustafa Dukeen, Tayseer Elamin El Faki, Asma Hsahim El Hassan, Ibrahim El Hassan, Limya El Yamani, Khalid Elmardi, Salah Elbin Elmubark, Homooda Totoa Kafy, Fatih Malik, Jaffar Mirghani, Alaa Moawia, Tasneem Moawia, Abdullah Sayied Mohammed, Maowia Mukhtar, Fazza Munim, Samia Seif Murghan, Ali Elamin Nasir, Abdisalan Mohamed Noor, Bakari Nour, Abdelhameed Elbirdiri Nugad, A. Omer, Abdala Sayaid, Jihad Eltaher Sulieman, Mohamad Tarig, Randa Youssef, Ghasem Zamani

Swaziland: Sabelo Dlamini, Frank Hansford, Simon Kunene, Joe Novotny, Graham Root, Brian Sharp

Tanzania: Salim Abdulla, Mike Bangs, Jubilate Barnard, James Beard, Michael Beasley, Anders Bjorkman, Teun Bousema, Ilona Carneiro, Frank Chacky , Prosper Chaki, Daniel Chandromohan, Mark Divall, Stefan Dongus, Chris Drakeley, Mathew Dukes, Yvonne Geissbühler, Nicodem J Govella, Francesco Grandesso, George Greer, Kara Hanson, Deus Ishengoma, Timona Jarha, Patrick Kachur, Rose Kibe, Gerry Killeen, Safari Kinunghi, William Kisinza, Samson Kiware, Astrid Knoblauch, Karen Kramer, Paul Lango, Martha Lemnge, Christian Lengeler, Neil F Lobo, Zudson Lucas, Rose Lusinde, John Lusingu, Zawadi Mageni, Stephen Magesa, Robert Malima, Alpha D Malishee, Renata Mandike, Tanya Marchant, Honoratia Masanja, Fabian Mashauri, Fabiano Massawe, Caroline Maxwell, Ben Mayala, Leonard Mboera, Peter McElroy, Clara Menendez, Sigsbert Mkude, Yeromin Mlacha, Bruno Mmbando, Ally Mohammed, Fabrizio Molteni, Daniel Msellemu, Hassan Mshinda, Frank Mtei, Zacharia J Mtema, Athuman Muhili, Theonest Mutabingwa, Victoria Mwakalinga, Dismas Mwalimu, Yusufu Mwita, Irene Mwoga, Isaack Namango, Kenneth Nchimbi, Rhita Njau, Fredros Okumo, John Owuor, Milka Owuor, Faith Patrick, Lynn Paxton, Hugh Reyburn, Tanya L Russell, Sunil Sazawa, David Schellenberg, Self Shekalaghe, Clive Shiff, Method Segeja, Rosemary Silaa, Thomas Smith, Paul Smithson, Bob Snow, José Sousa-Figueiredo, Thor Theander, Andrew Tomkins, Patrick Tungu, Jacobien Veenemans, Hans Verhoef, Ying Zhou

Togo: Kodjo Morgah, Anja Terlouw

Uganda: Jane Achan, Seraphine Adibaku, Miriam Akello, Paul Ametepi, Paul Bazongere, Lea Berrang-Ford, Martha Betson, Teun Bousema, Clare Chandler, Jessica Cohen, Jon Cox, Deborah DiLiberto, Grant Dorsey, Calvin Echodu, Dorothy Echodu, Thomas Egwang, Allison Elliot, Anthony Esenu, Samuel Gonahasa, Francesco Grandesso, Jean-Paul Guthmann, Stephen Hillier, Katy Hurd, Narcis Kabatereine, Rita Kabuleta-Luswata, Mark Kaddumuka, Ruth Kigozi, Simon Kigozi, Macklyn Kihembo, Fred Kironde, Steve Kiwuwa, Moses Kizza, Jan Kolaczinski, Mary Kyohere, Steve Lindsay, Myers Lugemwa, Caroline Lynch, Godfrey Magumba, Catherine Maiteki, Catherine Maiteki-



Sebuguzi, Edith Mbabazi, Monami Patrick, Levi Mugenyi, Lawrence Muhangi, Carolyn Nabasumba, Halima Naiwumbwe, Marjorie Najjengo, Zaria Nalumansi, Florence Nankya, Sussanne Nasr, Dida Manya, Florence Nankya, Sussann Nasr, Juliet Ndibazza, Gloria Oduru, Michael Okia, Jaffer Okiring, Peter Okui, Ambrose Onapa, Niels Ornbjerg, Erling Pedersen, Carla Proietti, Rachel Pullen, Andrea Rehman, Denis Rubahika, John Rwakimari, Indrani Saran, Paul Simonsen, Bob Snow, James Ssekitoleeko, Sarah Staedke, Claire Standley, Laura Steinhardt, Anna-Sofie Stensgaard, Russell Stothard, Ambrose Talisuna, James Tibenderana, Henry Wannume, Emily Webb, Adoke Yeka, Charlotte Muheki Zikusooka

Zambia: Pascalina Chanda, Elizabeth Chizema, Michael Coleman, Ruben Conner, Umberto D'Alessandro, Mark Divall, Thomas Eisele, Lindsey Everett, Jean-Pierre Van Geertruyden, David Hamer, Aniset Kamanga, Mulakwa Kamuliwo, Sera Karuiki, Joe Keating, Immo Kleinschmidt, Astrid Knoblauch, David Larsen, Sungano Mharakurwa, John Miller, Victor Mukona, Modest Mulenga, Boniface Mutombo, Mercie Mwanza, Michael Nambozi, Eric Njunju, Milka Owuor, Travis Porter, Richard Steketee, Philip Thuma

Zimbabwe: Tim Freeman, Nicholas Midzi, Francisca Mutapi, Graham Root, Crispin Lumbala



SUPPLEMENTARY INFORMATION - Nature · 1.2 Background to malaria parasite prevalence 1.3 Assembling malaria survey data 1.3.1 Archive searches ... states of Sao Tome and Principe,

Documents