Over 25 million people in sub-Saharan Africa live with HIV. After reporting its first AIDS case in 1985, Botswana is one of the most severely affected countries in the region, with one in five adults now living with HIV.

Movement of the population is likely to have contributed to a geographically dispersed, and high-prevalence, HIV epidemic in Botswana. Since 1985, urbanization, rapid economic and population growth, and migration have transformed Botswana. Yet, few studies have analyzed the role of population-level movement patterns in the spread of HIV during this time.

By studying micro-census data from Botswana between 1981 and 2011, Song et al. found that the country’s population was highly mobile during this period. Reconstructions of internal migration patterns show very high rates of rural-to-urban and urban-to-rural migration, with 10% of Botswana’s population moving each year. The first reported AIDS cases in Botswana occurred in mining towns and cities where high-risk behavior was prevalent. These areas were also migration hubs during this period and could have contributed to the rapid spread of HIV throughout the country as infected individuals moved back to rural districts.

Understanding human migration patterns and how they affect the spread of infectious diseases using current data could help public health authorities in Botswana and additional sub-Saharan African countries design control strategies for HIV and other important infections that occur in the region.

Over 25 million people live with HIV infection in sub-Saharan Africa. All of the HIV epidemics in this continent are generalized: in these types of epidemics, the epidemic is dispersed throughout the country. Therefore, a population needs to be highly mobile in order for a generalized epidemic to have developed. However, although the epidemiology of HIV in sub-Saharan Africa has been widely studied (Farley et al., 2022; Read et al., 2022), the role of population-level mobility patterns in the development of generalized HIV epidemics has not been assessed for any country on the continent. Botswana has one of the most severe HIV epidemics worldwide and reported its first AIDS case in the early 1980s. Since then, the epidemic has become generalized and hyperendemic: in 2021, HIV prevalence in adults (15–64) was 21% (Mine et al., 2022). Simultaneous to the development of its generalized HIV epidemic, Botswana has undergone rapid urbanization. It was predominantly rural in 1981, but had become predominantly urban by 2011 (Statistics Botswana, 2014); by that time Botswana’s HIV epidemic had stabilized (CSO Botswana, 2009). We hypothesize that – during the development of Botswana’s epidemic – the population was extremely mobile and highly connected by substantial urban-to-rural and rural-to-urban migratory flows. We test this mobility hypothesis by conducting a network analysis using a historical time series of micro-census data collected in Botswana. These data contain information both on migration and urbanization. The time series covers the time period from when the epidemic was first apparent to when it stabilized in 2011. We use these data to (i) estimate the annual incidence (at the national level) of internal migration over three decades (1981–2011), (ii) characterize migrants on the basis of gender and age, (iii) reconstruct internal migration networks (in order to identify large-scale population movements and connectivity patterns), (iv) identify migration hubs, (v) understand and visualize the role of migration networks in urbanization, and (vi) determine the extent to which internal migration led to a substantial geographic redistribution of the population. The Government of Botswana defines internal migration as residents changing their place of permanent residence within their home country. Finally, we discuss how the historical migration networks – that we identify in our analyses – had the potential to disperse HIV throughout Botswana and generate a hyperendemic epidemic.

After 80 y of colonial rule, Botswana gained independence from the United Kingdom in 1966 and has since been Africa’s longest uninterrupted democracy. One of the poorest countries in the world at the time of independence, Botswana has since undergone rapid urbanization and economic expansion; the diamond mining industry is the economic foundation of Botswana (Barnett et al., 2002). Rich diamond deposits were discovered immediately after independence; the first diamond mine (Orapa) became fully operational in July 1971, the second mine (Jwaneng) in August 1982. These two mines are now amongst the world’s richest diamond mines. Today, Botswana is an upper-middle-income country with a high demographic growth rate: the population approximately doubled from 1 to 2 million, from 1981 and 2011 (Statistics Botswana, 2014).

In 1981, the vast majority of the population lived in small rural villages. Only ~18% of the population were living in urban areas (Statistics Botswana, 2014): either in one of the two cities (Gaborone, the capital, or Francistown) or in one of the four towns (Lobatse, Selebi Phikwe, Orapa, and Jwaneng). Lobatse is an administrative center, and the other three are mining towns. Selebi Phikwe is based on copper and nickel mining, and was founded in the early 1970s. By 2011, only ~36% of the population were living in rural areas (Statistics Botswana, 2014). The remaining population were living in one of the three types of urban centers: a city (there were still only two cities), a town (there were now five towns following the 1991 addition of Sowa, a mining town for soda ash), or an urban village. In Botswana, an urban village is defined as a settlement with at least 5,000 individuals and 75% of the workforce engaged in non-agricultural economic activities (Statistics Botswana, 2014). The urban villages developed by in situ urbanization, which is defined as rural settlements transforming into urban areas by expanding their non-agricultural activities and increasing economic linkages with neighboring areas (Moriconi-Ebrard et al., 2020). Botswana now consists of 28 administrative districts (Okano et al., 2021). Each city and town are separate administrative districts, and the remaining 21 districts each contain at least one urban village and many rural villages.

A great deal is known about the epidemiology of HIV in Botswana in terms of the temporal increase, and geographic variation, in prevalence (Magosi et al., 2022; Novitsky et al., 2015; Novitsky et al., 2020; Okano et al., 2021). HIV prevalence rose quickly and spread fairly widely. The first case of AIDS was reported in 1985 from the mining town of Selebi Phikwe (African Natural Resources Center, 2016) in the late 1980s; additional cases were reported from the two diamond mining towns (Jwaneng and Orapa) and the city, Francistown. In 1990, the first HIV Sentinel Surveillance Survey of antenatal clinic (ANC) attendees (15–49years old) was conducted in Gaborone and in one rural district (Boteti): HIV prevalence was found to be 6% and 4%, respectively (UNAIDS and World Health Organization, 2004). In 1992, a national sentinel surveillance survey began and was conducted annually to 2011. National prevalence in pregnant women was found to have already reached a very high level (18%) by 1992 and continued to increase for the next 8 y: 23% by 1993, 32% by 1995, and 39% by 2000 (UNAIDS and World Health Organization, 2004). Based on the ANC data, HIV prevalence in the cities and the mines was the highest. In Francistown, prevalence was 8% in 1991, 24% by 1992, and 44% by 2000. In Gaborone, prevalence rose from 15% in 1992 to 36% by 2000. In the mining town of Selebi Phikwe, prevalence rose from 27% in 1994 to 50% by 2000. By the late 1990s, a high AIDS morbidity and mortality rate in miners led the diamond mining company Debswana to conduct an HIV prevalence survey of its (male) employees. The survey took place in 1999: 29% of miners were found to be infected with HIV (Barnett et al., 2002). The first population-level survey in Botswana to test participants for HIV (BAIS II, the Botswana AIDS Impact Survey) was conducted in 2004 (NACA and CSO Botswana, 2005). It found, at that time, that HIV prevalence in the general population was 29% in women and 20% in men, aged 15–49 years old. Gender-stratified prevalence was the same in 2008, when BAIS III was conducted; the considerable geographic variation in HIV prevalence amongst districts that existed at that time is shown in Figure 1. Notably, phylogenetic studies have shown that viral lineages have dispersed widely throughout Botswana, suggesting that large-scale population-level movements have occurred (Magosi et al., 2022; Novitsky et al., 2015; Novitsky et al., 2020).

Figure 1

Download asset Open asset

Our results support our mobility hypothesis that – during the development of Botswana’s generalized HIV epidemic (i.e., between 1981 and 2011) – the population was extremely mobile and the country was highly connected by urban-to-rural migratory flows and rural-to-urban counter-flows. Using our time series of historical data, we found ~10% of the population moved their residency in the 12 mo before each census in 1981, 1991, 2001, and 2011. The constancy of this value at each census suggests that the annual migration rate in Botswana was stable at ~10% for each year between 1981 and 2011. Notably, the type of migrants also remained constant over this time period: evenly split by gender, with younger people more likely to migrate. We found that migration occurred within districts, as well as between districts, but that migration between districts was more common. Even in 1981, at the beginning of the period of rapid urbanization, the migration network was highly connected and therefore large-scale population movements linked almost all of the districts in the country. Notably, we found at each census, that although the number of urbanized districts increased (as did their population size) the vast majority of migratory flows between the rural and developing urban districts were balanced by counter-flows.

Our results on migration networks show how a generalized epidemic had the potential to develop: HIV could have diffused, via the migration networks, throughout Botswana. The first cases of AIDS in Botswana were reported in the 1980s (African Natural Resources Center, 2016); this suggests that HIV may have been circulating in Botswana in the 1970s, as HIV has an incubation period of 10–12 y (Hendriks et al., 1993; Muñoz et al., 1989). One phylogenetic study indicates that HIV was introduced into Botswana in the 1960s, but exponential growth did not begin until the 1980s (Wilkinson et al., 2015). Another study suggests that HIV spread southward from Kinshasa (the epicenter of the HIV epidemic) in the Democratic Republic of the Congo via travelers on the railroad that was built in the 1800s (Faria et al., 2014). Notably, this railroad goes through Francistown, which is one of Botswana’s two cities, and where HIV prevalence quickly rose to high levels in the early 1990s (UNAIDS and World Health Organization, 2004). It is possible that HIV was introduced into the mining towns of Botswana from overseas or South Africa. In the 1970s and 1980s, many young men from Botswana went to work in the mines in South Africa, returning as the mining industry became established in their home country. Our results show that, by 1981, the migration networks (for women and men) linked almost every district in Botswana. This indicates that HIV-infected migrants from Francistown and/or the mining towns (where AIDS cases were first reported and HIV prevalence quickly rose to high levels) could have seeded HIV sub-epidemics in multiple rural districts throughout the country. HIV could then have spread widely within these districts due to the high rates of within-district migration that we have found. Due to the high and continuous migration rates, the sub-epidemics in rural districts were likely to have been seeded numerous times by HIV-infected migrants. When HIV was first introduced into rural districts, the sub-epidemics in these districts may have been maintained by source-sink dynamics (Okano et al., 2020): districts where transmission was high enough to be self-sustaining (e.g., mining towns) could have maintained sub-epidemics in rural districts where transmission was too low to be self-sustaining. Our results suggest that migratory flows, and counter-flows, between high and low prevalence areas (e.g., between the mining towns and rural areas, or between Botswana’s two cities and rural areas) are likely to have existed, at least from the early 1980s onward. These flows would have functioned as ‘transmission corridors’ (Okano et al., 2021), and created high-risk flows that crossed the country (Valdano et al., 2021), that is, HIV-infected people migrating from high-prevalence to lower-prevalence areas, and uninfected people migrating from low-prevalence areas to higher-prevalence areas. Taken together, our results show that the highly connected migration networks, with high migratory flows and counter-flows, could have contributed to the development of Botswana’s generalized epidemic by distributing HIV throughout the country in the general population.

Our results on migration networks potentially explain how Botswana’s HIV epidemic became hyperendemic. Migrants have been shown to be at high risk of acquiring HIV infection (Camlin and Charlebois, 2019; Dobra et al., 2017; Olawore et al., 2018); our results suggest that, due to migration alone, each year ~10% of the population of Botswana was at high risk. The migrants in Botswana were young women and men, a group generally at the highest risk of HIV infection through sexual transmission. Mining towns and cities would have been transmission hot-spots (for both women and men) due to high levels of risky sexual behavior (as evidenced by high HIV prevalence; Barnett et al., 2002; UNAIDS and World Health Organization, 2004) and a high probability of encountering an HIV-infected individual as a sex partner. Due to the high prevalence of HIV, high turnover rates, and a high level of connectivity to other districts, the mining towns and cities could have functioned as geographically defined core groups for sexual transmission (for both women and men). Some individuals could have been part of a high-risk core group when they were living in the cities and mining towns, and decreased their risky behavior when they returned to the rural districts. For example, FSWs in Botswana can have an extremely high number of sex partners in mining towns and urban areas (~7 per week) and remain as FSWs for ~4 y (Ministry of Health Botswana, 2013). Notably we found that the mining towns and cities were both migration in-flow hubs and migration out-flow hubs. The fact that they were migration in-flow hubs would have led to a high in-flow of uninfected individuals from rural districts where prevalence was low; the fact that they were migration out-flow hubs (and had a very high prevalence of HIV) would have led to a high out-flow of HIV-infected individuals that could have seeded, and reseeded, sub-epidemics in rural districts. Therefore, rural-to-urban migration coupled with urban-to-rural migration could have been a very important driver in the Botswana epidemic becoming hyperendemic. Without it, prevalence could have reached very high levels in the mining towns and cities, but remained fairly low in rural districts.

We have discussed the potential impact of mobility on Botswana’s HIV epidemic, and focused on migration, that is, one directional movement in terms of a permanent relocation of residency. This type of migration, by changing an individual’s sexual network and social environment, has been shown to increase the risk of acquiring HIV for both women and men (Anglewicz et al., 2018; Camlin and Charlebois, 2019; Dzomba et al., 2019; Low et al., 2021; Olawore et al., 2018; Thorp et al., 2023). Short-term mobility (e.g., short-term circular migration, where the trip can range in duration from overnight to an entire season) can also affect HIV transmission dynamics. Circular migrants have been shown to both have an increased risk of acquiring HIV, and of transmitting HIV (Camlin and Charlebois, 2019; Palk and Blower, 2015). The greater the number of trips and/or the duration of the trip, the greater the risk (Palk and Blower, 2015). We note that both migration and short-term mobility are important, and their relative importance to each other is likely to evolve over time as a generalized HIV epidemic diffuses through the population. Their relative importance is also likely to vary amongst countries in sub-Saharan Africa.

Our results also provide insights into the role of Botswana’s internal migration network during a period of rapid urbanization. The urbanization process was first studied quantitatively by Ravenstein, who analyzed the 1871 and 1881 censuses of Great Britain and Ireland, and calculated the flows of lifetime migrants to urban centers (Ravenstein, 1885). He showed that internal migration from rural areas was essential to the growth of industrial cities and towns. Additionally, he demonstrated that migration was extremely important in changing settlement patterns and geographically redistributing the population in both Great Britain and Ireland. In contrast, we found at each census, that generally the majority of rural-to-urban migratory flows were balanced by counter-flows from urban areas to rural villages. Consequently, even though there was a very high migration rate, we found that migration did not drive urbanization, and also had relatively little impact on geographically redistributing Botswana’s population. Differences between Botswana and high-income countries, in the role of internal migration on urbanization, were due to in situ urbanization occurring in Botswana (rather than the rapid growth of cities and towns as in high-income countries) and high rural-to-urban migratory flows being balanced by high urban-to-rural counter-flows. During the 30-year time period that we have analyzed (1981–2011), population growth in Botswana was high: the population doubled. Notably, the Organization for Economic Co-operation and Development recently reported that since 1990, in situ urbanization and high population growth rates have driven Africa’s rapid growth in urbanization (Moriconi-Ebrard et al., 2020). Countries are at different stages of urbanization, but urbanization is occurring in essentially all African countries (United Nations, Department of Economic and Social Affairs, & Population Division, 2019).

Our results on the role of migration on urbanization have significant implications for predicting migratory flows in Africa. Africa has high migration rates and is one of the least urbanized places in the world (United Nations, Department of Economic and Social Affairs, & Population Division, 2019): the number of urban residents on the continent approximately doubled between 1995 and 2015, and is projected to double again by 2040 (Lall et al., 2017). Consequently, it is important to be able to predict migration patterns and flows in Africa (Okano et al., 2021); for example, for urban planning and developing efficient resource allocation strategies. The mathematical models currently being used to predict migratory flows in Africa (Ciavarella and Ferguson, 2021; Marshall et al., 2018; Meredith et al., 2021) have been developed based on migratory processes that have been observed, and census data that have been collected, in high-income countries. The two most commonly used predictive models, the gravity model (Stewart, 1948) and the radiation model (Simini et al., 2012), make the central assumption that individuals are attracted to move to a new location based on the population size at the destination relative to the size at the origin, moving from the smaller to the larger population. Under this assumption, it is expected that we would have found asymmetry in migration flows between rural-to-urban and urban-to-rural areas, and a migration-induced geographic redistribution of the population. However, we found almost symmetric migratory counter-flows occurring throughout the country, and that migration had almost no impact on changing the geographic distribution of the population. Therefore, our results suggest that new mathematical models are needed to predict migration in Botswana, and – potentially – in other rapidly urbanizing countries in Africa where in situ urbanization is occurring.

Our study has several limitations. First, we have examined only internal migration (i.e., migration that takes place within a country) and have not included international migration. However, for the three-decade time period that we have investigated, the number of international migrants to Botswana was much smaller than the number of internal migrants (Statistics Botswana, 2014; Statistics Botswana, 2015). Second, all of our analyses are focused on Botswana in order to analyze the impact of migration networks on the development of a generalized hyperendemic epidemic; we have not evaluated migration networks in other African countries. We recommend that studies, such as we have conducted here, are conducted in these other countries by analyzing a time series of historical micro-census data, especially in other countries in sub-Saharan Africa that have generalized hyperendemic HIV epidemics, such as Eswatini and Lesotho. Third, a potential limitation is that the IPUMS-International (Integrated Public Use Microdata Series-International) data that we used are samples rather than the complete censuses; we used this approach as it is generally not possible to obtain complete census data due to privacy reasons. Fourth, while we have gained insights into understanding the development of the generalized hyperendemic HIV epidemic in Botswana, we have not conducted an analysis of the transmission dynamics of the epidemic in order to explore these insights quantitatively. This is the focus of current research in which we are analyzing a geospatial mathematical model of the epidemiological evolution of the HIV epidemic in Botswana. Finally, we have examined only one process (internal migration) that has affected the evolution of Botswana’s epidemic; other factors could also have been important.

All HIV epidemics in SSA are generalized; HIV can only diffuse through a population by the movement of people. To our knowledge, our study is the first to study population-level mobility patterns and rates in any country in sub-Saharan Africa. Taken together, our results identify particular characteristics of migration networks in Botswana that could explain why HIV prevalence rose to very high levels in many districts throughout the country. Migration networks in other countries with generalized HIV epidemics may be very different. If mining towns and/or cities in these countries were not migratory in-flow and out-flow hubs with high turnover of their populations, as in Botswana, they would not have served as major sources of infection for dispersing HIV throughout the general population. Botswana has recently achieved the 95-95-95 targets (Mine et al., 2022) that UNAIDS specified needed to be reached by 2030 in order to eliminate HIV: 95% of people living with HIV need to be diagnosed, 95% of the diagnosed need to be on treatment, and 95% of those on treatment need to be virally suppressed. However, transmission is continuing (Magosi et al., 2022), as is urbanization and large-scale population movements (Okano et al., 2021). The same processes that we have identified that may have contributed to the development of the generalized hyperendemic HIV epidemic in Botswana may prevent the elimination of HIV.

All data needed to evaluate the conclusions in the paper are presented in the paper, or freely available for registered users at the IPUMS International website: https://doi.org/10.18128/D020.V7.3. We note that we did not collect these data, nor are they permitted to be posted to other repositories. All code needed to reproduce all parts of this analysis are available from the first author's GitHub page: https://github.com/janetsong80/pop-mobility-botswana-hiv (copy archived at: https://zenodo.org/badge/latestdoi/594401539).

1. 1. Minnesota Population Center

   (2020) IPUMS: Version 7.3

   Integrated Public Use Microdata Series, International.

   https://doi.org/10.18128/D020.V7.3

1. Website

   1. African Natural Resources Center

   (2016) Debswana Diamond Company and the Government of Botswana: an HIV/AIDS public-private partnership programme

   Accessed October 17, 2022.

   https://www.afdb.org/fileadmin/uploads/afdb/Documents/Publications/anrc/DEBSWANA_ENG__HR_PAGES.pdf
2. 1. Anglewicz P
   2. VanLandingham M
   3. Manda-Taylor L
   4. Kohler HP

   (2018) Health Selection, Migration, and HIV Infection in Malawi

   Demography 55:979–1007.

   https://doi.org/10.1007/s13524-018-0668-5

   • PubMed
   • Google Scholar
3. Book

   1. Barnett T
   2. Fantan T
   3. Mbakile B
   4. Whiteside A

   (2002)

   The Private Sector Responds to the Epidemic: Debswana–a Global Benchmark

   Geneva, Switzerland: UNAIDS.

   • Google Scholar
4. 1. Bell M
   2. Blake M
   3. Boyle P
   4. Duke-Williams O
   5. Rees P
   6. Stillwell J
   7. Hugo G

   (2002) Cross-national comparison of internal migration: Issues and measures

   Journal of the Royal Statistical Society 165:435–464.

   https://doi.org/10.1111/1467-985X.t01-1-00247

   • Google Scholar
5. 1. Camlin CS
   2. Charlebois ED

   (2019) Mobility and its effects on HIV acquisition and treatment engagement: Recent theoretical and empirical advances

   Current HIV/AIDS Reports 16:314–323.

   https://doi.org/10.1007/s11904-019-00457-2

   • PubMed
   • Google Scholar
6. 1. Ciavarella C
   2. Ferguson NM

   (2021) Deriving fine-scale models of human mobility from aggregated origin-destination flow data

   PLOS Computational Biology 17:e1008588.

   https://doi.org/10.1371/journal.pcbi.1008588

   • PubMed
   • Google Scholar
7. Website

   1. CSO Botswana

   (2009) 2008 Botswana AIDS Impact Survey III: Statistical Report

   Accessed October 12, 2022.

   https://microdata.statsbots.org.bw/index.php/catalog/13/download/77
8. 1. Dobra A
   2. Bärnighausen T
   3. Vandormael A
   4. Tanser F

   (2017) Space-time migration patterns and risk of HIV acquisition in rural South Africa

   AIDS 31:137–145.

   https://doi.org/10.1097/QAD.0000000000001292

   • PubMed
   • Google Scholar
9. 1. Dzomba A
   2. Tomita A
   3. Govender K
   4. Tanser F

   (2019) Effects of migration on risky sexual behavior and HIV acquisition in South Africa: A systematic review and meta-analysis, 2000-2017

   AIDS and Behavior 23:1396–1430.

   https://doi.org/10.1007/s10461-018-2367-z

   • PubMed
   • Google Scholar
10. 1. Faria NR
    2. Rambaut A
    3. Suchard MA
    4. Baele G
    5. Bedford T
    6. Ward MJ
    7. Tatem AJ
    8. Sousa JD
    9. Arinaminpathy N
    10. Pépin J
    11. Posada D
    12. Peeters M
    13. Pybus OG
    14. Lemey P

    (2014) HIV epidemiology. The early spread and epidemic ignition of HIV-1 in human populations

    Science 346:56–61.

    https://doi.org/10.1126/science.1256739

    • PubMed
    • Google Scholar
11. 1. Farley SM
    2. Wang C
    3. Bray RM
    4. Low AJ
    5. Delgado S
    6. Hoos D
    7. Kakishozi AN
    8. Harris TG
    9. Nyirenda R
    10. Wadonda N
    11. Li M
    12. Amuri M
    13. Juma J
    14. Kancheya N
    15. Pietersen I
    16. Mutenda N
    17. Natanael S
    18. Aoko A
    19. Ngugi EW
    20. Asiimwe F
    21. Lecher S
    22. Ward J
    23. Chikwanda P
    24. Mugurungi O
    25. Moyo B
    26. Nkurunziza P
    27. Aibo D
    28. Kabala A
    29. Biraro S
    30. Ndagije F
    31. Musuka G
    32. Ndongmo C
    33. Shang J
    34. Dokubo EK
    35. Dimite LE
    36. McCullough-Sanden R
    37. Bissek A-C
    38. Getaneh Y
    39. Eshetu F
    40. Nkumbula T
    41. Tenthani L
    42. Kayigamba FR
    43. Kirungi W
    44. Musinguzi J
    45. Balachandra S
    46. Kayirangwa E
    47. Ayite A
    48. West CA
    49. Bodika S
    50. Sleeman K
    51. Patel HK
    52. Brown K
    53. Voetsch AC
    54. El-Sadr WM
    55. Justman JJ

    (2022) Progress towards the UNAIDS 90-90-90 targets among persons aged 50 and older living with HIV in 13 African countries

    Journal of the International AIDS Society 25 Suppl 4:e26005.

    https://doi.org/10.1002/jia2.26005

    • PubMed
    • Google Scholar
12. 1. Hendriks JC
    2. Medley GF
    3. van Griensven GJ
    4. Coutinho RA
    5. Heisterkamp SH
    6. van Druten HA

    (1993) The treatment-free incubation period of AIDS in a cohort of homosexual men

    AIDS 7:231–239.

    https://doi.org/10.1097/00002030-199302000-00012

    • PubMed
    • Google Scholar
13. Book

    1. Lall SV
    2. Henderson JV
    3. Venables AJ

    (2017) Africa’s Cities: Opening Doors to the World

    World Bank Publications.

    https://doi.org/10.1596/978-1-4648-1044-2

    • Google Scholar
14. 1. Low A
    2. Sachathep K
    3. Rutherford G
    4. Nitschke AM
    5. Wolkon A
    6. Banda K
    7. Miller LA
    8. Solmo C
    9. Jackson K
    10. Patel H
    11. McCracken S
    12. Findley S
    13. Mutenda N

    (2021) Migration in Namibia and its association with HIV acquisition and treatment outcomes

    PLOS ONE 16:e0256865.

    https://doi.org/10.1371/journal.pone.0256865

    • PubMed
    • Google Scholar
15. 1. Magosi LE
    2. Zhang Y
    3. Golubchik T
    4. DeGruttola V
    5. Tchetgen Tchetgen E
    6. Novitsky V
    7. Moore J
    8. Bachanas P
    9. Segolodi T
    10. Lebelonyane R
    11. Pretorius Holme M
    12. Moyo S
    13. Makhema J
    14. Lockman S
    15. Fraser C
    16. Essex MM
    17. Lipsitch M
    18. Botswana Combination Prevention Project and PANGEA consortium

    (2022) Deep-sequence phylogenetics to quantify patterns of HIV transmission in the context of a universal testing and treatment trial - BCPP/Ya Tsie trial

    eLife 11:e72657.

    https://doi.org/10.7554/eLife.72657

    • PubMed
    • Google Scholar
16. 1. Marshall JM
    2. Wu SL
    3. Sanchez C HM
    4. Kiware SS
    5. Ndhlovu M
    6. Ouédraogo AL
    7. Touré MB
    8. Sturrock HJ
    9. Ghani AC
    10. Ferguson NM

    (2018) Mathematical models of human mobility of relevance to malaria transmission in Africa

    Scientific Reports 8:7713.

    https://doi.org/10.1038/s41598-018-26023-1

    • PubMed
    • Google Scholar
17. 1. Meredith HR
    2. Giles JR
    3. Perez-Saez J
    4. Mande T
    5. Rinaldo A
    6. Mutembo S
    7. Kabalo EN
    8. Makungo K
    9. Buckee CO
    10. Tatem AJ
    11. Metcalf CJE
    12. Wesolowski A

    (2021) Characterizing human mobility patterns in rural settings of sub-Saharan Africa

    eLife 10:e68441.

    https://doi.org/10.7554/eLife.68441

    • PubMed
    • Google Scholar
18. Conference

    1. Mine M
    2. Stafford K
    3. Laws R
    4. Marima R
    5. Lekone P
    6. Ramaabya D
    7. Makhaola K
    8. Mapondera P
    9. Wray-Gordon F
    10. Agbakwuru C

    (2022)

    Botswana achieved the Joint United Nations Programme on HIV/AIDS (UNAIDS) 95-95-95 targets: results from the Fifth Botswana HIV/AIDS Impact Survey

    The 24th International AIDS Conference.

    • Google Scholar
19. Website

    1. Ministry of Health Botswana

    (2013) 2012 Mapping, Size Estimation & Behavioral and Biological Surveillance Survey (BBSS) of HIV/STI Among Select High-Risk Sub-Populations in Botswana

    Accessed October 3, 2022.

    https://www.fhi360.org/sites/default/files/media/documents/BBSS
20. Website

    1. Minnesota Population

    (2020) Integrated Public Use Microdata Series (IPUMS)

    Accessed June 4, 2020.

    https://doi.org/10.18128/D020.V7.3
21. Book

    1. Moriconi-Ebrard F
    2. Heinrigs P
    3. Trémolières M

    (2020) Africa’s Urbanisation Dynamics 2020: Africapolis, Mapping a New Urban Geography

    Paris: OECD.

    https://doi.org/10.1787/b6bccb81-en

    • Google Scholar
22. 1. Muñoz A
    2. Wang MC
    3. Bass S
    4. Taylor JM
    5. Kingsley LA
    6. Chmiel JS
    7. Polk BF

    (1989) Acquired immunodeficiency syndrome (AIDS)-free time after human immunodeficiency virus type 1 (HIV-1) seroconversion in homosexual men. Multicenter AIDS Cohort Study Group

    American Journal of Epidemiology 130:530–539.

    https://doi.org/10.1093/oxfordjournals.aje.a115367

    • PubMed
    • Google Scholar
23. Website

    1. NACA
    2. CSO Botswana

    (2005) Botswana AIDS Impact Survey II

    Accessed October 12, 2022.

    https://www.statsbots.org.bw/sites/default/files/stats_report%282%29.pdf
24. 1. Novitsky V
    2. Kühnert D
    3. Moyo S
    4. Widenfelt E
    5. Okui L
    6. Essex M

    (2015) Phylodynamic analysis of HIV sub-epidemics in Mochudi, Botswana

    Epidemics 13:44–55.

    https://doi.org/10.1016/j.epidem.2015.07.002

    • PubMed
    • Google Scholar
25. 1. Novitsky V
    2. Zahralban-Steele M
    3. Moyo S
    4. Nkhisang T
    5. Maruapula D
    6. McLane MF
    7. Leidner J
    8. Bennett K
    9. PANGEA Consortium
    10. Wirth KE
    11. Gaolathe T
    12. Kadima E
    13. Chakalisa U
    14. Pretorius Holme M
    15. Lockman S
    16. Mmalane M
    17. Makhema J
    18. Gaseitsiwe S
    19. DeGruttola V
    20. Essex M

    (2020) Mapping of HIV-1C transmission networks reveals extensive spread of viral lineages across villages in Botswana treatment-as-prevention trial

    The Journal of Infectious Diseases 222:1670–1680.

    https://doi.org/10.1093/infdis/jiaa276

    • PubMed
    • Google Scholar
26. 1. Okano JT
    2. Sharp K
    3. Valdano E
    4. Palk L
    5. Blower S

    (2020) HIV transmission and source-sink dynamics in sub-Saharan Africa

    The Lancet. HIV 7:e209–e214.

    https://doi.org/10.1016/S2352-3018(19)30407-2

    • PubMed
    • Google Scholar
27. 1. Okano JT
    2. Busang L
    3. Seipone K
    4. Valdano E
    5. Blower S

    (2021) The potential impact of country-level migration networks on HIV epidemics in sub-Saharan Africa: the case of Botswana

    The Lancet. HIV 8:e787–e792.

    https://doi.org/10.1016/S2352-3018(21)00267-8

    • PubMed
    • Google Scholar
28. 1. Olawore O
    2. Tobian AAR
    3. Kagaayi J
    4. Bazaale JM
    5. Nantume B
    6. Kigozi G
    7. Nankinga J
    8. Nalugoda F
    9. Nakigozi G
    10. Kigozi G
    11. Gray RH
    12. Wawer MJ
    13. Ssekubugu R
    14. Santelli JS
    15. Reynolds SJ
    16. Chang LW
    17. Serwadda D
    18. Grabowski MK

    (2018) Migration and risk of HIV acquisition in Rakai, Uganda: a population-based cohort study

    The Lancet. HIV 5:e181–e189.

    https://doi.org/10.1016/S2352-3018(18)30009-2

    • PubMed
    • Google Scholar
29. 1. Palk L
    2. Blower S

    (2015) Brief report: Mobility and circular migration in Lesotho: implications for transmission, treatment, and control of a severe HIV epidemic

    Journal of Acquired Immune Deficiency Syndromes 68:604–608.

    https://doi.org/10.1097/QAI.0000000000000526

    • PubMed
    • Google Scholar
30. 1. Ravenstein EG

    (1885) The Laws of Migration

    Journal of the Statistical Society of London 48:167.

    https://doi.org/10.2307/2979181

    • Google Scholar
31. 1. Read SW
    2. Kim P
    3. Marovich M
    4. Dieffenbach CW
    5. Fauci AS

    (2022) Forty years of investment in HIV research: progress towards ending the HIV pandemic and preparation for future pandemics

    Journal of the International AIDS Society 25:e26039.

    https://doi.org/10.1002/jia2.26039

    • PubMed
    • Google Scholar
32. 1. Ruggles S
    2. McCaa R
    3. Sobek M
    4. Cleveland L

    (2015)

    The IPUMS collaboration: Integrating and disseminating the world’s population microdata

    Journal of Demographic Economics 81:203–216.

    • PubMed
    • Google Scholar
33. Book

    1. Shryock HS
    2. Siegel JS
    3. Larmon EA

    (1975)

    The Methods and Materials of Demography

    San Diego: Academic Press.

    • Google Scholar
34. 1. Simini F
    2. González MC
    3. Maritan A
    4. Barabási AL

    (2012) A universal model for mobility and migration patterns

    Nature 484:96–100.

    https://doi.org/10.1038/nature10856

    • PubMed
    • Google Scholar
35. Website

    1. Statistics Botswana

    (2014) Population and Housing Census 2011: Analytical Report

    Accessed March 21, 2022.

    https://www.statsbots.org.bw/sites/default/files/publications/Population%20%26%20Housing%20Census%20Dissemination%20analytical%20report%20.pdf
36. Website

    1. Statistics Botswana

    (2015) Population and Housing Census 2011: National Statistical Tables

    Accessed April 20, 2022.

    https://www.statsbots.org.bw/sites/default/files/publications/national_statisticsreport.pdf
37. 1. Stewart JQ

    (1948) Demographic gravitation: Evidence and applications

    Sociometry 11:31.

    https://doi.org/10.2307/2785468

    • Google Scholar
38. 1. Thorp M
    2. Ayieko J
    3. Hoffman RM
    4. Balakasi K
    5. Camlin CS
    6. Dovel K

    (2023) Mobility and HIV care engagement: a research agenda

    Journal of the International AIDS Society 26:e26058.

    https://doi.org/10.1002/jia2.26058

    • PubMed
    • Google Scholar
39. Website

    1. UNAIDS
    2. World Health Organization

    (2004) Botswana: Epidemiological Fact Sheet - 2004 Update

    Accessed October 11, 2022.

    https://data.unaids.org/publications/fact-sheets01/botswana_en.pdf
40. Website

    1. United Nations, Department of Economic and Social Affairs, & Population Division

    (2019) World Urbanization Prospects: The 2018 Revision

    Accessed May 16, 2023.

    https://population.un.org/wup/publications/Files/WUP2018-Report.pdf
41. 1. Valdano E
    2. Okano JT
    3. Colizza V
    4. Mitonga HK
    5. Blower S

    (2021) Using mobile phone data to reveal risk flow networks underlying the HIV epidemic in Namibia

    Nature Communications 12:2837.

    https://doi.org/10.1038/s41467-021-23051-w

    • PubMed
    • Google Scholar
42. 1. Wilkinson E
    2. Engelbrecht S
    3. de Oliveira T

    (2015) History and origin of the HIV-1 subtype C epidemic in South Africa and the greater southern African region

    Scientific Reports 5:16897.

    https://doi.org/10.1038/srep16897

    • PubMed
    • Google Scholar

Author details

1. Janet Song

   Center for Biomedical Modeling, Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, United States

   Contribution

   Software, Formal analysis, Validation, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

2. Justin T Okano

   Center for Biomedical Modeling, Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, United States

   Contribution

   Software, Formal analysis, Validation, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

3. Joan Ponce

   Center for Biomedical Modeling, Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, United States

   Contribution

   Software, Formal analysis, Visualization, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

4. Lesego Busang

   The African Comprehensive HIV/AIDS Partnerships (ACHAP), Gaborone, Botswana

   Contribution

   Conceptualization, Writing - review and editing

   Competing interests

   No competing interests declared

5. Khumo Seipone

   The African Comprehensive HIV/AIDS Partnerships (ACHAP), Gaborone, Botswana

   Contribution

   Conceptualization, Writing - review and editing

   Competing interests

   No competing interests declared

6. Eugenio Valdano

   Sorbonne Université, INSERM, Institut Pierre Louis d’Epidémiologie et de Santé Publique, Paris, France

   Contribution

   Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

7. Sally Blower

   Center for Biomedical Modeling, Department of Psychiatry and Biobehavioral Sciences, Semel Institute for Neuroscience and Human Behavior, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, United States

   Contribution

   Conceptualization, Supervision, Funding acquisition, Writing - original draft, Writing - review and editing

   For correspondence

   sblower@mednet.ucla.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4342-3911

• 368

  views

• 58

  downloads

• 1

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.