During PM fieldwork, the study area is segmented into zones, and an exhaustive list of locations where FSWs congregate is systematically compiled within each zone. Subsequently, these spots are visited with the assistance of local community members (FSWs) to confirm onsite risk activities and the number of FSWs at each location. Population size is estimated at the site level, which is then adjusted to account for double counting and/or mobility across sites, to finally reach an estimated number of FSWs in the study area. Planning for and calculating coverage requires an understanding of where sex work happens and an estimate of various subtypes in each of these locations and their characteristics. One of the key strengths of this approach lies not only in its development of size estimates but also in providing a distribution of FSWs at different hot spots. For planning services and subsequently monitoring coverage, this approach has been proven valuable in ensuring high level coverage [22].

PM includes 3 main approaches: basic geographic mapping (GM), the Priorities for Local AIDS Control Efforts (PLACE) method, and progressive mapping. Selecting the best approach depends on the context and the availability of time and resources. However, all adhere to the principles of generating location and population size information to design and implement programs, and community ownership and leadership. Although basic GM has been used primarily for mapping and estimating size of KPs, PLACE has been used pragmatically to identify locations where persons meet new partners, and thus, networks where HIV transmission is most likely to occur. Both GM and PLACE have been used where mapping and size estimates are not available and/or more rigorous baseline or follow-up estimates of program coverage are required. Progressive mapping approach is implemented as a component of a more mature program, where it is used quickly by program teams to generate information to scale up programs. A practical difference between the methods is that GM and PLACE include systematic interviews with key informants across all study zones to develop an exhaustive list of sites, whereas progressive mapping uses a crude list of sites from previous mapping as the starting point, available program data, and discussions with peer educators and outreach workers. Each method includes visits to identified spots to verify they are operational, describe their characteristics, and collect information on estimates, typology, and other operational characteristics of the hot spot. This is done by interviewing a key informant, preferably an FSW who operates from the identified spot, or someone knowledgeable about the spot (eg, a bar manager). Both GM and PLACE have extended the value of basic PM by using the venue list as a sampling frame for venue-based biobehavioral surveillance. Examples have been provided in subsequent sections on how countries with large FSW populations such as India, Kenya, and Haiti have used these approaches to gather sufficient information in a timely manner with enough geographic specificity and at low cost to plan KP program scale-up [23-25]. Other critical information such as condom availability or service availability for FSWs in the specific spots can be also collected in all the methods if such information is needed for effective program planning.

Development partners (eg, The Global Fund, the World Bank, the Bill and Melinda Gates Foundation, WHO/UNAIDS, and PEPFAR) have supported various PM approaches in a number of countries. Figure 1 shows countries where various forms of PM have been implemented.

Conceptually, high impact programs aim to reach the highest number of FSWs with quality services. At the macro level, national and subnational programming [23] focuses on establishing HIV prevention and care services, upscaling programs where needed, and launching new programs where they are nonexistent. Effective program coverage can only be measured when a realistic estimate of the number of FSWs is available.

PM data in Kenya estimated 103,298 FSWs in 7 provinces [26], which was used by National AIDS and STIs Control Program as a denominator to set national coverage targets for the minimum package of services defined in the national guidelines [27]. PM results guided decisions about where to prioritize services to improve program coverage. Thus, Nairobi (with 27% of the FSW population) was the top priority for prevention efforts, followed by Coast, Rift valley, Nyanza, and Western province. Within Nairobi, the subcounty of Starehe, with 25% of the Nairobi FSW population, was prioritized over counties with lesser concentrations of FSWs [26]. PM data also inform progressive coverage based on resources and targets and provide insights into the heterogeneity of site typology and overall sex work so that intervention packages can be tailored to the local context [25,28]. For example, PM in Nigeria identified diverse sex work typologies based on the operational dynamics of sites and sex workers [29]. Thus, as 80% of the FSWs were venue-based in Abuja, the program focused on robust outreach and service delivery efforts at bars and night clubs and implemented less intensive programming with FSWs at other type of spots. On the basis of the number of FSWs (Lagos: 46,691 vs Anambra: 4846), 10-fold more resources were allocated for the FSW prevention program in Lagos. The geographical distribution of FSWs in Mauritius highlighted that expanding HIV prevention services in only 04 of the 09 districts would provide coverage to more than 80% of the FSW population in Mauritius, targeting specifically the high-intensity sites, thus, improving outreach and coverage within available resources [30].

Micro planning decentralizes outreach management and planning to a grass-root level of outreach workers (ORW) and peer educators (PEs) through a process that collects and uses data at an individual level to empower a community to make healthy decisions for themselves [31]. Although macro plans are made at the national or subnational levels, micro level planning takes place at the site level and becomes the cornerstone of peer-led programs [32].

Through PM, a list of all locations or hot spots where FSWs operate is developed, assigning a range of estimates for FSWs at each spot by subtypology (eg, street, venue, and public place). Furthermore, PM data characterize the spots by providing information on timing of sex work activity at the spot (busiest time of the day or busiest day of the week), which is used as the basis for micro planning. Steps in micro planning include “hot spot validation” identified by PM, followed by “spot load mapping” where FSWs’ peers validate and record peak times and days of operation along with other characteristics of the spot. This is done to build ownership of the PM data by the peers, update spot lists, and estimate the number of PEs required. “Contact mapping” is then done to understand FSW networks and select PEs and determine their allocation across the target area, taking spot clustering in account. “Peer plans” are developed by PEs for each allocated FSW based on her risk and vulnerability, prioritizing those with higher risk and vulnerability. Finally, “Peer calendars” are used to document services provided to each FSW based on individual peer plans and track them at each visit outreach and micro planning [31,32]. The entire process of micro planning was conducted in Haiti using the PM data generated using the PLACE approach. These tools assisted the program to recognize each sex worker as an individual, assess her HIV risk (client volume, age, duration in sex work) and vulnerability (violence), and help plan a prevention pathway for her [33].

Likewise, in Mombasa, Kenya, PM data supported successful scale-up of HIV programs for FSWs. PM identified 9208 FSWs with the highest concentration (5809) in Kisauni and Mvita [26]. Hot spot validation and spot load mapping informed that 60% of FSWs operated through venues such as bars and clubs and another 30% through street spots. The 233 validated spots were plotted on a map, highlighting the number of FSW congregating at each spot. Location, spot size, and typology were used to allocate spots to PEs within their area of operation. On the basis of a ratio of 1 PE for 60 to 75 FSWs, 68 PEs were recruited, and 12 ORWs were hired to supervise them. Within areas of high FSW concentrations, 2 drops in centers and clinics were established, and linkage with government and NGO clinics were established for FSWs in dispersed or far away spots. Through peer plans and peer calendars, cumulative enrollment increased to 80% of the estimated FSW population within the first 10 months. An average of 35 to 40 condoms per month were distributed to each FSW during outreach as per the client volume calculated at registration [24]. As shown in the example, sex worker programs benefit tremendously from PM data in its finest details to estimate the specific number of condoms, lubes, outreach testing supplies, and other materials needed. Not only does this improve the efficiency and effectiveness of programs, it has proven its value in strengthening community networks through a community-based approach, which adds to the ownership of such programs [34].

PM has been used extensively in a number of countries to produce size estimates of sex worker populations [25,26,29,30]. Mapping produces estimates of FSWs at each spot, assesses the amount of overlap between various spots, and then adjusts the total number for potential duplication to calculate a range of estimated number of FSWs in a focused geography. Although PM provides valid population size estimates in a timely manner to inform program design and supports monitoring and evaluation of ongoing interventions, a few limitations of the approach need to be discussed. As the inherent approach focusses on geolocation-based sex work, there is a probability to miss out FSWs who do not visit physical venues and solicit online or via social media. Thus, care needs to be taken while relying on size estimates derived from PM approaches alone. Data from PM are best used as a starting point for programs, and it is anticipated that as programs implement, they will validate and supplement the information and thereby periodically update the estimates [23,24,33]. Despite this limitation, PM has been used extensively across countries for estimating size of the KPs successfully as part of a National size estimation exercise [25,26,30].

Finally, PM data provides service delivery programs with denominators, which are crucial to be able to set goals and establish benchmarks for key outcomes indicators. Key indicators related to program coverage and utilization of programs and services by target populations serve as markers for program success. Coverage gaps at spot level can be evaluated through mapping of these spots on a continuous basis. PM data provide denominators for monitoring 90-90-90 targets and service uptake indicators across the HIV prevention, care, and treatment cascade among FSWs. In Kenya, the national program routinely uses PM data as a denominator to measure the KP program performance every quarter [35]. Under the LINKAGES project, the list of spots generated through PM contributed to the development of a sampling frame for conducting integrated bio-behavioral surveys (IBBS) of KPs in Haiti, Malawi, Angola, and Mozambique, which helped development of 90-90-90 indicators, including proportion of FSW on treatment with suppressed viral loads [25,36-38].

Likewise, Pakistan has extensively used PM data to sample KPs for IBBS. In each city, the IBBS sample was distributed based on weights derived from the respective number of KPs in each zone. Spots were randomly selected from spot lists generated by PM within each zone, and later, respondents were selected based on sampling weights assigned to each spot based on spot size, making the sample representative of the overall population [39,40]. A similar approach was employed across India, where PM data were used to select a random sample of hot spots and select respondents from the selected hot spots. As PM listed all hot spots in a selected geography, the list of spots served as a sampling frame, and it was possible to select a representative sample of KPs using a Time Location Cluster Sampling technique [41]. Thus, PM provides a sampling frame from which a representative sample of FSWs can be selected for survey. Table 1 provides the information available through mapping and its use for service delivery planning.