According to national estimates, Ethiopia has a substantial population of female sex workers (FSWs), particularly concentrated in urban centers and along major transport routes (7, 10). This study was conducted in 16 cities—including Addis Ababa, Bahir Dar, Dire Dawa, Hawassa, Adama, Arba Minch, Dessie–Kombolcha, Dilla, Gambella, Gonder, Harar, Jimma, and others—which serve as commercial, administrative, and transit hubs. These urban areas are marked by rapid urbanization and high mobility, attracting diverse populations and creating socio-economic conditions that influence sex work engagement. Exploring these varied settings supports a context-specific understanding of the economic, familial, and social drivers of sex work in Ethiopia.

A cross-sectional bio-behavioral survey was conducted from December 2019–April 2020 across 16 urban sites in Ethiopia. The study aimed to explore the underlying motivations for engagement in sex work among female sex workers (FSWs), alongside relevant socio-demographic and behavioral characteristics.

Female sex workers (FSWs) were defined as biologically female individuals aged 15 years or older who had exchanged sex for money or goods within the past month and had at least four clients in the last 30 days. This definition follows Ethiopia's national bio-behavioral surveillance protocol and aligns with UNAIDS standards (32). Eligible participants were those living or working in the 16 selected cities. Exclusion criteria included living outside the study areas, inability to provide informed consent, or refusal to participate in either the behavioral interview or biological sample collection. A sample refers to blood specimens collected for HIV, hepatitis B, and hepatitis C testing. Respondent-Driven Sampling (RDS) was used for recruitment, starting with 6–10 diverse seeds per site. Seeds were selected to reflect variation in age, work venue, and mobility. Each was given three coupons to recruit peers, enabling access to a hidden population and allowing for RDS-weighted analysis (6, 33, 34).

The sample size was estimated using the single population proportion formula, assuming a 95% confidence level, a 5% margin of error, and a 2% prevalence of FSWs in Ethiopia. Adjustments were made for design effects and non-response rates, resulting in a final sample size of 6,085 participants proportionally allocated across the 16 cities based on population size (35).

Respondent-driven sampling (RDS) was used for recruitment, beginning with carefully selected seed participants representing different sex work types, age groups, and geographic locations. Each seed received three recruitment coupons to distribute within their social networks. Recruitment continued until RDS equilibrium was reached to ensure a diverse and representative sample. RDS-adjusted weights were applied to minimize bias, and homophily indices and equilibrium diagnostics were assessed (34). An anonymous fingerprint-based coding system was implemented to prevent duplicate enrollments. Recruitment was conducted in hotspot locations such as bars, hotels, streets, and mobile phone networks with the support of local organizations working with FSWs. Key parameters, including HIV status, type of sex work, and condom use, were monitored to track progress toward equilibrium.

Data collection was conducted using a structured, pre-tested questionnaire developed in English and translated into local languages. The questionnaire covered demographics, socio-economic factors, family responsibilities, and motivations for engaging in sex work. Open Data Kit (ODK) software was used on tablet computers for real-time data collection. Data collectors underwent training on RDS methodology, ethical considerations, and survey administration, and a pilot survey was conducted in Bishoftu to refine the data collection procedures before full implementation.

Data collected using Open Data Kit (ODK) were cleaned in Excel and analyzed in R (version 4.3.0). Respondent-Driven Sampling (RDS) procedures—including recruitment tree construction, weight generation, and assumption checks—were performed using the RDS package. Descriptive statistics and chi-square tests assessed associations between categorical variables. Variables with p-values < 0.20 in bivariate analysis were included in a multinomial logistic regression to examine associations between motivation types and socio-demographic factors. This method was preferred for its suitability with unordered categorical outcomes. The model accounted for clustering within RDS networks, and results were reported as adjusted odds ratios (AORs) with 95% confidence intervals. Variables unrelated to the study objective, such as hepatitis B and C status, were excluded from the analysis. We assessed multicollinearity using variance inflation factors (VIFs) and evaluated model fit with McFadden's pseudo R² and likelihood ratio tests (33, 36).

The study received ethical approval from the Scientific and Ethical Research Office (SERO) of the Ethiopian Public Health Institute. Informed consent was obtained in participants' preferred language. For minors aged 15–17, an IRB-approved waiver of parental consent was used in line with international guidelines to protect adolescent autonomy. Given the stigma around sex work, strong confidentiality safeguards were in place. No names were recorded, interviews were private, and data were encrypted and de-identified. To prevent repeat participation, a fingerprint-based code system (non-identifiable and encrypted) was used with prior ethical approval and community sensitization. Participation was voluntary, and those declining fingerprinting were not excluded. Participants received compensation, and those testing positive for HIV or hepatitis were referred to care. The study adhered to the Declaration of Helsinki and national guidelines for research involving human participants.

Table 1 shows the distribution of primary motivations for sex work among 6,085 FSWs from 16 diverse urban sites in Ethiopia, serving as a basis for subsequent analyses. The highest representation came from Addis Ababa (18.1%), the centrally located capital and largest city in Ethiopia. Adama (11.1%), a major commercial hub situated southeast of the capital along the Addis Ababa–Djibouti transport corridor, also contributed significantly. Other key sites included Hawassa (8.6%), the capital of the Sidama region in southern Ethiopia; and Dire Dawa (7.1%), a chartered city in the eastern part of the country with high mobility and trade activity. Additional contributions came from Gambella (7.7%), located in the western lowlands near the South Sudan border; Jimma (4.2%) and Mizan Aman (4.2%) in the southwest; Nekemte (4.2%) in the western highlands; and other cities such as Arba Minch, Bahir Dar, Dessie/Kombolcha, Dilla, Logia/Semera, Shashemane, Harar, and Gondar (each contributing approximately 4%). Regarding motivations for entering sex work, economic reasons were most commonly cited (41.7%, n = 2,539), including income needs and a lack of alternative employment. Family-related factors—such as the need to support children, divorce, or household conflict—were reported by 22.7% (n = 1,382) of participants. Additionally, 21.0% (n = 1,278) indicated a combination of economic and family-related pressures, while 14.6% (n = 886) cited social or behavioral factors, including peer influence, coercion, or personal preference.

As shown in Table 1, the distribution of motivations varied across cities, suggesting that geographic location—and the associated socio-economic environment—may influence the primary reasons why women enter sex work. These descriptive findings provide important contextual background for interpreting the multivariable analysis presented in later sections.

Table 2 presents the distribution of primary motivations for engaging in sex work across socio-demographic characteristics among 6,085 female sex workers (FSWs) in Ethiopia. Economic motivations were most common overall (41.7%), followed by family-related (22.7%), combined economic and family (21.0%), and social or behavioral motivations (14.6%). Age was strongly associated with motivational differences. Among FSWs aged 15–19 years, 40.5% cited economic reasons and 25.2% family-related, while among those aged 35–59, 49.7% reported economic reasons and only 7.3% social motivations. Motivation also varied by education: FSWs with non-formal education most frequently reported economic reasons (44.9%), while those with secondary or post-secondary education had more varied motivations, including higher family-related (26.2%) and social reasons (13.2%). Marital status revealed distinct patterns. Never-partnered women more often cited economic motivations (44.9%) and social reasons (20.3%), while previously partnered women reported a higher share of combined motivations (31.2%). Among those currently partnered, motivations were more evenly distributed. Income-sharing practices influenced motivations. FSWs who did not share income with others were more likely to report economic motivations (42.3%). In contrast, those sharing with workplace owners or intermediaries reported higher proportions of family-related (up to 27.2%) and social motivations (up to 22.2%). Monthly income also showed a clear gradient. FSWs earning ≤1,500 ETB had the highest proportion reporting economic motivations (46.4%), while those earning >10,000 ETB were more likely to report family-related (30.6%) or social motivations (19.2%).

These patterns provide contextual insight into the socio-demographic variation in motivations for engaging in sex work, supporting interpretation of adjusted results in subsequent analyses.

As illustrated in Figure 2, motivations for engaging in sex work varied by both age and income level, with 95% confidence intervals indicating the precision of the estimates. Younger female sex workers (FSWs) were more likely to report family-related or social/behavioral motivations compared to economic ones, while older FSWs showed a gradual shift toward citing economic necessity. Similarly, higher income levels were modestly associated with increased reporting of family-related or social motivations over purely economic reasons. These trends suggest that motivations for sex work are influenced not only by financial hardship but also by evolving social, familial, and contextual factors across the life course (Figure 2).

While age and income offer important insights into FSWs' motivations, a fuller understanding requires examining behavioral and health-related factors. The next section explores how personal behaviors, health status, and service access relate to motivations for engaging in sex work.

Table 3 presents the distribution of behavioral and health-related characteristics of female sex workers (FSWs) by their primary motivation for engaging in sex work. Motivation types varied across levels of interest in sex work. Among those who reported always enjoying sex work, 48.8% cited economic reasons, compared to 41.7% of those who never enjoyed it. In contrast, 17.2% of the latter group reported social or behavioral reasons, suggesting lower interest was linked to non-economic motivations.

Age at first involvement showed similar variation. Among FSWs who entered the profession at age 25 or older, 48.5% reported economic reasons and only 6.5% cited social motivations. In contrast, among those who started before age 20, economic (37.9%), family (26.3%), and social (19.6%) motivations were more evenly distributed. Time in sex work followed a parallel trend. Among newer entrants (≤1 year), 44.3% cited economic reasons and 17.6% social reasons. With longer durations, the proportion citing economic motivations declined slightly, and combined motivations increased (e.g., 24.0% among those with 8+ years). Meeting location was also associated with motivational type. FSWs operating in commercial places cited economic reasons most often (40.1%), while those in private settings showed a higher percentage reporting combined motivations (28.0%). Regarding health-related characteristics, 42.2% of HIV-positive FSWs reported economic motivations and 23.6% reported combined reasons. FSWs with access to healthcare services were more likely to cite economic (43.3%) or combined motivations (24.0%) than those without access, who more often cited family (26.3%) or social reasons (18.9%). Motivations also varied by condom source awareness. Those relying on non-commercial or unknown sources were more likely to cite combined motivations (up to 25.2%) than those using health facilities or commercial sources.

These findings highlight the influence of behavioral, health, and environmental factors on sex work motivations, providing important context for tailored intervention strategies.

Table 4 presents the results of the multinomial logistic regression analysis examining factors associated with the reasons FSWs engage in sex work, categorized as family-related, economic combined with family, and other social/behavioral reasons (see Table 4). FSWs from Addis Ababa, Ethiopia's capital, were significantly less likely than those from Adama to report family-related [AOR = 0.52, 95% CI (0.40, 0.67), p < .001] or combined reasons [AOR = 0.72, 95% CI (0.53, 0.97), p = .03] relative to economic motivations. Similarly, those from Bahir Dar, the northern Amhara regional capital, were 60% less likely to cite family reasons [AOR = 0.40, 95% CI (0.27, 0.59), p < .001]. FSWs from Dilla (Southern Nations), Diredawa (eastern Ethiopia), and Gonder (northwest Amhara) also showed lower odds of family-related motivations compared to economic reasons (AORs 0.32–0.53, all p < .001). In contrast, FSWs from Dilla and Dire Dawa—major urban centers in southern and eastern Ethiopia—were significantly more likely than those from Adama to report combined economic and family reasons over economic reasons alone [Dilla: AOR = 1.90, 95% CI (1.28, 2.81); Dire Dawa: AOR = 1.86, 95% CI (1.33, 2.59)]; both p < .001]. FSWs from smaller or emerging towns in northern and western Ethiopia, such as Dessie—Kombolcha and Nekemte, had significantly higher odds of citing social or behavioral (“other”) motivations rather than economic ones (Dessie—Kombolcha: AOR = 2.02; Nekemte: AOR = 2.61; both p < .001). Elevated odds of reporting “other” motivations were also observed among FSWs in Hawassa, Semera—Logia (Afar), and Shashemene, with Semera—Logia showing the highest increase [AOR = 2.40, 95% CI (1.54, 3.74), p < .001]. Age also significantly predicted motivational patterns. Compared to FSWs aged 15–19, those aged 35–59 were less likely to report family-related motivations [AOR = 0.50, 95% CI (0.31, 0.81), p = .00] or social/behavioral motivations [AOR = 0.55, 95% CI (0.31, 0.97), p = .04], relative to economic motivations. The 30–34 age group also showed reduced odds of social/behavioral motivations, although the association was not statistically significant [AOR = 0.67, 95% CI (0.41, 1.11), p = .12].

Education showed mixed effects on motivations for engaging in sex work. Compared to FSWs with no formal education, those with secondary education were significantly more likely to report family-related reasons [AOR = 1.54, 95% CI (1.22, 1.93), p < .001], but less likely to cite social or behavioral motivations [AOR = 0.68, 95% CI (0.52, 0.89), p = .00]. Similarly, FSWs with primary second cycle (grades 5–8) and first cycle (grades 1–4) education were also more likely to report family-related reasons (AOR = 1.32 and 1.31, respectively). Marital status was strongly associated with motivational patterns. Previously partnered FSWs (divorced, separated, or widowed) were significantly more likely than currently partnered women to report family-related reasons [AOR = 1.76, 95% CI (1.20, 2.58), p = .00] and combined economic and family motivations [AOR = 2.57, 95% CI (1.73, 3.81), p = .00]. However, they were less likely to cite social or behavioral reasons [AOR = 0.64, 95% CI (0.43, 0.96), p = .03]. FSWs who had never been partnered showed no significant differences from those currently partnered. Higher monthly income was associated with increased odds of reporting non-economic motivations. FSWs earning more than 10,000 ETB per month were more likely to cite family-related (AOR = 1.56), combined (AOR = 1.57), and social or behavioral reasons (AOR = 2.39) compared to those reporting economic reasons (all p < .05). Similarly, those earning 5,001–7,500 ETB had higher odds of reporting combined (AOR = 1.46) and social/behavioral (AOR = 1.47) reasons compared to the lowest income group (≤1,500 ETB). FSWs earning 7,501–10,000 ETB per month were 72% more likely to report social and behavioral motivations [AOR = 1.72, 95% CI (1.22, 2.41), p = .00], suggesting that middle- to upper-income earners may have more complex or multidimensional motivations.

Age at initiation also influenced motivations. Compared to FSWs who began sex work at age ≤20, those initiating at age 25 or older were less likely to report family-related (AOR = 0.63) and social or behavioral reasons (AOR = 0.48). Those who started between ages 20–24 also had reduced odds of reporting family-related (AOR = 0.77) and social or behavioral motivations (AOR = 0.78). Duration of engagement in sex work was another significant factor. FSWs with 8 or more years of experience were 39% more likely to cite family-related reasons (AOR = 1.39, p = .05), while those with 5–7 years were 28% more likely to report combined economic and family motivations (AOR = 1.28, p = .04). These findings suggest that longer involvement in sex work may reflect increasingly complex or evolving life circumstances beyond economic need. FSWs obtaining condoms from commercial sources were 20% less likely to report social or behavioral motivations compared to those accessing condoms through health facilities (AOR = 0.80, p = .03). No significant differences were observed among those using non-commercial sources, suggesting that commercial access may provide a form of protective autonomy. Lack of access to healthcare was strongly associated with non-economic motivations: FSWs without access to health services were 41% more likely to cite family-related reasons (AOR = 1.41, p < .001) and 52% more likely to report social or behavioral factors (AOR = 1.52, p < .001) than those with healthcare access.

Overall, these findings underscore the complex and multifaceted motivations behind FSWs' engagement in sex work. While economic need remains central, many also cite family-related and social or behavioral reasons—shaped by factors such as income, age, education, duration in sex work, condom access, healthcare, and geographic region. Non-economic motivations were more commonly reported in southern, eastern, and pastoral areas, and less so in major urban centers and northern regions. These patterns emphasize the need for tailored, context-specific interventions that address the diverse realities of FSWs across Ethiopia.