Introduction

Stigma is a complex societal phenomenon [1,2,3]. Despite the increasing interest and understanding of the adverse health outcomes associated with stigma [4], there is an exigent need of narrowing down and highlighting the factor-specific societal contribution to stigma. These factors, i.e., societal beliefs and practices, play a major role in creating the perceived stigma which in essence is the fear of being discriminated against or the fear of enacted stigma where the stigmatized persons internalize prejudices and develop negative feelings about themselves [5]. Consequently, they end up feeling abashed and embarrassed about the practice [6]. On a larger scale, this perceived stigma has the potential of curtailing societal advancements in terms of academic achievements and decreasing the uptake of health and social services [7].

While many studies are being conducted among sex workers, they have extensively sought to delve into HIV, Sexually Transmitted Infections (STIs) and some behavioral risk factors like Sexual and Gender-Based Violence [8,9,10,11]. Many of these studies have also concentrated on HIV/AIDS-associated stigma as a major encumbrance to voluntary testing, counseling and the necessary prompt treatment [12, 13]. However, just a few studies have endeavored to scrutinize and provide a measure of the perceived stigma related to being a sex worker and identify the key variables of stigmatizing sources [14, 15].

Perceived stigma is usually related, contributed and exacerbated by society. This is profoundly in relation to the expected reactions towards the activities of these sex workers [16]. In most circumstances, these sex workers are predisposed to a plethora of stigmatizing forces in their daily lives through their interactions with relatives, neighbors, religious institutions, health providers and law enforcers. These societal prejudices may have deleterious effects on the health and well-being of the sex workers through obvious manifestations such as physical or verbal abuse and through slicker and subtle means such as those that propagate and immortalize vulnerability which then obliges these workers to initiate personal individualized mechanisms or collective ways of dealing with the stigma [17].

In some cases, this stigmatization is due to how sex work is perceived in the legal sphere. The different legal systems have advanced for either the full or partial criminalization of the trade. Interestingly, the partial criminalization creates a disconnect between society and sex workers, thus breeding the perceived stigma [18,19,20]. In the Kenyan context, such partial criminalization of sex work also exists [21, 22]. This disjointed relationship is often exhibited by denigratory labels on the female sex workers such as ‘prostitutes’, ‘hookers’, and ‘whores’ in describing the sex work [18].

A significant body of research has shown that female sex workers experience higher levels of stigma and related violence compared to Male Sex Workers [23]. However, few studies have endeavored to provide a measure of this perceived stigma among female sex workers. This paper aims to establish a thorough factor-structured index verifiable in exploratory and confirmatory factor analyses with acceptable goodness of fit of the societal contributory nature to the perceived stigma among sex workers.

The few developed Perceived Stigma Indices among FSWs [24,25,26] have all adopted different concepts and structures in their measurements of stigma. This entailed using factors pooled from bases that are sometimes different and peculiar to certain societies. As envisaged in previous research on stigma measurement, indices from different populations and settings can only be used in the scale development and initial validation; however, they cannot be equated to every new society. As unequivocally stated in the USAID 2006 report: “Constructing a Stigma and Discrimination Index: Hopes, Dreams, and Lessons Learned,” no standardized measure or index can be adopted to encompass all the relevant variables and factors in different settings.

Our analysis aimed to develop a standardized instrument among the female sex workers Perceived Stigma Index (PSI) to quantify stigma and measure the contributory capacity of the societal factors; and also to identify the key components of stigmatizing sources under the Social Practice Theory that will allow for tracking the levels of the Perceived Stigma and thus provide a mechanism for reduction interventions. This secondary analysis was conducted from the WHISPER or SHOUT study that was conducted among female sex workers in Mombasa, Kenya [27, 28].

Methodology

This paper describes a secondary analysis of the data from the WHISPER or SHOUT study: Women’s Health Intervention using SMS for Preventing Unintended Pregnancy (WHISPER) and SMS intervention to improve Nutritional Health Outcomes (SHOUT) study. The study methods are described in detail elsewhere [27, 28].

Social practice theory

Social Practice Theory offers a platform to understand the similitude and connection between the societal practices and stigma among female sex workers. Social practice refers to the typical and habitual everyday practices performed in society [29]. For the purpose of this paper, social practice theory is defined as a theory that calls out how people pursue diverse concerns, become aware of new possibilities for action as they move across different settings of practice; and learn as they adjust contributions to the flow of the ongoing activity and to fit the demands and structures of local institutions [30]. It seeks to understand and explain the social and cultural world by analyzing the basic bodily, knowledge-based practices that interconnect to form more complex social entities like groups’ lifestyles, social fields or entire societies [29].

SPTs have been considered important in offering alternative yet concrete ways of understanding human action in relation to health and well-being; and to further reconcile structure and urgency in the lived experience of everyday life. Health and several aspects of well-being are considered the outcome of participation in a set of Social Practices. Social Practice Theory hypothesizes that better designed and managed neighborhoods recruit residents into new practices or reconfigure the existing ones, resulting in observed increases in health and well-being [31].

Thus through social practices, this paper seeks to understand how the internalized stigma among the female sex workers is generated. Figure 1 below conceptualizes how FSWs, in the full awareness of the different social practices, perceive all the prejudices and discrimination against them, and as such develop new mechanisms or behaviours to cope in the society.

Fig. 1
figure 1

Conceptual framework on the connection between societal practices and perceived stigma among female sex workers

Full size image

Index development

Several steps were taken to develop the female sex workers’ Perceived Stigma Index (PSI). A thorough examination of the availability of existing stigma measures did not result in identifying an appropriate tool, further laying credence to the minimum or non-existing research work in this area.

Three main factor domains that contribute significantly to the Perceived Stigma were identified from the research questionnaire used during the primary research (Fig. 2). The index was designed to capture factors whose interplay was deemed to contribute to the perceived stigma, and included: (1) Social demographic factors, (2) Relationship control & sexual and gender-based violence, and (3) Sexual and reproductive history and social awareness.

Fig. 2
figure 2

Interaction of societal factor domains

Full size image

Data analysis

Factor analysis

In developing the female sex workers’ Perceived Stigma Index (PSI), the study employed various Structural Equation Modelling methods. A thorough and comprehensive analysis was conducted to interrogate the PSI’s reliability and validity. The internal consistency (how closely related the set of items are) of the Index and of the three domains therein was measured by the Cronbach’s alpha coefficients. Item Score Reliability test was used to assess the repeatability of an individual item score in the three groups. In this test, an item was deleted if its Item-rest correlation coefficient was lower than 0.40. At the end of the process, all the items in the three domains had Item-rest correlation coefficients higher than 0.40. This was very crucial for the purpose of satisfying the criteria of construct validity.

Primary Component Analysis, a subset of Exploratory Factor Analysis (EFA), was employed to verify the dimensionality of the index items. Other tools that include but are not limited to the Scree Plot Test, Parallel Analysis, Residual Variance, Cumulative Variance and Oblique Rotation were also performed as part of the EFA. This EFA was performed using R and subsequent Confirmatory Factor Analysis (CFA) was similarly done through R to verify the results.

The three-factor domains as envisaged from the Social Practice Theory were validated to be sufficient in providing a standard measure for the perceived Stigma among Female Sex Workers. Item endorsement was determined by calculating the median and the Inter-quartile ranges (IQRs) of the responses from each item. The scores for each scale were calculated separately by summing each weighted item in the index, with the weights being the factor loadings obtained from the Confirmatory Factor Analysis (CFA).

Regression analysis

A descriptive analysis was conducted on the three social domains conjured through the Social Practice Theory regarding the perceived stigma. The proceeding Linear Regression Analysis was conducted using R on each domain separately. The weighted scores on each scale were used to determine how the influence of each of the Social Practice components correlates and contributes to the perceived stigma levels from the three sources.

Results

Social practice domains

The study enrolled 882 Female Sex Workers, who had a median age of 26 years (IQR: 22–29; Table 1). Participating women were predominantly born in Kenya (98%) with 77.2% having changed residences in the last two years of the baseline study. Majority of the women (72.3%) were single, with only 6.2% either married or living with a partner. The sex workers supported a median of 3 people financially (IQR:1–4), with their main source of income being sex work (99.2%). Most of them (99.2%) relied on an income from the trade with most receiving a payment after sex with a client of $ 5 and above (54.9%) in a period of six months.

Relationship control and instances of sexual and gender based violence (Table 2) are clearly highlighted by the respondents; from unilateral decision making (54.5%), whether or not they should use a condom (32.9%), control of what the females should wear (44.2%) and whom they spend time with (44.5%).

Sexual and reproductive history of the participants in relation to the societal awareness (Table 3) indicate the non-willingness of disclosure by the female sex workers of their trade to neither their spouses (40%) nor their parents or siblings (71.9%).

Table 1 Socio-Demographic characteristics of 882 female sex workers enrolled in the WHISPER or SHOUT study between 2016 and 2017 in Mombasa, Kenya, N = 882
Full size table
Table 2 Shows the different forms of relationship control and instances of sexual and gender-based violence experienced by the female sex workers, N = 717
Full size table
Table 3 Shows the sexual and reproductive history of the participants in relation to the societal awareness of the same
Full size table

Item reduction and index reliability

The reliability of the Perceived Stigma Index was high, with a Cronbach’s alpha coefficient of 0.86 (Table 4). All the factors in the three domains satisfied the item-rest correlation of 0.4 and above, and so they were all retained in the perceived stigma index. The results from the primary component analysis, parallel analysis, and Screeplots laid credence to the social practice theory approach. We observed that our Social Practice Theory Approach only accounted for the 40% explained variance (Table 5). This illustrates the multi-dimensional construct of the perceived stigma .

Table 4 Internal consistency of items in Perceived stigma: item statement, corrected item-related correlation.alpha of deleted item and factor loadings (CFA).
Full size table

Exploratory factor analysis

We performed a parallel analysis (maximum likelihood) using a polychoric correlation matrix which suggested that the social practice theory structure had only one factor with an eigen- value > 1.0 (Fig. 3) which also accounted for 40% of the variance.

Fig. 3
figure 3

Scree plot from the EFA analysis

Full size image
Table 5 Shows the descriptive statistics for the Perceived stigms: Item statement and factor coefficients
Full size table

Regression analysis

Tables 6, 7 and 8 indicate the relationships between the Female Sex workers’ perceived stigma and the three societal domains as envisaged through the social practice theory. The key results from the regression analysis indicate highest correlation between perceived stigma and income, relationship control and societal awareness; factors derived from the three domains.The results also show that the personal characteristics of the female sex worker did not contribute much to the perceived stigma compared to the other two domains.

Table 6 A multi variable linear model showing the relationships between socio-demographic predictor variables and Female Sex Workers Perceived stigma
Full size table
Table 7 A multi-variable linear model showing the relationships between the predictor variables (Relationship control and Sexual and Gender-Based Violence) and female sex workers’ perceived stigma
Full size table
Table 8 A multi-variable linear model showing the relationships between the predictor variables (Sexual and reproductive history of the participants & societal awareness) and Female Sex Workers Perceived stigma
Full size table

Discussion

Our study aimed to develop a standardized instrument i.e., the PSI, to measure perceived stigma among female sex workers in Mombasa, Kenya. There being a very limited number of studies done in this space and context, and no comparable tool to quantify the perceptions of FSWs in the region, we endeavored to further consider the appraisal of specific factors that contribute to such perceptions. We adopted the Social Practice Theory by coming up with three societal domains that in some way were deemed to induce the perceived stigma. The factors from the three domains were subjected to both exploratory and confirmatory factor analyses and the multi-dimensional build-up of perceived stigma was proven. Linear regression models further provided a more distinctive and clear answer to the significant factors that contribute inordinately to perceived stigma among female sex workers. These major factors include: Income, Relationship Control, Sexual and Gender based violence and Parental and/ or siblings awareness of the person’s involvement in sex work.

An increase in income and family support

We observed that an increase in income correlated to perceived stigma among the FSWs. An increase from the weekly income range of $ 10 – $ 20 to $ 20 and above created a significant increase in the PSI (0.95 to 1.69 change in the regression coefficients). This was despite the fact that the female sex workers had a median of 3 people they supported financially. This can simply be connected to another factor, i.e., societal awareness, where an increase in income would mean increased financial support to family members hence risking the knowledge of their sexual and reproductive history. In this case, financial stability or independence influences how women internalize the experienced perceived stigma [32].

Societal awareness of the sexual activities

This poses the greatest contribution to perceived stigma among the aforementioned significant factors. A substantial 71.9% of the participants indicated that their parents and/or siblings were not aware that they were sex workers, and a further regression generated a very high correlation coefficient (3.54).

Relationship control

The FSWs indicated that their husbands or partners had some form of control over their mode of dressing and decision-making. This, in turn, enacted the fear among the FSWs that their societal rights to be involved in such a trade weren’t in their control. This further informs the fact that a majority of the FSWs are single (72.3%); perhaps choosing to maintain their freedom that is unhinged from any kind of control by potential or prospective partners.

Sexual and gender-based violence (SGBV)

The FSWs had experienced or observed some form of physical or sexual violence. Such dehumanizing acts had left an indelible mark on them, hence the high perceived stigma. These findings are in line with studies that have shown violence against the FSWs by either the partners or clients provokes this internalized stigma that has direct connections to suicidal thoughts and other depressive symptoms [33,34,35]. The violence meted against them was both sexual and physical in which the latter entailed the partners using kicks, fists and other objects on the FSWs.

Conclusion

The developed PSI takes cognizance of the social practices in the community, which can be particularly useful in providing the requisite interventions and tracking the level of perceived stigma in different societies. This study provides the first instrument through which perceived stigma can be measured and controlled by only focusing on key societal variables: income and family support, societal awareness of the FSW and their trade, relationship control, and sexual and gender-based violence.

Strengths and limitations of this study

The study’s strengths are that it was based on a large study population that provided a large data set that was really important for the psychometric analysis. Rigorous analytical tools were also employed in the measurement of invariance, an important aspect of structural validity. However, there were several limitations to this study. Out of the entire population that had been enrolled in the primary WHISPER or SHOUT study, 11% faced communication challenges in terms of network and language problems hence their data was not captured. This data would have offered more analytical base in the secondary analysis. Additionally, the 40% invariance indicates the need for further induction of other factors contributing to the perceived stigma.