Introduction

The neglected tropical disease (NTD) schistosomiasis, also known as snail fever and bilharziasis, is a water-borne parasitic disease caused by parasitic flatworms [1]. Infection occurs as a result of contact with fresh water contaminated with parasites released from infected fresh water snails. The causative agents of schistosomiasis are blood trematodes–with two different species types being present in Sub-Saharan-Africa: Schistosoma mansoni causing intestinal schistosomiasis and Schistosoma haematobium resulting in urogenital schistosomiasis which is characterized by haematuria (blood in urine), pain during urination and lower abdominal pain [2]. Schistosomiasis is considered a disease associated with poverty, as it predominantly affects rural communities characterized by low socio-economic status that lack adequate access to clean water and proper sanitation [1]. Therefore marginalized, poverty-affected communities living in schistosomiasis endemic areas are considered vulnerable to an infection and, as a result, prone to its disease-related stigma [3–6].

Stigma related to a health condition refers to any negative attitude towards persons suffering from a health condition, that can severely impact the quality of life of the persons stigmatized [3,7,8]. Health-related stigma is often characterized by social disqualification processes such as labeling, stereotyping, status loss and discrimination [9]. Even though the health conditions may vary from mental health problems to neglected tropical diseases (NTDs), the extent and forms of stigmatizing are largely similar [3,10]. Stigma towards neglected tropical diseases has received increasing attention in the last two decades [3,4,11–13]. Not only physical impairments and disfigurements, as a result of an infection with NTDs, are associated with stigma [3–6]. The whole intersection of stigma based on socio-economic status, ethnicity and a health condition amongst others can result in an enormous burden for those affected [3,14].

Previous studies assessing attitudes and perceptions towards schistosomiasis have rarely addressed stigma towards schistosomiasis directly. However, several forms of stigma associated with the disease have been reported and suggest that schistosomiasis is considered a shameful disease [6,15–18]. Intrinsic negative attitudes towards the disease within the person affected, known as internalized stigma [19], were expressed in the form of shame, embarrassment [15,17,18,20] and the fear of being a financial burden to the family [15]. Anticipated stigma outlines the expectation of stigma by the affected person [19] and was apparent in the form of fear of disclosure and fear of stigma [6,15,17,18]. The expression of discrimination due to schistosomiasis, called enacted stigma, [19] was described as avoidance, negative attitudes, labelling and making fun of schistosomiasis patients [15,17,18,20]. However, reports of schistosomiasis being considered shameful contradicts research describing schistosomiasis as being perceived as a normal disease [16,21,22]. Two qualitative studies addressing stigma related to female genital schistosomiasis have described stigmatizing attitudes similar to experiences when having a sexually transmitted disease [23,24]. Stigma towards schistosomiasis has not yet been investigated by means of a quantitative approach neither in Tanzania nor in another country with endemic communities. In order to understand the extent, the forms and the consequences of schistosomiasis-related stigma a standardized tool is needed. Assuming a generic concept of health-related stigma, the use of similar stigma assessment tools across different health conditions and countries is recommended [8,10]. The Explanatory Model Interview Catalogue–Community Stigma Scale is one of the most frequently used and validated questionnaires for the assessment of public stigma in health conditions [10,12,25–28].

The original EMIC used a mixed methods approach including quantitative and qualitative parts in order to examine the experience of illness of affected people [29]. Recent research largely focused on the quantitative part of the catalogue, the Community Stigma Scale, which was adapted to enable measurement of stigmatizing attitudes and behaviors in the communities in which it was applied [10]. As a result, multiple versions of the scale exist with the number of items of the scale varying between 8 to 25 [10]. The response-scale contains four options ‘Yes’ [2]; ‘Possibly’ [1]; ‘No’ (0) and ‘Don’t know’ (0) [10]. Stigma related to various NTDs, mental health and other conditions has been examined using the EMIC-CSS questionnaire in Nigeria, Cameroon, Ghana, Uganda, India, Nepal, Indonesia and Thailand. [10,12,20,25–28,30]. For the 15-item-version a mean sum score of 8 and higher is perceived to be a meaningful indication of stigma for health-related stigma [25]. The 15-item-scale was adapted for the context of leprosy [27] in Indonesia following the framework for cross-cultural equivalence of Herdman and Fox-Rushby [31] and recommendations of Terwee et al. [32] and Stevelink and van Brakel [33]. The cultural equivalence framework of Herdman and Fox-Rushby comprises conceptual, item, semantic, operational and measurement equivalence and is displayed in Table 1 [31].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Overview of conceptual, item, semantic, operational and measurement equivalence.

https://doi.org/10.1371/journal.pntd.0011534.t001

If all categories of the cultural equivalence framework are properly reflected in the scale, cultural equivalence is achieved and the scale can be considered validated in the new culture [31]. In their large study of cultural adaption and validation (N > 200), Peters et al. [27] tested on convergent validity of the EMC-CSS with the Social Distance Scale of r = 0.41 and showed Cronbach’s alpha α = 0.83 and a retest reliability coefficient of 0.84. The process of validating the 15-item version of the EMIC-CSS across various cultures and health conditions has demonstrated different understandings of the items, proposed new response scales and has shown different sub-components in factor analysis [27,34]. These findings underline the importance of cultural validation prior to widespread use of health questionnaires in diverse health and cultural settings. In this study, the English version of the EMIC-CSS for the context of leprosy was adapted to the context of schistosomiasis and translated to Kiswahili. In the Tanzanian culture, the EMIC-CSS has been used in Kiswahili language for the assessment of stigma related to albinism [34,35]. A self-assessment version of the scale was tested and adapted on cultural equivalence [35]. This Kiswahili version was not yet accessible when adapting and validating for schistosomiasis.

In this study, the 15-item version of the EMIC-CSS with the above mentioned four-option scale was used. It was supplemented with demographic items and a questionnaire on knowledge and perceptions on schistosomiasis. Yet, due to the focus on the validation of the EMIC-CSS, the additional survey parts will not be addressed in this article.

Methods

Design of the cultural validation process

The adjusted version of the EMIC-CSS was applied and validated following the cultural equivalence framework of Herdman and Fox-Rushby [31]. Specifically, the aim was to assure conceptual, item, semantic, operational and measurement equivalence of the scale. Table 2 provides an overview of the taken steps according to the framework of cultural equivalence of Herdman and Fox-Rushby [31].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Overview of the strategies to test for conceptual, item, semantic, operational and measurement equivalence.

https://doi.org/10.1371/journal.pntd.0011534.t002

Results

Sociodemographic characteristics

The mean age in the pilot was 37 years (SD = 16); in the focus group discussions 44 years (SD = 13) and in the main implementation 41 years (SD = 14). The data of 97 female and 103 male participants were included in the main implementation, 30 female and 11 male participants in the pilot study; 40 male and 41 female participants in the focus group discussions. The majority of the participants were from Sukuma ethnicity (55.28%), were married (75.16%) and were Christians (77.33%). Characteristics of participants of the pilot study, focus group discussions and main implementation of the EMIC-CSS are depicted in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Sociodemographic characteristics of participants in the pilot study, the FGDs and the main study.

https://doi.org/10.1371/journal.pntd.0011534.t003

Descriptive data

Main survey implementation of the EMIC-CSS resulted in mean sum scores of M = 7.24 (SD = 5.66) for women and M = 9.28 (SD = 7.24) for men. Thus, the mean sum score of the EMIC-CSS in the pilot and in the main study exceeds the score of 8 which is perceived to be a meaningful indication of stigma for health-related stigma [25]. The range of each item is 2.0 (0 = no; 1 = possibly; 2 = yes) and describes the degree of agreement per item. With 15 items total, the maximum sum score possible is 30. 43% (86 participants) reached the defined cutoff-value and thus reported stigma [25], 45% (90 participants) did not reach the cut-off value. Participants answering single items with “I don’t know” (12%) were excluded when considering the EMIC-CSS sum score. The overall mean sum score of M = 8.35 (SD = 6.63) indicates the existence of stigmatizing attitudes and perceptions among participants. The mean sum scores and matching standard deviations of each item of the adapted EMIC-CSS of the pilot study and main survey are displayed in Table 4.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Mean scores and standard deviations of pilot study and main study.

https://doi.org/10.1371/journal.pntd.0011534.t004

Results of the cultural validation process

The results of the analysis of the five aspects of cultural equivalence (as described in Table 2) are displayed in the following.

Conceptual equivalence.

Coded and clustered topics discussed during the FGDs confirm the relevance of aspects addressed in the adapted EMIC-CSS. The topic categories of the EMIC-CSS derived from the validation study of Peters et al. [27] were met in the coded and clustered subjects. Statements made in the FGDs supporting and conflicting the different aspects of the EMIC-CSS are provided in Table 5. Problems in marriage and therefore the aspect of concealment were addressed repeatedly, even though schistosomiasis was also described as a “normal disease” several times.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Match of topics of the EMIC-CSS with addressed aspects in the FGDs.

https://doi.org/10.1371/journal.pntd.0011534.t005

Item equivalence.

The items containing the expression “thinking less of oneself” (EMIC2, EMIC4 and EMIC8) lead to misunderstandings in the pilot study described by the interviewers. The misunderstandings were reflected in false-positive responses, i.e. shown in the second highest mean score for EMIC2 with M = 1.24 (SD = 0.97). These misunderstandings raised the question whether these items are culturally relevant, useful and accepted. After correcting translational misunderstandings, as described in the next paragraph semantic equivalence, the items were understood and accepted. The relevance of the items and their contribution to the scale can also be seen in the item-total-statistics. These are displayed under the measurement equivalence.

Semantic equivalence.

In expert discussions with the interviewers after the pilot run, some misunderstandings could be identified that will be outlined in some detail. First, for the expression “thinking less of oneself” (EMIC2, EMIC4 and EMIC8) there is no equivalent expression in Kiswahili. Thus, the term was occasionally falsely understood as “thinking less often of someone” in the items EMIC4 and EMIC8. In item EMIC2, instead of the actual meaning “If a member of your family had schistosomiasis, would you think less of yourself?” the item was expressed the following way: “Would you feel bad/sorry, if somebody of your family had schistosomiasis?” leading to the above-reported false-positive answers in the pilot run. Accordingly, questions with the concept of “thinking less of oneself” were translated into

• If anyone in your family suffers from schistosomiasis, does it impair your worth? (EMIC2),
• Would others think of a person with schistosomiasis being worthless/“being no human being”? (EMIC4)
• Would people in your village think that families with schistosomiasis patients are impaired in their worth? (EMIC8)

According to the language expert conducting the backtranslation, the translation of worth in Kiswahili language “thamani” refers to an economical value, i.e. a price, which may also spark some misunderstanding.

Second, item EMIC5 “Would knowing that someone has schistosomiasis have an adverse effect on others?” was unclear to the participants as the “adverse effect” was not described in a concrete way. The “adverse effect” was translated into “being dangerous” according to the backtranslation.

Finally, a challenge in the translation of “possibly” could be overcome. According to a consulted Kiswahili language expert, “inawezekana” has the meaning of “it is possible” which may always be perceived to be true. This may lead to false-positive responses. Fortunately, this response option was ticked very seldom. Thus, the translation with “inawezekana” appears to be suitable. Yet, clearly presenting the response options in the beginning is recommended.

Operational equivalence.

The interviewers who conducted the EMIC-CSS reported in the pilot study that they sometimes explained the items of the EMIC-CSS in more detail when the participants asked. To ensure a standardized procedure in the interviewing, it was agreed on having as little explanation on the items in the interview situation as possible.

Due to the numbers of illiterate participants (43.9% in pilot study, 27.5% in main study) the interview format proved to be a good approach. Further, it was very helpful that all interviewers spoke basic Kisukuma language as the majority of the participants in all study parts spoke Kisukuma language as mother tongue. Greetings in Kisukuma language allowed a welcoming atmosphere as reported by the interviewers. Still, interviewees not proficiently fluent in Kiswahili language were excluded from the assessment ensuring standardized interviews.

Measurement equivalence.

• Floor and ceiling effects
• Reliability analysis: internal consistency
  1. ○ Cronbach’s alpha
  2. ○ item-total correlations
• Factor analysis
• Differences between the interviewers: Kruskal-Wallis test for independent samples

Floor and ceiling effects

Less than 15% of the data of the EMIC-CSS sum score were in the highest or lowest scores. Floor and ceiling effects (i.e., many participants responding in the lowest (floor) or highest (ceiling) numbers of scale) may limit content validity and reliability. If floor and ceiling effects occur, participants within highest or within lowest scores cannot be distinguished from each other. According to Terwee et al., if less than 15% score highest or lowest in a sample size of at least 50 participants, one can assume sufficient variability in the scale [32].

Reliability analysis: Internal consistency and item selectivity

The internal consistency of the EMIC-CSS scale was satisfying with a Cronbach’s alpha of α = 0.86 [43]. Item-total correlations were calculated to assess item selectivity of each item. Good item selectivity is described in literature to be .40 ≤ r ≤ .70 [44] and indicates that the item is effective in measuring stigma. Two items showed selectivity below .40: EMIC15 with r = .26 and EMIC1 with r = .32. Because Cronbach’s alpha would only increase from .857 to .861 or to .856, respectively if the items EMIC15 and EMIC1 were excluded, the low item selectivity of EMIC15 and EMIC1 could be neglected. Item-total-statistics can be seen in Table 6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Item-total-statistics.

https://doi.org/10.1371/journal.pntd.0011534.t006

Factor analysis

The principal component analysis with varimax rotation identified four factors according to the Kaiser-criterion, while the Scree test favors a one-factor solution [45,46]. This pattern is due to the fact that a first factor explains 35.5% of the variance, while factors 2 to 4 merely explain between 11.2% and 7.4% of the variance. While the four factors could in principle reflect different subscales (e.g. distinct types of stigma), their interpretation based on the content of the items does not yield clearly distinct concepts. The contents of the items in the third and fourth factor seem to be rather heterogeneous and therefore do not confirm the four-factor-structure (Table 7). In addition, the small number of items per factor renders it difficult to formulate clear subscales. Finally, order effects can be seen in the distribution of the items on the four factors combining those items to a factor which are assessed after each other, e.g. item 3, 4 and item 5. Accordingly, we support the one-factor-solution suggested by previous literature [27].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Rotated component matrix.

https://doi.org/10.1371/journal.pntd.0011534.t007

Differences between the mean scores of the interviewers

The Kruskal-Wallis-test demonstrated a statistically significant overall difference in the EMIC-sum score between the different interviewers: χ2(3) = 32.129, p < .0001. The significant difference appeared between the second male interviewer and the other interviewers, as revealed by post-hoc Dunn Bonferroni tests. There were no significant differences between the other three interviewers. The second male interviewer shows a higher mean rank EMIC-sum score of 118.25, whereas the second female interviewer has 85.92 as mean rank EMIC-sum score, the first female interviewer 74.61 and the first male interviewer 66.00. The effect size for this difference ranges from moderate effect r = .31 between the second male interviewer and the first female interviewer up to a large effect between the second male interviewer and the first male interviewer r = .52. As there were no indications as to why sum scores were higher for the second male interviewer—e.g., whether he committed mistakes during the procedure or whether he happened to talk to people with higher overall scores—the data of the second interviewer were not excluded.

Discussion

The aim of the present study was to validate the suitability of the adapted 15-item version of the EMIC-CSS to assess stigma towards schistosomiasis in the Mwanza region, Tanzania, using the cultural framework of Herdman and Fox-Rushby as reference [31].

When considering conceptual and item equivalence, all forms of stigma addressed in the EMIC-CSS (concealment, shame, avoidance, process of discrediting, problems with getting married or ongoing marriage, problems with work, problems for family or other people) were referred to and described repeatedly by the participants in the focus group discussions. These qualitative results were supported by quantitative results: 43% of the participants reported stigma. Hence, perceptions and attitudes of stigma towards schistosomiasis are existent under the survey participants [25]. A mean EMIC-CSS sum score of 8.35 underlines the notion that stigma related to schistosomiasis is a relevant construct in Mwanza region, Tanzania. However, the FGDs did not provide an equally clear picture for all aspects of stigma. For instance, while shame and embarrassment were mentioned in relation to schistosomiasis, it was also described by some as a normal disease. Additional research is therefore needed to address how these individual aspects contribute to stigma towards schistosomiasis. The aspects problems with getting married or ongoing marriage were perceived to be major topics for the participants of the FGDs. This is consistent with the scale assessing marriage problems with three different items. Further, problems with work were reported as major problem when suffering from schistosomiasis. However, it was not entirely clear whether work-related problems were due to schistosomiasis itself or to poor living conditions. The intersection of multiple stigmatizing conditions, especially that of stigma related to health conditions and stigma related to poverty, is an often-reported phenomenon [8,14,47] and thus, it does not question conceptional equivalence.

The concept of stigma itself deserves a closer look in the framework of the Tanzanian culture. Stigma in Kiswahili is “unyanyapaa” and reflects a well-known concept. The use of the word and concept “unyanyapaa” is especially prevalent in the context of HIV/AIDS and other highly stigmatized conditions. When looking at research on “unyanyapaa”, it is apparent that most studies focused on stigma towards HIV/AIDS and homosexuality, especially between men, also considering sexually transmitted diseases such as HIV/AIDS [48,49]. This inspires the question whether “unyanyapaa” is associated with severe forms of stigma. In the EMIC-CSS, the term “unyanyapaa” was not used, but it was referred to in explanations by the facilitators and participants in the FGDs. The associations with extreme forms of stigma would explain some negations of stigma and discrimination in the FGDs. Moreover, there is apparently no word for the concept of stigma in Kisukuma language which may contribute to a decreased understanding of the concept of stigma.

In general, our data support conceptual equivalence as well as item equivalence.

Semantic equivalence was confirmed in large parts in the expert discussions evaluating the interview process and in the backtranslation by a language expert. However, the expression “thinking less of oneself” referred to in three items (EMIC2, EMIC4, EMIC8) does not seem to have equivalent terms in Kiswahili culture. In the evaluation of the translation after the assessment, the items could be compared with the Kiswahili version of the EMIC-CSS used for albinism by De Groot and colleagues [34,35]. De Groot faced the same challenge and translated the term in EMIC2 different from the items EMIC4 and EMIC8. The term “thamani” used in this study in EMIC2, EMIC4 and EMIC8 was also selected by De Groot in EMIC2 in a broad assessment with several runs. Though the term is often referred to in an economic sense and thus may be misunderstood, the items EMIC2, EMIC4, EMIC8 did not show larger standard deviations than other items, suggesting that respondents did not understand these questions in very different ways. Hence, it may be well understood by the participants despite the connotation. The verb “kudharau” (“to despise sb.”/”to ignore sb.”) which is strongly associated with discrimination due to racism and used by De Groot and colleagues [34,50] for EMIC4 and EMIC8, or “kunyanyapaa” which is used in the context of HIV/AIDS-related stigma, was not investigated in this study.

The translation of the term “adverse effect” in EMIC5 was also challenging. The backtranslation showed that it was understood as being “dangerous”. In the Kiswahili EMIC-CSS for albinism by De Groot and colleagues, this item was removed entirely as it is related to the aspect of concealment which is not possible for people with albinism [34]. As the term is also relatively vague in English, this may be an item which leaves much room for interpretation. Yet, the translation into “athari mbaya” seems to be suitable, as it meets this broad meaning of an “influence/impact/effect” which is “bad” or even “evil”.

The response scale brought up questions on the translation of “possibly” into “inawezekana” and the translation of “no” into “hapana”. “Inawezekana” was selected as it is closest to the word “possibly”, though our language expert remarked that the response “it is possible” may be confirmed for almost any question. Yet, it was seldomly selected in this study and thus can be maintained in the response scale. “No” was translated with “hapana” which stands for a very strong negation with the connotation “never”/” under no circumstances”. A translation with “hamna” (“no/there is nothing”) could be a less strong negation which is used more often in everyday communication. To have a clear distance between the responses, “hapana” was used and is further recommended.

When considering operational equivalence in the EMIC-CSS implementation, particular aspects which have been discussed in previous studies were addressed: the interview format and its challenges and the response scale [27,34,35]. In order to include illiterate participants, the assessment with the EMIC-CSS was conducted via interviews, following the example of the EMIC-CSS used for stigma towards leprosy [27]. Interviews instead of self-administered questionnaires bear the risk of influence due to the social desirability bias. The social desirability bias describes the tendency of the interviewee to respond in a way they assume to be desired by the interviewer or society [51]. During FGDs, the facilitators were perceived to be experts on schistosomiasis, as reflected, for instance, in the participants’ requests for educational training etc. In case the participants considered the interviewers to be experts, this could further increase the chances of introducing social desirability bias, while for instance the sex, race and ethnicity and language use of the interviewers could introduce interview bias [52,53]. The measured effect of interviewers will be discussed when reflecting measurement equivalence.

A response scale with four options was chosen. First, this ensures comparability with previous research, as it has been the common response scale for the EMIC-CSS [10]. In addition, memorizing four response options seems ideal in a verbal interview situation in which the options cannot be re-read by participants. Yet, the response scale with the four options (“Yes”, “Possibly”, “No”, “I don’t know”) bears two risks: On the one hand, the low number of response options may lead to extreme responses. On the other hand, the responses “Possibly” and “I don’t know” may lead to indecisive response behavior. De Groot et al. used a five-point-scale for the assessment of stigma related to albinism and specifically asked about the frequency of the behaviors or experiences described in the items (i.e., it rarely/sometimes/often happens) [34]. The added value of these responses by De Groot is that they may represent the gray area of the phenomenon in society better. This may make the questions easier to answer.

The closed response format of the EMIC-CSS shows to be very useful for controlling bias due to language barriers, e.g. a different mother tongue. Research suggests that sensitive subjects like emotions can be expressed most openly in mother tongue [54,55]. Since in Mwanza region many communities rather speak Kisukuma language as mother tongue [56], this may have influenced results of the FGDs with participants being hesitant to disclose themselves or connotations of statements being lost.

The validation has shown sufficient psychometric properties ensuring measurement equivalence. Internal consistency of the adapted stigma scale as a whole was confirmed and no floor and ceiling effects were found. In contrast to some previous findings, our results suggest the existence of four subscales when following the Kaiser-criterion [46]. The four factors of the EMIC-CSS discovered in this study may reflect different aspects of stigma related to schistosomiasis. These are problems in marriage, family and work (EMIC10,11,12,13,14), social disclosure (EMIC6,7,8), shame and process of discrediting (EMIC3,4,5,9) and concealment and social distance (EMIC1,2,15). As the first factor explains 35.5% of the variance and the item number for second, third and fourth factor is too small for independent subscales, the one-dimensionality is presumed. Further research on the factor-structure of the EMIC-CSS is recommended.

Since the interviewer bias was explicitly considered in the present investigation, a highly standardized procedure in the interview situations was discussed in briefings and established in the interview situations. To avoid race and ethnicity interviewer effects, White persons were excluded from interview situations. Davis and colleagues [53] already stated the importance of measuring and controlling interviewer bias in public health surveys, e.g. race and ethnicity effects [53]. For instance, a White person interviewing a Black respondent may lead to a higher social desirability bias and to reduced openness [53]. Yet, there is little research on the question whether matching of interviewer and interviewee would improve data quality [53].

Conclusion

The present results suggest cultural equivalence of the EMIC-CSS and thus introduces a valid instrument for public stigma related to schistosomiasis. Future studies should additionally address the particular roles of specific aspects, e.g. the burden for the family, or even different forms of schistosomiasis in the development of stigmatization, e.g. female genital schistosomiasis. The use of additional questionnaires that assess participants’ knowledge on schistosomiasis and its symptoms in order to get a clearer idea on the preconditions of stigmatization is recommended. To achieve a full picture of stigma related to schistosomiasis, experienced and internalized stigma should be addressed via self-stigma scales in future research.

The use of the EMIC-CSS has been proven to be useful in the context of stigma assessment towards schistosomiasis in Mwanza region. The scale was validated and the results confirm the existence of stigma towards schistosomiasis.

Supporting information

Acknowledgments

We wish to express our special thanks to the German Leprosy and Tuberculosis Relief Association for making it possible to shed light not only on a neglected (tropical) disease, but also on the underestimated field of stigmatization. Due to the fact that quality of research is always based on the quality of data collection, we want to express our gratitude for the wonderful Tanzanian research team; Anna Samson, Grace Peter, Rogers Kwengulia, Sunday Deogratias, and Caritas Kitoga. Special thanks to all participants for sharing their thoughts and stories.