Study population

The study population included all MisSA, 18 years of age or above, living in one of the six study sites in Germany, regardless of immigration status. The six study sites were chosen because they had the highest numbers of officially registered MisSA at foreigners’ registration offices at the end of 2013. These include Berlin, Frankfurt, Munich, Cologne, the region of Hannover (including Hildesheim and Brunswick), and the Rhine-Ruhr region (including Duesseldorf, Essen, Oberhausen, Muelheim an der Ruhr and Duisburg). In 2013 officially 43,207 migrants of sSA-nationalities lived in the six study sites, out of a total of 184,440 MisSA registered in Germany [26].

Study design

We conducted a quantitative cross-sectional survey among MisSA in six study sites between August 2014 and December 2016. The study design was created through a participatory process with an expert group including MisSA living in Germany, and different stakeholders in HIV prevention and researchers and is described in detail elsewhere [24]. We piloted the study design in order to test its feasibility in Hamburg, Germany, and adapted the main study accordingly [25]. Details about the research process that initiated the pilot study, the pilot itself, and the subsequent adaptations made to the study can be found in previous publications [24,25,27].

Sampling procedure

We used convenience sampling, as a sampling frame that would allow random sampling was not available. Only limited information is available regarding the sSA-population living in Germany [26]. Official statistics only include MisSA with sSA-nationalities, but not those with German citizenship or without regular (legal) status.

MisSA peer researchers with heterogenic ethnical, linguistic, religious and socio-demographic backgrounds recruited a diverse sample of MisSA in their communities. Previous CBPHR studies have shown feasibility and acceptability of participant recruitment through peer researchers, i.e. trained members from the MisSA communities under study [17,28,29]. Peer researchers were selected based on their established contacts with different MisSA communities at the study site and their interest in sexual health topics; full selection criteria are available in the study protocol [24]. In order to recruit a heterogeneous sample of MisSA living in the respective study site, we aimed to adequately represent the biggest subpopulations by countries of birth among the group of peer researchers. In total, 99 peer researchers (10–25 per site) recruited participants over a period of ten to twelve weeks in each study site. During recruitment, we analysed the sociodemographic characteristics of the participants on a weekly basis and provided timely feedback to the local study coordinator and peer researchers about the recruited participants’ sex, age, country of birth, religion, length of stay in Germany, health insurance status, and level of education. We aimed to represent the sSA community in each study site within our sample with regards to known characteristics such as sex and nationality [26]. The peer researchers adapted their future recruitment according to the weekly feedback to assure the sample represented the MisSA community of the study site with respect to sex and nationality and was diverse with regards to other characteristics. This practice also allowed the collection of a heterogeneous sample regarding the other characteristics mentioned above. For example, if the weekly feedback revealed a very small proportion of Muslims in the sample, the peer researchers with appropriate background increased recruitment in mosques.

Sample size

We aimed to detect differences in proportions of 10% (45% vs. 55%) between men and women with a significance level of 0.05, accepting a beta-error of 0.2. Additionally, migrants from southern Africa, the smallest groups among all MisSA in Germany (about 10%), should make up at least 8.5% of the sample for stratified analysis, accepting a beta-error of 0.2 and alpha-error of 0.05 (both-sided). The total sample size consisting of all six study sites was calculated at 3,009 participants accordingly.

Data collection

In each study site, an expert group identified a local partner organisation, that was active in HIV prevention among MisSA, and a local study coordinator, MisSA him- or herself. To overcome common language and culture-sampling barriers, the local partner organization selected paid MisSA peer researchers for the recruitment of study participants. Peer researchers received a two-day training on HIV and survey methods, and ethics before participant recruitment started. Peer researchers promoted the study within their community and to other contacts from sSA. Peer researchers recruited participants to fill in a paper-based standardized questionnaire in German, English, or French. The peer researchers obtained consent and either handed out the questionnaire for self-completion or went through the questionnaire with the participant. Participants did not receive any reimbursement for participation in the study. As an incentive, participants received a key chain, a leaflet that outlined local providers of HIV testing and support for PLHIV, and a condom. Questionnaires were sent back in pre-paid envelopes that were addressed to the study coordinators at the Robert Koch Institute (RKI).

The questionnaire was developed collectively by an expert group consisting of MisSA community members, experts working in HIV prevention or HIV clinics, and researchers [24]. Content was based on indicators suggested by the ECDC for behavioural surveillance in migrant populations [30] and the survey instrument of a sexual health survey for African communities from the United Kingdom [11]. The resulting questionnaire was tested prior to the pilot study, and adapted according to the feedback. It was then piloted and adapted again based on feedback from the pilot, pretested again and then finally implemented in this study [24,25,27]. The final questionnaire consisted of 51 questions and 23 knowledge items including: sociodemographic information; HIV, Hepatitis and STI knowledge and testing behaviour; sexual behaviour and risk factors; stigmatization; and preferred means of information. All covered items are available at [24,31].

Items about HIV, Hepatitis and STI knowledge were framed in a way that true information about the respective topic was presented (e. g. “You cannot tell from someone’s appearance whether he or she has HIV or not”) and the participant had the choice to select one of the following answers: “I knew this already”, “I was not sure if that was true or not”, “I didn’t know this,” or “I don’t understand this statement” [24].

Definitions

Together with the expert group of MisSA and stakeholders in HIV prevention, we defined MisSA as persons who at the time of the study lived in Germany but were born in an sSA country (first generation migrant) or with at least one parent originating from sSA (second generation migrant).

To determine uptake of HIV testing we defined a positive answer as participants who reported having had an HIV test at some point in their life. A negative answer was defined as participants who reported never having had an HIV test or who were not sure if they had ever been tested for HIV.

To determine uptake of HIV testing in Germany we considered participants who stated they had a test for HIV and had their last HIV test in Germany. We did not consider participants who reported to never have had an HIV test, or who were not sure if they were ever tested for HIV, or those who reported that their last test was not in Germany but in their country of origin, in another country, or they did not remember where they were last tested.

For general HIV knowledge, the number of HIV statements that were answered with “I knew this already” was counted, with a possible range of 0–8. The following statements were included: “HIV and AIDS also exist in Germany”, “AIDS is caused by a virus called HIV”, “You cannot tell from someone’s appearance whether he or she has HIV or not”, “There is a test which shows whether someone is HIV positive or not”, “HIV is not transmitted through kissing or shaking hands”, “HIV can be transmitted through sexual intercourse”, “There is no cure for HIV infection”, and “There are medications that can help people with HIV stay healthy”.

Two additional statements on HIV were analysed separately from the count of general HIV knowledge questions, the answers are presented in the descriptive analysis “Africans are NOT deported from Germany just for having HIV” and “In [study site] you can get tested for HIV anonymously and for free, e.g. at the local public health department”. In addition, one question on whether HIV is a topic that is discussed in their communities with the answer options “Yes” and “No” is displayed as it appears in the questionnaire.

To assess HIV related stigma we grouped answers to the question “How would you behave towards a person who is HIV positive?” in two groups. The possible answers to this question included: “I treat them like any other person”, “I avoid physical contact”, “I avoid being seen with this person”, “I blame this person secretly.”, “I behave differently: __”. In the category “I treat persons living with HIV like any other person” all participants that chose the response “I treat them like any other person” were categorized as “Yes” if none of the following statements were chosen in addition “I avoid physical contact”, “I avoid being seen with this person”, “I blame this person secretly.”, “I behave differently: __”. If any of the latter statements were chosen the answer was categorised as “No”.

Data analysis

All data was entered in Voxco Interviewer Web™ (an online survey and data collection software) and imported to STATA version 14 for analysis.

Participants were eligible to be included in the analysis if they had non-missing and unambiguous answers regarding sex, they or at least one of their parents was born in a sSA-country, and at least 60% of the questionnaire was filled in.

For the descriptive analysis we included only participants that had non-missing and unambiguous answers to the question “Have you ever had an HIV Test?”, i.e. participants that answered “Yes”, “No”, or “I don’t know”. We further excluded participants with missing data on education, length of stay in Germany, and German language skills. We conducted a descriptive analysis on the sociodemographic characteristics, knowledge, attitude, behaviour, and HIV risk factors amongst the study population, and compared those among males and females using a chi-squared test.

We conducted univariable (UVA) and multivariable analyses (MVA) to identify determinants associated with ‘ever having had an HIV test in Germany or abroad’; and a second MVA with the outcome ‘having had the last HIV test in Germany’. For UVA and MVA we excluded participants that reported an HIV test without consent at work, during the visa process, or during the asylum process to ensure that the results were relevant to voluntary uptake of HIV testing. Determinants considered in the UVA and MVA were sociodemographic characteristics (age, sex, education, income, health insurance, religion, length of stay in Germany and German language skills), HIV knowledge (general HIV knowledge, knowledge of the lack of association of HIV diagnosis with deportation, knowledge that LPHD offer testing as defined above) HIV stigma (treating people with HIV like anyone else as defined above and the answer to the statement “HIV is a topic discussed in my community”) and risk factors for HIV (sexual behaviour and STI diagnosis).

Protection of human subjects

No names, addresses, or other personal identifiers were documented on the questionnaires. Prior to data entry all questionnaires were given a unique identifier. Data protection procedures were cleared by the RKI data protection officer.

All peer researchers received training on the meaning and importance of informed consent. Prior to any recruitment, the study participants were informed of the study objective and voluntary and anonymous nature of the study. The questionnaire was only filled in if informed consent was verbally obtained.

Full ethical clearance was granted for this specific study (full title: “KABP (knowledge, attitudes, behaviour and practices)-study with optional HIV/STI-testing among migrants from sub-Saharan Africa living in Germany (MisSA-study)” by the Ethical Committee of the Charité Berlin in November 2014 (EA4/105/14).

HIV knowledge and stigma impact on the uptake of HIV testing

As has been shown in previous research, people who have had an HIV test were more often female, older than 25 years, had secondary education or higher, and earned more than 1,000 Euros per month [11,12,15]: Women are tested for HIV as part of antenatal care in many countries and therefore have higher odds of having had an HIV test [12]. Older adults accumulated more possibilities for testing during their lifespan, so their odds of ever having been tested has previously been described to be higher than the odds of younger adults [11,12].

Additionally, we were able to show that participants who declared treating people living with HIV like anyone else–as a proxy for non-stigmatization–had 62% higher odds of having had a test, which corroborates previous findings on the impact of stigma on the uptake of HIV testing [13,14]. Persons that perceived HIV stigma to be high were shown to be more likely to present late to care [14] and were less likely to be successfully linked to care [13]. Furthermore, we found that living in communities that discussed HIV almost doubled the odds of having had an HIV test, i.e. living in communities where HIV is actively addressed seems to have a strong effect on the willingness to get tested.

Participants who had varied sexual partners with inconsistent condom use had 28% lower odds of ever having had an HIV test, a result that might be due to fear of being positive and of being stigmatized. However, there is a possibility that both risky sexual behaviour and having never undertaken an HIV test are due to the same underlying factor, namely lack of knowledge about HIV. We were able to show that each general fact that was known about HIV increased uptake of HIV testing by 19%. Additionally, knowing one could not be deported from Germany just for having HIV increased uptake of HIV testing by 37%; and knowledge of LPHD’s VCT offer increased uptake of HIV testing by 52%. These results are in line with a study in sSA that identified knowledge and education as a driver to increase VCT uptake rather than sexual risk behaviour [32].

HIV testing in Germany shows least uptake among recent migrants

Among participants that did the last HIV test in Germany, being female was more strongly associated with having had an HIV test, possibly because most female participants were of reproductive age and HIV testing is a standard test offered as part of antenatal care to pregnant woman in Germany [33].

Interestingly, income and education were not associated with having had the last HIV test in Germany, whereas knowing about VCT availability at LPHD doubled the odds of having last had an HIV test in Germany. A possible explanation for these findings could be that VCT is available free of charge for people with no financial means at LPHD in Germany. However, not knowing about the offer of VCT at LPHD decreases the odds of being tested for HIV, indicating that there is a need to increase awareness of LPHD services and eligibility.

Migrants who came to Germany less than a year ago had, among all subgroups, the lowest odds of having had their last HIV test in Germany. This is in accordance to previous findings indicating that recent migrants are less likely to reach (or be reached by) health care as they do not (yet) know where to find preventive and curative care, how to access it, or both [34–36]. Second generation migrants, i.e. migrants who were born in Germany, showed the second lowest odds for having had their last HIV test in Germany. There is possibly less awareness of the risk of HIV infection among second generation migrants than among first generation migrants, as the HIV incidence is higher in sSA countries and more extensive HIV campaigns have been carried out in their countries of origin.

The results indicate that people who had an HIV test in Germany were more likely to be female, older than 25 years, living in Germany for more than five years (but not born in Germany), knew more about HIV, knew that migrants cannot be deported for having HIV, knew where to go for VCT and had sex only with steady partners. While those least reached by HIV testing services in Germany were more likely to be male MisSA, MisSA with a history of risky sexual behaviour, young people, and recent migrants.

Also, an STI diagnosis in the past did increase the odds of having had the last HIV test in Germany, but not to the extent it increased the odds for ever having had an HIV test anywhere. This could indicate that an STI diagnosis in Germany less often prompts the offer of an HIV test than elsewhere, and may present a missed chance for HIV testing.

Gaps in knowledge and action

The low uptake of HIV testing among young migrants and people with risky sexual behaviour requires innovative interventions that include outreach efforts targeting African communities. Our study shows that discussing HIV increases the uptake of HIV testing; thus, prevention efforts focusing on conversation and de-stigmatization regarding HIV are crucial. Our participatory research project indicates that community-owned intervention can have multiple advantages: a) it increases acceptance of studies or interventions, b) it can reach people in their daily life, in communities that would not be reached by passive offers expecting people to actively go to a certain institution for information, testing, and treatment, and c) it can directly increase the uptake of HIV testing. Previous research has already shown that the active involvement of MisSA community members when planning interventions is highly effective, as their participation increases adoption of preventive measures [37–41].

In addition, the introduction of routine opt-in VCT at clinics, e.g. general practitioner practices, that were mentioned as the preferred option for testing by participants, might increase the uptake of VCT in general–not only among migrants.

Further research will be needed to identify reasons for the very low uptake of HIV testing among second generation migrants in Germany and to identify strategies to improve their uptake of VCT. Involvement of MisSA communities in all steps of the research process will be key to determine research questions that matter to the communities as well as to researchers. The feeling of “ownership” by community members and their collaborative work with peer researchers were shown to be effective in addressing relevant questions with communities and in sparking discussions about HIV with and within the communities [11,15,17,38].

Limitations

Convenience sampling may have had an impact on representativeness, as may have the weekly feedback and adaptation process during sampling. However, we have no indication that the sample systematically differed from the population with regards to known characteristics of the study population, such as sex and country of origin. However, we do not know how well our sample reflects the local communities and the general MisSA community with regard to unknown characteristics.

We did not apply a standardized instrument to collect information with regards to stigma, as the collective development of a questionnaire with MisSA community members and stakeholders in HIV prevention was given priority over the use of standardized tools. The questions that assessed stigma covered dimensions of fear of infection and social judgement but it did not assess other equally important aspects of HIV-related stigma [42,43].

Recall bias and social desirability bias may further have biased obtained results. We did not use medical records to verify any self-reported information. In addition, the use of true statements (instead of questions) that allowed people to say “I knew this before” may have led to overestimation of knowledge.

As all information was obtained at the same time in this cross-sectional survey, no inference on causality can be made in the MVA with regards to the association between knowledge and behaviour with HIV testing.

We may have underestimated the proportion of those tested in Germany as we only obtained information if the last HIV test was taken in Germany. There is a possibility that participants had not had their last test in Germany but had an HIV test in Germany before.