Impact of migration origin on individual protection strategies against sexual transmission of HIV in Paris metropolitan area, SIRS cohort study, France

Our results show that people’s HIV prevention attitudes differ according to their country of origin. First, models adjusting only for demographic features show that Maghrebi women and French women of Maghrebi origin try less to protect themselves from HIV, and rely less on condom for protection against HIV than French women with French parents. This result suggests that Maghrebi women are insufficiently reached by prevention campaigns. It may be because they have a limited access to the messages, e.g. because they are not an actual target group in a context where most new HIV infections in foreign women are diagnosed in women originating from Sub-Saharan Africa. It may also be due to messages being insufficiently tailored for these persons, that are then not clearly understood or don’t translate into a change in attitude or behaviour, e.g. because of the women’s status in some of the Maghrebi immigrant subgroups that hinders communication about sexuality or because they distance themselves from risk groups. As noted already 10 years ago in France [27], Maghrebi population –originally from countries where the HIV epidemic is concentrated in high-risk groups [37] – still miss prevention interventions as compared to the rest of the general population in France. This may result from a low level of acculturation in France, which is a barrier to health-related communication and specifically prevention messages [38] and deserve the reinforcement of specific, targeted, information and prevention programs. These associations between Maghrebi citizenship or origin and a reduced intention for protection against HIV disappear when variables reflecting the women’s socioeconomic situation and sexual biography are included into the models: it seems that these characteristics outweigh migration origin and/or – at least partially - explain the origin-related differences previously observed.

A very similar situation was observed among the few foreign women from non-African and non-European countries included in our sample. The small number of these women (a heterogeneous population structured in small communities in the Parisian population) and, also, the fact that our sample missed the neighbourhoods inhabited by Asian communities (although numerous in the Parisian area) make this result difficult to interpret and infer.

The second foremost result is that, in fully adjusted models, foreigners from SSA and French men of SSA origin declare more often that they do something to protect themselves from HIV than the French with French parents. This can be due to recent prevention campaigns specifically addressed to SSA populations in France (and/or those in their countries of origin) [13, 15] since perceiving oneself as being part of a risk group influence prevention behaviours [21, 22]. It can also be related to certain forms of positive discrimination in individual HIV prevention promotion performed by health care workers, similarly to what has been previously described for HIV testing [22, 26]. Alternatively it may be related to a social desirability bias that is known to be more frequent in cultures based on oral tradition such as those in SSA [39]. It is worth noting that this more frequent declaration of protection did not translate into a more frequent recourse to the use of condoms as a means of protection in any of these 3 subgroups. If, generally speaking, most of the individual characteristics associated with the intention to protect oneself against HIV were also associated with a condom-based approach, this was less true for the origin from SSA, possibly because people in this category opt more for other means such as HIV testing, as described in the Paris area [40]. Concerning women, it has been shown that they are used to resort to other means of protection than condoms only [34]. These various results confirm the need for a culture-specific approach of HIV prevention and in the interpretation of results of HIV related surveys [11, 22, 24, 26, 40].

Globally, our results show that French people with foreign origin are in an intermediate situation between French people with French parents and foreigners from the same continent. This likely reflects the acculturation process [4] and, more precisely, the maintenance – or the re-appropriation - of cultural norms inherited from the culture of origin. Given the definition of the different categories of French people with foreign origins used here, we cannot go further in this cultural interpretation and distinguish between the situation of people born French in France to foreign parents and those of the immigrants who acquired the French citizenship (each representing about half of the above mentioned categories). Conversely, even if limited in their interpretation, our results are the first quantitative results ever published in France in a representative, population-based sample about the situation of the French people of foreign origin in the field of HIV prevention (and of HIV epidemiology in general). We think that replicating such approaches in other European countries would be very useful at the time when European public health institutions, such as the European Centre for Disease Prevention and Control, emphasize that prevention efforts should not only aim at the general population but also at groups such as immigrants or minorities [41, 42]. Their results may usefully precede and/or trigger the development of HIV prevention interventions targeted to immigrants at a larger scale or with more significant resources.

Looking at covariables, some results deserve being discussed as they confirm or infirm previous findings. We found that people with a religious affiliation were less likely to intend for protecting themselves against HIV, similarly to what has been observed previously in France [40, 43] but this influence was not observable when we adjusted for variables related to sexual biography. This suggests that the impact of religion on means of prevention may be limited to (or mediated by) an influence on the sexual biography even if further analysis would be necessary to explore this pathway. Conversely, our results clearly show that affective and marital status has a major impact on the individual’s definition of protection means for oneself and plea, with others, for a systematic adjustment for the couple status to study these HIV prevention strategies, which was not done in the above mentioned study [43]. We also observed that individuals stigmatizing HIV+ people were less likely to intend to protect themselves against HIV and to opt for condom-based approaches for protection. This result was expected since self-protective attitudes and stigma share common determinants such as perceived severity of AIDS and perceived self-susceptibility to HIV infection [44]. Nevertheless the fact that the stigma and the perceived susceptibility to HIV infection are both independent factors in our models and a restricted mediation analysis (not presented) showing that only 12.3 % of the effect of stigma on the intention of protection is mediated by the perceived susceptibility, suggest that stigma and perceived susceptibility are related with intended prevention behaviours mostly through unconnected pathways, unlike suggested elsewhere [45]. The importance of other determinants, such as stereotypes, has not been evaluated in our study and it is difficult to do so in a quantitative study like this one. Obviously, qualitative studies make invaluable contributions to a deep and comprehensive knowledge of the various determinants of these attitudes and behaviours [11].

This study presents some limitations. Our sample is restricted to Paris metropolitan area and our results cannot be extrapolated to the whole France. However since this region bears the most important part of the HIV burden in mainland France, our conclusions are pertinent to help in the management of the epidemic in France but also in foreign settings with similar features [46]. The attrition rate between 2005 and 2010 might have influenced the representativeness of the sample, but has few chance to affect the relationships between variables in a sample of this size [47] and we must remind here that the non-reinterviewed people have been replaced by people selected by the same random procedure (with a quite good response rate for a population-based survey). Also the sampling design didn’t allow for reaching very vulnerable populations such as homeless people, among which immigrants are overrepresented [48] and we interviewed only people speaking French when non-French speaking people accounted for 6.8 % of individuals contacted for the study. If those people could have been included and interviewed in SIRS, the difference between French people and foreigners might have been accentuated (by construction, our sample does not allow us either to test the influence of speaking French or not). The questionnaire of the SIRS surveys didn’t allow for the investigation of ways of transmission of HIV other than sexual and was also limited by the measurement of multipartnership over the last 5 years only, while it is usually calculated on the sexual activity of the previous year [19, 49]. Also, several factors known to impact the strategies of protection against STDs were not measured, such as the social context in which a sexual relationship takes place [50]. Moreover, since our questions investigated attitudes and declared behaviours but not practices (e.g. the utilization of condoms during last sexual intercourses or according to different kinds of sexual partners), we are unable to evaluate the gap between intention and effective practices [19].