The average age was 30 ± 10.8 years. The most represented age group was 20 - 25. Depending on years, the average age of CADI’s clients ranged between 27.9-year-old and 31.8-year-old with extremes varying from 1.0-year-old to 87.0-year-old. Significant progress was made from one year to the other (p < 0.0001). Women were mostly represented (55.5%) with a sex ratio of 0.80. But between 1996 and 2000, male was predominant. Subsequently, the number of women attending the centre increased and was higher than that of men. Most clients were living in couple (79.5% [78.6 - 80.4]). From 1996 to 2000, attendance rates increased although between 2001 and 2007, numbers decreased. The number of widowers attending the centre considerably declined since 1996, whereas the bachelors were stable except in 2003, 2004, and 2007 (Figure 1).

Among centre’s clients, 52.90% [52 - 53.8] had no occupational activity. Over three-quarters were educated representing 78.3% [77.5 - 79.1]. From 1996 to 2014, educated people attended the centre more often (Figure 2).

Almost half of the educated people (46.3%) attended secondary school, 19.47% primary school, and 11.43% high school. 23.82% of clients always used condoms. When testing was conducted, 21.5% stated that they have several sexual partners.

The desire to know their serological status was the first reason for screening reported by 36.4% [35.35 - 37.41] of the clients.

The reasons for HIV testing have improved in 5 steps (Figure 3).

The first stage was the first two years during which testing was determined by the appearance of symptoms. Between 1998 and 2000, the desire to know one’s status was more common. From 2001 to 2007, testing was mostly requested by a third person (physician), and then from 2008 to 2014 seeking to know one’s status became the main reason put forward (Figure 4).

The 16.4% (15.7 - 17.1) of the cases from the 11,451 clients included in our study were tested HIV positive. The results were indeterminate for 299 clients (2.6%); 9277 clients were tested HIV negative. Figure 4 shows the progression of HIV prevalence from 1996 to 2014, with the peaks of 22.9%, 34.1%, and 22.8% noted in 1999, 2001, and 2006 respectively. In 1996, almost one out of two clients had a positive HIV result against 4.2% in 2014. From 1996 to 2014, HIV prevalence among the clients considerably decreased. From 1997 to 1998, the average prevalence rate was about 15%. From 2007 to 2011, the average of the HIV sero-positivity rate was stable (13%). The year 2013 was characterized by the lowest HIV prevalence rate (3%) despite high attendance at the centre where 23,279 voluntary screenings. In 2014, an increase in HIV prevalence (11.5%) was recorded compared to 2013. From 1996 to 2003, the number of undetermined results was growing. Since 2003, this number has been decreasing. The peaks of indeterminate HIV results were noted in 2000, in 2003, and in 2006 with 17, 13, and 06 cases respectively, for 100 screenings performed. The number of indeterminate results since 2007 was relatively less than 1%.

Clients aged from 30-year-old to 35-year-old were more affected by HIV infection (21.9%; p < 0.0001). Sero-positivity rate was significantly higher among women (66.4%; p < 0.0001). Widowers were more HIV-affected (41%; p < 0.0001). The sero-positivity rates were higher among uneducated clients representing 36.73%, against 31.88% for those who attended secondary school; it was 26.32% for those with primary school background and 16. 81% for high school background (p < 0.0001). The sero-positivity rate was significantly higher among clients with several sexual partners (19.82%; p < 0.0001) and those who never or occasionally use condoms (83.3%; p < 0.0001). Sharing information related to the serological status with another person was significantly lower among HIV positive patients (84.3%; p < 0.0001) compared to HIV negative clients (93.2%; p < 0.0001).

In 2006, the average age of screened people in our centre was higher than the average age of the general population (21.8-year-old) with the most represented age group from 20-year-old to 25-year-old [7] . In a study carried out among the general population in Ouagadougou (in Burkina Faso) targeting attendants of a screening centre [8] , Kirakoya et al. noted an average age (29 ± 8 years) closed to our study.

Related to the gender, our population was mainly composed of women which is similar to the structure of the general population in Burkina Faso [7] . Somé et al. reported a considerable number of women in fixed sites [2] . This was also noted by Biadglegne et al. [8] from Ethiopia and Dagnan et al. [9] from Côte d’Ivoire who respectively recorded 56.4% and 6 times more HIV-positive women. Such data express the increasing number of women with HIV due to their biological, socio-cultural and economical vulnerability. From 1996 to 2000 where there was a predominance of men and more women from 2001 to 2014, that could be explained is the fact that, since 2000, the national policy for HIV control gives more opportunities to women for counselling relating to the voluntary HIV screening, especially during prenatal visits. In addition, women tend to get themselves tested as part of a fixed strategy compared to men [2] .

The education rate in our study (78.3%) was higher than that of the general population (26.8%) [7] . Such data probably result from the various sensitization campaigns in schools [10] . Our study showed that more educated people get themselves tested compared to the uneducated. In addition, more uneducated clients were HIV-infected while most of Burkinabe people were uneducated [11] . From Dillnessa et al in Addis Ababa [12] , people with primary school level were four times more likely to use voluntary testing facilities.

From 1996 to 2000, most clients (79.5%) were in couple because in the 1996’s, HIV infection was considered as occurring among individuals at risk (singles and couples) [13] . This was noted by Kirakoya et al. who highlighted that 48.6% of HIV screening Centre’s clients were married couples living in Burkina Faso [14] . In the same line, Fiorillo et al. [15] from Tanzania and Hensen et al. [16] from Zambia noted a high testing rate among married people and widowers. We also noted that widowers were significantly more HIV-affected (41%). On the other hand, divorced people were the least HIV-affected (9.03%). Biadglegne et al. [8] observed that married people (2.8 OR; 95% IC: 1.70 to 4.71, p < 0.001) and those in couple (3.8 OR; 95% IC: 2.47 to 6.09, p < 0.001) were the most exposed to HIV infection. Such data draw attention to the importance of preventing HIV infection among couples considered by some authors as high-risk persons. An infected person can contaminate his/her partner and this raises challenging ethical, human rights and public health issues [17] [18] as well as the sharing of serological information within the couple.

Few bachelors attended the centre since its opening though between 2002 and 2008 their number slightly increased. Such results raise concerns because this group of people is considered at risk by some authors [19] . Yet in the USA, Schechter-Perkins et al. reported that bachelors were people more likely to accept being tested [20] .

Less than half of the people tested at the centre (23.8%) said they regularly use condoms. Sero-positivity rate was considerably higher among clients with multiple sexual partners (21.5%). More than 83% of the clients said that they never or occasionally use condoms were HIV positive. These results show high rates in risky behaviours especially among the youth. It should be reminded that it is very delicate asking questions on one’s sexual behaviour; so risky behaviours might not be reported. Is risk taking an indication of slackening vigilance or trivialization of the HIV infection? Studies show that in the general public reactions changed from one extreme to the other, that is, from a deep fear to be contaminated to the fact that HIV has become very common since efficient therapy have been introduced [19] . An increasing number of people are no longer scared of the disease as it is considered chronic. In a prospective study conducted in New York Calderon et al. [21] noted that 24.3% of patients admitted to the emergency unit who accepted to be HIV tested did not use condoms. Sherr et al. [22] in a Zimbabwean rural cohort found no relationship between risky behaviours and screening counselling. People want to know their serological status, therefore the first reason for them to get tested [14] . In the same line, for Ndiaye et al. [23] in 69% cases, curiosity to know one’s status was the first reason for testing [14] whereas in Great Britain Mc Garrigle CA et al. [24] pointed out that attendance to screening centres was closely related to having many sexual partners and new foreign sexual partners. Fiorillo et al. [25] reported that infidelity and the existence of new sexual partners were the main reasons for screening. They also considered reasons for leading people who have already been tested to ask for another testing. Decision to share one’s serological status with another person was considerably low among HIV-positive patients (84.3%) compared to negative ones (93.2%; p < 0.0001). Disclosing one’s serological status depends on the testing results. Very few of those tested positive share the information in this regard. In Burkina Faso, Guiro et al. [26] reported that only 22% of patients share their serological information with their partner. Stigmatization and moral conflicts are apparently the major reasons for which information is not shared, which is an obstacle to HIV infection control, especially optimal treatment management.

The general prevalence of HIV infection among CADI clients was 16.4%, close to 18.4% reported by Fiorillo et al. [15] in Tanzania. In Uganda, Johnston et al. noted a prevalence of 12.3% [16] . In 2013, Dagnan et al. [9] reported a lower rate of 5.30% [17] which can be explained by the size of the sample, the duration and site of the study [10] .

HIV prevalence among clients of the centre considerably reduced from 46.15% to 11.5% between 1996 and 2014. Progression in the prevalence stems from the increasing number of people attending the centre. At the beginning of HIV infection in Burkina Faso, fear and inaccessibility of screening centres were the major causes for a very late screening, often at the AIDS stage [26] . Most clients were therefore sick by the time they attend the centre. In 1996, almost one over two clients (42.7%) showed HIV symptoms with high prevalence (46.1%). From 1997 to the 20s, mainly high-risk people attended the centre, i.e. those having unprotected sex very often with several sexual partners, those whose partner is seropositive, or died of Acquired immunodeficiency syndrome (AIDS). Later, with sensitization conducted by the various actors of HIV infection control, prevalence reduced and reasons for screening changed. Therefore, prevalence reduced after 2001; from 2007 to 2011, sero-positivity rates remained unchanged about 13%. From 2006 to 2014 no other reason than the wish to know one’s serological status motivated CADI clients to get tested. Since 2006, HIV testing seems to be performed as a routine though attendance rate has remained low.

Over time some authors reported different variations in HIV prevalence. Indeed, during voluntary screening campaigns held from 2006 to 2010 in Burkina Faso, Somé et al. reported a 24.6% prevalence rate [2] .

Kee et al. indicated that HIV prevalence among patients attending health centres in Korea increased from 2000 to 2005 (1.3% for 10,000 clients to 5.3 for 10,000 clients) [27] . Yet, they noted that HIV prevalence remained stable since 2005. Baryarama et al. [28] noted that HIV prevalence reduced from 1999 (23%) to 2000 (13%), then went up to 15% in 2013.

United Nations Programme on HIV and AIDS (UNAIDS) reports on HIV epidemic indicated a decrease in the prevalence [1] . Such results show the progress achieved in the world and particularly in Sub-Saharan Africa as part of HIV control.