Late presentation for HIV care in Southwest Ethiopia in 2003–2015: prevalence, trend, outcomes and risk factors

The Joint United Nations Program on HIV and AIDS (UNAIDS) has developed a 90–90–90 treatment target framework in order to end AIDS globally which aims at 90% of people living with HIV knowing their HIV status, 90% of HIV diagnosed patients receiving sustained treatment, and 90% of those on HIV treatment achieving viral suppression [1]. While diagnosing HIV infection is vital as the initial tread in the 90–90­90 targets, diagnosis per se is no longer sufficient [2]. Early diagnosis and access to treatment helps people with HIV to timely get and appropriately use HIV treatment [3] that further reduces the virus load and risk of morbidity and mortality. Nevertheless, late presentation for HIV care (LP) has been recognized as an impediment to meet the above mentioned UNAIDS targets.

LP is the result of being late in HIV diagnosis and/or late in linking with or in accessing HIV care [4]. The definition of LP is disparate and is contextualized using the threshold for ART eligibility [5]. To date, numerous definitions have been used including: i) when the baseline CD4 count is < 200 or < 350 cells/μl and/or with an AIDS defining disease [3, 6, 7], ii) when AIDS defining conditions are diagnosed either before or during the period to an HIV diagnosis [8], iii) when AIDS defining conditions are diagnosed in the subsequent 6 months period to an HIV diagnosis [9], or iv) when AIDS defining conditions are diagnosed 12 months period to an HIV diagnosis [10].

LP is associated with increased risk of HIV transmission [11], ART drug resistance [11], and health care expenses [12]. Additionally, LP has been acknowledged as a challenge for the achievement of the ambitious UNAIDS 90–90-90 targets [13, 14]. For the first 90, high magnitude of late HIV diagnosis reflects that there are a number of people who did not know their HIV status. For the second 90, LP results in poor health outcomes and this interrupts the sustainable uptake of the treatment [13]. Furthermore, LP significantly contributes for pre-ART deaths and, this in turn, reduces the number of HIV diagnosed patients on ART [15]. For example, a study from South Africa reported that ART initiation at the time of first presentation to ART clinic boosted treatment uptake by 36% [16]. For the third 90, LP lowers the number of CD4 cells and increases the number of viral counts, and this causes clinical, immunological or virological failure [14, 17]. Previous studies have shown that late diagnosis appeared to be the main reason for virological failure, and ART initiation at first visit increased viral suppression by 26% [16].

LP has been reported to be a significant problem across the globe. In Europe, overall prevalence of 15–66% has been reported [18, 19]. The magnitude of LP in Asia was very significant (72–83.3%) [20], and in Africa, the overall prevalence has been reported to be between 35 and 65% [21, 22]. Nonetheless, heterogeneity in its measurement limited direct comparisons [23]. In Ethiopia in 2015, there were 39,140 new infections, 768, 040 people living with HIV and 28, 650 HIV/ADS deaths [24]. Universal access to ART in the country was launched in 2005 [25], and to date, the coverage of ART—the percentage of people on ART among those in need of treatment [26]— is 52% [24]. However, the status of timely presentation for HIV care is yet to be assessed. One cross-sectional study from northern part of the country has reported a LP prevalence of 68.8% [27].

Demographic, behavioural and clinical factors contributed for LP [6, 20, 28, 29]. For instance, being female, older age, rural dwellers, alcohol users, ‘Khat’ chewers, cigarette smokers, being diagnosed with sever co-morbidities, perceiving HIV related stigma, having contact with commercial sex workers and being exposed to risky sexual behavior were the factors associated with LP [6, 20, 28, 29]. In Ethiopia, other studies have assessed factors affecting LP [6, 27, 29], and all except one were from the northern part of the country.

However, it is well known that the southwestern part of the nation has different cultural and socioeconomic characteristics. It also has the highest HIV prevalence (6.5%) in the country [30] and may have different LP factors which need to be understood to address HIV in these settings. In addition, for patients who started ART, no study has been conducted to assess the outcome and trends of LP. The prevalence of LP among children has also not been determined. Given the above gaps, and the clinical and public health importance of early HIV diagnosis on timely ART commencement, it is imperative to comprehend the LP situation and recommend effective programs that facilitate early presentation for HIV care in Ethiopia. Furthermore, addressing LP may have a substantial contribution for SDG-3 to have good health and wellbeing, and particularly for SDG-3.3 to end HIV epidemics by 2030. This paper examines the prevalence, trend, outcomes and risk factors of LP among children and adults enrolled for ART in Jimma University Teaching Hospital (JUTH), Southwest Ethiopia.

Of the 8172 patients enrolled for HIV care between 21 June 2003 and 15 March 2015, 5299 (64.8%) patients on ART, the study population for the study, were included. Of the total patients enrolled for ART, 4900 (92.5%) were adults and 399 (7.5%) were children.

Table 2 demonstrates the characteristics of HIV patients on ART. Among the children, 58.1% were aged 5–15 years, 52.4% were males, and 73.9% were Christians. The majority of the children (79.4%) had moderate or severe immune-suppression, and half of the children had baseline WHO clinical stage 3 or 4. A total of 114 (28.6%) children had Tb/HIV co-infection. The median time on ART and estimated survival time, respectively, was 40 and 104.2 (99.8–108.5) months. Among adults, three fifth (59.8%) were females, about half (48.7%) were married and 39% had primary school education. Two fifth (41.6%) of adults had no history of HIV testing. A significant number of adults (73.6%) had baseline CD4 count below 200 cells/mm³, and 54.3% had WHO clinical stage 3 or 4. Over a quarter of adults (27.9%) had Tb/HIV co-infection. Twenty nine (0.9%) adults changed ART regimen during the course of the period. The median time on ART was 49 months, and the median estimated survival time was 121.9 (95%CI: 120.3–123.5) months.

LP has been described as a sizable obstacle to attaining the UNAIDS 90–90-90 and 95–95-95 targets [7, 14]. This study has shed light on the general problems of late HIV care—magnitude, trend, outcomes and its risk factors. In the current study, the overall prevalence of LP was considerably high (65.5%). Furthermore, the trends of LP had shown persistently elevated prevalence (between 54% and 83%) although a lessening pattern was observed. This finding is consistent with another finding conducted in the country [27]. The prevalence of LP in the current study (65.5%) was lower (72–83.3%) than the prevalence from studies conducted in Asia [20], but higher than the findings from other studies in Africa that reported between 35 and 65% [21, 22]. This implies that LP in Ethiopia is still highly prevalent even after the introduction of universal ART.

The high and persistent LP prevalence may be due to: i) lack of information [35], ii) persistently high level of HIV related stigma [3, 27], iii) low HIV risk perception [27, 35] especially among high risk groups [36], iv) use of traditional treatment [27], v) poor integration between modern medicine and traditional healers, and vi) phasing out of international funding agencies [37]. Additionally, it could also be attributed to poor access to HIV services [37, 38]. For example, only 79% of the total health facilities in Ethiopia deliver HIV counseling and testing services [39]. Primary Health Care principles and scholars describe characteristics of accessible health systems to be approachable, acceptable, available, affordable and appropriate for the target population [40, 41], and thus raising a question whether HIV services are accessible to all HIV patients in Ethiopia. As such, LP issues should be given top priority if Ethiopia is to meet the UNAIDS targets.

Several strategies including the use of technology have been recommended to reduce LP prevalence in developing countries. For example, in collaboration with UNICEF, Amukele and colleagues successfully piloted unmanned aerial systems (drones) for transporting laboratory specimens to reduce late infant diagnosis in Malawi [42]. Other programs such as using mobile text messages [14], home [43] and community-based HIV testing [44, 45] have also been recommended to meet the HIV diagnosis target. Furthermore, encouraging repeat HIV testing [46, 47], HIV testing services delivery by lay workers [48], and self-testing [49] can tackle the substantial gap in HIV diagnosis in low-income countries like Ethiopia. The mandatory HIV testing strategy that has been implemented in China since 2005 for testing at-risk groups such as drug users, inmates, and commercial sex workers along with their clients was found to be an effective strategy to heighten early HIV diagnosis [7] that Ethiopia could consider. Interventions such as lay counselor home visits [50], home visits by peer supporters [51] and informational brochure provision [52] were also important programs in linking patients to ART clinics timely after HIV testing. In addition, the application of rapid or point of care CD4 count technology has shown to enhance the number of eligible patients for ART whereby the frequency of appointments is reduced, and early ART initiation is increased [53].

Compared to the early presenters, great majority of late presenters had died in the current study, and this is similar with findings from other studies conducted elsewhere [54, 55]. It is plausible therefore to argue that late presentation leads to a greater risk of: rapid progression to advanced AIDS stage, compromised immune response, poor treatment response, and finally death [3]. Similarly, consistent with findings of other studies [56, 57], the majority of adults who presented late for care had discontinued and developed immunologic failure. This could be highly possible since late presenters progress easily to advanced AIDS stage, a stage characterized by marked CD4 reduction, multiple comorbidities and poor overall functional status [58, 59]. Subsequently, this leads to poor immune recovery even after treatment initiation, and increases the likelihood of ART toxicity that deters patients to take the treatment regularly [60]. The prevalence of LP among adults was higher than children. This might largely be due to the ‘opt out’ screening programs for pregnant women and delivery of HIV care (testing and treatment) to children born to affected mothers timely [61].

Adult late presenters were more likely to be younger, females, Tb/HIV co-infected, with no history of HIV testing and enrolled to HIV care in 2011 and before. Unlike other studies in Africa [62, 63], older adults were less likely to delay for HIV care than their younger counterparts. We found the finding surprising. HIV disease progresses with time, and it would be expected, that individuals diagnosed with HIV at a higher age would also have advanced disease progression (lower CD4 cell count) because they, on average, had a longer time span between time of infection and time of diagnosis. However, the presence of high HIV related stigma among young adults [64] hampers HIV testing and may be linked to delays in seeking HIV care. In addition, it is also possible that older adults assume a caring responsibility for their family and might realise the need to access HIV care service consistently increase their longevity and to achieve the self-imposed caring responsibility. Unlike in some others [56, 63], females were more likely than males to delay for HIV care. This might be because females have low understanding and comprehensive HIV care knowledge [3]. The high probability of females for LP might also be explained by the fact that 62% of females in care did not have a previous history of HIV testing. Females are also known to use traditional healers more than males, which may lead to commencing the HIV care late [3, 35].HIV related stigma is higher among females than males [65]. It is also known that the health seeking behaviours of females in urban and rural southwest Ethiopia are lower compared to males [66].

The association between Tb/HIV co-infection with increased LP replicate findings from other studies [4]. It has been stated that Tb is inextricably linked with HIV, causes a synergistic combination of illness with HIV, facilitates the progression of HIV disease to advanced stage, and thereby deters patients from linking to care timely [31]. Furthermore, focus has to be given for Tb/HIV co-infection, as Tb remains the highest mortality risk among HIV infected patients [13]. The absence of previous history of HIV testing in association with LP could be linked with poor awareness of the care [3], less access to HIV testing and/or ART clinic [39], high HIV related stigma [67], fear of diagnosis [68] and feeling of wellbeing [69].HIV patients who were enrolled for HIV care in 2012 and after were less likely to present late for HIV care as compared to those enrolled in 2011 and before. This may be attributed to: i) improving awareness to HIV care; ii) improving access and availability to HIV care; and iii) reducing perceived HIV related stigma.

The study has the following limitations. Firstly, data were collected from JUTH, a referral hospital that also receives referrals of patients with advanced stage. Hence, these study participants are not necessarily representative of HIV patients who attend their follow up in health centers or lower health care setups. Secondly, we did not assess the annual proportions of LP across HIV testing strategies (voluntarily counseling and testing, provider initiated HIV testing and counseling (PITC), Outreach or ‘opt out). Previous studies have shown that PITC was not found more effective program for early HIV diagnosis than targeted HIV counseling [7]. Thirdly, the use of conservative definitions for LP is not able to differentiate whether the late presentation is before diagnosis, between diagnosis and first entry to care, and between first entry to care and ART initiation. A gold standard definition for LP among general HIV positive population and special groups such as HIV positive mothers and Tb/HIV co-infected patients for low-income countries is yet to be established. Fourthly, we found no statistically significant predictor for LP among children, and this could be due to small sample size. Finally, the presence of incomplete data may bias the precision of estimates.

In summary, three of five HIV infected people presented late for HIV care, and the annual proportions of LP persistently remained high across the 10 year period. The majority of HIV infected children and adults who presented late for care had discontinued, transferred out and developed immunologic failure. Late presenters (adults) were more likely to be younger, females, Tb/HIV co-infected, no history of HIV testing before diagnosis, and enrolled to HIV care before 2012. A large sample size should be considered to assess factors influencing for LP among children. Prioritizing the aforementioned risk factors, tremendous efforts are necessary to curb LP and further achieve the UNAIDS goal. Some of the strategies and programs that help to decline LP are use of innovative technologies, home and community based HIV testing, encouraging repeat and self-HIV testing, mandatory HIV testing and effective linking strategies to care after HIV diagnosis. HIV related stigma should also contextually be tackled since it continues to be a lingering issue among HIV infected people.