Study screening and selection processes
About 4736 articles were searched from different data sources. Of which, 29 articles from PubMed, 129 articles from Google scholar, 2094 from HINARI, 2480 from Google, and 4 articles from university repositories. However, 3154 articles due to duplicates; 1362 articles due to irrelevant titles and abstracts; 140 articles due to study setting and design were removed. About 80 articles were selected for full-text review. Of these, 62 articles were excluded after full-text review. Finally, 18 studies were included in this systematic review and meta-analysis to estimate HIV positive status disclosure and identify the determinant factors of HIV positive status disclosure among PLWHA in Ethiopia (Fig.1).
Characteristics of the included studies:
In this systematic review and meta-analysis, a total of 18 articles with a total of 7084 study participants were included. Regarding the region/location of studies conducted, about one third of the studies were from Oromia region (1, 20-24), Amhara region accounts the second largest segment of the included studies 22.22%(13, 25-27), Tigray region has contributed three studies (8, 28, 29), two regions SNNPs(30, 31) and Addis Abeba(32, 33) each supplied a couple of studies and one study was found from Hareri region(34). Regarding the study design all the included studies were cross-sectional. Concerning the JBI assessment nineteen studies were review and the highest score was 100% which has occurred among six studies(13, 20, 23, 25, 28, 30, 33) and the lowest value was 33.33% from a study conducted in Metu and Gore towns Illubabur Zone Southwest Ethiopia(35) which is subjected to exclusion from the review; However, from the included studies the lowest value was 66.66% among two articles(26, 32). In addition to this, the lowest and the highest response rate of the included studies was between 95.5 and 100. The highest prevalence of HIV positive status disclosure among people living with HIV was reported in Amhara region, Kemissie zone(93.1%) (13) and the lowest was from Oromia region, Bale Zone (52%)(36). Based on the JBI quality assessment the average quality of the included studies was 62%. Nineteen articles were involved into the systematic review and Meta-analysis as their JBI score is 50% and above. One study conducted in Gore and Metu towns Oromia region(37) has been excluded from the review as it accounts about 37.5% of the JBI quality assessment(Table 1). Moreover, JBI quality assessment for studies reviewed to determinant factors affecting HIV positive status disclosure has been performed and almost all studied scored more than 80%.
Prevalence of HIV positive status disclosure:
In the random-effects model, HIV positive status disclosure of people living with HIV in Ethiopia was found 75.95% (95% CI: 69.93-81.98%, I2 =98%, P<0.001) (Fig 2.).
Subgroup Analysis:
Subgroup analysis has been carried out by the type of population and regions of the studies. As a result, the output of the subgroup analysis with three categories of the population such as ART users, women on ART and HIV positive pregnant women attending the PMTCT care unit revealed a significant difference has been observed between PMTCT users (82.74%) and ART users. However, there was no difference between the adult PLWHA (74.84%) and women on ART (74.25%) (Fig.3). On the other hand, noticeable variation among study regions has been detected. The highest and lowest pooled estimated HIV status disclosure was at Amhara region (82.78%) and Tigray region (54.31%) respectively (Fig 4).
Publication bias Test:
The presence of publication bias was evaluated both subjectively as well as objectively. The symmetrical alignment of studies in the funnel plot (Fig 5) and eggers regression graph (Fig 6) revealed the absence of publication bias among the included studies. However, the statistical value of eggers regression test proved the presence of publication bias (P=0.001).
Sensitivity analysis:
There was no influential study that affects the pooled estimate of HIV positive status and caused disparity between studies, according to the sensitivity analysis (Fig 7).
Determinant Factors of HIV Positive status disclosure to sexual partner
There were a total of ten studies that reported the effect of knowing the HIV status of sexual partner towards HIV status disclosure of PLWHA. Of these, three of the studies were from Tigray region(8, 28, 29). Amhara(13, 27), Oromia(22, 24) and SNNPs(30, 31) each contributed a couple of studies and one study was from Addis Abeba(33). All the included studies demonstrate a significant association with the HIV status disclosure. About 70% of the involved studies showed an AOR greater than 10. Hence knowing HIV positive status of a sexual partner was almost 20 times more likely to disclose their HIV test result to their sexual partner than those who didn’t know the status of their sexual partner 19.66(10.19,37.91; I2= 98.8%; P<0.001)(Fig 8). The sensitivity analysis revealed no small study effect that distorts the pooled estimate. Furthermore, both the subjective and objective publication assessment confirmed the absence of bias (P-value = 0.297).
The second outcome variable tested to verify the presence of association with HIV positive status disclosure was a discussion with a sexual partner before HIV testing. As a result, a total of seven studies from four regions were identified reporting effect of prior discussion about HIV testing with a sexual partner with HIV positive status disclosure. From these, three of these studies were from Tigray region(8, 28, 29), two from Oromia region(22, 23) and Addis Abeba(32) and Amhara region(13) each contributed a single study. Of the included studies one study(28) does not show an association between having prior discussion about HIV testing with a partner and HIV positive status disclosure. The rest of the studies showed a positive association with a minimum AOR of 2.99(13) and a maximum AOR value of 12.28(32). Hence, the likely hood of disclosing HIV positive status to their sexual partner among clients who had prior discussion about HIV testing was nine-fold higher than their counterparts 9.18(95% CI = 5.53, 15.24; I2= 93.5%; P<0.001) (Fig 9). The effect of small studies has not been observed. Furthermore, the eggers regression test revealed the absence of publication bias (P = 0.852).
The third variable identified to examine the presence of association was getting pre-test HIV counseling. Six primary studies that showed the effect of HIV counseling on HIV positive status disclosure were identified and reviewed. Amhara(25, 27) and Tigray(8, 29) regions each contributed two studies, while Oromia(24) and Southern Nations Nationalities and Peoples SNNPs(30) each added a single study to the review. Originally, getting pretest counseling service demonstrated a significant positive association with HIV positive status disclosure with an AOR ranging between 2.8 and 6.25. On top of this, the meta-analysis revealed that PLWHA who got pretest counseling service has about four times higher chance of disclosing their HIV positive status to their sexual partner/s 4.29(95% CI = 2.56, 7.21; I2= 98.6%; P<0.001)(Fig.10). Moreover, the sensitivity analysis and eggers regression test testified the absence of influential study and publication bias (P = 0.349).
Lastly, this systematic review and meta analysis (SRMA) also evaluates the effect of participation in the HIV Association on HIV positive status disclosure. Hence, about four articles from three regions; two from Tigray region(8, 28), one from Amhara(13) and the other from Oromia(21) regions were analyzed. Each study independently showed a significant association with an odd ration ranging between 2.09-5.2. The meta-analysis revealed that PLWHA who are members of HIV associations have more than a threefold chance of disclosing their HIV positive status to their sexual partner/s than their counterparts 3.34(95% CI: 2.17-5.12;I2 =91.3%; P<0.00)(Fig.11). Furthermore, the sensitivity analysis and eggers regression test testified the absence of influential study as well as publication bias (P = 0.147).