Assessment of the proportion and the factors associated with partner and family based index case HIV testing in Woliso town, Oromia, Ethiopia: a cross-sectional study

Background: Despite the enormous expansion of HIV testing services (HTS), an estimated 40% of people with HIV infection remain undiagnosed. In Ethiopia, the current working UNAIDS spectrum estimate for PLHIV is 649,264, of the estimated PLHIV, only about 72% know their status. Methods: A facility based cross-sectional study design with internal comparison was conducted among randomly selected 346 people currently on ART in Woliso town. Data entry carried out by Epi Info™ version 7.2.3.1 and analyzed using SPSS version 21.0 statistical software for Windows. Results: Among 345 study participants, 333 (96.5%) with 95 % CI (94.5 - 98.3) of index cases have tested families. The odds of HIV testing were 7.22 times higher among those who disclosed their HIV status (AOR=7.22 95% CI: 1.45, 35.82) compared to those who did not disclosed. Those who have stayed <12 months on ART are 87% less likely to have tested families (AOR=0.13 95% CI: 0.03, 0.63) compared to those who stayed ≥ 12 months on ART. Conclusion: this study nding shows that higher proportions of families of index cases have been tested for HIV. It also shows that partner and family based index case HIV testing has signicant association with HIV status disclosure of index cases and the length of the duration that the index cases stayed on ART. It is essential to sustain the platform of partner and family based index case HIV testing service through strengthening disclosure counselling and assisting HIV status disclosure in health facilities with fully trained provider and qualied health providers. It also suggests the need to focus on those who received ART for less than 12 months duration and communicate on the timing of HIV testing for HIV negative families with ongoing risk of acquiring HIV.

Almost half of HIV infected patients enrolled in ART had untested family members, many of whom were children (5). Family testing is an index case nding strategy to identify family members with unknown HIV status (6). Index case HIV testing strategy uses a known HIV-infected person receiving HIV care as an index reference to target partners for HIV testing (5,7). In 2014, UNAIDS set the goal to end the AIDS epidemic by 2030 and also stated that worldwide 95% of all people living with HIV should be aware of their HIV status while an estimated 40% still remain undiagnosed (8). HIV partner noti cation enables to identify people with undiagnosed HIV infection (9,10) and linking clients to ART (11).
In 2017, the UNAIDS estimated that 75% of PLHIV were aware of their HIV status, leaving 9.4 million persons with undiagnosed (12). Assisted partner services are rarely available in sub Saharan Africa (13) even though pediatric HIV testing barriers are distinct from adult barriers (14). However, partner noti cation strategies must be feasible in healthcare setting and acceptable to the population (2, 15) as well to children's HIV testing during the routine activities (16). Healthcare settings screen broadly for HIV use social network and partner testing to select high-risk individuals based on their contacts (17).
Efforts should be made to pursue partner noti cation to identify people living with HIV infection (18). In Tanzania, 249 partners reaching the facilities, 96% tested for HIV, 148 (61.9%) tested HIV positive, and 104 (70.3%) of partners testing positive were enrolled into ART (19). In Kenya, for each index patient, 61% of family members identi ed and tested were children (20). In Zimbabwe, the mean monthly HIV positivity rate from index testing (32.6%) was signi cantly higher than that achieved in other HIV testing modalities (21).
The South African PMTCT programs had reduced perinatal HIV transmission at 6 weeks of age to 2.7% and timely maternal diagnosis enables PMTCT uptake to improve infant outcomes (22). Integrating HIV testing into all pediatric health services can assist in identifying HIV infected children (23).
In Ethiopia, the HIV diagnosis is affected by different factors. From study in Harar, clients aged 35-44 years were less likely to accept PITC services than those whose ages were 15-24 years (AOR: 0.17 [0.07-0.41]) (24). In Northern Ethiopia, subjects who initiated by friends, families and other socials to undertake HIV testing (AOR = 0.65; 95% CI = 0.29, 1.48) (25). HIV stigma has an important role in the spread of the AIDS epidemic. Fear of being identi ed as having HIV may discourage a person from getting tested (26). In Awi Zone, Northwest Ethiopia, respondents unwilling to disclose themselves utilized VCT services about 38% less than their counterparts (AOR: 0.62; 95% CI: 0.44, 0.88) (27). Many children living with HIV in resource-limited settings remain undiagnosed and at risk for HIV-related mortality and morbidity (28) and several factors were associated with non-testing of partners and there remains numerous serious challenges to reaching full global epidemic control (95-95-95) (29,30).
Financial incentives have overcome nancial barriers to identify HIV-infected children in other populations by offsetting direct and indirect costs (31). In Zambia, emotional intimate partner violence and HIV status disclosure to the male partner, may play an important role in maternal uptake of early infant HIV testing (32).
An estimated 30% of HIV cases in the EU are also not aware of their HIV status (37). In Georgia during 2013 revealed that 82% of PLHIV knew their diagnosis (38).
This health facility based cross-sectional study with internal comparison was aimed to assess the proportion and the factors associated with partner and family based index case HIV testing services in Woliso town, Oromia, Ethiopia.

Study Setting
The study was conducted in Woliso town which is located in Oromia Regional National State 114KM from Addis Ababa on the road from Addis Ababa to Jimma. It has a latitude and longitude of 8°32′N 37°58′E with an elevation of 2063 meters above sea level (39). The population of the town is 64,681 of which 51% are females and the rest are males. In 2015 the town was selected as one of the twenty towns with high burden of HIV in Ethiopia and one of the ten towns with high burden of HIV in Oromia. At the end of September 2019 the town has 2,376 PLWHIV who are currently on ART. The town has three health facilities with two ART site and one PMTCT site.

Study design
Facility based cross sectional study with internal comparison was conducted to assess the proportion and the factors associated with partner and family based index case HIV testing services in Woliso town, Oromia, Ethiopia.

Source And Study Population:
The source population of the study is all 2,376 people living with HIV (PLWHIV) in Woliso Town who are currently receiving anti-retroviral therapy (ART) in Woliso Town health facilities.
The study population was those clients currently receiving anti-retroviral therapy (ART) who are eligible for Partner and Family Based Index Case HIV Testing (P&FBICT) service during the study period. The study was conducted among eligible clients selected by systematic random sampling.

Inclusion criteria
All people living with HIV (PLWHIV) in Woliso Town who are currently receiving anti-retroviral therapy (ART) and eligible for Partner and Family Based Index Case HIV Testing (P&FBICT) service during the data collection period in Woliso Town health facilities.

Exclusion criteria
Eligible people living with HIV who are seriously ill during data collection period or who are physically un t for interview.
Sample Size: The sample size for people living with HIV who are currently on ART and eligible for Partner and Family Based Index Case HIV Testing (P&FBICT) services were determined using single proportion formula of initial sample size [n= (Z∝/2) 2 p(1-p)/d 2 ] (40) by considering the following assumptions: n= (Z∝/2) 2 p(1p)/d 2 , 95% con dence level (Z∝/2 = 1.96), assuming that 50% of eligible clients have tested families (P = 0.5), 5% marginal error between the sample (d = 0.05). The sample size for the rst objective, since the total number of population is less than 10,000 (N = 1563) and relatively small population Since the total number of population is less than 10,000 (N = 1563) and relatively small population the sample size will be But, from Epi Info Stat Calculator, the nal sample size for the rst objective at two-sided con dence level of 95%, 80% power, ratio of unexposed to exposed equal to 1, 70% outcome in unexposed group, risk ratio (0.78), OR (0.52) and 54.8% outcome in exposed group was 346.
The sample size for the second objective was calculated using Epi Info Stat Calculator for un matched case control using different associated factors included in the conceptual framework (41). Since the sample size for the second objective is less than that of the rst objective, 346 study participants were included in the study.

Sampling procedures
The sample was drawn from people living with HIV in Woliso Town who are currently receiving antiretroviral therapy (ART) and eligible for Partner and Family Based Index Case HIV Testing service during the study period.

Page 6/21
The total sample size was 346 which was generated using systematic random sampling method from line list of eligible PLHIV for Partner and Family Based Index Case HIV Testing (P&FBICT) service and which is small population size that is (N = 1563) because the total population of this study population is n < 10,000. Then, 346 eligible for Partner and Family Based Index Case HIV Testing (P&FBICT) during the study period were included in order to approach study subjects during data collection period in two public health facilities in Woliso Town based on their population proportion.
To get the interval value (K) the study population was divided for study subjects (N/n). By dividing 594 eligible clients expected to come monthly to 226 clients that was taken from St Luke Hospital which is (594/226 = 2.6 ~ 3), from Woliso Health Center is (336/120 = 2.8 ~ 3) then one was selected using lottery method and the number selected was the rst number to start. Thus, every 3rd clients were selected to get all study subjects. Likewise, the proportional allocation for distribution of study population was 226 (65.3%) for St. Luke Hospital and 120 (34.7%) for Woliso Health Center. By taking the study population of N = 1563 (for St. Luke hospital and Woliso Health center) and n = 346 (the calculated sample size).

Data collection procedures
Structured questionnaire was used to capture data relevant to the study's objective and research questions. The questionnaire is formulated to capture the Proportion and the Factors Associated with Partner and Family Based Index Case HIV Testing (P&FBICT) services. The data was collected by directly interviewing eligible individuals and reviewing some records as secondary data. The questionnaire is initially prepared in English and translated into the local languages, Amharic and Afan Oromo, and then it was translated back to English to check the comprehension and consistency.
Information generated through interview was entered into Epi Info™ version 7.2.3.1 and exported to SPSS version 21.0 software for analysis. The data was collected at health facility during the time period from December 2019 to January 2020.
To assure the quality of the data appropriately designed questionnaire was pre-tested on 5 percent of the sample size, in health facility outside the study area. The principal investigator trained the data collectors. During the data collection, principal investigator supervised over the data collection process. Data collected was reviewed and checked for completeness and consistency of the response.
The majority of the study participants were from urban area 209 (60.6%) and the rest 136 (39.4%) were from rural area. The majority of study participants were other unemployed 142 (41.2%) and farmers were 94 (27.2%), while merchants and government employees were 74 (21.4%) and 35 (10.1%), respectively.

Data Management
Data collected from the assessment were checked for consistency to assess the data quality. The data were entered into Epi Info™ version 7.2.3.1 and exported to SPSS version 21.0 software for analysis.

Data Analysis procedures
Data was quantitatively analyzed and it was done by running frequencies, percentages and cross tabulation of categorical data. Binary logistic regression was used to examine association between selected exposure variables and outcome variable HIV testing. All variables with < 0.2, in bivariate analysis were entered into multivariate analysis to identify factors independently associated with HIV testing and for controlling some potential confounders. The

Factors Associated with Partner and Family Based Index Case HIV Testing services
In multivariate analysis, HIV status disclosure and duration of ART were signi cantly associated with partner and family based index case HIV testing. The odds of partner and family based index case HIV testing was 7.22 times higher among study participants who have disclosed their HIV status (AOR = 7.22 95% CI: 1.45, 35.82) compared to those who did not disclosed. Those participants who have stayed < 12 months on ART are 87% less likely to have families for partner and family based index case HIV testing (AOR = 0.13 95% CI: 0.03, 0.63) compared to those who stayed ≥ 12 months on ART (Table 4).   Knowledge on P&FBICT Frequency Percentage Note: The modi ed Bloom's scale of HIV/AIDS knowledge adopted from Nyasinde Mujumali's KAP study was used and the cutoff points for the score was 75-100%, 50-74%, and < 50% placed the respondents in good, moderate, and poor knowledge groups on the bene ts and facilitators of P&FBICT services, respectively.

Discussion
HIV testing is an essential gateway to treatment and care, through provision of ART and counseling on avoiding risky sexual behavior which is a key part of preventing transmission. In its 95-95-95 targets, UNAIDS has called for 95% of PLHIV to know their status, 95% of those who know their status to receive ART, and 95% of ART recipients to achieve viral suppression by 2030 (30).
This study nding shows that higher proportion of families of index cases have been tested for HIV. It also shows that partner and family based index case HIV testing has signi cant association with HIV status disclosure of index cases and also signi cantly associated with the length of the duration that the index cases stayed on ART.
In this study, about more than nine in ten (96.5%) with 95% CI (94.5-98.3) of index cases have brought either partner or children less than 15 years of age for partner and family based index case HIV testing. This result is relatively similar to study in Tanzania which was 96% (19). However, it was relatively lower than study in Abuja, Nigeria which was 98.7% (42). On the other hand, higher than study in Hawassa, Kenya and India, which was 84.1%, 61% and 77%, respectively (20,29,35). Likewise, it was higher than studies conducted in Harar, Barcelona, Peru and Burkina Faso, which was 70.6%, 70.8%, 60% and 88.1%, respectively (16,24,33,37). It is also higher than study in Gondar and Southern Ethiopia, which was 81.7% and 68.8%, respectively (43,44  This study result also showed that index cases those stayed on ART for < 12 months were signi cantly 87% less likely to bring their partner and children less than 15 years of age for HIV testing (AOR = 0.13 95% CI: 0.03, 0.63) compared to those who stayed on ART for ≥ 12 months. In Ethiopian context, it is expected that the longer time index cases stayed on ART, the higher chance to get families tested (2). This nding is consistent with the ndings from a mixed methods evaluation on the HIV basic care package including family HIV testing in Kenya and Uganda, the HIV basic care package receipt was signi cantly associated with using ART [OR = 1.1, 95% CI: 1.0-1.1] (45).
The results of this study should be interpreted within several limitations. First, it is cross-sectional and cannot establish causality or the timing of HIV testing. Second, HIV status disclosure is based on selfreport, which can be vulnerable to recall and social desirability biases. The sample is also nonrepresentative and health facility based, limiting the generalizability of the ndings. This study may also have selection bias because all participants were coming for ART, which may not be representative of all partners and children less than 15 years of age in needing partner and family based index case HIV testing. We did not measure the level of communication among the couple or partner and their matured children, which may affect the acceptance HIV testing. Third, the interview was conducted based on the client's perspectives, which may not re ect the whole picture of HIV testing service. Therefore, institutional related characteristics should be included and evaluated in the future study.

Conclusion
In conclusion, this study nding shows that higher proportions of families of index cases have been tested for HIV. It also shows that partner and family based index case HIV testing has signi cant association with HIV status disclosure of index cases and also signi cantly associated with the length of the duration that the index cases stayed on ART. It is essential to sustain the platform of partner and family based index case HIV testing service through strengthening disclosure counselling and assisting HIV status disclosure in health facilities with fully trained provider and quali ed health providers. It also suggests the need to focus on those who received ART for less than 12 months duration.

Declarations
Ethics approval and consent to participate Ethical clearance was obtained from the Research Ethical Review Committee of the Addis Continental Institute of Public Health and letter was obtained with Ref. No.:ACIPH-MPH/026/12. O cial letter of cooperation was submitted to respective health facilities. The aim of the study, the con dentiality, and the right to withdraw at any time without facing any consequences was explained to each participant at the time of recruitment. Written informed consent was obtained from each participant before enrollment to the study.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during this study are available from the corresponding author on reasonable request.

Funding
Not applicable.