In all, 1405 eligible studies were identified from a total of 855,117 studies from our initial search (Additional file 1). Of these 1405 studies, 1095 duplicates were deleted. A total of 255 were also excluded after abstracts screening. Full article screening resulted in an exclusion of an additional 38 articles. Of the 38 excluded studies following full article screening, 1 study was conducted in a high-income country (33); 5 studies were conduct before 2013 (9, 34–37), 11 Studies were conducted in LMIC’s but did not report on our intervention which was home-based HTC (38–48) and 21 studies reported no evidence of primary outcome of this study (use of home-based HTC) (7, 10, 34, 49–65).
Finally, a total of 17 articles were included for data extraction as indicate in Fig. 1. Level of agreement between reviewers was 74.65% versus 71.55% expected by chance which constitutes lack of agreement (Kappa statistic = 0.11, p-value> 0.05). However, the McNemar’s chi-square statistic suggests that there was no statistically significant difference in the proportions of yes/no answers by the reviewer (Additional file 2). Discrepancies between reviewers’ responses were resolved by involving a third reviewer.
Level of bias for included studies
All included studies underwent methodological quality assessment (Additional file 3) using the Mixed Methods Appraisal tool (MMAT)-Version 2011 . All studies scored between 37.5 and 100%. Eight of the 17 included studies scored the highest quality score of 100% [31,32,33,34,35,36,37,38]. Three of the included studies score 50% [39,40,41]; one study scored the lowest quality score of 37.5%  and five of the remaining studies scored between 62.5 and 88.9% [43,44,45,46,47].
Characteristics of included studies
A total of seventeen studies were eligible for data extraction. Of these, two studies were conducted at a semi-urban setting [31, 48], 10 in rural settings [32, 34,35,36, 39,40,41,42, 46, 47] and five studies did not specify their setting [33, 38, 44, 45, 49]. Of the included studies, three were conducted in south Africa [32, 41, 46], five in Kenya [35, 36, 42, 44, 45], two in Uganda [43, 49], two in Malawi [31, 33], one each in Zambia , Botswana , Ethiopia , Pakistan  and Swaziland . All included studies were published between the year 2013 and 2017. The total number of study participants reported in included studies was 9447,025. This number excludes two systematic reviews that did not specify the total number of their population [41, 49]. The age of participants in all included studies ranged from 18 months to 70 years.
The majority (58.6%) of the population included in the study were females. All included studies reported evidence on the use of home-based HTC in LMICs. The included studies used the following study designs: one was a qualitative study ; eight were randomized control trials [32, 34, 39, 42, 44,45,46,47], five were cross-sectional studies [31, 35, 36, 38, 40]; two were systematic reviews [41, 49], and one cohort studies .
Usage of home-based HTC
All 17 included studies showed evidence of use of home-based HTC. They all reported on supervised HTC intervention, that is, HTC done by qualified healthcare practitioners. The included studies reported usage of home-based HTC on different HIV tests. A total of 47% of the rapid tests were from Unigold and Determine, 5. 8% were from SD Bio line and SENSA Triline, 5.8% from Wilcoxon Mann-Whitney test, 29% did not specify the type and manufacture and 11.7% used rapid test but did not specify the manufacture. Of the 17 included studies, two study reported on couples home-based HTC [31, 38], one study reported on community-based HTC for Prevention of Mother-to-Child Transmission (PMTCT) , one study reported on family-based HTC , three studies reported on home-based HTC on pregnant women and their partners [36, 44, 45], five studies reported on the effects of home-based HTC intervention [32, 34, 39, 42, 47]. Of the 17 included studies, seven reported usage of home-based HTC in LMICs [31, 35, 36, 38, 44, 45, 50]. Four studies reported about feasibility of home-based HTC [40, 43, 46, 50]. Lastly, six studies reported on effectiveness of home-based HTC [32,33,34, 39, 42, 47].
Home-based HTC for couples
A study conducted in Malawi reported on couple-home-based HTC . The aim of the study was to estimate the uptake of couple home-based HTC and couple family planning (CFP) services delivered to couples in their homes. A pair of male and female counselors subsequently visited each couple and offered a HIV test . This study reported that 97 (58%) consented to couple home-based HTC, 4 (2%) to couple family planning only and 18 (11%) declined any intervention . Shahid et al., 2016 conducted a study to explore the utility of home and community-based testing of undiagnosed HIV among spouses and children of HIV-positive people who inject drugs . The results showed that, of the 1959 spouses of HIV-positive people who inject drugs, 1896 (96.7%) consented to home and community-based HTC . The study further found that home-based HTC was an effective way of expanding access and identifying cases of undiagnosed HIV among spouses of people who inject drugs (PWID) . Adopting a family centered approach for HIV testing and counseling would help to reinforce health promoting messaged and protective behaviors. However, even though home and community-based HIV testing was reported to be an effective means of accessing children of PWID, majority of spouses who were diagnosed positive were reluctant to consent to testing for their children. This is an alarm that children of PWID might be at risk and need to be tested and referred for HIV treatment and care.
Three studies reported evidence on the use of home-based HTC on pregnant women and their partners [36, 44, 45]. Krakowiak et al., 2015 conducted a study on home-based HTC among male partners of expectant mothers aimed at assessing its effectiveness in Kenya . The study showed evidence that home-based HTC improved substantially the uptake of HIV testing among male partners of expectant mothers, improved HIV status disclosure between couples, and increased sero-discordant partner identification [36, 44]. Krakowiak et al., 2016 conducted another study aimed at comparing the efficacy of planned home-based visit with expectant mother and their male partners to written invitations extended to male partners to accompany their wives for HIV couple counselling and testing at the next antenatal visit . Home-based HTC was conducted by a team of 2 health advisors trained in HIV counseling and testing, 1 male and 1 female . The study results demonstrated the efficacy, acceptability, and practicability of planned home-based visits strategy compared to HIV couple counselling and for expectant mothers and their male partners . It was also reported that twelve women in the home-based HTC strategy and six women who participated in the written invitations strategy experienced physical intimate partner violence but did not attribute this to their participation in the study . The linkage between home-based HTC on pregnant women and their partners and intimate partner violence need to be investigated.
Home-based HTC for PMTCT
In 2014, Kohler and colleagues conducted a community-based HTC study in Malawi to assess a community’s perspective on the barriers of accessibility to PMTCT and coverage . Their study findings showed that majority of the study participants who were registered as HIV- positive in the home-based HTC register failed to disclose their HIV-positive status . Their study also reported that very few expectant mothers received a completed course although majority of HIV-positive mothers were place on antiretroviral drugs besides late started of antenatal care by many of the pregnant women . Ndege et al., 2016, study also showed that home-based HTC contributes significantly to reduce the incidence of HIV among newborns and additionally, increases PMTCT coverage . There is a need for more studies on home-based HTC for PMTCT in LMICs to help demonstrate the effectiveness of this intervention on PMTCT in these settings.
Home-based HTC for families
Van Rooyen et al., 2016 study conducted with the aim to test all family members for HIV, encourage disclosure and facilitate linkage to care showed that family-based HTC treats the family as a social environment (not just a location for service delivery), through which HIV prevention, treatment, adherence, and support could be achieved . The study further showed that family based HIV testing and counseling increases the identification of HIV-positive children before they become sick and enabling early linkage to care . Despite the increase in the identification of HIV-positive and linkage to care home-based HTC has on the family, nothing has been said on the consequences of this intervention on the family well-being.
Feasibility of home-based HTC
Four studies reported the feasibility of home-based HTC [40, 41, 43, 46]. Parker et al., 2015 conducted a study in Swaziland to assess the feasibility and effectiveness of two community-based HTC models in rural Swaziland . The results of this study found mobile-and home-based HTC to be feasible and affordable ways to reach a substantial number of people . Bogart et al. 2016 showed that home-based testing is a feasible and acceptable model for fisher folk communities that can complement existing event-based testing models by encouraging testing among different types of clients . Studies conducted in South Africa, also reported the feasibility of home-based HTC in LMICs [41, 46]. Despite the reported feasibility of the home-based HTC, linkage to care needs to put more emphasis on initiation of anti- retroviral to reduce attrition between testing and treatment initiation.
Effectiveness of home-based HTC on patient centered outcomes
Five studies showed evidence on effectiveness of home-based HTC on patient-centered outcomes [32, 34, 39, 42, 47]. Jürgensen et al., (2013) conducted a study to investigate the impact of HIV testing on stigma and to investigate whether home-based HTC has a larger impact on stigma than standard testing services in Malawi . The study observed a reduction in overall reported stigma over time in this rural community. This reduction was most prominent in the items measuring individual attitudes regarding equal rights and respectful treatment for people with HIV/AIDS, that is symbolic stigma . Helleringer et al., (2013) study confirms that home-based HTC campaigns may be one of the most effective strategies to make progress towards universal access targets in Sub-Saharan settings .
A study reported that testing rates (supervised VCT) were higher in communities reporting higher stigma, and individuals from high-stigma communities were less likely to have a previous test (Supervised VCT) . It was also demonstrated that although home-based HTC testing increased feelings of anger among HIV-positive individuals, it lowered the sense that having HIV was a sign of immoral behavior . In rural South Africa, Doherty et al., (2013) conducted a study to assess the effect of home-based HTC on the prevalence of HIV testing and reported behavioral changes . This study reveal a moderately high levels of stigma, with over a third of control participants reporting that people with HIV are treated badly owing to their status and almost half observed stigmatizing behavior towards someone with HIV/AIDS . Novitsky et al., 2015 study also reported a lower HIV prevalence rate, 10.9% (95% CI 9.5–12.5%) among individuals tested for the first time using home-based HIV testing in a peri-urban community in Botswana.
Kim et al., 2014 conducted a study in Ethiopia with the aim of studying the causal effects of HIV education, home-based voluntary counseling and testing, and conditional cash transfer for facility-based VCT on HIV/AIDS knowledge and demand for HIV testing . The results showed home-based HTC increases test uptake to a limited extent . However, when HIV/AIDS education is combined with either home-based VCT or CCT for facility-based VCT, testing uptake increased substantially by about 63 and 57 percentage points, respectively . The effectiveness of combining HIV/AIDS education and VCT simultaneously need to be examined.