Study, quality score, study type | Study objective | Study area, type of clinic/program | Year of data collection | Study design (comparison between population a and b), population sizes, sampling method/inclusion criteria | Statistical analysis | Outcome on association as reported per equity criteria |
---|---|---|---|---|---|---|
Cleary 2011 *** Observational [25] | To evaluate whether the distribution of ART services in the public system reflects the distribution of people in need among adults in the urban population | Urban area: poor communities in Mitchells Plain (Cape Town, Western Cape province) and Soweto township (Johannesburg, Gauteng province), public clinics | National survey: 2008. Urban clinic data: unknown | a. Population in need for ART (n = 742): national survey (2008, HIV + residents), sampling unknown, | Comparison distribution of equity criteria (i.e. patients characteristics) | Sex (not associated): percentage of HIV + women in national survey is same as in ART users in urban clinic; 67.4% [95% CI: 61.5-72.9] versus 65.7% [95% CI: 60.6-70.7], p >0.05. Socioeconomic status (not associated): no significant differences in SES distribution between HIV + in need for ART and ART patients in urban clinics; independence partition Pearson’s chi-square test: 8 [p = 0.43] Race/ethnicity (inconclusive results): percentage of non-African is 2,5% in population HIV + in need versus 4.3% of ART users in urban clinics, authors state that sample size of non-African is too small to draw conclusions on equity |
b. ART patients in urban public clinics (n = 635): data from ART users (>18 yrs, >14 days on treatment) in three clinics in Mitchells Plain (selected proportional to the number of ART patients in facility) and three in Soweto (stratified random sampling) | ||||||
Cooke 2010 *** Observational [32] | To investigate factors associated with uptake of ART through a primary health care system in rural South Africa | Rural, peri-urban and urban areas: Hlabisa sub-district, Umkhanyakude district, Northern KwaZulu-Natal province, public clinics supported by NGOs | Aug 2004 – Dec 2008 | a. HIV + residents not on ART (n = 1,003): population-based surveillance in 6 catchment areas, | Multivariate logistic regression | Sex (not associated): no significant association between gender and receiving treatment: aOR men 0.875 [95% CI: 0.708-1.081, p = 0.216] Age (associated, younger (15–19 yrs) < older (>19 yrs)): compared to age 15–19 (reference) all higher 5-year-age-groups [20–24, 25–29, 30–34, 35–40, 40–45, 45–50, 50–54, 55–60, >60] have significant higher aOR [ranging between 4.9-14.0, p < 0.05] for receiving treatment Area of living (not associated): no significant differences in aORs between peri-urban [1.042, 95% CI: 0.699-1.554, p = 0.838], rural [0.941, 95% CI: 0.628-1.410, p = 0.768] and urban (reference) areas for receiving treatment Socioeconomic status (not associated): no significant differences in aORs between index profiles 1 (reference), 2 [0.932, 95% CI: 0.688-1.262, p = 0.649], 3 [0.842, 95% CI: 0.624-1.135, p = 0.258], 4 [0.829, 95% CI: 0.607-1.131, p = 0.237] and 5 [0.984, 95% CI: 0.702-1.379, p = 0.927] for receiving treatment Education (not associated): no significant association between years of education and receiving treatment; aOR years of education: 1.022 [95% CI: 0.995-1.063, p = 0.128] |
b. HIV + residents on ART (n = 1,251): population based 2008 cohort (HIV+, > 15 yrs, on ART) | ||||||
Govindasamy 2011 *** Observational [26] | To assess the proportion and characteristics of individuals who accessed HIV care after testing HIV + in a mobile testing unit | Rural area: Cape Metropolitan region, Western Cape province, type of clinic not clearly reported | Tested HIV+: 2008–2009. Interviewed: Apr-Jun 2010. | Patients tested HIV + in mobile testing units that: a. linked to ART care (i.e. receiving CD4 test result), b. not linked, A random sample of patients tested HIV + between August 2008 – December 2009, ≥18 yrs, CD4 < 350, received CD4 test results, available socio-demographic variables was selected using mobile testing unit records (n = 77) | Binomial univariate and bivariate regression analysis | Sex (not associated): same likelihood to link to care for female as male patients; **RR female: 1.18 [95% CI: 0.81-1.72, p = not reported, 1.0 falls within CI] Age (not associated): same likelihood to link to care for younger (≤30 years) as older patients (≥30 years) to link to care; **RR ≥30 years: 1.21 [95% CI: 0.83-1.77, p = not reported, 1.0 falls within CI] Severity of disease (contradicting results, CD4 cell count associated and WHO stages not associated): significantly lower likelihood to link to care for patients with high (>350) compared to low (≤350) CD4 cell count;*RR CD4 > 350: 0.49 [95% CI: 0.27-0.87, p = 0.014] / same likelihood to link to care for patients in WHO stage I as WHO stage II, III or IV; **RR WHO clinical stage I: 0.88 [95% CI: 0.65-1.18, p = not reported, 1.0 in CI] Education (not associated): same likelihood to link to care for patient completed primary school as patients that have not; **RR completed primary school: 1.17 [95% CI: 0.66-2.08, p = not reported, 1.0 falls within CI] Employment (not associated, employed < unemployed): likely lower likelihood to link to care for employed compared to unemployed patients; **RR employed: 0.72 [95% CI: 0.51-1.01, p = 0.056]. * = univariate ** = bivariate analysis |
Tsai 2009 ** Observational [33] | To assess differences in socioeconomic profiles between those who access HIV-related clinical services and the HIV-infected individuals living in the wider community | Rural area: Limpopo province, public hospital | Community survey: 2004-2005. Clinic survey: Jan 2003 – Nov 2005 | a. community sample, HIV + not on ART (n = 242): household survey, random sampled from eight rural villages in the province (14–35 yrs, HIV+), | Uni-variate comparison and multiple regression | Sex (not associated): no significant difference percentage women in the community vs. clinic sample: 79% vs. 79% [p = 0.78] Age (associated, younger (18–30 yrs) < older (30–35 yrs)): significant difference in age distribution between community and clinic sample: 18–20 yrs: 13% vs 3.6%; 21–25 yrs: 33% vs. 16%; 26–30 yrs: 36% vs 33%; 31–35 yrs: 18% vs. 47%; X2 = 85 [p < 0.001*] Education (associated, higher education > lower education): significant difference in distribution educational attainment between community and clinic sample: in clinic less likely to completed secondary education [p < 0.001], but more likely to completed matric or tertiary education [p = 0.04] X2 42 [p < 0.001*] Employment (associated, not having salaried employment < having salaried employment, unemployed < employed): significant difference percentage having salaried employment between community and clinic sample: 6.2% vs. 11%, X2 3.8 [p = 0.05] and in percentage unemployed and able to work: 57% vs. 37%; X2 26 [p < 0.001*] Marital status (associated, never married < married or cohabiting): significant difference distribution marital status between community and clinic sample: never married: 78% vs. 43%; married/ cohabiting: 16% vs. 30%; X2 83 [p < 0.001*] *also significant after multivariable regression |
b. clinical sample, HIV + on ART (n = 534): convenience sample of patients (18–35 yrs) in primary HIV/AIDS provider hospital, referred by 45 primary health care clinics. Note: samples were not taken from identical sub-districts | ||||||
Adam 2009 ** Observational [34] | To quantify the coverage in South Africa up to the middle of 2008, according to various definitions of antiretroviral treatment eligibility | Rural and urban: National/ nine provinces, public clinics | 2008 | For nine provinces: a. number of HIV + in need for ART: Markov model on HIV progression using different CD4 count compartments | Comparison ART coverage data | Area of living (associated, unequal coverage among nine provinces): unequal ART coverage in 2008 among 9 provinces: Eastern Cape 32.4%, Free State 25.8%, Gauteng 43.5%, KwaZulu-Natal 39.4%, Limpopo 32.2%, Mpumalanga 31.2%, Northern Cape 61.1%, North West 35.4%, Western Cape 71.1% |
b. number of HIV + on ART: estimates of patients starting ART in public health facilities using Department of Health unpublished internal report (7 May 2009) | ||||||
Muula 2007 * Systematic review [24] | To describe the gender distribution of patients accessing ART in Southern Africa | Rural and urban: National (1999–2004), Khayelisha township in Capetown (2001–2), Eastern cape province 2001–4), Northern cape province (2001–5), public clinics | 2000 – 2006 | a. National HIV + prevalence female/male ratio in 2005, | Comparison female/male ratios | Sex (associated, male < female): female have higher access than men to ART: HIV prevalence female/male ratio = 1.2, while 4 studies report access to ART female/male ratio of 1.9, 2.3, 1.8 and 1.5 |
b. access to ART female/male ratio. Sampling methods not reported | ||||||
Nattrass 2006 * Critical assessment [23] | To compare ART roll-out in public sector between provinces in 2003-2005 | Rural and urban: National (nine provinces), public clinics | 2003 - 2005 | For nine provinces: | Comparison ART coverage data | Area of living (associated, unequal coverage among 9 provinces): unequal ART coverage at the end of 2005 among 9 provinces: Eastern Cape 21.8%, Free State 21.0%, Gauteng 29.6%, KwaZulu-Natal 20.0%, Limpopo 27.3%, Mpumalanga 20.9%, Northern Cape 32.3%, North West 24.5%, Western Cape 55.7% |
a. number of HIV + in need for ART, | ||||||
b. number of HIV + on ART, estimates of ART coverage based on ASSA2003 demographic model (includes public, NGOs and private sector providers) |