Efforts to expand HIV infection treatment programmes in sub-Saharan Africa (SSA) have demonstrated improvement in access to HAART. The estimated number of HIV-infected individuals in low-income and middle-income countries (LMICs) on antiretroviral therapy (ART) increased five-fold from 1.3 million in 2006 to over 6.6 million in 2013 [1,2], with robust and sustained immune responses for those patients who remain on treatment, especially when starting ART with higher baseline CD4+ cell counts . Overall AIDS-related morbidity and death have declined markedly in countries where access to ART has significantly improved [3–5]. However, the persistent problem of ART initiation in the advanced stages of HIV infection threatens to limit the full potential of HIV scale-up programmes in SSA , and is associated with high mortality rates after ART initiation and further HIV transmission [6–8]. The excess mortality is due in large part to ART initiation at low CD4+ cell counts . In resource-limited settings with weak healthcare infrastructure, many operational challenges beside social, economic and clinical determinants are associated with lower CD4+ cell count at enrollment and at ART initiation, all of which mitigate the benefit of HIV treatment programmes [9,10].
With an HIV prevalence of 3.0% , Rwanda has been scaling up access to HIV care since 2004, and the mature national programme has achieved more than 93% ART coverage  with high rates of retention  and medication adherence with viral suppression . We therefore examined the proportion of persons enrolling in HIV care and initiating ART in the advanced stages of HIV infection among patients in Kigali, Rwanda, before and after the first expansion of national ART guidelines from a CD4+ threshold of 200 cells/μl to 350 cells/μl. In order to assess barriers and enablers to timely ART initiation in the context of high ART coverage and high median CD4+ at ART initiation, we examined correlates of advanced HIV infection at ART initiation among persons initiating ART in 2012.
Study population and data source
The study population included HIV-infected adult patients (≥15 years) enrolling in care, and initiating ART at five urban HIV care clinics [Women's Equity in Access to Care and Treatment, Carrefour, Bethsaida, Kicukiro and Makasa] in Kigali from January 2003 to June 2012 participating in the International epidemiologic Databases to Evaluate AIDS (IeDEA) consortium (http://http://www.iedea.org). We used electronic patient-level data collected as part of routine service delivery from clinics that are part of the national HIV care and treatment programme in Rwanda. Patient information at enrollment and/or ART initiation was collected routinely by nurses and trained data clerks using standard national patient medical records and then entered into an electronic database. Data were de-identified prior to being shared with researchers for analysis.
We defined advanced HIV disease at enrollment into care or ART initiation as either having a CD4+ cell count of less than 200 cells/μl or a WHO stage IV condition at the time of enrollment or ART initiation, respectively. The CD4+ cell count at ART initiation included CD4+ results measured up to 90 days before or 30 days after the date of ART initiation, with priority given to CD4+ cell counts prior to ART initiation and otherwise closest to the ART initiation date. A small proportion of patients (<5%) who did not have WHO staging information and CD4+ cell counts at enrollment into care or ART initiation were included in analyses and categorized as missing/unknown.
HIV-infected patients who enrolled into care, and were not eligible for treatment were offered various services that included monitoring of their CD4+ cell counts on a 6-month basis and prophylaxis for opportunistic infections and prevention of mother-to-child transmission (PMTCT) according to guidelines in place at the time. Patients received HIV voluntary counseling and testing (VCT) or provider-initiated voluntary counseling and testing (PVCT) and were enrolled into care. Enrollment into care often occurred at the same testing site as all five study clinics routinely offer HIV prevention, diagnostic and treatment services including CD4+ cell count determination. Patients who enrolled into care from 2003 to 2006 were considered ART eligible under the 2004 national ART guidelines in Rwanda  (WHO stage IV diagnosis irrespective of CD4+ cell count, or a WHO clinical stage of III with a CD4+ cell count of <350 cells/μl, or WHO stage I or II with a CD4+ cell count of <200 cells/μl). Patients enrolled into care from 2007 to 2011 were considered ART eligible under the 2007 national guidelines (WHO stage IV or a CD4+ cell count of less than 350 cells/μl irrespective of WHO stage). Patients enrolling into care in 2012 were considered ART eligible under the 2012 national ART guidelines (CD4+ cell count of <500 cells/μl irrespective of WHO stage). Women who were pregnant and not already on ART were referred to PMTCT services. Also, those with active tuberculosis disease, WHO stage III and IV and, beginning in 2012, those in a serodiscordant relationships were initiated on ART regardless of their CD4+ cell count.
We examined trends in the median CD4+ cell count and percentage of patients enrolling in care and initiating ART with CD4+ cell count more than 200, more than 350 and more than 500 cells/μl during 2003–2012, before and after the first expansion of ART eligibility criteria in the national guidelines. Relative risk (RR) ratios for the proportion of patients with advanced disease at enrollment into care and ART initiation were calculated, comparing the risk before 2007 with that in 2011–2012 within patient characteristics. For patients initiating ART in 2011–2012, we examined factors associated with advanced HIV disease (CD4+ <200 cells/μl or WHO stage IV). Adjusted odds ratios (AOR) and 95% confidence intervals (CIs) were generated using generalized estimating equations to account for clustering of patients within sites. Statistical analyses were carried out using SAS version 9.2 (SAS Institute, Cary, North Carolina, USA).
Ethical approvals for this study were obtained from the Rwandan National Ethics Committee and the Institutional Review Board of Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, New York, USA.
Characteristics of patients at enrollment into care and at antiretroviral therapy initiation
Sociodemographic characteristics of the 6326 patients at least 15 years enrolled into care from 2003 to 2012 are presented in Table 1. Median age was 35 years [interquartile range (IQR): 29–42] and 62% were women, of whom 3.4% were pregnant. Twenty percent of patients were married and 64% were enrolled through a voluntary testing and counseling center. About 71% (n = 4486) of patients initiated ART during the study follow-up to 2012, with 2.0% of women (n = 87) pregnant at ART initiation.
The proportion of patients with advanced HIV disease enrolling into care and initiating ART was 23.8 and 45.7%, respectively (Table 2). Median CD4+ cell count for all patients at enrollment was 349 cells/μl (IQR: 194–543) and for those who initiated ART was 211 cells/μl (IQR: 131–300) at the point of ART initiation. For those with data available at enrollment (n = 5280), the proportion of patients with WHO stage III or IV at enrollment into care was 21% compared with 27% at ART initiation (Table 2). Approximately 48% of all patients were immediately eligible for ART at enrollment based on the Rwanda national guidelines in effect at the time of their enrollment.
Trends in CD4+ cell count among patients at enrollment into care and at antiretroviral therapy initiation
From 2003 to 2012, there was a near doubling in median CD4+ cell count at enrollment in care for all patients from 185 (IQR: 109–378) in 2003 to 340 cells/μl (IQR: 184–480) in 2012 (Fig. 1a). Women had higher median CD4+ cell counts at enrollment than men for the entire time period of 2003–2012. Overall, the median CD4+ cell count at enrollment into care for men and women were 317 cells/μl (IQR: 168–508) and 372 cells/μl (IQR: 217–571) respectively, P < 0.001. Men on average initiated ART at a lower CD4+ cell count than women; 198 cells/μl, (IQR 120–290) vs. 223 cells/μl (IQR: 140–305) for men and women, respectively, P < 0.001 (data not shown). The median CD4+ cell count at ART initiation increased steadily during 2003–2012. During the first Rwandan national guidelines expansion in 2007–2008, which increased the CD4+ cell count for treatment initiation to less than 350 cells/μl, the median CD4+ cell count for patients in our study increased from 183 to 246 cells/μl within 1 year of national guideline expansion, and ultimately reached 293 cells/μl in 2011 (Fig. 1b).
The proportion of patients who initiated ART at CD4+ cell count of more than 200, more than 350 and more than 500 cells/μl over time is shown in Fig. 1c. Whereas there was a large increase in the proportion of patients initiating ART with CD4+ cell count of greater than 200 cells/μl from 38.2 to 67.3% after the 2007 national guideline expansion, the proportion of patients in 2012 with CD4+ cell count of greater than 350 cells/μl at ART initiation was 19.7%. Patients initiating ART at a CD4+ cell count of greater than 500 cells/μl rose from 5.3% in 2011 to 9.8% in 2012 during the year where national guidelines expanded to include treatment for all patients with a CD4+ cell count of less than 500 cells/μl. The median CD4+ cell counts of patients enrolling in HIV care who ultimately initiated ART appeared to increase more in women than men, with an overall median CD4+ cell count of 377 cells/μl by 2012, with women at 430 cells/μl vs. men 322 cells/μl P < 0.001 (data not shown).
Advanced HIV disease at antiretroviral therapy initiation before and after Rwanda national guideline expansion
Data on the proportion of patients of enrolling in care and initiating ART with advanced HIV disease before 2007 vs. 2011–2012 and RR for different sociodemographic and clinical categories are summarized in Table 3. Overall, there was a decrease in the proportion of patients enrolling into care with advanced HIV disease from 28.2% for those who enrolled before 2007 to 22.6% in 2011–2012. An even more dramatic decrease was found in the proportion of patients initiating ART with advanced HIV disease before 2007 and in 2011–2012 at 66.2 to 29.4%, respectively (Table 3). The decrease in the risk during 2007 to 2011–2012 was also observed in most subgroups. A decreased risk was found in both men and women [RRmen = 0.56 (95% confidence interval or CI: 0.48–0.65), RRwomen = 0.35 (95% CI: 0.30–0.42)], those who were younger [RR15–25 = 0.30 (95% CI: 0.20–0.45)] and older [RR46–55 = 0.64 (95% CI: 0.50–0.84)], those married [RRmarried = 0.52 (95% CI: 0.41–0.67)] or single [RRsingle = 0.51 (95% CI: 0.39–0.68)], entering enrollment through the voluntary counseling and testing center: VCT [RRVCT = 0.44 (95% CI: 0.37–0.50)], or those with clinical WHO stage I at enrollment [RRWHOstageI = 0.39 (95% CI: 0.33–0.47)] among other factors.
Factors associated with initiating antiretroviral therapy with advanced HIV disease in 2011–2012
The multivariable model of factors associated with higher odds of initiating ART with advanced HIV disease is presented in Table 4. Male sex (AORmenvs.women = 1.66, 95% CI: 1.34–2.05) and older age: (AOR36–45vs.<25 = 1.68, 95% CI: 1.11–2.56, and AOR46–55vs.<25 = 2.26, 95% CI: 1.20–4.30) were associated with higher odds of having advanced HIV disease at ART initiation after adjusting for covariates. Among those initiating ART more than 1 year after enrollment in care, those who had a gap in care of 12 or more months prior to ART initiation had higher odds of advanced HIV disease (AOR = 5.23, 95% CI: 1.23–21.13). Living with a partner and unknown eligibility for ART at enrollment were associated with lower odds of having advanced HIV disease at ART initiation AOR livingwithapartnervs.married = 0.75, 95% CI: 0.58–0.98 and AOR unknowneligibilityvs.eligibleforART = 0.16, 95% CI: 0.04–0.63, respectively.
Our analyses showed substantial improvements in the median CD4+ cell counts at ART initiation in our sample of five clinics from Rwanda, where the ART coverage has risen to over 93%  and national ART eligibility guidelines have been expanded, often in advance of other countries in the region. The five HIV clinics included in our study comprised of two private clinics and three government-owned clinics where IeDEA programmes initially started. They are among the HIV clinics in the country with a large volume of patients, and are located in both urban and periurban settings. The median CD4+ cell count at ART initiation in our study of HIV-infected Rwandan patients is higher than reported in other studies in the region [10,15,16]. A global analysis of data from 48 countries, including those in North America, Europe, and SSA, found that in 2009 the United States had the highest mean CD4+ cell count at ART initiation (307 cells/μl) followed, remarkably, by Rwanda (287 cells/μl) [17,18]. Given this and other HIV programmatic scale-up successes in Rwanda [13,14,19], there are likely important lessons to be learned by comparing Rwanda's experience with those of other countries in SSA.
Surprisingly, although this was an analysis of specific HIV sites in an urban setting, our reported median CD4+ cell count at ART initiation reached up to 293 cells/μl in 2011, and is considerably higher than the recent estimates of median CD4+ of 155 cells/μl in 2009 from clinical sites in various LMICs . Therefore, our observation of higher median CD4+ cell count at ART initiation is different from the consistently reported low median CD4+ cell count at ART initiation in other SSA settings [3,4,7–10]. One recent study of trends in CD4+ cell count at ART initiation in four SSA studies, including Rwanda, found an increase in median CD4+ of 125–185 cells/μl during 2006–2011, or only about 10 cells/per year . We believe the larger increases in median CD4+ cell count at ART initiation in Rwanda demonstrates rapid achievement of very high ART programme coverage, achievable in part because of lower HIV prevalence than that of other East and Southern African countries. Systematically identifying specific reasons for this progress that may be replicable in other settings is an important research priority for the region.
We observed an overall improvement in the timeliness of ART initiation, with risk of initiating ART with CD4+ cell count less than 200 cells/μl falling from 58.7% in 2007 to less than 32.0% in 2008–2012. More timely ART initiation has several benefits associated with improved individual patient survival, lower burden on the healthcare system, and decreased likelihood of onward HIV transmission [15,21]. The observed increase in the proportion of people initiating ART with CD4+ cell count more than 200 cells/μl was restricted to those with 200< CD4+< 350 cells/μl is further evidence that the guideline change drove the increase. The reason for the increase followed by the decline in the median CD4+ cell count at enrollment into HIV care beginning in 2010 is less clear, but likely because of factors outside the HIV care system such as improvements in testing coverage and linkage. For example, increases testing coverage could have the effect of identifying a large bolus of prevalent positives in the short term. Over the long term, those undiagnosed people living with HIV (PLWH) who are more marginalized and thus less likely to be tested before development of symptoms could be driving declines in the median CD4+ cell count at enrollment (possibly representing a return to baseline/steady state).
In this study, several factors were independently associated with greater likelihood of late ART initiation in 2011/12, including male sex and older age. Other studies from SSA have also reported that men start ART at lower median CD4+ cell count, and the rate of change in CD4+ cell count over time is significantly lower in men than women [22–25]. Our study found that the disparity persists even in Rwanda where the ART coverage is high, and that the sex difference is largely explained by men being more likely to be ART eligible (i.e. sicker) at enrollment than women, likely because of later diagnosis and/or delayed linkage to HIV care . This in turn suggests that guideline changes and a strong HIV care system alone will not eliminate sex disparities in late treatment initiation, and more research is needed on how to address this important sex disparity in the timeliness of ART initiation, which appears to be worsening with time .
In addition to earlier ART initiation, we observed an increase in median CD4+ cell count at enrollment in care from 185 to 340 cells/μl. This finding is likely a result of improved timeliness in diagnosing and linking HIV-positive patients to care, and not a direct result of expanding treatment guidelines, which tend to affect the timeliness with which treatment is initiated among those already enrolled in HIV care. The finding of increasing median CD4+ cell count at enrollment into HIV care may be explained in part by several synergistic factors associated with improved primary healthcare access in Rwanda. This includes a strong healthcare system and infrastructure which is being rapidly developed, with community-based health insurance [19,26,27,28], improved workforce skills promotion of performance-based financing and strong leadership commitment [19,29,30].
Drivers of timely ART initiation include timely HIV diagnosis following infection, followed by timely linkage to care and ART initiation . There are several measurable proximal factors that could explain why Rwanda appears to have superior outcomes with regard to the timeliness of enrollment into HIV care and ART initiation. In countrywide Demographic Health Survey surveys, 92.0% of adult HIV-infected Rwandans report having previously been tested for HIV, compared with only 74.9% of Ethiopians and 74.3% of Tanzanians . Perhaps as a result of greater testing coverage, HIV-infected Rwandans appear to exhibit less stigma against PLWH infection (94.9% would buy vegetables from a HIV-infected vendor, compared with 71.5% of Ethiopians and 77.6% of Tanzanians), have better prevention knowledge (95.9% endorse condoms as a risk reduction method vs. 78.2% in Ethiopia and 82.1% in Tanzania) and greater understanding of PMTCT (98.5% know transmission can occur during delivery vs. 83.9% in Ethiopia and 79.2% in Tanzania) . Overall, 62% of HIV-infected Rwandans have comprehensive HIV knowledge, compared with only 31.2% of Ethiopians and 53.1% of Tanzanians, with a particularly stark contrast among women (65.1 vs. 25.1 and 48.1%, respectively) .
It is likely that other more upstream factors are responsible for these HIV-specific outcomes. In addition to the increasing proportion of Rwandans with national community-based health insurance, Rwanda's healthcare system and infrastructure has been improving over time . Despite being one of the poorest countries in the world, with approximately 63% of its population living on less than $1.25 USD per day , the country has harmonized donor support for its HIV programme, retaining ownership of all developmental programmes to ensure that support funds are sustainable in the medium to long term [19,26]. The availability of trained community health workers or animateurs de santé to deliver preventive, diagnostic and therapeutic services at each village level or Umudugudu, further help link patients to healthcare facilities in a timely manner [12,26].
Our study has some limitations. First, our sample was limited to patients enrolled in HIV care programmes in urban settings. Urban patients may have greater opportunity and capacity than rural patients to access HIV services, with fewer barriers to access, for instance, shorter distance to clinic, higher education and better socioeconomic demographic characteristics. Our analysis also included only sites that have electronic data systems, which may not be representative of HIV programmes without such systems. Additionally, the five urban sites included in our study are not representative of all HIV care sites in Kigali or Rwanda. Finally, the observational nature of our study limits our ability to assess and quantify any underlying causal effect of the risk factors and correlates examined.
The strengths of our study include use of routinely collected patient data from several urban private and government-owned HIV sites. This allows for analysis of routine clinical data rather than data collected for research purposes only. Secondly, our analyses include broad data for the sites over 10 years from 2003 to 2012, a period when the Rwandan national guidelines eligibility criteria for ART initiation changed in 2007–2008 from CD4+ cell counts of less than 200 cells/μl +/− WHO stage IV or WHO stage III with 200–350 cells/μl to CD4+ cell counts less than 350 cells/μl irrespective of WHO stage.
In conclusion, we observed a marked increase in the median CD4+ cell count at ART initiation, and a dramatic increase in the proportion of patients initiating ART with CD4+ more than 200 cells/μl following the expansion of ART eligibility criteria in 2007–2008. Although the increase in the median CD4+ cell count at ART initiation was observed in both women and men, men were significantly more likely than women to initiate ART at lower CD4+ cell counts, and to initiate ART with CD4+ cell counts less than 200 cells/μl or WHO IV. Despite high levels of ART coverage in Rwanda, a substantial proportion of persons continue to initiate ART well below the CD4+ thresholds specified in national guidelines, even with near universal coverage. Additional research is therefore needed on the determinants of late diagnosis, enrollment into care, and treatment initiation, even in the context of very high ART coverage. It is also important to better understand the mechanisms by which men initiate ART later than women, and to diagnose and link them to HIV care earlier through VCT and PMTCT with partner testing. As Rwanda is among the few countries in SSA with rapid HIV scale up programmes that has achieved high treatment coverage, further studies are needed to examine whether programme quality can be maintained, and to understand how best to replicate the successes of Rwanda in other settings.
The authors acknowledge the efforts of the entire IeDEA community and clinic staff at various HIV sites tasked with collecting the data used for this analysis.
E.M., K.A., D.R.H. and D.N. designed the study. D.A., D.R.H., and D.N., analyzed the data. E.M., D.A. and D.N. drafted the first manuscript. K.A., D.R.H, J.C.D., B.K., I.I., L.M. and S.N. contributed to the first manuscript preparation. E.M., D.A., K.A., D.R.H., J.C.D., B.K., I.I., L.M., S.N. and D.N. were responsible and contributed and approved the final manuscript writing.
Presentation: This work was presented by E.M. at the 20th Conference on Retroviruses and Opportunistic Infections (CROI, March 2013) (poster #1090), 3–6 March, Atlanta, Georgia USA.
Conflicts of interest
There are no conflicts of interest.
1. Ojikutu B, Makadzange AT, Gaolathe T. Scaling up ART treatment capacity: lessons learned from South Africa, Zimbabwe, and Botswana
. Curr HIV/AIDS Rep
2. WHO. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector: progress report 2010
. Geneva, Switzerland: WHO; October 2010.
3. Nash D, Katyal M, Brinkhof MW, Keiser O, May M, Hughes R, et al. Long-term immunologic response to antiretroviral therapy in low-income countries: a collaborative analysis of prospective studies
4. Jahn A, Floyd S, Crampin AC, Mwaungulu F, Mvula H, Munthali F, et al. Population-level effect of HIV on adult mortality and early evidence of reversal after introduction of antiretroviral therapy in Malawi
5. Reniers G, Araya T, Davey G, Nagelkerke N, Berhane Y, Coutinho R, et al. Steep declines in population level AIDS mortality following the introduction of antiretroviral therapy in Addis Ababa
6. Dlodlo RA, Fujiwara PI, Hwalima ZE, Mungofa S, Harries AD. Adult mortality in the cities of Bulawayo and Harare, Zimbabwe: 1979–2008
. J Int AIDS Soc
2011; 14 (Suppl 1):S2.
7. Lahuerta M, Ue F, Hoffman S, Elul B, Kulkarni SG, Wu Y, et al. The problem of late ART initiation in sub-Saharan Africa: a transient aspect of scale-up or a long-term phenomenon?
. J Healthcare Poor Underserved
8. Brinkhof MW, Boulle A, Weigel R, Messou E, Mathers C, Orrell C, et al. Mortality of HIV-infected patients starting antiretroviral therapy in sub-Saharan Africa: comparison with HIV-unrelated mortality
. PloS Med
9. Nash D, Wu Y, Elul B, Hoos D, El Sadr W. Program-level and contextual-level determinants of low-median CD4+ cell count in cohorts of persons initiating ART in eight sub-Saharan African countries
10. Lahuerta M, Lima J, Nuwagaba-Biribonwaha H, Okamura M, Alvim MF, Fernandes R, et al. Factors associated with late antiretroviral therapy initiation among adults in Mozambique
. PLoS One
Demographic and Health Survey 2014 Final Report. National Institute of Statistics of Rwanda
. Ministry of Finance and Economic Planning, Kigali, Rwanda
; February 2012, pp. 15–18.
12. Nsanzimana S, Ruton H, Lowrance DW, Cishahayo S, Nyemazi JP, Muhayimpundu R, et al. Cell phone-based and internet-based monitoring and evaluation of the national antiretroviral treatment program during rapid scale-up in Rwanda: TRACnet, 2004–2010
. J Acquir Immune Defic Syndr
13. Lowrance DW, Ndamage F, Kayirangwa E, Ndagije F, Lo W, Hoover DR, et al. Adult clinical and immunologic outcomes of the national antiretroviral treatment program in Rwanda during 2004–2005
. J Acquir Immune Defic Syndr
14. Elul B, Basinga P, Nuwagaba-Biribonwaha H, Saito S, Horowitz D, Nash D, et al. High levels of adherence and viral suppression in a nationally representative sample of HIV-infected adults on antiretroviral therapy for 6, 12 and 18 months in Rwanda
. PLoS One
15. Geng EH, Hunt PW, Diero LO, Kimaiyo S, Somi GR, Okong P, et al. Trends in the clinical characteristics of HIV-infected patients initiating antiretroviral therapy in Kenya, Uganda and Tanzania between 2002 and 2009
. J Int AIDS Soc
16. Cornell M, Grimsrud A, Fairall L, Fox MP, van Cutsem G, Giddy J, et al. Temporal changes in programme outcomes among adult patients initiating antiretroviral therapy across South Africa, 2002–2007
17. Mugglin C, Althoff K, Wools-Kaloustian K, Sterne J, Nash D, Dabis F, et al. Immunodeficiency at the start of ART: global view. In: 19th Conference on Retroviruses and Opportunistic Infections (CROI); 3–8 March 2012; Seattle.
18. Avila D, Althoff KN, Mugglin C, Wools-Kaloustian K, Koller M, Dabis F, et al. Immunodeficiency at the start of combination antiretroviral therapy in low-, middle-, and high-income countries
. J Acquir Immune Defic Syndr
19. Binagwaho A, Farmer PE, Nsanzimana S, Karema C, Gasana M, de Dieu Ngirabega J, et al. Rwanda 20 years on: investing in life
20. Lahuerta M, Wu Y, Hoffman S, Elul B, Kulkarni SG, Remien RH, et al. Advanced HIV disease at entry into HIV care and initiation of antiretroviral therapy during 2006–2011: findings from four sub-Saharan African countries
. Clin Infect Dis
21. Johansson KA, Robberstad B, Norheim OF. Further benefits by early start of HIV treatment in low income countries: survival estimates of early versus deferred antiretroviral therapy
. AIDS Res Ther
22. Taylor-Smith K, Tweya H, Harries A, Schoutene E, Jahn A. Gender differences in retention and survival on antiretroviral therapy of HIV-1 infected adults in Malawi
. Malawi Med J
23. Hawkins C, Chalamilla G, Okuma J, Spiegelman D, Hertzmark E, Aris E, et al. Sex differences in antiretroviral treatment outcomes among HIV-infected adults in an urban Tanzanian setting
24. Alibhai A, Kipp W, Saunders LD, Senthilselvan A, Kaler A, Houston S, et al. Gender-related mortality for HIV-infected patients on highly active antiretroviral therapy (HAART) in rural Uganda
. Int J Womens Health
25. Maman D, Pujades-Rodriguez M, Subtil F, Pinoges L, McGuire M, Ecochard R, et al. Gender differences in immune reconstitution: a multicentric cohort analysis in sub-Saharan Africa
. PLoS One
26. Logie DE, Rowson M, Ndagije F. Innovations in Rwanda's health system: looking to the future
27. Twahira A. Sharing the burden of sickness: mutual health insurance in Rwanda
. Bull World Health Organ
28. Lu C, Chin B, Lewandowski JL, Basinga P, Hirschhorn LR, Hill K, et al. Towards universal health coverage: an evaluation of Rwanda mutuelles in its first eight years
. PLoS One
29. Bucagu M, Kagubare JM, Basinga P, Ngabo F, Timmons BK, Lee AC. Impact of health systems strengthening on coverage of maternal health services in Rwanda, 2000–2010: a systematic review
. Reprod Health Matters
30. Rodriquez PR, Samuels F. Rwanda's progress in health: leadership, performance and insurance
. London: ODI publications, 111 Westminster Bridge Road, London SE1 7JD, UK Overseas Development Institute, 2011.
31. Demographic and Health Surveys. 2014. http://www.measuredhs.com/
[Accessed 25 April 2014]
32. The World Bank. Data: Poverty headcount ratio at $1.25 a day (PPP) (% of population). (n.d.) Washington, DC. Retrieved November 10, 2014, from http://data.worldbank.org/indicator/SI.POV.DDAY