The adult population impact of HIV care and antiretroviral therapy in a resource poor setting, 2003–2008 : AIDS

Secondary Logo

Journal Logo


The adult population impact of HIV care and antiretroviral therapy in a resource poor setting, 2003–2008

Gargano, Julia W.a,b; Laserson, Kaylac; Muttai, Hellend; Odhiambo, Frankc; Orimba, Vincentc; Adamu-Zeh, Mirabellec; Williamson, Johnd; Sewe, Maquinsc; Nyabiage, Lennahe; Owuor, Karenc; Broz, Ditaa,d; Marston, Barbarad; Ackers, Martad

Author Information
AIDS 26(12):p 1545-1554, July 31, 2012. | DOI: 10.1097/QAD.0b013e328353b7b9



Worldwide, 33 million people are infected with HIV, and two-thirds of the infection burden is in sub-Saharan Africa [1]. Since 2002–2003, global funding initiatives have enabled the rapid initiation and expansion of HIV care and antiretroviral therapy (ART) services in many developing countries, including most African nations [2]. Over 5 million people worldwide have enrolled in HIV care and treatment programs and have initiated effective treatment regimens that can prolong their lives [3].

Although the effectiveness of HIV care and ART at the individual level has been well documented in patient cohorts, only a few studies in sub-Saharan Africa have demonstrated declines in population-level adult mortality rates after introduction of ART programs [4–9]. To date, such studies have had limited information on characteristics of patients enrolling in care that may relate to access to care, treatment adherence, and survival. Changes in the demographics or clinical status of patients enrolled in HIV clinics during scale-up may have implications for assessing the impact of HIV care and treatment at the population level. Details on patient characteristics such as socioeconomic status, access to care, and clinical circumstances may enable a fuller understanding of the experience and impact of scale-up experience in a defined geographic area.

In Kenya, the 2007 adult HIV prevalence was 7.1%, which translated to 1.4 million adults living with HIV/AIDS [10]. The predominantly rural Nyanza Province in western Kenya has the nation's highest HIV burden, with a 14.9% adult HIV prevalence and ∼500 000 people living with HIV/AIDS [10]. HIV care and ART services were introduced, and with decentralization from provincial to health center facilities, clinical officers rapidly expanded population access to services during the years 2003–2008. Since 2001, the Kenya Medical Research Institute (KEMRI) and US Centers for Disease Control and Prevention (CDC) have maintained a Health and Demographic Surveillance System (HDSS) to capture population characteristics and mortality data from approximately 140 000 residents of a rural area of Nyanza Province with high rates of HIV, malaria, and tuberculosis (TB) [11]. We linked HIV clinic records of adult HDSS residents to longitudinal population-level HDSS data to describe the uptake of HIV care and ART services, identify trends in the clinical and sociodemographic characteristics of HIV clinic patients enrolling in care, and explore changes in adult mortality rates in the population during a time of rapid expansion of HIV clinical services.


Health and Demographic Surveillance System population

Details of the setting, design, and conduct of the HDSS have been published previously [11]. Multihousehold residential compound locations and health facilities have been mapped using global positioning system methods. Trained staff conduct household interviews three times annually to collect birth, death, and migration information for HDSS residents, defined as persons resident in HDSS households for at least 4 months prior to the interview or newborns born to residents. Demographic data such as age/date of birth, education, and occupation of head of household are gathered during initial population enumeration; household assets and occupation are updated annually. HDSS data used in this study included wealth index (quintiles of all HDSS households), household water chlorine treatment, education (≤primary, ≥secondary), and residential location.

A trained interviewer conducts a detailed verbal autopsy interview with a primary caregiver 1–4 months after notification of a resident death, using standardized methodology [12]. Clinicians review the verbal autopsy data and assign a probable cause of death; in 2007, the Sample Vital Registration with Verbal Autopsy coding methods were adopted [13]. For this analysis, deaths were placed into one of four categories: HIV/TB (combined due to their overlap in reporting [5,8,14]); other infectious diseases (e.g. malaria, meningitis, pneumonia, gastroenteritis, typhoid); noninfectious diseases (e.g. trauma, pregnancy-related, cancer, asthma); or unknown/unclassifiable.

Clinic cohort

We assessed service receipt at clinics providing HIV care to HIV-infected HDSS residents who enrolled in HIV care and treatment services. From September 2008 to October 2009 medical chart abstractions were conducted at all nine health facilities providing HIV care/ART within the HDSS area and at an additional eight health facilities within 10 km of the HDSS boundaries. Eligibility criteria for chart abstraction included adult (age ≥18 years), enrollment into HIV care/ART in 2003–2008, and HDSS residency at enrollment. Demographic, clinical [WHO stage, TB diagnosis, medications (e.g. antiretrovirals [ARVs]), and services provided], and laboratory (CD4 cell count) data were abstracted from patient charts, along with dates of visits, deaths, and clinic transfers. Patients were classified as lost to follow-up if their most recent visit occurred more than 90 days before record abstraction. Except where noted, we defined ART eligibility at enrollment consistent with the Kenya National AIDS & STI Control Programme (NASCOP) 2007 criteria: WHO stage 4 (regardless of CD4 cell count); WHO stage 3 (CD4 cell count unavailable); WHO stage 3 and CD4 cell count less than 350 cells/μl; and WHO stage 1 or stage 2 and CD4 cell count less than 250 cells/μl, or CD4 cell count less than 250 cells/μl regardless of WHO stage. First-line ART regimens consisted of zidovudine or stavudine + lamivudine + efavirenz or nevirapine.

Clinical and HDSS data were linked using patient name, sex, age at enrollment (±5 years), and residential location. Of 5421 adult clinic patients who reported HDSS residency and enrolled in HIV care during 2003–2008, 4004 (73.9%) linked to HDSS. Of these, 673 (16.8%) did not have an HDSS data point after clinic enrollment. In most cases (66%) this was attributable to enrollment near the end of the HDSS follow-up period; documented out-migration from HDSS before clinic enrollment and loss to follow-up from HDSS explain a minority of cases. The 3331 linked and followed up patients comprise the clinic cohort.

Data analysis

We calculated the straight-line distance between patients’ residence and their HIV care facility of enrollment (distanced traveled by roads unavailable). We evaluated crude secular trends in patient characteristics across enrollment years using the Cochran–Armitage test for variables with two categories, the Cochran–Mantel–Haenszel test with modified ridit scores for variables with three or more categories, and Spearman's rank correlation for continuous variables [15].

We calculated 12-month clinic cohort patient mortality rates per 100 person-years and 95% confidence intervals (CIs) from the date of enrollment in care to date of death or censoring (31 December 2008 or last HDSS contact, whichever came first) using Poisson regression. For these calculations, we excluded patients who enrolled in 2008 because they were followed for less than 12 months.

We calculated HDSS population adult mortality rates per 1000 person-years and 95% CI using 2-year time periods corresponding to the history of HIV care/ART scale-up (i.e., 2003–2004, early program implementation; 2005–2006, increased implementation; and 2007–2008, widespread implementation). We calculated adult (age 18–64 years) mortality rates for each period and estimated unadjusted rate ratios for the two later time periods compared with the earliest.

We estimated the proportion of population eligible for ART based on Nyanza Province data obtained in a national survey [10]: approximately 24% of HIV-infected adults were ART-eligible based solely on CD4 cell count less than 250 (or ∼30% with CD4 cell count <350). Using clinic cohort patients active in care/on ART at midyear as numerators and midyear HDSS population as denominators, we calculated proportions of HIV-positive patients receiving services and the proportion of ART-eligible patients on ART for each year, and adjusted estimates for missing data due to incomplete linkage or lack of follow-up (see Table, Supplemental Digital Content 1,, which shows derivation of estimates of HIV Care and ART coverage).

All statistical analyses were conducted in SAS version 9.2 (Statistical Analysis Software, Cary, North Carolina, USA).

Ethical review

This study was approved by CDC and KEMRI Institutional Review Boards.


Figure 1 illustrates the rapid scale-up of HIV care sites through clinic cohort patient enrollment in the HDSS and bordering areas from 2003 to 2008. As of 2003, only 11 HDSS residents were documented as being enrolled in HIV care at one facility. At the end of 2004, 210 residents had enrolled at six facilities in or near the HDSS, although ART initiation was not yet available within these HDSS facilities (Fig. 1a). By the end of 2006, HIV care services had expanded to an additional five health facilities (n = 11), and eight (73%) offered ART; a total of 1143 HDSS residents had accessed care and 399 (25%) had initiated or continued ART (Fig. 1b). At the end of 2008, 12 (70%) of the 17 facilities now providing HIV care were offering ART and 3331 HIV-infected adult HDSS residents had enrolled in HIV care; 1248 (37%) had received ART (Fig. 1c). In 2008, the median number of new patient enrollments per clinic was 41 (range 1–315); five facilities had cumulative enrollments of more than 400 HDSS residents.

Fig. 1:
Scale-up of HIV Care, Asembo and Gem, 2003–2008.Cumulative clinic cohort enrollment through (a) December 31, 2004; (b) December 31, 2006; (c) December 31, 2008.

Compared to the 2090 noncohort clinic patients, the 3331 clinic cohort patients were older (median age 35 vs. 31 years), were more likely to be married, to have enrolled before 2008, or to have died according to clinic records; however, the two clinic groups differed by less than 5% in terms of sex and WHO stage (not shown). Compared to the entire adult HDSS population in 2006 (n = 65 944), the 3331 clinic cohort patients were more often female (65.8 vs. 56.7%) and more often age 25–49 (71.9 vs. 43.1%). Clinic cohort education and wealth were similar to HDSS residents (19.7 vs. 22.8% with secondary education, 31.5 vs. 31.4% in poorest two quintiles of households), whereas patients had greater access to chlorine water treatment (27.2 vs. 21.0%).

Trends in enrollment characteristics and service receipt of clinic cohort patients by enrollment year are shown in Table 1. During 2003–2008, the proportion of new patients with a primary education or less and home access to safe water treatment increased significantly. Wealth index distribution remained stable (not shown), and the median distance from patient residence to enrollment clinic declined. The proportion of patients referred from voluntary counseling and testing sites decreased, while other clinic-based referral sources increased. Among 2186 women, 6% were referred from Prevention of Mother to Child Transmission programs and antenatal clinics (PMTCT/ANC). Among all patients, 4.4% (n = 147) had documented exposure to ARVs before their HDSS clinic enrollment, primarily for ART initiated at another facility, although 25 women (1.1% of all women) received ARVs for PMTCT. Most patients (92.9%) received cotrimoxazole prophylaxis. Only 3% of patients had condom receipt documented; family planning, bednets, and safe water systems were documented for less than 10 patients (not shown). WHO stage documentation declined as documentation of CD4 cell count at enrollment increased. Among patients with data available, the proportion enrolling at WHO stage 4 declined from 20.4 to 1.9%, and median enrollment CD4 cell counts increased from 192 cells/μl [IQR 111–279 cells/μl] (2005) to 256 cells/μl [IQR 130–457 cells/μl] (2008). Correspondingly, the proportion of patients eligible for ART at enrollment declined from 61.9 to 44.1%.

Table 1:
Trends in enrollment characteristics and HIV care services, by year of enrollment.

When limiting analysis to medical chart data for the 2298 clinic cohort patients who enrolled before 2008, the median patient follow-up time was 12 months (range 0–66 months), and 12-month patient retention improved slightly over time, from 51.0 (2003–2004) to 57.2% (2007). However, by using available HDSS outcome information for patients whose clinic data recorded them as lost to follow-up or transfers, we increased the median documented patient follow-up to 18 months and identified approximately four times more patient deaths (from 171 to 790). For example, as of the end of 2008, of the 210 patients who enrolled in 2003–2004, 89 (42%) were alive and residing within the HDSS area [although only 41 (20%) were reported as active in their HIV clinic of enrollment]; 99 (47%) had died, and 22 (10%) migrated out. The 12-month patient mortality rates per 100 person-years declined (P-trend < 0.0001): 32.9 (95% CI 27.7–39.2) in 2003–2004, 36.9 (95% CI 28.6–47.6) in 2005, 24.7 (95% CI 21.9–27.9) in 2006, and 22.3 (95% CI 17.9–27.8) in 2007.

From 2003 to 2008, the HDSS documented 7137 deaths in 345 308 person-years of observation among adults (Table 2). A total of 3960 deaths (55.5%) were attributed to HIV/TB, 1033 (14.5%) were attributed to other infectious disease, 838 (11.7%) were attributed to noninfectious disease, and 1306 (18.3%) were unknown/unclassifiable. All-cause adult mortality rates declined from 24.6 per 1000 person-years in 2003–2004 (i.e., the early program implementation period) to 21.4 per 1000 person-years in 2005–2006 after increased implementation, a 13% decrease (i.e., rate ratio = 0.87, 95% CI 0.83–0.92); rates further declined to 16.3 per 1000 person-years in 2007–2008 after widespread program implementation, a 34% decrease from 2003–2004. HIV/TB mortality rates increased 8% in 2005–2006 compared with 2003–2004, then declined in 2007–2008, for a 26% decrease from 2003–2004. Other infectious disease and unknown/unclassifiable mortality rates also declined. Noninfectious disease mortality rates did not significantly change.

Table 2:
Population-level all-cause and cause-specific mortality rates per 1000 person-years and mortality rate ratios, KEMRI/CDC Demographic Surveillance System, 2003–2008.

Figure 2 shows population-level mortality rates stratified by age and sex. In males, all-cause and HIV/TB mortality rates were highest in the 35–49-year age group, whereas in females, these rates peaked in the 25–34-year age group. Noninfectious disease mortality rates increased with age. Other infectious disease and unknown/unclassifiable mortality rates were similar, increasing with age for males and peaking at age 25–34 for females. From 2003–2004 to 2007–2008, the largest declines in all-cause and HIV/TB mortality rates occurred among residents age 25–34; the largest declines in other infectious disease mortality rates occurred among males age 25–34 years and females age 35–49 years (see Table, Supplemental Digital Content 2,, which shows age-stratified and sex-stratified cause-specific mortality rates and mortality rate ratios). Noninfectious disease mortality rates did not change significantly in any age/sex stratum.

Fig. 2:
Population-level all-cause and cause-specific mortality rates by age group and sex.Health and Demographic Surveillance System, 2003–2008. ID = infectious disease.

Figure 3 shows the scale-up of HIV care as an estimated proportion of HIV-positive adults accessing care at mid-year, overlaid with mortality rates for TB/HIV, other infectious disease, and noninfectious disease causes. In 2003–2004, less than 2% of the HIV-positive population was enrolled in care. This proportion increased to 29.5% in 2008. ART was first offered in these HDSS facilities in 2005 to 5% of clinic patients; by mid-2008, more than half of clinic patients were on ART. ART coverage (based on CD4 cell counts <250 cells/μl for eligibility) increased from less than 1% in 2003–2005 to 6.6% in 2006 to 64.0% in 2008. Changing the ART eligibility threshold to CD4 cell count less than 350 cells/μl reduced our estimate of ART coverage to 5.3% (2006) and 51.9% (2008).

Fig. 3:
Population cause-specific mortality rates and 95% confidence intervals (lines, y-axis on left), and estimated proportion of HIV-positive patients in care and receiving ART (stacked bars, y-axis on right), by year.Estimates are based on the mid-year adult (age 18–64) population of Health and Demographic Surveillance System (HDSS), Nyanza Province, Kenya and HIV clinic patients active at mid-year within the HDSS areas (details of calculations in text and Supplemental Digital Content 1,


Prior to 2003, residents of a rural area of Kenya with high HIV prevalence lacked access to nearby HIV care or ART services; however, 5 years later, nearly 30% of HIV-positive adult residents were receiving HIV care and almost two-thirds of eligible patients initiated ART. Geographic coverage was achieved relatively uniformly, over time the distance from household residence to facility of HIV care decreased, and residents presented earlier in the course of their illness. We documented declining mortality rates in clinic patients and a 34% reduction in all-cause mortality rates in the surrounding adult population during this period of rapid services expansion. Although studies in three sub-Saharan African countries (Malawi [4,6,7], South Africa [5]; and Ethiopia [8]) have reported reductions in population-level adult mortality rates following scale-up of HIV care and ART programs, our study is the first to also comprehensively describe changes in enrollment and clinical characteristics among HIV patients (pre-ART and ART) over time, thus illustrating the expansion of HIV program services.

Although our study does not directly demonstrate that the increased availability of HIV care and ART caused declines in all-cause mortality and deaths attributed to HIV/TB and to other infectious diseases at the population level, we believe this is the most likely explanation. First, although several other health interventions were implemented in the HDSS area during the study period, they were mostly directed toward mortality reduction among children less than 5 years of age or pregnant women (e.g., insecticide-treated bed nets, malaria intermittent preventive treatment during pregnancy [16–18]). To our knowledge, no other major public health interventions to control or prevent illness were implemented either by the Ministry of Health or through research studies during this period, nor were there any major socioeconomic interventions in this area, and HIV prevalence remained relatively stable [10,19]. Second, the greatest mortality rate reductions occurred in the age groups documented to have the highest HIV prevalence [10,19,20]. Third, verbal autopsy data indicated that adult deaths attributed to noninfectious disease processes did not decline during this time, suggesting the presence of an intervention directed primarily at infectious diseases. Finally, the consistency between our findings and those from other countries showing declines in AIDS-related mortality following the widespread provision of ART in other high HIV prevalence areas of sub-Saharan Africa [4–6,8] contributes to our confidence that HIV care and ART coverage largely contributed to the declines we observed.

Providing HIV care and ART to individual patients has direct health benefits at the individual level [21–25], but the population impact of providing HIV care might be greater than aggregated individual-level effects. Although ART has been shown to have significant health impacts, other components of HIV care (e.g., cotrimoxazole prophylaxis, TB screening, bednets, and safe water treatment systems) can also reduce individual morbidity and mortality due to other infectious diseases, and these benefits might accrue to other household members and the community [26–32]. Indeed, not only did we observe high rates of cotrimoxazole prophylaxis (consistently >90%), but HDSS data indicated that HIV patients had greater access to chlorine water treatment than the general HDSS population suggesting that their families may also have experienced less illness during HIV care. In addition, because ART has been shown to reduce the risk of HIV transmission to uninfected partners [33], and treatment of other illnesses can also reduce viral load [34], the administration of chemotherapeutic agents provided through ongoing HIV care services may have decreased HIV transmission within some households.

Although we believe our estimates of care and treatment coverage (29.5% of HIV-infected in care, 64.0% of CD4 cell count less than 250 cells/μl on ART in 2008) are reasonable, the actual coverage might be higher because some HDSS residents received HIV care at clinics where chart abstraction did not occur. However, based on the limited economic resources of the HDSS population and a 2007 survey indicating that patients traveled an average of 5 km for HIV clinical services (M. Ackers, personal communication), we estimate that, at most, 20% of HDSS HIV-infected residents may have received HIV care at nonabstracted facilities. If so, mid-2008 coverage may have been as high as 37% of HIV-positive patients in care and 80% of eligible adults on ART. On the contrary, expanding the ART-eligibility criteria to CD4 cell count less than 350 cells/μl would reduce the 2008 coverage estimate to 51.9%. We were unable to incorporate data on clinical staging into our estimates, which might have further increased the pool of eligible residents and decreased our estimated coverage.

A key strength of this study – linkage between clinic data and population surveillance – allowed us to evaluate declining mortality at both the patient and population levels during expansion of services. Nevertheless, limitations in the study design and data sources merit consideration. We were unable to stratify HDSS mortality rates by HIV serostatus because HDSS does not routinely collect biologic measures; additionally, we could not directly evaluate the mortality benefit accruing to HDSS residents not enrolled in HIV care programs. Although ascertainment of patient outcomes was greatly improved through linkage with HDSS, patient and population mortality rates might still be underestimates if residents lost to follow-up from the HDSS died; however, we would not expect this to impact trends over time. Differences between HDSS clinic cohort and nonlinked HDSS patients suggest some selection bias in favor of more stable patients in the clinic cohort (i.e. those who were older or married) who might be more likely to access services or have better outcomes. The clinical data were abstracted from patient charts and were limited by incomplete documentation, including missing dates, that may have led to nonrandom misclassification. Although ART was not documented for any patients until 2005, a few patients might have started ART at another facility in 2003–2004, however, clinic records do not include any reliable data on such treatment. Trends showing increases in median enrollment CD4 cell counts must be interpreted cautiously as these data were only available for a subgroup of patients and may not be representative, although the trend appears substantiated by WHO staging data (increases in the proportions of patients with earlier disease stages in successive patient cohorts). Finally, as verbal autopsy data have inherent limitations [35] and 18.6% of HDSS adult deaths lacked a defined etiology, cause-specific mortality rates may be inaccurate. Although the validity of the KEMRI/CDC HDSS verbal autopsy system has not been formally evaluated, other verbal autopsy systems in sub-Saharan Africa have reported variable sensitivity (31–86%) and specificity (78–94%) for HIV or HIV/TB deaths [36–39]. Coding changes implemented during the study period might have influenced HIV/TB mortality rates; nevertheless, validity is supported by observations that the highest HIV/TB and infectious disease mortality rates and greatest rate reductions were observed in the age and sex strata with the highest population HIV prevalence [10,19] and the greatest enrollment in HIV care and treatment services.

HIV program services changed substantially from 2003 to 2008. The quantity of clinical sites and volume of patients increased substantially. With the decentralization of HIV services, patients were able to access care closer to their homes, which was reflected in the declining distance to health facilities with successive cohorts of patients. Over time, patients with lower education levels and less advanced disease began entering care. Expansion in service delivery and system efficiencies improved the quality of HIV care as evidenced by the increasing availability of CD4 cell count testing through equipment placement and laboratory networks and reduced time to ART initiation following enrollment.

Despite the programmatic advancements and expanded population access to ART, this study highlights the need for continued improvements in HIV service delivery to further improve HIV-related morbidity and mortality and reduce HIV transmission. Although PMTCT services expanded during this time period, referral from PMTCT to care and treatment services remained low (5.1%) in 2008, and few women (n = 25) reported PMTCT ARV prophylaxis. While patient clinical status at enrollment improved, the median CD4 cell count of patients enrolling in 2008 was still low at 256 and 44% of patients qualified for ART at enrollment. Thus, efforts to identify HIV-positive individuals earlier and achieve rapid enrollment into clinical care with high rates of retention and ART adherence are warranted to optimize individual and population-level benefits of HIV care and ART services.


All authors revised the article for important intellectual content and approved the final version. In addition, J.W.G. developed the data analysis plan, analyzed the data, and drafted the manuscript; K.L. contributed to study design, interpretation of the data, and preparation of the manuscript; H.M. contributed to data collection, data interpretation and preparation of the manuscript; F.O. oversaw the HDSS and contributed to data collection and the final draft of the manuscript; V.O. managed the verbal autopsy system and assisted with interpretation of cause-specific mortality data; M.A.-Z. oversaw the clinic data collection and contributed to the final manuscript; J.W. assisted in developing the data analysis plan and assisted with interpretation of the statistical methods; M.S. analyzed portions of the HDSS data and assisted in its interpretation; L.N. contributed to clinical data collection and final manuscript preparation; K.O. contributed to the clinic data collection and linkage between the clinic and HDSS data; D.B. developed portions of the data analysis plan and assisted with analysis and interpretation of the data; B.M. contributed to the study design and the final manuscript preparation; M.A. provided oversight for the study as a whole, contributed to the study design, interpretation of the data, and preparation of the manuscript. The authors would like to acknowledge the contributions of the late Dr Kubaje Adazu to the HDSS, and Drs Kevin De Cock, Laurence Slutsker, Mary Hamel, and Allen Hightower for their contributions to the HDSS and critical comments made on the manuscript. The authors also wish to thank Alan Rubin, Maurice Ombok, and Paul Lee for their assistance with map design.

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the funding institutions.

This work received funding support through the President's Emergency Plan for AIDS Relief (PEPFAR) through the U.S. Centers for Disease Control and Prevention (cooperative agreement number 5U19CI000323-05).

Conflicts of interest

There are no conflicts of interest.


1. UNAIDS, World Health Organization. AIDS epidemic update – 2009. 2009.
2. Biesma RG, Brugha R, Harmer A, Walsh A, Spicer N, Walt G. The effects of global health initiatives on country health systems: a review of the evidence from HIV/AIDS control. Health Policy Plan 2009; 24:239–252.
3. World Health Organization. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector: progress report 2010. Geneva: World Healh Organization; 2010.
4. Floyd S, Molesworth A, Dube A, Banda E, Jahn A, Mwafulirwa C, et al. Population-level reduction in adult mortality after extension of free antiretroviral therapy provision into rural areas in northern Malawi. PLoS One 2010; 5:e13499.
5. Herbst AJ, Cooke GS, Barnighausen T, KanyKany A, Tanser F, Newell ML. Adult mortality and antiretroviral treatment roll-out in rural KwaZulu-Natal, South Africa. Bull World Health Organ 2009; 87:754–762.
6. 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. Lancet 2008; 371:1603–1611.
7. Mwagomba B, Zachariah R, Massaquoi M, Misindi D, Manzi M, Mandere BC, et al. Mortality reduction associated with HIV/AIDS care and antiretroviral treatment in rural Malawi: evidence from registers, coffin sales and funerals. PLoS One 2010; 5:e10452.
8. 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, Ethiopia. AIDS 2009; 23:511–518.
9. Mills EJ, Bakanda C, Birungi J, Chan K, Ford N, Cooper CL, et al. Life expectancy of persons receiving combination antiretroviral therapy in low-income countries: a cohort analysis from Uganda. Ann Intern Med 2011; 155:209–216.
10. Kenya National AIDS/STI Control Programme (NASCOP). 2007 Kenya AIDS indicator survey: final report. Nairobi, Kenya: NASCOP; 2009.
11. Adazu K, Lindblade KA, Rosen DH, Odhiambo F, Ofware P, Kwach J, et al. Health and demographic surveillance in rural western Kenya: a platform for evaluating interventions to reduce morbidity and mortality from infectious diseases. Am J Trop Med Hyg 2005; 73:1151–1158.
12. World Health Organization. Verbal autopsy standards: ascertaining and attributing cause of death. Geneva: WHO Press; 2007.
13. MEASURE Evaluation, US Census Bureau. SAVVY: sample vital registration with verbal autopsy: verbal autopsy supervisor's manual. Chapel Hill, NC: US Census Bureau; 2007.
14. van Eijk AM, Adazu K, Ofware P, Vulule J, Hamel M, Slutsker L. Causes of deaths using verbal autopsy among adolescents and adults in rural western Kenya. Trop Med Int Health 2008; 13:1314–1324.
15. SAS Institute Inc. SAS/STAT 9.2 user's guide, Second Edition. Cary, NC: SAS Institute Inc.; 2009.
16. Hightower A, Kiptui R, Manya A, Wolkon A, Vanden Eng JL, Hamel M, et al. Bed net ownership in Kenya: the impact of 3.4 million free bed nets. Malar J 2010; 9:183.
17. Odhiambo FO, Hamel MJ, Williamson J, Lindblade K, ter Kuile FO, Peterson E, et al. Intermittent preventive treatment in infants for the prevention of malaria in rural Western kenya: a randomized, double-blind placebo-controlled trial. PLoS One 2010; 5:e10016.
18. Ouma PO, van Eijk AM, Hamel MJ, Sikuku ES, Odhiambo FO, Munguti KM, et al. Antenatal and delivery care in rural western Kenya: the effect of training healthcare workers to provide ‘focused antenatal care’. Reprod Health 2010; 7:1.
19. Kenya National Bureau of Statistics, ICF Macro. 2008-09 Kenya demographic and health survey: key findings. Calverton, Maryland, USA; 2010.
20. Amornkul PN, Vandenhoudt H, Nasokho P, Odhiambo F, Mwaengo D, Hightower A, et al. HIV prevalence and associated risk factors among individuals aged 13-34 years in Rural Western Kenya. PLoS One 2009; 4:e6470.
21. Hawkins C, Achenbach C, Fryda W, Ngare D, Murphy R. Antiretroviral durability and tolerability in HIV-infected adults living in urban Kenya. J Acquir Immune Defic Syndr 2007; 45:304–310.
22. Sanne IM, Westreich D, Macphail AP, Rubel D, Majuba P, van Rie A. Long-term outcomes of antiretroviral therapy in a large HIV/AIDS care clinic in urban South Africa: a prospective cohort study. J Int AIDS Soc 2009; 12:38.
23. Siika AM, Ayuo PO, Sidle MJ, Wools-Kaloustian K, Kimaiyo SN, Tierney WM. Admission characteristics, diagnoses and outcomes of HIV-infected patients registered in an ambulatory HIV-care programme in western Kenya. East Afr Med J 2008; 85:523–528.
24. Stringer JS, Zulu I, Levy J, Stringer EM, Mwango A, Chi BH, et al. Rapid scale-up of antiretroviral therapy at primary care sites in Zambia: feasibility and early outcomes. JAMA 2006; 296:782–793.
25. Wools-Kaloustian K, Kimaiyo S, Diero L, Siika A, Sidle J, Yiannoutsos CT, et al. Viability and effectiveness of large-scale HIV treatment initiatives in sub-Saharan Africa: experience from western Kenya. AIDS 2006; 20:41–48.
26. Lule JR, Mermin J, Ekwaru JP, Malamba S, Downing R, Ransom R, et al. Effect of home-based water chlorination and safe storage on diarrhea among persons with human immunodeficiency virus in Uganda. Am J Trop Med Hyg 2005; 73:926–933.
27. Mermin J, Were W, Ekwaru JP, Moore D, Downing R, Behumbiize P, et al. Mortality in HIV-infected Ugandan adults receiving antiretroviral treatment and survival of their HIV-uninfected children: a prospective cohort study. Lancet 2008; 371:752–759.
28. Dworkin MS, Williamson J, Jones JL, Kaplan JE. Prophylaxis with trimethoprim-sulfamethoxazole for human immunodeficiency virus-infected patients: impact on risk for infectious diseases. Clin Infect Dis 2001; 33:393–398.
29. Grimwade K, Sturm AW, Nunn AJ, Mbatha D, Zungu D, Gilks CF. Effectiveness of cotrimoxazole prophylaxis on mortality in adults with tuberculosis in rural South Africa. AIDS 2005; 19:163–168.
30. Khoo S, Back D, Winstanley P. The potential for interactions between antimalarial and antiretroviral drugs. AIDS 2005; 19:995–1005.
31. Mermin J, Lule J, Ekwaru JP, Downing R, Hughes P, Bunnell R, et al. Cotrimoxazole prophylaxis by HIV-infected persons in Uganda reduces morbidity and mortality among HIV-uninfected family members. AIDS 2005; 19:1035–1042.
32. Wiktor SZ, Sassan-Morokro M, Grant AD, Abouya L, Karon JM, Maurice C, et al. Efficacy of trimethoprim-sulphamethoxazole prophylaxis to decrease morbidity and mortality in HIV-1-infected patients with tuberculosis in Abidjan, Cote d’Ivoire: a randomised controlled trial. Lancet 1999; 353:1469–1475.
33. Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med 2011; 365:493–505.
34. Barnabas RV, Webb EL, Weiss HA, Wasserheit JN. The role of co-infections in HIV epidemic trajectory and positive prevention: a systematic review and meta-analysis. AIDS 2011; 25:1559–1573.
35. Chandramohan D, Setel P, Quigley M. Effect of misclassification of causes of death in verbal autopsy: can it be adjusted?. Int J Epidemiol 2001; 30:509–514.
36. Chandramohan D, Maude GH, Rodrigues LC, Hayes RJ. Verbal autopsies for adult deaths: their development and validation in a multicentre study. Trop Med Int Health 1998; 3:436–446.
37. Kamali A, Wagner HU, Nakiyingi J, Sabiiti I, Kengeya-Kayondo JF, Mulder DW. Verbal autopsy as a tool for diagnosing HIV-related adult deaths in rural Uganda. Int J Epidemiol 1996; 25:679–684.
38. Setel PW, Whiting DR, Hemed Y, Chandramohan D, Wolfson LJ, Alberti KG, et al. Validity of verbal autopsy procedures for determining cause of death in Tanzania. Trop Med Int Health 2006; 11:681–696.
39. Lopman BA, Barnabas RV, Boerma JT, Chawira G, Gaitskell K, Harrop T, et al. Creating and validating an algorithm to measure AIDS mortality in the adult population using verbal autopsy. PLoS Med 2006; 3:e312.

AIDS/HIV; antiretroviral therapy; mortality; population surveillance

Supplemental Digital Content

© 2012 Lippincott Williams & Wilkins, Inc.