The President's Emergency Plan for AIDS Relief (PEPFAR) was a response by the US government to the significant public health impact of HIV infection in the developing world . Signed in May 2003, $15 billion were committed to fund this program for 5 years, 55% of which was earmarked for expanding treatment access for individuals living with HIV/AIDS. The ultimate goal of these treatment funds was to extend combination antiretroviral therapy (cART) to at least 2 000 000 individuals by the end of 2006.
The Indiana University School of Medicine has had a collaborative partnership with Moi University in Eldoret Kenya since 1990 . In 2001, the Academic Model for the Prevention and Treatment of HIV/AIDS (AMPATH) was created as a joint initiative between Moi University, Moi Teaching and Referral Hospital and the Indiana University School of Medicine [3–5]. The goal of AMPATH was to establish an HIV care system to serve the needs of both urban and rural patients, create a partnership to assess and overcome barriers to care, and improve outcomes of cART in a resource-poor setting [4,6,7]. The AMPATH electronic medical record system (AMRS) was developed to support care and program evaluation [8,9].
AMPATH has experienced three distinct periods of operation as defined by funding for and availability of cART. Period 1 began in November 2001 with the establishment of its first two HIV clinics when cART was funded through patient self-pay and a handful of private donations and includes AMPATH's entry into the MTCT-plus initiative (supported treatment for 350 women and their families), leading to available funding of approximately $650 000 during this period. Period 2 started in June 2003 with the wide availability of a generic combination antiretroviral (Triomune), which lowered monthly per patient cART costs from $40 to 23. During this period, AMPATH received over a million dollars in donations to support cART provision. Period 3 was launched in March 2004 with the receipt of PEPFAR funding. During this period, PEPFAR provided approximately $6.5 million for HIV care in addition to supplying antiretrovirals at no cost to the program.
The present study uses AMRS data to assess the impact of PEPFAR funding on the expansion of AMPATH services to nonpregnant adult patients and its impact on patient characteristics at enrollment and at the time of cART initiation.
The present study was approved by the ethical bodies of both the Indiana University and Moi University Schools of Medicine. This cohort study used de-identified data extracted from the AMRS.
Adult patients were eligible for inclusion, if they were enrolled at any AMPATH clinic between 27 November 2001 and 8 May 2006. Pregnant women were excluded because their initiation and duration of cART differs from that of the standard protocol. All AMPATH clinics are located in western Kenya within health facilities operated by the Kenyan Ministry of Health. Patients were assigned to one of the three periods on the basis of their date of enrollment in AMPATH.
Detailed algorithms consistent with the 2004 WHO guidelines for the care of HIV-infected patients were developed locally and followed throughout the study period [7,10]. Patients meet eligibility for cART on the basis of a combination of clinical and immunologic criteria. All WHO clinical stage 4 patients, stage 3 patients with a CD4 cell count less than 350 cells/μl and stage 1 and stage 2 patients with a CD4 cell count less than 200 cells/μl were eligible for cART. In period 1, some patients were initiated on cART on the basis of an absolute lymphocyte count of less than 1000 cells/μl. Patients who did not meet criteria for cART were maintained in the care program and received routine care and laboratory monitoring through the cART clinic. Patients initiated on cART were seen 2 weeks after treatment initiation and then monthly thereafter, patients not receiving cART were seen every 1–3 months depending on comorbidities and patient travel time. Mortality data were captured using a passive surveillance system during periods 1, 2, and most of period 3. Increasing utilization of outreach workers to track patients who missed clinic appointments characterized the end of period 3.
Two standard adherence questions were added to the follow-up clinic in June 2003 to allow the assessment of adherence at every visit. The first question asked was ‘during the last month has the patient missed any medications?’ The responses were ‘yes’ or ‘no’; and if yes, proceeded to ask which medications were missed: antiretrovirals, tuberculosis prophylaxis, tuberculosis treatment or pneumocystis prophylaxis. The second question asked was ‘during the last seven days how many of his/her pills did the patient take?’ The available responses were ‘none’, ‘few’, ‘half’, ‘most’, and ‘all’.
Laboratory testing was based on local protocols and clinical necessity. A CD4 cell count was performed at baseline and every 6 months both on patients initiating cART and those who did not yet meet eligibility criteria for treatment. However, because of cost constraints, during period 1 and to a lesser extent period 2, self-pay patients frequently did not have CD4 cell counts drawn at the specified intervals. In these individuals, initiation of therapy was based on clinical signs, symptoms, and absolute lymphocyte counts. Because of the limitations in funding during period 1 and 2, first priority for treatment was given to ill patients who were thought by their clinician to be capable of responding to cART. Thus, some patients who were clinically stable (i.e. CD4 cell count <200 cells/μl but WHO clinical stage 1) or profoundly ill (i.e. with metastatic Kaposi's sarcoma) and thus felt to be less likely to benefit from cART were not provided antiretrovirals despite meeting criteria for treatment . The cART regimen used throughout the study period consisted of stavudine, lamivudine, and nevirapine. Efavirenz was substituted for nevirapine in patients receiving rifampin for treatment of tuberculosis.
Data collection and management
The data elements collected on the AMRS encounter forms (available as online supplemental information) at the initial encounter included standard demographic, historical, psychosocial, physical, and laboratory data as well as medications provided (antiretrovirals and opportunistic infection prophylaxis). Follow-up data were collected on intercurrent symptoms, medication adherence, new diagnoses, laboratory data, and additions or changes in drug regimens. Dedicated data entry clerks entered this information into the AMRS [7–9].
Median time from enrollment to ART initiation was estimated via the method of Kaplan and Meier and 95% confidence intervals (CIs) were generated via the Greenwood variance formula. Association between categorical factors was assessed through the χ 2-test. Continuous or ordinal factors were compared using Kruskal–Wallis nonparametric test and comparisons between periods in terms of length of time to an event (e.g., time from enrollment to cART initiation, time from enrollment to cART eligibility, and time from eligibility to cART to cART initiation) were performed by the log–rank test. The visit for the 6-month CD4 cell count was the closest patient encounter to the derived 6-month visit and within a 3-month window of this visit date). Adherence was measured by a summary score that assigned perfect adherence to a patient if all visit-wise self-reported adherence measures signified perfect adherence. Because patients with longer follow-up (and thus more individual adherence evaluations) are less likely to exhibit perfect overall adherence (analyses not shown), we adjusted for the number of adherence evaluations by adding the log-transformed number of evaluations as a predictor in a logistic-regression model of the summary adherence measure (perfect/not perfect adherence) already containing the period effect as the other predictor. All analyses were performed with the SAS system version 9.1 (SAS Institute, Cary, North Carolina, USA).
As of May 2006, 23 539 adult nonpregnant patients had been enrolled in AMPATH: 2181 were enrolled during the first two periods and 21 357 during period 3. The percentage of men enrolling during period 3 (34.0%; P < 0.001) was lower than in the initial two periods (37.3% period 1 and 38.8% period 2). The median age at enrollment of approximately 36 years remained unchanged across all periods (P > 0.1).
Infrastructure support and expanded enrollment
Two clinics (one urban and one rural) were established during period 1 with expansion to two additional rural clinics in period 2. During period 3, HIV clinical services were extended to 13 clinics, bringing the total number to 17 (four urban or semi-urban and 13 rural). The proportion of individuals traveling more than 2 h to receive care declined from 28% during the initial periods to 20% during period 3 (P < 0.001).
During period 1, AMPATH enrolled 1210 patients at a rate of approximately 64.2 patients per month. During period 2, the rate of enrollment almost doubled, with 971 new patients enrolled at a rate of 118 new patients per month. Between receipt of PEPFAR funding in March 2004 and the end of the study period (May 2006), 21 358 new patients were enrolled at an average enrollment rate of 817 patients per month (Fig. 1).
Patient clinical characteristics during the three periods
WHO stage at enrollment differed significantly between period 3 and both prior periods, with only 6.7% of patients having stage 4 disease during period 3 compared with 13.8 and 14.8% during periods 1 and 2, respectively (P < 0.001) (Table 1). The median CD4 cell count within 6 months of enrollment for period 3 was 172 cells/μl (95% CI: 4–807), which is significantly higher than that for period 2, 119 cells/μl (95% CI 1–726; P < 0.001) and period 1, 146 cells/μl (95% CI 2–715; P < 0.001).
The documentation of cART eligibility within 30 days of enrollment (including CD4 cell count criteria) improved significantly during period 3 with 53% of enrolled patients missing data on eligibility in period 1, 51% in period 2, and only 19% in period 3 (P < 0.0001). For those patients with eligibility criteria documented within 30 days of enrollment, 59.8% were eligible for cART in period 1, 72% in period 2, and 58% during period 3 (P < 0.001). Among patients documented as eligible for cART, 62.1% initiated therapy within 1 month after documentation of eligibility during period 1, 67.5% during period 2, and 56.5% during period 3 (P < 0.001). Although the increase in median time from documented cART eligibility to antiretroviral initiation over the three periods is 2.0, 1.9, and 4.0 weeks during period 1, 2, and 3, respectively, there was a significant decrease in the time from enrollment to cART initiation during the three periods, moving from a median of 63.7 weeks (95% CI 56.0–75.0) during period 1 to 26.3 (95% CI 20.7–30.3) in period 2 and 12.0 (95% CI 11.6–12.6) in period 3 (P < 0.001; Fig. 2a–c). The proportion of patients with WHO stage 4 disease at initiation of therapy was lower during period 3 (11%) than during period 1 (14%) and period 2 (18%; P < 0.001). At initiation of cART, the median CD4 cell counts were not significantly different between the periods at 112 cells/μl (95% CI 2–49.0) for period 1, 92 cells/μl (95% CI 1–47.8) for period 2, and 103 cells/μl (95% CI 3–40.4) for period 3. A CD4 cell count was available by 6 months after cART initiation for 46.5% of patients during period 1, 59.8% in period 2, and 59.4% in period 3. The median increase in CD4 cell count at 6 months after cART initiation is significantly different between period 2 and period 3, with a median increase of 110 cells/μl [interquartile range (IQR) 49–188] in period 1, 101 cells/μl (IQR 50–176) in period 2, and 119 cells/μl (IQR 55–189) (P = 0.0118).
The passively collected 6-month mortality rate for patients enrolled was 1.5% in period 1, 1.4% in period 2, and 2.7% in period 3. For patients enrolled during period 1, the 6-month and 12-month loss to follow-up (LTFU) rates were15.9 and 20.2%, respectively. When compared with LTFU in period 1, the LTFU increased significantly during period 2 to 22.0 and 27.9% at 6 and 12 months, respectively (P < 0.001). During period 3, LTFU declined significantly in comparison with period 2, leading to 18.0 and 24.3% LTFU rates at 6 and 12 months, respectively (P = 0.004). After adjusting for the number of adherence evaluations, adherence did not differ significantly during the three periods, with 73.0, 73.3, and 81.4% of the patients achieving perfect adherence during periods 1, 2, and 3, respectively.
Although our program, as well as many others in sub-Saharan Africa, has documented the feasibility and effectiveness of large-scale rollout of ART, none of these programs have previously assessed the impact of PEPFAR on program expansion and effectiveness [11–16]. The receipt of PEPFAR funds allowed AMPATH to triple its number of treatment sites and accelerated the enrollment of HIV-infected patients by a factor of six. Expansion of services to additional sites has led to a significant decrease in patient's travel time, which, in turn, likely lowered the economic burden of travel to the clinic. Documentation of cART eligibility increased significantly during the three periods, with 81% patients having documentation of their eligibility status within 30 days of program enrollment. Although time from documentation of cART eligibility to treatment initiation increased during period 3, we hypothesize that this related to the criteria (i.e., increased use of immunological criteria) by which a patient was deemed eligible for treatment. On the basis of WHO stage at enrollment, period 1 had a significantly greater number of patients meeting the eligibility for cART based on clinical criteria than had period 3. The fact that clinical criteria were immediately available, although there was a lag between the date of CD4 cell count collection (documentation date of eligibility) and the date that they were available to the clinician for decision making, likely explains the increase in time from documentation of cART eligibility to initiation of therapy. Despite the increase in time between documented cART eligibility and initiation, there was a significant overall decrease in the overall time from enrollment to cART initiation, which we attribute to an increase in documented cART eligibility status due to the greater availability of CD4 cell counts during period 3. We further anticipate that increasing clinic access and decreasing time to cART initiation will lead to improved patient outcomes.
The improved patient access to AMPATH services resulting from PEPFAR funding has had some impact on the clinical stage at enrollment, with patients presenting with higher median CD4 cell counts and less advanced disease during period 3 than during the previous periods. It is likely that the effects of PEPFAR funding on clinical stage at presentation and at cART initiation is greater than what we observed, because expansion to new clinic sites has attracted local patients who were already symptomatic but had no previous ability to access care.
High LTFU rates throughout all periods within this program are of great concern and are reflective of the reality for the majority of HIV care and treatment programs in sub-Saharan Africa [13,15,17]. In Rosen's recent systematic review of LTFU in sub-Saharan African ART programs, retention was not found to be associated with cohort size, date of program initiation, baseline CD4 cell count, or sex distribution. Retention, however, was found to be significantly better in cohorts, which did not require payment for services . In addition, a home-based care program in Uganda, which conducts weekly patient home visits, has reported nearly 100% retention . Within our program, the highest retention rates were seen during period 1 when the cohort was relatively small and the cost of cART was covered for the vast majority of patients. The highest LTFU rates were seen in our program during period 2 when the greatest number of patients enrolled in the program were self-pay and resources were unavailable for patient tracking. LTFU rates, though lower than the rates in period 2, remain higher in period 3 and may be related to several factors including the evolution of a less selective patient population as the program matured, increasing patient panels leading to less provider time to reinforce adherence, and an increasing number of alternative ART care sites within the country that provide patients with opportunities to move their care closer to their home or place of work. Our program is currently addressing retention issues using a multipronged approach, which includes enhanced outreach from the ART clinic, use of patient mentors in the community, streamline clinic visits for stable patients, and integrating the information systems between the primary care clinics and the ART clinics (so as to capture information on ART patients who are accessing primary care service). In addition, in contrast to Rosen's  findings, recent work done within our program indicates that men are more likely to be LTFU than WOMEN . As such, USAID-AMPATH is working on identifying strategies to better retain men within our program. Our findings as well as those of other programs indicate a pressing need to identify and integrate innovative retention and patient-tracking strategies into existing ART program as well as to incorporate them into ART roll-out plans.
Because data are collected as part of routine clinical care, the documented data elements are dependent on the limitations of the clinic environment. This could account for the fact that a significant number of patients during period 1 and 2 did not have CD4 cell count data documented within 30 days of enrollment and 6 months after cART initiation. As such, the median CD4 cell count data presented in this paper are reflective of a subset of patients during each of the three periods. During the first two periods, these individuals were likely to differ from patients without documented CD4 cell counts with regard to their socioeconomic status; thus, because of issues related to access to care, the median CD4 cell counts of these individuals may not be representative of the entire AMPATH population at that time point. Although this may also be considered a significant limitation, it does reflect the realities of HIV programs in sub-Saharan Africa, and as such provide valuable information for cART programs in the process of ramp-up.
PEPFAR funds have significantly increased the ability of individuals residing in western Kenya to access HIV care. It is anticipated that this access will, over time, lead to earlier HIV diagnosis, and initiation of treatment, thus allowing for more prolonged maintenance of health and economic well being. Despite this improvement in access, LTFU continues to be a significant issue for the majority of ART programs and must be addressed as programs ramp-up. Further strides forward in cART roll-out can only be maintained with the continued financial commitment by United States and other international donors to support the expansion of HIV treatment services.
We thank the staff and patients of the AMPATH HIV clinics and our collaborators from Indiana University, Brown University and Columbia University. The Purpleville Foundation, MTCT-plus, the United States Agency for International Development (USAID) and the President's Emergency Plan for AIDS Relief (PEPFAR) have funded patient care activities including the purchase of antiretroviral drugs. The Rockefeller Foundation and Fogarty International Center (grant 1-D43-TW01082) contributed funding to the development of the Electronic Medical Records System. All of the co-authors of this paper receive some salary support through PEPFAR and acknowledge that this may be considered a conflict of interest with regard to this manuscript. K.W.-K. was responsible for project inception, data interpretation, and drafting of the manuscript; S.K. and R.E. are responsible for the structure and administration of the USAID-AMPATH Partnership as well as being involved with interpretation of the data and intellectual contributions to the manuscript; B.M. was responsible for project inception, data analysis and intellectual contributions to the manuscript; J.S., A.S., W.N. and W.M.T. were involved in data interpretation and intellectual contributions to the manuscript; C.Y. was responsible for project inception, data analysis and interpretation and intellectual contributions to the manuscript. C.Y. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
2. Einterz RM, Kelley CR, Mamlin JJ, Van Reken DE. Partnerships in international health. The Indiana University-Moi University experience. Infect Dis Clin North Am 1995; 9:453–455.
3. Voelker R. Conquering HIV
and stigma in Kenya. JAMA 2004; 292:157–159.
4. Mamlin JJ, Kimaiyo S, Tierney W. Academic institutions linking acess to treatment and prevention: case study. Geneva: World Health Organization; 2004.
5. Einterz RM, Kimaiyo S, Mengech HN, et al
. Responding to the HIV
pandemic: the power of an academic medical partnership. Acad Med 2007; 82:812–818.
6. Wools-Kaloustian K, Kimaiyo S. Extending HIV
care in resource-limited settings. Curr HIV
Rep 2006; 3:182–186.
7. Wools-Kaloustian K, Kimaiyo S, Diero L, et al
. Viability and effectiveness of large-scale HIV
treatment initiatives in sub-Saharan Africa
: experience from western Kenya. AIDS
8. Siika AM, Rotich JK, Simiyu CJ, et al
. An electronic medical record system for ambulatory care of HIV
-infected patients in Kenya. Int J Med Inform 2005; 74:345–355.
9. Tierney WM, Rotich JK, Hannan TJ, et al
. The AMPATH medical record system: creating, implementing, and sustaining an electronic medical record system to support HIV
care in western Kenya. Medinfo 2007; 12(Pt 1):372–376.
10. WHO. Chronic HIV care with ARV therapy: integrated management of adolescent and adult illness
. Geneva: WHO; 2004.
11. Weidle PJ, Wamai N, Solberg P, Liechty C, Sendagala S, Were W, et al
. Adherence to antiretroviral therapy in a home-based AIDS
care programme in rural Uganda. Lancet 2006; 368:1587–1594.
12. Hammond R, Harry TC. Efficacy of antiretroviral therapy in Africa
: effect on immunological and virological outcome measures – a meta-analysis. Int J STD AIDS
13. Toure S, Kouadio B, Seyler C, Traore M, Dakoury-Dogbo N, Duvignac J, et al
, The Aconda Study Group. Rapid scaling-up of antiretroviral therapy in 10,000 adults in Cote d'Ivoire: 2-year outcomes and determinants. AIDS
14. Fairall LR, Bachmann M, Louwagie GMC, van Vuuren C, Chikobvu P, Steyn D, et al
. P Effectiveness of antiretroviral treatment in a South African program: a cohort study. Arch Inter Med 2008; 168:86–93.
15. 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.
16. Ferradini L, Jeannin A, Pinoges L, Izopet J, Odhiambo D, Mankhambo L, et al
. Scaling up of highly active antiretroviral therapy in a rural district of Malawi: an effectiveness assessment. Lancet 2006; 367:1335–1342.
17. Rosen S, Fox M, Gill C. Patient retention in antiretroviral therapy programs in sub- Saharan Africa
: a systematic review. PLoS Med 2007; 4:0001–0010.
18. Ochieng V, Ochieng D, Sidle JE, Holdsworth M, Siika AM, Yiannoutsos C, et al
. Gender and loss-to follow-up (LTFU) from a large HIV
treatment program in Western Kenya. Mexico City: International AIDS