Antiretroviral therapy in antenatal care to increase treatment initiation in HIV-infected pregnant women: a stepped-wedge evaluation : AIDS

Secondary Logo

Journal Logo


Antiretroviral therapy in antenatal care to increase treatment initiation in HIV-infected pregnant women: a stepped-wedge evaluation

Killam, William Pa,b; Tambatamba, Bushimbwa Cc; Chintu, Namwingab; Rouse, Dwighta; Stringer, Elizabetha,b; Bweupe, Maximillianc; Yu, Yonga,b; Stringer, Jeffrey SAa,b

Author Information
AIDS 24(1):p 85-91, January 2, 2010. | DOI: 10.1097/QAD.0b013e32833298be
  • Free



Zambia and other sub-Saharan African countries have made tremendous progress in expanding antenatal HIV testing of pregnant women and provision of antiretroviral prophylaxis to reduce mother-to-child transmission of HIV. Although antiretroviral prophylaxis regimens, such as intrapartum single-dose nevirapine (sd NVP) to the mother and infant, effectively reduce perinatal HIV transmission, the WHO and others recommend that HIV-infected pregnant women who meet eligibility criteria for antiretroviral therapy (ART) be offered triple drug combination regimens not only to protect their infants from HIV infection but also to reduce maternal morbidity and mortality [1–6]. However, our experience in Lusaka, Zambia, resembles that of other settings in the region: the mere presence of ART services in the public sector has not been sufficient to ensure that all eligible pregnant women avail themselves of it [7,8]. Until recently in Lusaka, HIV-infected mothers have been referred to separate ART services provision of triple therapy if eligible. With this model of care, only a tiny proportion (<3%) of HIV-infected women actually initiated ART during pregnancy.

In October 2007, in an effort to increase utilization of ART among pregnant women, we implemented a program of ART provision integrated into the Lusaka district public sector antenatal care (ANC) clinics (ART in ANC). The implementation was phased, one clinic at a time, and thus provided an opportunity to evaluate in a controlled manner whether the new strategy of ART in ANC would result in a greater proportion of treatment-eligible women initiating ART during pregnancy, compared with the existing strategy of referral for ART care.


Standard of care

The public sector ANC clinics in the Lusaka urban district are funded by the Zambian Ministry of Health and receive technical support from the Centre for Infectious Disease Research in Zambia. All Lusaka public ANC clinics used a ‘reflex CD4+’ testing and referral strategy for linking prevention of mother-to-child HIV transmission (PMTCT) and ART services. With this approach, women found to be HIV seropositive through antenatal testing had a specimen routinely sent for a CD4+ cell count. Post-test counseling stressed the importance of returning for the CD4+ results within 2 weeks as well as the benefits of ART if the woman was found to be eligible for therapy. Women in whom advanced HIV disease was suspected based on a WHO symptoms screen (simplified for use by midwives and other ANC personnel) and those whose CD4+ cell count was less than or equal to 350 cells/μl were referred urgently to the ART clinic. The ART clinic in each site was located on the same premises as the antenatal clinic, but physically separate and separately staffed. Local peer educators provided additional education and support to women who qualified for ART and were asked to escort them to the ART clinic. Women with a negative WHO symptoms screen and whose CD4+ count was more than 350 cells/μl were provided antiretroviral prophylaxis for PMTCT and a nonurgent appointment to the ART clinic for long-term care and follow-up.


Between October 2007 and May 2008, one new site per month upgraded its services to provide ART in the ANC clinic; reflex CD4+ testing remained in place at all sites throughout the evaluation period. Integration of ART in ANC was planned in the eight Lusaka district clinics with the largest volume of HIV-infected antenatal women. The eight sites were matched into four pairs based on the number of HIV-infected pregnant women expected in each site. We then rolled out the intervention to one member of the least busy pair, one member of the second busiest pair, one member of the third busiest pair, and one member of the busiest pair. Rollout to the remaining pairs proceeded in reverse order (i.e., the busiest next, ending with the least busy remaining site). ART services in ANC were provided 1–2 days per week, depending on the expected number of treatment-eligible women (predicted range 4–22 per month) in the clinic. Women already receiving ART at the general ART clinic were encouraged to continue receiving their services in the general ART clinic.

If the patient's CD4+ cell count was less than or equal to 250 cells/μl, she was considered eligible for ART and enrolled into ART care on the day she returned to receive her CD4 results. Standard written protocols and a team approach were used. During the enrollment visit, the clinical officer performed a detailed history and physical examination, WHO staging, and treatment of any opportunistic infections if present. A nurse midwife provided health education and antenatal services. Counseling on antiretroviral drugs, including the need for lifelong adherence, was provided by the peer educator and any other member of the team. At the enrollment visit, patients were started on cotrimoxazole prophylaxis, multivitamins, and iron and were asked to return in 2 weeks for initiation of ART. However, if the patient was late in gestation (i.e., 34–36 weeks estimated gestational age), ART initiation was usually, at the discretion of the treating clinician, recommended at the enrollment visit. If the patient's CD4+ cell count was more than 250 cells/μl, she was referred to the general ART clinic for enrollment, where, although ART might not be initiated immediately, she would be followed for long-term monitoring, care, and support.

Both the general and ANC-integrated ART clinics used the same schedule of visits, laboratory evaluations, record systems, and quality assurance systems. They were staffed by the same cadres of providers: a clinical officer, a nurse, and a peer educator. Nurses and clinical officers staffing both the general and integrated ART in ANC clinics received a ministry-approved ART training. All patients initiating treatment but not returning for subsequent visits were actively followed-up with phone calls or home visits. Women received ART in the ANC clinic until approximately 6 weeks postpartum and then were referred to the general ART clinic. At approximately 6 weeks postpartum, infant cotrimoxazol prophylaxis and infant testing for HIV DNA were recommended.

We rolled out the ART in ANC integration in a stepped-wedge fashion (Fig. 1). In this design, all participating clinics began collecting data simultaneously, while providing the standard of care intervention (referral to ART services). Then, one-by-one, each clinic crossed over to the intervention (ART in ANC). This design allowed each clinic to provide patients to both arms of the comparison (i.e., each clinic acted as its own control), while also controlling for time trends by allowing contemporaneous comparisons across clinics at fixed points in time [9]. After the programmatic implementation, we retrospectively reviewed the existing, routinely collected data to ascertain our study outcomes. The study population included all women initiating ANC in the eight clinic sites between 16 July 2007 and 31 July 2008 and found to be eligible for ART. Women already on ART were excluded. We also excluded women who initiated ANC at other clinic sites in an effort to mitigate the effects of contamination [10].

Fig. 1:
Timing of the intervention rollout in a stepped-wedge evaluation of two strategies to enroll HIV-infected women into antiretroviral therapy, Lusaka, Zambia July 2007 to 2008.

The CD4+ cell count threshold for ART eligibility in the Zambian national PMTCT guidelines changed from 250 to 350 cells/μl during the study period [11]. As women with a CD4+ cell count less than or equal to 250 cells/μl were considered eligible throughout the entire study period in all sites, and because our study protocol identified the 250 cells/μl threshold as the study's primary outcome, we used this value to define ART eligible in this study. We defined ART initiation as having collected from the pharmacy a combination regimen of at least three antiretroviral medications. Enrollment into ART was defined as having a documented enrollment visit described above, including a complete history and physical examination, WHO staging, and counseling.

Because clinic sites implemented the ART in ANC services on a discrete date, women might be identified as ART eligible in the control period but actually enroll into and initiate ART in the intervention period. To deal with the ambiguity of this transition period, we excluded from analysis any patient with an initial antenatal visit more than 60 days prior to the date of implementing the ART in ANC intervention. Furthermore, for analysis of the primary outcome, we required enrollment into ART to occur within 60 days of the first ANC visit, thus categorizing patients into three mutually exclusive categories: a referral to ART cohort, an integrated ART in ANC cohort, and a transition cohort. The study's primary aim is to compare ART enrollment and initiation rates between the first two categories. The primary outcome was defined as the proportion of ART eligible women (HIV seropositive and CD4+ cell count ≤250 cells/μl) enrolling within 60 days of the CD4+ cell count and initiating ART before delivery [or estimated date of delivery (EDD) for those in whom delivery data were not available].

As starting ART earlier in pregnancy is associated with lower rates of perinatal transmission, we investigated timeliness of ART initiation in two ways: the average gestational age when ART was initiated and the average number of weeks ART was received prior to delivery. Patient retention was defined as the number of women initiated on treatment and still on treatment (not lost to follow-up, dead, or transferred out) in the first 90 days of treatment.

We collected study data through our electronic medical record systems used in the Lusaka public sector. Antenatal care data on all patients initiating ANC at all study sites were recorded in the Zambia Electronic Perinatal Record System (ZEPRS) [12], whereas the ART treatment data on all patients enrolled and initiated on ART within the Lusaka public health clinics were captured in a separate electronic record, called SmartCare [13]. The electronic systems are linked to each other through a common identification number and are fully integrated into care.

The analysis of this pre-planned stepped-wedge study was guided by the methodology of Hughes et al.[14] working with our group and later revised [10]. Crude odds ratios and 95% confidence intervals (CIs) were calculated for enrollment and initiation proportions using the referral to ART cohort as the referent value. Adjusted odds ratios (AORs) and their CIs were generated with generalized estimating equations to account for the clinical site cluster and time effects. Categorical variables were compared using the chi-squared test and continuous variables were compared using two-sample t-test and Wilcoxon test for mean estimation. For analysis of patient retention, a weighted t-test for cluster-level comparison was performed. A P value less than 0.05 was considered statistically significant. Analysis was performed using SAS statistical software version 9.1 (SAS Institute Inc., Cary, North Carolina, USA). This protocol underwent continuing ethical review by the relevant authorities at the University of Alabama-Birmingham and the University of Zambia.


Between 16 July 2007 and 31 July 2008, a total of 37 203 patients initiated antenatal care in the eight study clinics. Of these women, 5667 initiated ANC in the 60 days prior to the intervention start date and were thus considered members of the transition cohort; 13 917 initiated ANC more than 60 days before the intervention rollout and were classified as members of the referral to ART (control) cohort; and 17 619 initiated ANC after the intervention start date and thus were considered members of the integrated ART in ANC (intervention) cohort (Fig. 2). The control and intervention cohorts were similar with respect to their proportion tested for HIV, HIV seroprevalence, proportion tested for CD4+ cell count, and proportion with a CD4+ cell count less than 250 cells/μl (Table 1). Likewise, patients found eligible for ART in the two cohorts were comparable in terms of age, weight, BMI, education level, marital status, parity, RPR test status, mean initial CD4+ cell count, and mean initial hemoglobin level (Table 2).

Fig. 2:
Flow diagram demonstrating testing and outcome cascade by patient cohort.
Table 1:
HIV testing characteristics of patients attending antenatal care by strategy cohort.
Table 2:
Characteristics of patients eligible for ART by strategy cohort.

A greater proportion of ART-eligible women enrolled into ART within the outcome period (60 days of starting ANC and prior to delivery or EDD) (AOR 2.06, 95% CI 1.27–3.34) and initiated ART (AOR 2.01, 95% CI 1.37–2.95) in the intervention compared with the control cohort (Table 3). Excluding the ‘within 60 days of HIV diagnosis’ criterion, 52.2% (442/846) of eligible women in the intervention cohort enrolled into ART before delivery or EDD and 38.4% (325/846) initiated ART before delivery or EDD.

Table 3:
Enrollment and ART initiation outcome of ART eligible patients by strategy cohort.

Among women who initiated ART, the mean gestational age of ART initiation and duration of ART use in pregnancy did not differ significantly between the two cohorts. The mean estimated gestational age (EGA) at first ANC visit was approximately 22 weeks in both cohorts (22.2 weeks in the intervention arm and 21.7 weeks in the control arm; P = 0.2). The mean gestational age at ART initiation (27.7 versus 27.1 weeks EGA, P = 0.4) and mean weeks of ART initiation before delivery (10.0 versus 10.8 weeks, P = 0.3) were not significantly different in the intervention versus control cohorts.

Of the patients who initiated ART, the 90-day retention was 87.8% (244/278) in the intervention and 91.3% (94/103) in the control cohort, but was not statistically different when analyzed on the individual level (P = 0.3) or by weighted t-test for cluster-level comparison (P = 0.2). Reasons for nonretention on ART were loss to follow-up, transfer out of the district, and death (10.8, 0.7, and 0.07%, respectively, in the intervention and 6.8, 1.9, and 0% in the control cohorts).


In this large, public sector cohort in Zambia, a strategy of providing ART integrated in the antenatal clinic doubled the percentage of treatment-eligible pregnant women initiating ART prior to delivery compared with a strategy of active referral to the ART clinic. The integrated ART in ANC strategy did not affect the timeliness of ART initiation (i.e., it did not get women onto ART earlier in pregnancy); however, the average length on ART before delivery in both cohorts was at least 10 weeks. Viral load suppression and risk of perinatal transmission of HIV are related to length of therapy [15], and in a similar African setting, use of nevirapine-based regimens for greater than 7 weeks prior to delivery has been associated with an extremely low (<1%) risk of perinatal transmission [7]. The strengths of this study are that it included a large cohort of antenatal attendees assessed for ART eligibility across eight busy, public sector PMTCT program sites. Because of our integrated electronic patient record systems, we were able to capture comprehensive information related to a patient's ART eligibility, enrollment into care, initiation of ART, and retention. The stepped-wedge rollout of the intervention allowed a controlled evaluation, unbiased by time trends, while ultimately allowing all sites to participate in the enhanced ART in ANC intervention.

A weakness of the study is that we did not directly report incidence of infant HIV infection or HIV-free survival. However, other studies have consistently demonstrated that providing ART to pregnant women is safe, well tolerated, and extremely effective in preventing infant HIV infection [3,7,8,16,17]. In fact, the major challenge in PMTCT is not the safety or efficacy of ART, but instead extending coverage to eligible women [4,18]. Toward that end, the focus of this study was to identify strategies to maximize ART provision to eligible pregnant women. Although most women initiating therapy took it at least until their delivery, we did experience losses to follow-up as noted in similar settings [7,17]. It is possible that by encouraging patients who feel well to initiate ART, the long-term retention after delivery was reduced. On the basis of these findings, we have made a programmatic decision to keep women in the integrated clinic until weaning at approximately 6 months postpartum.

We hypothesize several explanations why providing ART in ANC was more effective than referral to ART services. Pregnant women may wish to avoid the inconvenience of enrolling into often crowded ART clinics and the high visit burden associated with separate ART and antenatal care. Staff attitudes have also been reported as a barrier to mothers receiving PMTCT [19,20]. It is possible that the antenatal clinic staff took more ownership and initiative in counseling and following eligible patients when ART provision was integrated into ANC. The integrated program also provided greater focus and interest in providing ART to pregnant women. Although we successfully initiated 38% of eligible pregnant women on ART before delivery, there were 62% who did not initiate during pregnancy. Future studies will try to elucidate reasons for their not accessing ART and will target strategies to improve uptake further.

This study demonstrates that provision of ART in antenatal care is feasible in resource-limited settings, although it may involve greater investment in laboratory capacity, drugs, and staff [21,22]. The cost and human resources involved in implementation of these strategies are areas for future analysis; however, in our setting, we have committed to deploying this strategy to other district clinics, believing it to be an essential step along the pathway to our ultimate goals of eradicating pediatric HIV and promoting maternal health.


We wish to thank Ms Katie Boos for her noteworthy assistance in the programmatic rollout. We also thank Mrs Emelda Kabwe, Mr Justin Msonda, Mr David Phiri, and Mr Joseph Mwale for their contributions to making the programmatic intervention a success.

The funds for the study were received from the Doris Duke Charitable Foundation (ORACTA grant 2007016), and the University of Alabama/Centre for Infectious Disease Research in Zambia.

The Doris Duke Charitable Foundation provided funding but was not otherwise involved in study planning, implementation, data collection, or data interpretation. The decision to submit this manuscript for publication was solely that of the authors.


1. World Health Organization. Antiretroviral drugs for treating pregnant women and preventing HIV infection in infants in resource-limited settings: towards universal access. Geneva: WHO; 2006.
2. U.S. Public Health Service Task Force. Recommendations for use of antiretroviral drugs in pregnant HIV-1-infected women for maternal health and interventions to reduce perinatal HIV-1 transmission to reduce perinatal HIV-1 transmission in the United States. Updated June 23, 2004. [Accessed 10 December 2004]
3. Cooper ER, Charurat M, Mofenson L, Hanson IC, Pitt J, Diaz C, et al. Combination antiretroviral strategies for the treatment of pregnant HIV-1-infected women and prevention of perinatal HIV-1 transmission. J Acquir Immune Defic Syndr 2002; 29:484–494.
4. De Cock KM, Fowler MG, Mercier E, de Vincenzi I, Saba J, Hoff E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice. JAMA 2000; 283:1175–1182.
5. Colebunders R, Kolsteren P, Ryder R. Giving antiretrovirals in the peripartum period to prevent mother-to-child HIV transmission in low-income countries: only a short-term stopgap measure [editorial]. Trop Med Int Health 2003; 8:375–377.
6. Maclean CC, Stringer JS. Potential cost-effectiveness of maternal and infant antiretroviral interventions to prevent mother-to-child transmission during breast-feeding. J Acquir Immune Defic Syndr 2005; 38:570–577.
7. Black V, Hoffman RM, Sugar CA, Menon P, Venter F, Currier JS, et al. Safety and efficacy of initiating highly active antiretroviral therapy in an integrated antenatal and HIV clinic in Johannesburg, South Africa. J Acquir Immune Defic Syndr 2008; 49:276–281.
8. van der Merwe K, Chersich MF, Technau K, Umurungi Y, Conradie F, Coovadia A. Integration of antiretroviral treatment within antenatal care in Gauteng Province, South Africa. J Acquir Immune Defic Syndr 2006; 43:577–581.
9. Brown CA, Lilford RJ. The stepped wedge trial design: a systematic review. BMC Med Res Methodol 2006; 6:54.
10. Hussey MA, Hughes JP. Design and analysis of stepped wedge cluster randomized trials. Contemp Clin Trials 2007; 28:182–191.
11. Ministry of Health Zambia. National Protocol Guidelines for Integrated Prevention of Mother-to-Child Transmission of HIV/AIDS. Vol. 53. Printec Press; 2007.
12. Kelley C. The ZEPRS Project. In Cressman G, editor. Raleigh, NC, USA: Research Triangle Institute; 2006. [this Web site provides documentation about the ZEPRS project and system]
13. 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.
14. Hughes JP, Goldenberg RL, Wilfert CM, Valentine M, Kasonde G, Guay LA, et al. Design of the HIV Prevention Trials Network (HPTN) Protocol 054: A cluster randomized crossover trial to evaluate combined access to nevirapine in developing countries. UW Biostatistics Working Paper Series. Seattle, WA: University of Washington; 2003. p. 15.
15. Patel D, Cortina-Borja M, Thorne C, Newell ML. Time to undetectable viral load after highly active antiretroviral therapy initiation among HIV-infected pregnant women. Clin Infect Dis 2007; 44:1647–1656.
16. Watts DH, Balasubramanian R, Maupin RT Jr, Delke I, Dorenbaum A, Fiore S, et al. Maternal toxicity and pregnancy complications in human immunodeficiency virus-infected women receiving antiretroviral therapy: PACTG 316. Am J Obstet Gynecol 2004; 190:506–516.
17. Marazzi CM, Germano P, Liotta G, Guidotti G, Loureiro S, Gomes Ada C, et al. Implementing antiretroviral triple therapy to prevent HIV mother-to-child transmission: a public health approach in resource-limited settings. Eur J Pediatr 2007; 166:1305–1307.
18. Luzuriaga K, Sullivan JL. Prevention of mother-to-child transmission of HIV infection. Clin Infect Dis 2005; 40:466–467.
19. Painter TM, Diaby KL, Matia DM, Lin LS, Sibailly TS, Kouassi MK, et al. Women's reasons for not participating in follow up visits before starting short course antiretroviral prophylaxis for prevention of mother to child transmission of HIV: qualitative interview study. BMJ 2004; 329:543.
20. Chi BH, Chansa K, Gardner MO, Sangi-Haghpeykar H, Goldenberg RL, Sinkala M, et al. Perceptions toward HIV, HIV screening, and the use of antiretroviral medications: a survey of maternity-based healthcare providers in Zambia. Int J STD AIDS 2004; 15:685–690.
21. Chi BH, Fusco H, Sinkala M, Goldenberg RL, Stringer JS. Cost and enrollment implications of targeting different source population for an HIV treatment program. J Acquir Immune Defic Syndr 2005; 40:350–355.
22. Chi BH, Sinkala M, Stringer EM, McFarlane Y, Ng'uni C, Myzece E, et al. Employment of off-duty staff: a strategy to meet the human resource needs for a large PMTCT program in Zambia. J Acquir Immune Defic Syndr 2005; 40:381–382.

antiretroviral treatment; ART; HIV; mother-to-child transmission; pregnancy; resource-limited countries

© 2010 Lippincott Williams & Wilkins, Inc.