The prevention of mother-to-child HIV transmission programme in Lilongwe, Malawi: why do so many women drop out

Abstract

Mother-to-child transmission of HIV constitutes a substantial burden of new HIV infections in sub-Saharan Africa, and losses to follow-up continue to undermine prevention of mother-to-child transmission of HIV (PMTCT) programmes. This qualitative study sought to clarify why some women who were enrolled in a PMTCT programme in Lilongwe, Malawi, did not fully participate in follow-up visits in the first six months after testing HIV-positive. Twenty-eight women, 14 who participated fully in the programme and 14 who dropped out, were purposively selected for in-depth interview at two clinics. Focus group discussions with 15 previously interviewed and 13 newly recruited women were also conducted. Discussions with 12 of the women's husbands were also carried out. Although the proportion of women being tested has reportedly increased, losses to follow-up have shifted and are occurring at every step after testing. Major emerging themes associated with dropping out of the PMTCT programme within six months after delivery were to avoid involuntary HIV disclosure and negative community reactions, unequal gender relations, difficulties accessing care and treatment, and lack of support from husbands. The whole approach to the delivery of the PMTCT programme and home visits must be reconsidered, to improve confidentiality and minimise stigmatization. Women need to be empowered economically and supported to access HIV treatment and care with their partners, to benefit their whole family.

Résumé

La transmission mère-enfant du VIH représente une part importante des nouvelles infections en Afrique subsaharienne et les pertes du suivi continuent de saper les programmes de prévention de la transmission mère-enfant (PTME). Cette étude qualitative souhaitait déterminer pourquoi certaines participantes d'un programme de PTME à Lilongwe, Malawi, avaient manqué des visites de suivi les six premiers mois après avoir découvert leur séropositivité lors d'un dépistage. Vingt-huit femmes – 14 participant pleinement au programme et 14 l'ayant abandonné – ont fait l'objet d'un entretien approfondi dans deux dispensaires. Des discussions en groupe avec 15 femmes interrogées précédemment et 13 nouvelles recrues ont aussi été organisées, de même qu'avec 12 conjoints des femmes. Bien que la proportion déclarée de femmes se faisant dépister ait augmenté, les pertes pour le suivi ont évolué et se produisent désormais à toutes les étapes après le test. Les principales raisons de l'abandon du programme de PTME dans les six mois après l'accouchement étaient la crainte d'une révélation involontaire de la séropositivité et des réactions communautaires négatives, le manque de parité entre les sexes, les difficultés d'accès aux soins et au traitement, et le soutien insuffisant des maris. Toute l'approche de l'application du programme de PTME et des visites à domicile doit être revue, pour améliorer la confidentialité et minimiser la stigmatisation. Il faut donner aux femmes les moyens économiques de s'autonomiser et les aider à avoir accès au traitement et aux soins du VIH avec leur partenaire et leurs enfants ce qui profitera à toute la famille.

Resumen

En Ãfrica subsahariana, gran parte de nuevas infecciones por VIH se deben a la transmisión materno-infantil, y las pérdidas de seguimiento continúan socavando los programas de prevención de la transmisión materno-infantil del VIH (PTMI). El objetivo de este estudio cualitativo era aclarar por qué algunas de las mujeres inscritas en un programa de PTMI en Lilongwe, Malaui, no participaron al máximo en las consultas de control durante los primeros seis meses posteriores a haber sido diagnosticadas VIH-positivas. Se seleccionaron 28 mujeres para entrevistas a profundidad en dos clínicas: 14 participaron al máximo en el programa y 14 lo abandonaron. Además, se realizaron discusiones en grupos focales con 15 mujeres entrevistadas anteriormente y 13 recién reclutadas, así como discusiones con 12 de los esposos. Aunque la proporción de mujeres que se someten a la prueba del VIH ha aumentado, las pérdidas de seguimiento continúan ocurriendo en cada paso posterior a las pruebas. Los principales motivos emergentes asociados con abandonar el programa de PTMI dentro de los seis meses posteriores al parto fueron: para evitar la divulgación involuntaria del estado de VIH y reacciones negativas de la comunidad, relaciones desiguales entre los sexos, dificultades accediendo a los servicios y el tratamiento, y la falta de apoyo del esposo. Para mejorar la confidencialidad y minimizar el estigma, es necesario reconsiderar la estrategia de ejecución del programa y las visitas a domicilio. Las mujeres necesitan autonomía económica y apoyo para poder acceder al tratamiento y los servicios de VIH con sus parejas, a fin de beneficiar a toda su familia.

Key Words:

• HIV prevention of mother-to-child transmission
• HIV treatment programmes
• stigma
• provider–patient relations
• gender roles and issues
• Malawi

Approximately 370,000 children under 15 years became HIV-infected in 2007 contributing to nearly two million children living with HIV that year. Most of them became infected during pregnancy, birth or breastfeeding; 90% live in sub-Saharan Africa.¹ Antiretroviral prophylaxis has minimised mother-to-child transmission of HIV in developed countries. In most developing settings, initial efforts focused primarily on the use of single-dose nevirapine for the mother at delivery and the child at birth, infant feeding alternatives where safe, or exclusive breastfeeding in the first six months of life.² Multiple drug therapy is slowly replacing single-dose nevirapine. In 2007, of the 60 countries with disaggregated data on antiretrovirals for prevention of mother-to-child transmission (PMTCT) of HIV, 29% and 8% of HIV-positive women received dual and triple therapy, respectively.³ However, losses to follow-up from delivery and beyond have plagued these programmes.

Cumulative losses in sub-Saharan African PMTCT programmes, including in Malawi, Côte d'Ivoire and South Africa, range from 28% during antenatal care, up to 70% four months after delivery and close to 81% by six months post-partum.^4–6 HIV-related stigma, negative experiences of women while interacting with programme staff and lack of support from partners are some documented constraints. Not much is reported from Malawi on the perceptions and experiences of women who successfully participate and those who do not.

Prevention of mother-to-child transmission of HIV in Malawi

Malawi, with 13.6 million people, has an HIV seroprevalence of 15% among pregnant women. In 2007, 91,000 children under 15 years had HIV.³ Nevirapine-based PMTCT interventions were initiated in 2002 and by early 2008, 73% of the 542 maternal and child health facilities were providing these services.⁷ Of the 575,000 women who were pregnant in 2007, over 90% attended antenatal care at least once but only 49% of the total received HIV counselling and testing, of whom 24,653 (8.8%) tested positive. Of these, 18,983 (77%) women and 12,039 of their exposed infants received nevirapine.⁸ Limited coordination of adult and paediatric HIV care and treatment programmes resulted in non-continuation of care, as many found the process too cumbersome.

PMTCT sites in Lilongwe

The University of North Carolina (UNC) project began providing PMTCT services in Lilongwe in 2002 as a vertical programme at four district government health centres. Technical assistance to 12 other centres in the district was started in 2004. Before April 2005, an opt-in testing approach was used in the four centres but less than 80% of pregnant women attending antenatal care accepted an HIV test. After April 2005, opt-out testing was instituted and nearly 100% of women are now being tested. Currently, the UNC project tests over 20,000 pregnant women annually and in 2007, it reached 9% of pregnant women countrywide.⁹

PMTCT interventions since July 2006 have included HIV education, group counselling and rapid testing, individual post-test counselling, clinical staging and CD4 cell count testing. Those with CD4 cell counts below 250 or who are in WHO clinical stage III or IV are referred for antiretroviral therapy. Those above 250 or in clinical stage I or II, get zidovudine prophylaxis from 28 weeks gestation, with monthly visits for zidovudine supply until delivery, and single-dose nevirapine to take when labour starts, even in cases of home births. Between post-test and delivery, women receive routine antenatal care, including cotrimoxazole prophylaxis for opportunistic infections and treatment for minor illnesses, and they are referred to the outpatient department or tertiary medical facilities as needed. The programme continues to follow HIV-positive women and their infants monthly from six weeks after birth until the infant is 18 months old. During this time, women continue to receive infant feeding support and risk reduction counselling, including for family planning and dual protection. Their CD4 counts are taken and clinical staging repeated every six months. They are referred for antiretroviral therapy when indicated and are linked to support groups. Their children are tested for HIV at six weeks after birth and again at 18 months. Children testing positive get referred for paediatric care while those who are negative are discharged. It is assumed that all HIV-positive women agree to attend these visits.

In 2004, 20% of UNC programme participants missed at least one of their follow-up visits in the period from post-test until they exited the programme at 18 months post delivery. Only 1,721 of 3,405 (51%) women testing positive in 2007 delivered in a PMTCT facility, of whom 88% of the women and 1,700 of their exposed infants actually took nevirapine.^9,10 Many women were suspected not to have disclosed their HIV status during labour, and might not have been documented as positive in records. UNC and the Ministry of Health requested this study to clarify why some women did not fully attend follow-up visits in the PMTCT programme in the first six months after getting a positive diagnosis.

Methods

We conducted in-depth interviews between 1 July and 31 December 2005 with a sub-sample of women from a larger prospective cohort study at two of the four UNC-supported clinics in Lilongwe. Data were collected in the first six months after they tested HIV-positive. The clinics were selected to reflect different socio-economic profiles, rural-urban settings and size of patient load, factors that have been shown to affect women's PMTCT participation in Côte d'Ivoire and South Africa.^6,11,12 Site A, semi-urban, served over 174,000 people, provided antenatal care to 11.5% of pregnant women in Lilongwe and conducted approximately 3,000 deliveries annually. Site B, rural, served over 100,000 people and provided antenatal care to 6.5% of pregnant women in Lilongwe.

Research staff, who were also health care providers, initially contacted all programme enrollees post-HIV test. Those who accepted to attend follow-up visits to the programme were invited to join the study. Furthermore, they were asked to consent to home visits if they had missed a scheduled programme visit and not attended within three weeks of the appointment date. Their records were then monitored by visit for six months. If a woman failed to attend, she was classified as a drop-out and a home visit was scheduled for an individual interview. Matching participants by appointment visit were recruited through random selection of their records. Recruitment continued until we enrolled those we could realistically manage. After concluding individual interviews, focus group discussions were arranged with previously interviewed and newly recruited women at later dates, to enrich and validate obtained information.

Pregnant women met inclusion criteria if they were 18–49 years old, were HIV-positive, had joined the PMTCT programme at any of the study clinics and were staying within the UNC project's catchment area. We interviewed 28 women, including seven participants of follow-up visits and seven drop-outs at each of the four sites. Four discussion sessions involving 28 women, six to nine per group, were also conducted. Fifteen of the 28 had previously participated in individual interviews, of whom nine were from site A (four participants and five drop-outs) and six from site B (three participants and three drop-outs). The other 13 were recruited through monthly support group meetings at the clinics (two participants from site A and six from site B and one drop-out from site A and four from site B). All invited women agreed to participate.

Husbands of women attending services at site A, who had participated in the study and who said they had disclosed their status, were invited to participate through a letter sent via their wives. Of 15 men invited, 12 consented.

Women were asked how their lifestyles had changed after their diagnosis and joining the programme, what responses they had from their family and community and what they perceived as barriers or facilitators to further participation. Husbands were asked questions to validate what their wives had said about them. The study was approved by the institutional review boards of the Malawi National Health Sciences Committee and University of Oslo, Norway.

All participants were interviewed in the local language, Chichewa, in a private place at the study clinics. Drop-outs were interviewed at home by two researchers at times that coincided with programme processes. Women dropped out at various stages, including at labour and delivery, six weeks post-partum or afterwards in the 18-month period. Many of the women who dropped out began missing the monthly visits when their babies were 3–6 months old.

Interviews and discussions were conducted using semi-structured open-ended question guides. They lasted between 45 minutes and an hour and were audio-taped. All recorded information was translated and transcribed into English. The transcripts were checked to identify issues requiring clarification during subsequent interviews. To maintain confidentiality, interview and discussion transcripts were labelled with identifying numbers. Pseudonyms were used on all records and in this paper.

Two researchers independently analyzed the transcripts, identifying key categories and recurrent themes. Sections of text were marked manually and linked to similar sections in other interviews and discussions. This process continued until no new themes or ideas emerged. Emerging themes were jointly reviewed to reach consensus on data interpretation.

Findings

The mean age of the entire PMTCT cohort was 28 years. Participants in follow-up visits were slightly older than drop-outs (32 years vs. 30 years) and had on average three children against four for drop-outs. Over a third (39.3%) were not literate in both groups, lower than the 69% national average.¹³ All women were married and exclusively housewives except two, who were in informal employment. All attended antenatal care at least once and said they had disclosed their positive status to their husbands. All but one woman dropped out prior to delivery.

Major emerging themes associated with dropping out of the PMTCT programme within the first six months after delivery were fear of involuntary HIV disclosure and negative community reactions, unequal gender relations, difficulty accessing services, including long walking distances, and lack of support from husbands. Social support, on the other hand, facilitated participation.

Feat of involuntary HIV disclosure and negative community reactions

Twenty of the 28 women interviewed expressed concern that others would learn about their HIV-positive status if they participated in the programme, received home visits from clinic staff, stopped breastfeeding at six months, and used or even possessed breastmilk substitutes. In some instances, they also worried that signs like weight loss would be tell-tale.

“My friends have been asking me why I am not gaining weight after delivery. They say they thought it was the pregnancy that was eating up my weight. ‘Are you not eating? It should not be AIDS’.” (Bertha, interview, dropped out 3 months post-partum)

“I have lost several of my friends because of weight loss after HIV-positive diagnosis.” (Annie, interview, dropped out 1 week post-partum)

Although HIV-positive women are given several infant feeding options, the most frequently promoted is exclusive breastfeeding for the first six months.¹⁴ Stopping breastfeeding early created or confirmed suspicions about the woman's HIV status.

“My neighbours have been asking me why I stopped breastfeeding such a small child. ‘Have you been found with HIV?’” (Monica, interview, dropped out 6 months post-partum)

“I am already having fear about what people might say when I stop breastfeeding my child at six months. What will I tell them?” (Mable, 3-month-old baby, interview, fully participated)

Maternal food baskets containing soya flour, donated by the World Food Programme for use only by HIV-positive women and offered in the first year after delivery, were also a potential source of disclosure.

“One woman asked me why I receive soya. I told her that I am sick. She asked me what disease I was suffering from. After telling her that I was weak, she said: ‘No, soya is given to people who have got AIDS’.” (Bika, interview, dropped out 5 months post-partum)

Women's opinions of home visits by programme staff, whose aim was to encourage women who had dropped out to continue, differed. To some, they were seen as a source of potential HIV disclosure. Two of the 14 drop-outs said that to avoid these visits, they would restart going for programme visits. Others were more positive.

“The programme nurse visited me and two boys followed her. When she left, the boys started telling my neighbours that the nurse visits those with AIDS. The whole community is talking about me and my neighbours are disclosing to all my visitors.” (Charity, interview, dropped out 4 months post-partum)

“I will ensure that I come for my appointments because I do not want to be home visited.” (Chikondi, 2-month-old baby, interview, fully participated).

“The visits are made to encourage us to be coming back to the clinic. It is your [health worker's] duty to take care of our lives. As for me, I will from now onwards continue attending my appointments.” (Bessie, interview, dropped out 5 months post-partum)

In some cases, the physical layout of the service delivery site caused difficulties. The PMTCT clinic rooms at site B are situated apart from other units physically and women waiting to be seen must wait outside.

“…Everything is discussed on the benches outside within the hearing of our colleagues attending antenatal care nearby. This is not good because our friends ask us why we get services different from theirs.” (Chandilanga, interview, dropped out 3 months post-partum)

Unequal gender relations

Eighteen of the 28 women interviewed said that their partner's refusal to disclose his HIV status or to be tested made them suspect that they could be positive also. If he refused to use condoms it was a cause for concern because they would not be protected. Of the 14 drop-outs, nine said they had done so because of their husband's refusal to engage in protective behaviour. Many felt unable to negotiate condom use, which they feared potentially threatened their marriage.

“Even when I tell my husband about HIV risks, he does not listen to me. As you know, men do not listen to what women say. If my husband had agreed to use condoms, maybe I would be protected, but as things are, it is difficult.” (Victoria, interview, dropped out 3 months post-partum)

“My husband refuses condoms. He says he is already dead. I may be against having unprotected sex, but as a woman, this does not carry any weight.” (Monica, interview, dropped out 6 months post-partum)

“If you insist on condom use, the argument may reach the marriage counsellor who might say there is something you are doing behind his back. To avoid these confrontations, you just accept unprotected sex.” (Grace, 6-month-old baby, interview, fully participated)

“Condom issues are difficult. We know we can prolong our lives if we do not infect each other. On the other hand, marriage is also important. Everything is equally important. We cannot survive without men. Who will help us meet our needs?” (Focus group discussion 1, drop-out, site B)

“We are trying hard, but our husbands are pulling us down. We come to the clinic to get condoms but they refuse to use them. We therefore get disappointed, and probably that is why some of our friends drop out.” (Grace, 6-month-old baby, interview, fully participated)

Difficulties engaging men in protective behaviour

The men interviewed acknowledged their refusal to use condoms and their preference for unprotected sex, which they described as a sign of love. They blamed the health workers for disseminating conflicting HIV messages, and also said their wives did not tell them much about what happened at the programme, which they attributed to fear.

“You (health workers) have been saying that HIV/AIDS is a killer and it has no cure. You now tell us that we can live longer with HIV. What advice should we follow?” (Focus group discussion 2, men, site A)

Only two of the 14 drop-outs compared to eight of the 14 programme participants reported positive reactions on the part of their husbands when they disclosed their positive HIV status.

“He just kept quiet and next morning he told me he was going to work but he never came back. He left me when I was six months pregnant and there was no one to help me. I decided to go to my mother and went back to my home a month after delivery. My husband arrived three days ago and is still not talking to me.” (Victoria, interview, dropped out 3 months post-partum)

“When I disclosed to my husband, he didn't believe it. He just started laughing and told me that I would die within that year.” (Misonzi, interview, dropped out 3 months post-partum)

Six drop-outs said their husbands assumed that if their wives tested positive, they were also positive. As a result, some men felt there was no need to stop being promiscuous to avoid losing both sexual pleasure and life. One woman described her husband's reaction when she confronted him over promiscuity, which she argued brought HIV into their family:

“If you want life, go back to your parents' home. As for me, I am already dead. I do not want to lose out on two things: sexual enjoyment with my girlfriend and life.” (Taona, interview, dropped out 2 months post-partum)

Other husbands of drop-outs did not believe the test results and told their wives to stop attending the programme. Responses of husbands of women who fully participated in programme vists were by contrast supportive:

“If this [positive serostatus] is what hospital personnel have said, let us just accept, live positively and follow their advice.” (Dolibe, interview, 2-month-old child, fully participated)

“This is our problem; let us continue staying normally.” (Chrissie, interview, 2-month-old child, fully participated)

Four husbands assured their wives that they would not divorce them and encouraged them to be prayerful and to continue attending the programme to get medicine for their children. These men also decided to go for HIV testing.

Difficulties accessing treatment and in interactions with clinic staff

Five women who were interviewed and who also attended focus group discussions described difficulties accessing PMTCT services, despite knowing where to get free antiretroviral treatment for PMTCT and for their own health. They attributed this both to difficulties being seen once they got to the HIV clinic and to how programme staff treated them there. Others had difficulties getting to the clinic. Some women said they were afraid of programme staff.

“Who will be looking after my young ones when I go to the clinic? Each time I go there, my first born has to be absent from school.” (Victoria, interview, dropped out 3 months post-partum)

“We visited this clinic on two occasions but in both cases, service providers told us to come back next time because they had reached the maximum number they manage daily. On our next visit, we even slept there and got numbers three and four in the queue. Nevertheless, we were sent away so we just gave up…The problem is that people book in advance and some workers beckon their affluent friends to enter the consultation rooms. It is often us, the poor, who get deprived of these services.” (Focus group discussion participants 2 and 6, fully participated, site B)

“When I went for follow-up, the nurse said the clinic was busy, so she gave me another appointment date. When I went again, it was the same, and when this happened for the third time, I just gave up. [They] do not know that I have to walk 5–8 kms and often depart from home around 4 am.” (Taona, interview, dropped out 2 months post-partum)

“We were told priority is given to those who come first. We may arrive at the clinic around 9 am; at 10, 11 and 12 we are still there. They attend us around 1 pm. We get tired, our children become weak. Next time, we just give up.” (Mercy, 6-month-old child, interview, fully participated)

“I got tired of walking the long distances because each time, my legs swelled up. And I usually do not have transport money.” (Zondapi, interview, dropped out 2 months post-partum)

“I was hearing from friends that there is Nurse X who is harsh to patients but never believed it until it happened to me.” (Zinenani, 8 months pregnant, interview, dropped out)

None of the women from site B reported these problems. One participant suggested employing HIV-positive staff:

“HIV-positive staff would understand what we are going through and probably treat us better. If they are there, let them start work. Those who are HIV-negative lack sympathy, they just look for salaries.” (Mercy, interview, 6-month-old child, fully participated)

Two drop-outs said they had attended the programme support groups only because they found staff to be supportive. Three drop-outs stopped going because they could not remember the clinic's instructions or when to come back. One refused to believe her HIV-positive test result.

Discussion

While the problems experienced and the views of the women who participated in this study may not be generalizable, insights can be gleaned of relevance to HIV programmes for women and their infants in Malawi and elsewhere. The women in this study were struggling with problems of disclosure of their HIV status, which made participation in the PMTCT programme and follow-up visits difficult for them. Signs like weight loss and actions such as weaning children at six months, home visits by staff to drop-outs and receiving HIV-linked food assistance were visible signs of HIV-positive status. As in much of sub-Saharan Africa, despite high HIV prevalence and having lived with HIV for over two decades, the stigma of HIV in Malawi has not decreased, partly because positive status continues to be associated with immoral behaviour.¹⁵

Pregnant women were routinely tested for HIV as standard practice within antenatal programmes, unless they specifically declined.¹⁶ Many, however, became preoccupied with concealing a positive diagnosis for fear of community interrogation. In Botswana, similarly, lack of community support made many decline HIV testing when an opt-in approach was used, resulting in unsuccessful implementation of the initial government-led PMTCT programme.¹⁷ Although the current opt-out approach, like the availability of antiretroviral treatment, has reportedly increased the proportion of women being tested,¹⁸¹⁹ losses to follow-up observed have simply shifted and occur at every step after testing.

While communities are blamed for propagating stigma, certain activities at health facilities also violate confidentiality and disclose people's serostatus. The physical location of PMTCT services apart from other maternal and child health units noted in this study does not effectively protect confidentiality. Those attending for other services overhear or identify programme participants, who are then often presumed to be HIV-positive. In rural Uganda, HIV-positive people avoid seeking appropriate health care for fear of ostracism due to lack of privacy.²⁰ PMTCT services therefore become less attractive to women. Because of unequal relationships of power and authority between providers and patients in countries like Malawi,²¹ health workers may not recognise these constraints.

There are also structural shortfalls in the programme. Patients utilising maternal and child health services are often considered to be “healthy” because pregnancy is an ordinary physiological process. Given even the slightest constraints, services become under-utilised, as in this study, because providers advise women to wait and come back at a later date.

Inadequate access to antiretroviral therapy has also compromised health care for PMTCT participants, resulting in losses to follow-up. Of the estimated 73,000 positive pregnant women in 2007, only 32% received antiretroviral therapy for PMTCT and only 4% of the 9,150 that started treatment in the last quarter of that year were referred from PMTCT programmes.³ Although an HIV clinic was opened at Lilongwe district hospital in 2006 to provide antiretroviral therapy primarily to PMTCT and tuberculosis patients, the clinic is dominated by men and non-pregnant women seeking tuberculosis treatment.

This study found that more participants who attended follow-up visits after delivery reported having partner support than women who dropped out. The approach of linking PMTCT services to maternal and child health units, although logical, excludes women's partners, who viewed antenatal care as well as PMTCT programmes as “women's spaces”. Inviting them to participate was seen to jeopardize their masculinity, and precluded them from accepting advice or information from their partners. The evidence shows, however, that male involvement is key to ensuring preventive behaviour in relation to HIV.²² Several African countries, such as Tanzania, Malawi and Zambia, have begun counselling and testing for couples and families but the outcomes are yet to be evaluated.^9,23,24

Marriage is considered to be for life in Malawi¹⁵²⁵ and marital breakdown brings shame with it. Among economically less empowered women, as those in this study, many feared divorce because of their economic dependence on men. The men therefore dominated in decision-making, including on sexual matters. Spousal communication about HIV risk and condom use was also problematic. Culturally in Malawi, women are not permitted to communicate HIV and condom use with their spouses because this is associated with women having extra marital affairs.²⁵ To avoid raising tensions, which many in this study were not prepared to handle, the women succumbed to unprotected sex. Poverty was the primary and determining condition of the women's lives, which outweighed the risks of unprotected sex, given their subordinate position in marriage.

This study identified a range of factors that restrict Malawian women from participating in PMTCT follow-up. To address these factors, the whole approach to the delivery of the PMTCT programme and home visits must be reconsidered, to improve confidentiality and minimise stigma. Home visits, for example, could be to all mothers with new babies, irrespective of their HIV status, by integrating the visits into overall child welfare.

Recognizing the important role of men in HIV prevention and treatment, the separation of maternal and child health services from general health care needs rethinking. We should look at the possibility of integration, given that so many health workers have been trained to provide HIV and AIDS care. This might increase acceptance of HIV as an ordinary disease and mitigate stigma. Continuous educational campaigns at all levels remain crucial to change discriminatory attitudes. Above all, women need to be empowered economically and supported to negotiate safer sex and access health care, including HIV treatment and care, that will benefit them as well as their partners and children.

Acknowledgements

The authors thank all study participants, the Ministry of Health, Lilongwe District Health Office, and the UNC project. We also acknowledge the commitment of our research assistant, Alaizi Nkhoma, and programme staff. Thomas Painter provided technical advice on data analysis and interpretation, structuring and review of this paper. This research was funded by the Norwegian government through a NORAD grant and the Helles Legacy Foundation (ISAM).