Counsellors’ Perspectives on Antenatal HIV Testing and Infant Feeding Dilemmas Facing Women with HIV in Northern Tanzania

Abstract

This study investigated the infant feeding advice that counsellors were giving HIV-infected pregnant women in Moshi, Tanzania, the factors they thought had an impact on women’s infant feeding choices and their role in influencing these decisions. The data are drawn from in-depth interviews with 16 nurses working as counsellors in their spare time in an antenatal trial of prevention of mother-to-child transmission, five local HIV/AIDS counsellors and two medical doctors, whose counselling experience ranged from less than six months to nine years. Informed choice of infant feeding method by HIV-infected women, as recommended by UNAIDS/WHO/UNICEF Guidelines, was seriously compromised by the actual advice given, directive counselling, lack of time to cope with a positive HIV test result, and lack of follow-up support, regardless of socio-economic status. Infant feeding options were not always accurately explained, but counsellors believed most women had little choice but to breastfeed and were unlikely to exclusively breastfeed, despite advice. It was apparent that the risks and benefits of the options open to HIV-infected women were complicated for the counsellors, not only the women. Counsellors needed additional training in non-directive counselling and infant feeding options to ensure a better quality of advice-giving and support to follow-up women at home.

Résumé

Cette recherche analyse les conseils donnés par les consultants aux femmes enceintes séropositives sur l’alimentation des nourrissons à Moshi, Tanzanie, les facteurs pouvant selon eux influencer le choix des femmes et leur rôle dans ces décisions. Les données proviennent d’entretiens avec 16 infirmières travaillant comme consultantes pendant leur temps libre dans un projet de prévention anténatale de la transmission de mère à enfant, cinq conseillers locaux sur le VIH/SIDA et deux médecins, dont l’expérience dans les consultations allait de moins de six mois à neuf ans. Le choix informé de la méthode d’alimentation par les femmes séropositives, recommandé par les directives ONUSIDA/OMS/UNICEF, était compromis par les conseils donnés, le dirigisme des consultations, le manque de temps pour faire face à un test positif de dépistage du VIH et le suivi insuffisant, quelles que soient les circonstances socio-économiques des femmes. Les consultants n’expliquaient pas toujours précisément les options, mais pensaient que la plupart des femmes n’avaient guère d’autre choix que d’allaiter et que, malgré les conseils, elles ne nourriraient probablement pas leur bébé exclusivement au sein. Les risques et les avantages des options étaient compliqués pour les consultants, pas seulement pour les femmes. Les consultants devaient être formés à conseiller sans diriger et à présenter les différentes options pour garantir des conseils de qualité et un suivi des femmes à domicile.

Resumen

Este estudio investigó los consejos sobre la alimentación infantil impartidos a mujeres embarazadas e infectadas con VIH en Moshi, Tanzania, los factores que en la opinión de los consejeros incidieran en las decisiones que tomaran las mujeres acerca de la alimentación infantil, y la influencia de los consejeros sobre dichas decisiones. Se tomaron los datos de entrevistas con 16 enfermeras que trabajaban como consejeras durante su tiempo libre como parte de un ensayo antenatal de prevención de transmisión materno-infantil; cinco consejeros locales de prevención del VIH/SIDA; y dos médı́cos cuya experiencia como consejeros variaba de entre menos de seis meses y nueve años. La opción de un método de alimentación infantil recomendada para las mujeres infectadas con VIH por UNAIDS/OMS/UNICEF fue limitada por los consejos impartidos, la consejerı́a directiva, la falta de tiempo para asimilar el resultado positivo de una prueba de VIH, y la falta de un seguimiento que apoyara a las mujeres, irrespecto de su nivel socioeconómico. Evidentemente, los riesgos y beneficios de las opciones recomendadas para las mujeres infectadas con VIH eran complicadas para los consejeros y para las mujeres mismas. Los consejeros necesitan más entrenamiento en la consejerı́a no-directiva, y acerca de las opciones de alimentación infantil, para asegurar una consejerı́a de calidad y un seguimiento que apoye a las mujeres.

Key Words:

• mother-to-child transmission of HIV
• breastfeeding
• infant feeding
• HIV/AIDS
• counselling
• Tanzania

Mother-to-child transmission of HIV is an urgent and growing problem in resource-poor settings. It has been estimated that with no preventive measures taken, the risk of transmission in sub-Saharan Africa is 21–45% [1], with breastfeeding accounting for at least a third of that risk [2]. Voluntary counselling and testing (VCT) for HIV are recommended during pregnancy [3] so that, among other reasons, women can be offered antiretroviral drugs and infant feeding advice. Guidelines on breastfeeding and HIV prepared by UNAIDS, WHO and UNICEF in 1998 promote fully informed choice of infant feeding method for HIV-positive mothers (Box 1) and recommend that breastfeeding should continue to be protected, promoted and supported among HIV-negative mothers and among mothers of unknown HIV status [4].

However, studies in resource-poor settings have identified constraints and barriers for pregnant mothers who are offered VCT antenatally [5][6][7], and the best infant feeding method in these settings for infants of HIV-positive mothers is still being debated. A Kenyan randomised trial comparing breastfeeding with formula feeding found that breastfeeding approximately doubles the rate of infection in infants [8]. Exclusive breastfeeding was shown to reduce the risk of HIV transmission compared to mixed feeding in one observational study in South Africa [9][10]. Nicoll and colleagues argue that breastfeeding is a significant factor in the transmission of HIV to infants, but that formula feeding is rarely appropriate for most infants in resource-poor settings, regardless of whether or not they are exposed to HIV infection [11]. It has been argued that the promotion of exclusive breastfeeding for the general population would have numerous positive effects in addition to possible reduction of HIV in infants [12][13][14]. The concept of “safer breastfeeding practices” has been introduced for this reason (Box 2) [14][15][16]. Several studies have demonstrated that it is possible to support women to exclusively breastfeed [17][18][19][20].

Box 1. UNAIDS/WHO/UNICEF Guidelines for Preventing MTCT [21]

•	Early access to adequate antenatal care.
•	Voluntary and confidential counselling and HIV testing for women and their partners.
•	A short course of perinatal antiretroviral treatment (AZT) given to HIV positive women in the last weeks of pregnancy through delivery (and possibly also to their newborn infants).
•	Improved care during labour and delivery.
•	Counselling for HIV-infected pregnant women should include the best available information on the benefits of breastfeeding, the risk of HIV transmission through breastfeeding and the risks and possible advantages of alternative methods of infant feeding.
•	Support for HIV-infected mothers who choose not to breastfeed, to enable them to use breastfeeding replacements safely, without violating the International Code of Marketing of Breast Milk Substitutes and related resolutions of the World Health Assembly.

In Tanzania breastfeeding is universal and prolonged partial breastfeeding is widely practised [22][23]. According to the Tanzania Demographic & Health Survey 1996, the median duration of breastfeeding was 21.6 months nationally and 22.6 months in the Kilimanjaro region [24]. In a prior study in the Kilimanjaro region we found that few infants were exclusively breastfed and the benefits of exclusive breastfeeding by HIV-positive mothers were being questioned by pregnant women [25].

Box 2. Safer Breastfeeding Practices [14][15][16]

•	Exclusive breastfeeding for 4–6 months, followed by rapid weaning.
•	Reducing the duration of breastfeeding.
•	Proper positioning and latching during breastfeeding.
•	Practising safe sex during breastfeeding period.
•	Seeking medical care immediately for breast problems or when babies have mouth problems (e.g. oral thrush).
•	Avoiding breastfeeding from bleeding or fissured nipples.
•	Providing post-natal antiretroviral prophylaxis to the infants.
•	Heat treatment of breastmilk (especially when other foods are introduced).

For prevention of mother-to-child transmission (PMTCT) of HIV to be successful, it is important to have skilled counselling to provide guidance and help HIV-infected women make choices appropriate to their situation and to which they can adhere. It has been suggested that many weaknesses exist in the counselling currently being provided by most PMTCT programmes, and that resources should be invested to address these [26]. A number of studies have investigated women’s attitudes towards VCT, existing barriers and choice of infant feeding method [7][27][28][29][30]. However, only a few studies have described counsellors’ perceptions and examined how effective they are at conveying the complex issues involved in VCT [31][32].

The purpose of this study was to investigate how HIV/AIDS counsellors perceived the dilemmas that HIV-infected women face when making decisions about infant feeding options. Specifically, we investigated:

•	the kind of infant feeding advice that counsellors give HIV-infected women,
•	the factors that counsellors think will affect women’s feeding decisions,
•	counsellors’ perceptions of the feasibility of implementing the UNAIDS/WHO/UNICEF Guidelines on breastfeeding and HIV, and
•	counsellors’ perceptions of their own role vis-à-vis pregnant women.

Participants and methods

The participants recruited for this study were all 16 of the HIV/AIDS counsellors who had been trained in 1999–2000 specifically for a PMTCT trial in Moshi, the regional capital of an urban district of the Kilimanjaro region. We also included five local HIV/AIDS counsellors and two medical doctors with different training backgrounds, who had been working with HIV VCT for many years (Table 1

Table 1 Characteristics of participants (n=23)

Characteristic	No.
Age
20–29	2
30–39	2
40–49	9
50–59	8
60–69	2
Marital status
Single	8
Married/co-habiting	12
Divorced	2
Widowed	1
Breastfeeding experience^a
Yes	19
No	3
Occupation
Nurse	21
Doctor	2
Counselling training
1–2 weeks general training+PMTCT sensitisation	7
1–2 weeks general training+4 weeks PMTCT training	2
1–2 weeks general training+8 weeks PMTCT training	1
4–8 weeks PMTCT training+2 weeks breastfeeding course	3
4 weeks PMTCT training	5
8 weeks PMTCT training	3
1 year on-going training with supervision+4 weeks PMTCT training	2
Work experience
0–5 months (PMTCT only)	11
1–5 years (VCT)	3
6–9 years (VCT)	9

a Excluding the male counsellor.

). All but two participants were from the Kilimanjaro Christian Medical College (KCMC), while the two were from the Kilimanjaro Women’s Group against AIDS, a local NGO [33] collaborating with KCMC in the ongoing PMTCT trial. Only one was a man, which reflects the gender balance among HIV/AIDS counsellors locally. In this paper we will refer to all 23 participants as “counsellors”.

The data presented are drawn from in-depth interviews with these counsellors. One counsellor functioned as the main informant and outcomes from each interview were discussed with her. She was interviewed twice, the second time in order to summarise and clarify the findings at the end of the study. The interviews were mainly conducted in Kiswahili. English was used only when the interviewer (the first author) lacked sufficient Kiswahili to express herself.

Each participant signed a statement giving informed consent, and for the analysis each was given a fictitious name, used also in this paper. Ethical clearances and research permission were obtained from the Tanzanian National AIDS Control Programme, the Ministry of Health, the Tanzanian Commission for Science and Technology, the Tanzanian Food and Nutrition Centre, the KCMC Ethical Research Committee and the Norwegian Committee for Medical Research Ethics.

A semi-structured interview design was employed to allow the sensitive area of HIV transmission to be discussed as openly as possible. The interview guide covered a broad range of topics and allowed for additional questions on emerging themes. A continuous validation process took place whereby emerging themes were included in subsequent interviews and discussed with one main informant; this has improved the validity of the findings. The interviews ranged in length from 45 minutes to two hours. The field work was conducted from August 2000 through January 2001.

The interviews were tape-recorded, transcribed and coded using the qualitative programme Open Code [34]. Emerging and recurrent themes were identified, and views in agreement with or opposition to these themes were juxtaposed to determine the extent to which particular themes were present or absent in the accounts of other respondents [35].

Study site

Infant mortality in the Northern Highlands, of which the Kilimanjaro region is a part, was reported to be 41 per 1000 live births in 1996 and the under-five mortality rate 69 per 1000 live births [24]. In Moshi, the HIV sero-prevalence of pregnant women attending antenatal clinics was reported to be 20% in 1998, three times higher than in 1992 [36].

KCMC is one of five hospitals in Tanzania that since 1999 has been involved in a UNICEF-funded VCT and PMTCT trial. All of the staff at this site were said to have gone through awareness-raising workshops on PMTCT issues. Guidelines for the integration of PMTCT at these sites include VCT, obstetric care and monitoring and evaluation [37][38][39]. Guidelines on infant feeding specifically for HIV-infected women in Tanzania were also developed with support from UNICEF [40]. At KCMC the VCT consisted of one pre-test counselling session, HIV testing and one post-test counselling session for women who test HIV positive, including infant feeding advice. Rapid on-site HIV testing was used and confirmed by Western Blot if the result was positive or indeterminate. Results were available the same day. A short course of zidovudine was provided for HIV-infected mothers antenatally (weekly from 36 weeks) and during labour. Modified obstetric care and advice on infant feeding were also provided. The infant feeding methods recommended for HIV-infected women were exclusive breastfeeding with abrupt weaning, modified cow’s milk or formula feeding. HIV-infected mothers giving birth at KCMC who opted for infant formula were given supplies only while in the clinic.

Monitoring of HIV-infected infants, infection rate and follow-up of zidovudine adherence were in the protocol, but depended on mothers returning for postpartum care at KCMC. Male partners were also offered VCT, and this was especially recommended if the woman had HIV. We used the components of this trial as a point of reference for the interviews. However, our study was not designed to evaluate the training of the counsellors recruited for the PMTCT trial or any other component of the PMTCT trial.

In the following pages, we have combined presentation of the results of the study with discussion of their meaning and implications.

The “good counsellor”

Counselling can be defined as a directive process of helping someone to accept and use information, and give advice for solving or coping with a problem, or as a non-directive process of helping someone make a decision and plan how to solve or cope with a problem [41].

Proper counselling and support for HIV-infected mothers is crucial [42][43], but is regarded as problematic at most PMTCT sites [16]. The counsellors in this PMTCT trial were nurses on different wards in the hospital, working as counsellors on their days off for a small sum of money. Most of the PMTCT counsellors had had only four weeks of training for this work; 11 of the 16 had been doing PMTCT counselling for less than five months, one of them had less than five years of experience with VCT, while four of them had six to nine years of experience (Table 1). They reported job related stress, including from a heavy and mentally demanding workload in off-duty time, an unpredictable client flow and long hours.

According to the counsellors, most of the women offered VCT (70–80%) agreed to be tested for HIV. They believed that their main role was to prepare women for HIV testing and receiving the results. They felt that the most important personal qualities needed for the job were to be sensitive, non-judgmental, self-sacrificing and understanding. On the one hand, they saw their role as not imposing their own values on women. On the other, they commonly described a “good” counsellor as one able to convince pregnant women to “do what the counsellor wants” and accept VCT, and a “bad” counsellor as one with whom many women refused VCT. Several counsellors regarded it as a failure if they could not convince the pregnant woman to have an HIV test and disclose her status, and some descriptions of “counselling” appeared to contain a coercive element:

“You give the woman a choice but you tell her that the best she can do is to be tested. When I do pre-test counselling and she disagrees, I encourage her to be tested until she agrees.” (Gloria)

“The HIV-infected mothers often tell me that they would not like to disclose their status to husbands. In that situation the only thing I can do is to continue to advise her until she agrees to disclose her status.” (Halima)

“It is necessary that the counsellor is understanding and has the knowledge to convince the mother until she agrees to be tested for HIV. A counsellor will know she has succeeded in her work when the woman has agreed to be tested.” (Jessie)

Williams et al. express doubts about non-directive counselling and conclude that the boundary between choice and coercion is not clear-cut for many practitioners [44]. Although the counsellors showed understanding as to why a pregnant woman would choose not to be tested, this empathy was lost in the face of their own interest in being successful.

Two counsellors with more education and training in counselling issues were concerned about the inadequacy of the support given to women in the trial:

“We are giving the mother a big burden to carry. Finding out that she is HIV-infected is like an earthquake, and changes her life. There is often pressure to disclose her status to her partner.” (Adrophina)

“Pregnant mothers should be better prepared. They come to the antenatal clinic for one problem (pregnancy), and they may leave with a different problem (HIV-infection) … without knowing if they will get any support at home.” (Irene)

The meaning of informed consent is uncertain if women have had no time to consider what to do if they turn out to be HIV positive. Research from Tanzania also suggests that disclosure should not be regarded as a one-time event, but rather as a process of decisions [45] which cannot all be taken in one session. Our findings support this conclusion– that a one-time pre-counselling session may not be sufficient for pregnant women who visit an antenatal clinic, unaware that VCT will be offered.

These counsellors gave most of their attention to the HIV-positive women. Little time seemed to be spent advising uninfected women on how to prevent infection, the higher risk of transmitting HIV to a nursing child if a woman is infected during pregnancy or the breastfeeding period, or the feasibility of demanding safe sex. For those who were not infected, post-test counselling can be summarised as:

“Congratulations! Now go home and demand safe sex! Good-bye!”

To breastfeed or not?

UN-supported infant feeding advice for women with HIV is shown in Box 3. Most of the counsellors believed that women who decided not to breastfeed risked the stigmatisation of being identified as HIV-infected. This was given as a major reason why many of the HIV-infected women decided to breastfeed.

“If the neighbours find out that you are not breastfeeding, they will suspect that you are HIV-infected. That will have a lot of consequences in most communities.” (Elibariki)

Box 3. Infant Feeding Methods

• Commercial infant formula – the UNAIDS/WHO/UNICEF Guidelines list commercial infant formula as the first recommended feeding alternative for HIV-infected mothers who choose not to breastfeed. Other options are:

• Home-prepared formula of cow’s milk – the milk is diluted by mixing 100 ml of milk + 50 ml water + 2 teaspoons sugar and the mixture is boiled. The infant will need 150 ml per kg of body weight per day and feeding should be done using a cup.

• Micronutrient supplementation – recommended because animal milk may have insufficient zinc, iron, vitamins A and C and folic acid.

• Exclusive breastfeeding + abrupt weaning – exclusive breastfeeding is defined as breastfeeding in the absence of all other fluids and solids and is recommended for up to six months of age, during which time breastmilk alone can satisfy all the infant’s nutritional and fluid needs. Drops or syrups containing vitamins, mineral supplements or medicines can be given in addition to breastmilk. Abrupt weaning hinges on the theory that the reason why mixed feeding is associated with the highest rates of MTCT of HIV is because other foods in some way damage the infant gut epithelium, which thus promotes the establishment of HIV infection if infected breastmilk is ingested at the same time.

• Expressed and heat-treated breastmilk - heat treatment of expressed breastmilk will kill HIV in the milk. To pasteurise the milk, it should be heated to 62.5 degrees centigrade for 30 minutes. Or it can be boiled and then cooled immediately, by putting it in a refrigerator or standing the container in cold water.

Another major reason cited for breastfeeding was poverty, i.e. lacking the means to buy cow’s milk or milk powder. Therefore, only women who could disclose their status to their husbands and whose husbands had the money, could choose an alternative to breastfeeding. Since most women were both socially and economically dependent on men, the counsellors regarded the involvement of men as crucial but complicated.

“If she does not disclose her status to her husband, where will she get support and what will she buy milk with?” (Jessie)

“Many women are scared, especially those who do not have their own source of income. They know that they will get no help if they are isolated, rejected or sent away.” (Zainabu)

“The thing is, even if the HIV-infected woman has been given advice and knows how the virus is transmitted and what to do, she tells me that she cannot tell anybody… So you see, we are talking to the one with the least power, the bread-winner for the child who has the infected milk. But you see, it is very often the only little thing she has to enable the child to survive. Very little power to decide on her own, poverty, yet a responsibility to take care of the family. What little the husband has is diverted to drinking, and after he drinks, he is the most powerful person. So we may be talking to the wrong one.” (Elibariki)

The counsellors realised that HIV-infected mothers had to make plans in order to take care of all contingencies and often to hide the truth.

“I realised that this mother would have problems if she was not breastfeeding. The mother herself said that she would need time to plan what to do and where to go if she disclosed her HIV-status and her husband threw her out. ‘If he decides to throw me out now, I will not know where to go with my child, and where will I deliver? It will be very difficult, and I will not get any food.’ After listening to her, I decided that it was best for her to deliver first. Maybe we could tell the husband that she was HIV-infected after she had given birth.” (Zainabu)

Thus, breastfeeding was seen as the only possible option for the majority of the women.

“Oh, this trial is not for the housewives or the rural women. The target is the educated women with economic means because they have money and education; it will not be a problem for them. When you counsel them and give them infant feeding advice, they will understand and they can afford it. So far I have not counselled any poor HIV-infected woman who decided not to breastfeed her infant.” (Zainabu)

“There are a minority, only those with permanent jobs, who can afford to practise an infant feeding option. But they are not really typical. That means we are targeting a minority who are able to take our advice, and disregarding the majority, who are not able to afford the substitutes. This is a form of discrimination.” (Elibariki)

Most of the counsellors assumed that those who chose an alternative to breastfeeding also had all the necessary knowledge, skills and domestic facilities to prepare the milk in a safe and hygienic way.

‘If the mother does not breastfeed her child, feeds it cows’ milk, dilutes the milk according to the instructions and has clean utensils… then there is no problem.’ (Annie)

On further probing, however, the same counsellor added: ‘Some will manage, and some will not manage. A minority will manage.”

Exclusive breastfeeding

Most of the counsellors realised that the infant feeding advice could be difficult to implement, especially in rural settings. In this region, mothers-in-law are very respected, and it was considered a dilemma to oppose them by not breastfeeding. Indeed, there was a consensus among the counsellors that exclusive breastfeeding was the best option for HIV-infected women, even though most mothers were thought only to partially breastfeed. According to the counsellors, although the majority of HIV-infected mothers who received counselling said they would practise exclusive breastfeeding, it was questionable whether they would do so, not least because of common beliefs about the necessity of giving water to infants.

“Although most mothers say that they will practise exclusive breastfeeding, it may not be true. Let’s think about it! The poor HIV-infected mother says that she will practise exclusive breastfeeding, because she has no money to buy cow’s milk or milk powder. But what about her own nutrition? Is she healthy enough to produce all that milk? It can be difficult to breastfeed. That’s what I mean when I say we are only singing theory when giving advice.” (Edith)

“Even though we are teaching them about exclusive breastfeeding, really, in my opinion, based on my experience, there is no one who will resist and not give the child anything other than breastmilk for the first four months! No one! What we teach the mothers is something very different from what they practise.” (Gloria)

“It is a common belief that water is necessary for quenching thirst.” (Zainabu)

“Many of them believe that if you only breastfeed the infant it will cause a hard stool and that water will prevent constipation.” (Edith)

“The mothers don’t know what to do when their child is crying a lot. Most of the time a child is crying for the breast, but they will give water because they believe that their milk is not sufficient.” (Irene)

Many of the counsellors reported that mothers questioned the safety of exclusive breastfeeding. However, a few were of the opinion that an HIV-infected mother was more likely to practise exclusive breastfeeding having been counselled to do so.

“The HIV-infected mother may practise exclusive breastfeeding. She knows that she is infected and her incentive is for the child to survive. This incentive is the reason why it may be easier for an HIV-infected mother to practise exclusive breastfeeding” (Daniela)

None of the counsellors had themselves practised exclusive breastfeeding. All of the women counsellors said that if they themselves had HIV, they would not breastfeed but would use milk powder or cow’s milk. Most of them questioned the safety of exclusive breastfeeding, and thought it would not be a viable or safe option.

“If I myself were HIV-infected, I would only disclose my status to close family members. I would not be afraid of people in the community, but I would not tell them that I was infected! Instead I would tell them that my breasts have some problems. You can even start pretending during pregnancy that there is something wrong. You tell them there is a sore on one breast, put a fake bandage on it and tell them that your breast milk is not good. When you have given birth, no one will question why you don’t breastfeed.” (Consuela)

Expressed, heat-treated breastmilk

Although expressed, heat-treated breastmilk was not one of the options included in the protocol on infant feeding for HIV-infected mothers at this site, we explored the counsellors’ views on this method. Three of the counsellors who had received three weeks of breastfeeding counselling training included this option in their infant feeding advice. Most of the others expressed disbelief that heating could inactivate the virus. Since they were not convinced themselves and believed it was a difficult concept to understand, they thought it would be even more difficult to instruct and convince HIV-infected mothers to try it.

“We have been told that the HIV-virus is killed by heating the breast milk to 60–62 degrees. But even I am not convinced that the HIV-virus will be killed. Furthermore, to try to tell this to a mother, to educate her until she understands and agrees that if you heat treat breastmilk the HIV-virus will be killed. That will be a very difficult task.” (Daniela)

It is unfortunate that most of the counsellors disregarded expressed, heat-treated breast milk as a feeding methods since it has been demonstrated to be a feasible option for HIV-infected women in Tanzania [46], and is an option that could be pursued especially during the weaning period.

Conflicting messages

In antenatal clinics in Tanzania, breastfeeding has been promoted as the best infant feeding method since the early 1990s. However, since the UNAIDS Guidelines were amended in 1998, the risk of HIV transmission through breastfeeding has been included in the health information provided at antenatal clinics. There was concern among the counsellors that mothers would find it difficult to understand how something can be best and hazardous at the same time. One commented that the slogan “Breast is best” was not valid anymore.

“Really things are a bit vice versa. In 1998, the year before last, there was a seminar where we were taught about the benefits of breastfeeding. We educated all mothers to practise optimal and exclusive breastfeeding. But now during health education we mostly tell them about the dangers of breastfeeding, that HIV may pass through the milk. We used to tell them to start breastfeeding within 30 minutes after birth, but now it is all changed. Therefore, you shouldn’t be surprised if breastfeeding is going to be a practice of the past. Mothers will continue with these tins of milk…” (Pamela)

Counsellors need to be knowledgeable and understand how to put this change of message across. One of the more experienced counsellors felt that the training received by health care workers had become inadequate and we were told about some of the double messages health educators can give.

“The quality of the knowledge we are receiving is quite limited. Not only the counsellors, but even public health educators themselves sometimes have incorrect knowledge… In fact, it is fairly dangerous.” (Elibariki)

“I attended a breastfeeding campaign where mothers were told to exclusively breastfeed. Later, I overheard a conversation where one of the educators, a nurse, was saying to a mother that it was not so bad if once in a while she gave the child boiled water. I asked her why she was now telling the mother that she could add water but during the campaign said not to. She replied that teaching about exclusive breastfeeding was just theory and had nothing to do with practice. So, even the nurses who are teaching exclusive breastfeeding are sometimes encouraging mothers to give water.” (Joyce)

Conveying facts or creating confusion?

In Tanzania, as in other African countries [27], health care workers’ words are often taken as the final word. It is therefore essential that counsellors are thoroughly trained and have high-quality, up-to-date information to impart. A few counsellors, like Gloria, did not have a clear understanding of what exclusive breastfeeding meant or that there should not be a prolonged period of weaning. Other counsellors also questioned whether abrupt weaning was practical for the lactating mother.

“We advise them to exclusively breastfeed and at three months to terminate breastfeeding. They are told to gradually reduce breastfeeding already from the second month, maybe to 5–6 feeds a day, to gradually attune the child. As the amount of breastmilk is reduced from the second month, they are told to supplement with formula so that the child will not get ill.” (Gloria)

When it was indirectly suggested to her that this was mixed feeding, she replied:

“This is what I teach them, and it is based on my own experience. If you stop breastfeeding instantly, the child will get ill. That’s why you gradually reduce breastmilk from the second month, and at the end of the third month they should stop breastfeeding.” (Gloria)

None of the counsellors had been informed during training how infants should be fed after the abrupt cessation of exclusive breastfeeding. The advice they gave to dilute cow’s milk with water so that it would be nutritionally adequate for infants of various ages also varied.

‘We tell them about how to mix modified cow’s milk, you take half cow’s milk and mix it with a quarter of water. Then you add 800 g of sugar, we show them the amount with a spoon. We are also advising them to give vitamin and mineral supplements.” (Rehema)

“We tell them to mix two parts of milk with one part of water.” (Gloria)

“The amount of milk, water and sugar is according to the weight of the child. We have all this information in a notebook and can calculate the needs of each child.” (Halima)

The counsellors said they informed mothers about the risks and benefits of each infant feeding method. The risks of malnutrition and diarrhoea if the infant was not breastfed were regarded as small by these counsellors. They perceived the risk of HIV transmission through infected breastmilk as much greater. Yet most of them did not know the actual risk of HIV transmission through breastfeeding.

“We tell them the fact that a newborn child has a 60% chance of not being HIV-infected. Then we inform them that if they decide to breastfeed the child will most certainly be HIV-infected.” (Irene)

“We have been told that maybe 3% of children born to HIV-infected mothers will not be HIV-infected.” (Virginia)

Follow-up and support

The counsellors had high expectations of what they should achieve. They recognised that HIV-positive women need support in their choice of infant feeding method and safeguarding against stigmatisation by the community. They knew it was time-consuming to give proper instructions and provide support, especially in relation to infant feeding.

“It is necessary to instruct the mother carefully about how to prepare the milk, taking into consideration the utensils she has. She also needs to know how much cow’s milk and sugar the child should be fed for each meal, to practise at home and then to come back and show you. This has to be repeated two or three times in order to make sure she has understood. When she shows you, then you will know if she has understood or not. If she makes a mistake you have to correct her. If the mother is not instructed, the child will either get too much or too little milk. It is necessary to spend one hour to instruct her how to prepare the milk.” (Annie)

The counsellors emphasised the necessity of following up HIV-infected mothers at home, to see how they were getting on. However, no hospital transport was available to facilitate this; if the counsellors wanted to do home visits, this was considered voluntary and outside their working hours, at their own expense. Thus, home visits were seldom undertaken.

Furthermore, if mothers did not return for follow-up, the counsellors did not know how they ended up feeding their infants, missing the opportunity to get feedback on how their advice worked in practice. From the counsellors’ point of view, this was a major weakness of the PMTCT trial. It contributed to their feelings of lack of control and reduced the possibility of them doing their work adequately.

Beyond mother and counsellor

Pregnant women attending antenatal clinics where VCT is being offered tend to be the first member of the family to be tested.

“Most African men will not agree to be HIV tested, but will blame their wives for being the source of infection.” (Zainabu)

“Women can even be blamed for a pregnancy by their partners, if that pregnancy is not wanted. Anything that you blame the man for, they will deny and say that it has nothing to do with them. There is not an open discussion of how to solve this problem. Most of the time it is the woman’s responsibility to prevent pregnancy and to solve their other reproductive problems. Within a marriage it is difficult to prevent MTCT of HIV if the husbands are not willing to participate. The best thing would be to counsel the couple together so they can try to implement the advice they are given together.” (Irene)

Targeting only the pregnant woman at the antenatal clinic in countries where men do not usually accompany their wives makes male involvement difficult. According to the counsellors, only a few men had taken the opportunity to come to the antenatal clinic for testing and counselling.

Many of the counsellors were of the opinion that communication on HIV and infant feeding should be aimed at the community at large and not just HIV-infected mothers. The support and motivation of both partners and the community were seen to be essential to the successful feeding of infants born to infected mothers [47][48][49][50], but women first had to risk disclosure [51]. Strategies for how to do this may need pilot testing first [52]. This was beyond what this PMTCT trial was able to offer at the time of our study, however, and counsellors received no training on how to facilitate partner involvement, support disclosure or deal with existing barriers. Lie points out that hospital settings are usually not a very supportive context for relationship-building and do not easily facilitate openness with regard to taboos and socially sensitive issues [41].

Methodological considerations

Counsellors’ attitudes, perceptions and self-reported practices in giving advice on infant feeding methods are important. While their views may differ from those of the women they counsel and may be skewed by their own social and professional backgrounds, they are the crucial link between policy and practice.

During the interviews the reports of the counsellors seemed to be truthful and credible, and they showed no signs of wanting to impress the interviewer. We did not take into account that some of them may have been HIV-positive themselves, although this awareness might have strengthened our understanding of their responses.

We included all the PMTCT counsellors at this antenatal site, and believe they are representative of counsellors elsewhere. Eleven of them were newly trained and their views most probably reflect their inexperience. We did not analyse the data in relation to the difference between them and those with more experience.

The views of the counsellors were based on their experience of giving advice to HIV positive women. It is important to note that pregnant women attending this antenatal clinic were probably not representative of all women in this part of Tanzania, but were usually urban or peri-urban, with at least primary schooling. They were either more financially secure, or referral patients, or willing to pay out-of-pocket for the better facilities and services provided at this referral hospital. This may not be the case in other antenatal settings in Tanzania where women are of lower socio-economic status, have less access to health care and lower levels of education. There is therefore no reason to believe that the problems revealed in this study are less pronounced in other settings.

Conclusions

The UNAIDS/WHO/UNICEF Guidelines were meant to ensure that HIV-infected pregnant women would be given an opportunity to make an informed choice as to how they would feed their babies and to enable them to adhere to this choice. This study found that choice was seriously compromised by the actual advice given, directive counselling, lack of time to cope with a positive HIV test result, and lack of follow-up support, regardless of socio-economic status. Infant feeding options were not always accurately explained, but the counsellors believed most women had little choice but to breastfeed and were unlikely to exclusively breastfeed, despite advice.

It was apparent that the risks and benefits of the options open to HIV-infected women were complicated for the counsellors, not only the women. Many counsellors’ inexperience in counselling and/or limited knowledge of alternative feeding methods was critical. The fact that the majority of pregnant women who were offered VCT accepted it seems to be explained, at least in part, by counsellors pressuring women into saying yes, by their own reports. The fact that counsellors themselves were not supported to give support to women after their babies were born, can be considered a major failing.

Health care workers appointed as antenatal PMTCT counsellors need to do counselling as part of their jobs, not in their free time, and to be paid accordingly. To ensure a good quality of advice-giving, counsellors need additional in-service training in non-directive counselling and regularly updated, accurate information on infant feeding options. If exclusive breastfeeding and infant feeding alternatives are to be viable options for HIV-infected mothers in this setting, intensive efforts to promote understanding of these practices, not least among health care workers and other role models, is necessary. Counsellors need guidelines on weaning and support to be able to follow up women at home after their babies are born, both for the women’s sake and so that counsellors themselves are aware of the impact and outcomes of their counselling. Finally, it must be recognised that women can only make a real choice of infant feeding method and receive follow-up support at home if they can be open about their HIV-status and know they can count on the help of their partners and communities.

Acknowledgements

We thank all the counsellors who willingly participated in this study, especially the late Bertha Marenga. We also thank Prof Gro Th. Lie for providing valuable feedback on earlier drafts of this paper. This project was funded by a grant from the Norwegian Research Council and completed as part of a Health System Research Project, supported by the Norwegian Council of Universities Programme for Development Research and Education.