There are approximately 529 000 maternal deaths worldwide each year and more than 1 million children are left motherless1–3. The main burden (99%) of maternal mortality is in developing countries, and over half of the deaths occur in Africa1. Vitamin A deficiency (VAD) is a well-recognised nutritional problem in children in many developing countries4 and it is well established that even moderate deficiency increases a child's risk of mortality and severe infectionReference Beaton, Martorell, Aronson, Edmonston, McCabe and Ross5. It is also possible that VAD contributes to maternal mortality. Evidence comes from a field trial in Nepal, which reported a 44% reduction in maternal mortality following weekly vitamin A or β-carotene supplementation of women of reproductive ageReference West, Katz, Khatry, LeClerq, Pradhan and Shrestha6, and plausible biological pathways have been identifiedReference Faisel and Pittrof7. If the Nepal finding is verified by trials in Bangladesh and Ghana, vitamin A supplementation (VAS) could become an important component of maternal health programmes such as The Safe Motherhood Initiative, particularly as current interventions are of questionable effectivenessReference Luck8–Reference Sloan, Winikoff and Fikree10. For example, data from the few developing countries with vital registration indicate no significant change in maternal mortality rates since the start of The Safe Motherhood Initiative in 1987Reference AbouZahr and Wardlaw9. VAS could therefore be a potential breakthrough in reducing maternal deaths.
Iron supplementation programmes have shown the need for culturally acceptable information, education and communication (IEC) campaignsReference Galloway and McGuire11–Reference Yip13, especially those that focus on community perspectives, behaviour change and increasing demandReference Yip13–Reference Griffiths15, and have been criticised for only recognising the need for IEC after evaluations show they are ineffectiveReference Griffiths15. Any Safe Motherhood VAS programme must learn from the lessons of iron supplementation and ensure that barriers and facilitators to VAS are explored prior to the initiation of programmes and that programmes include appropriate IEC strategies.
The present study aimed to identify regimen, individual, community and cultural factors that affect the adoption and maintenance of weekly VAS among women of childbearing age in Ghana. This formative research was done as a prelude to the implementation of a large-scale, randomised, double-blind, placebo-controlled field trial exploring the impact of vitamin A on maternal mortality In Ghana. The estimated maternal mortality ratio for Ghana is 549 per 100 000 live births1.
Methods
The study was conducted in 2000 in Kintampo District, which is situated in the centre of Ghana and lies within the forest–savannah transitional ecological zone. Kintampo District has an estimated population of 147 000 people; the district is overwhelmingly rural and consequently farming is the most important economic activity. Only the district town and eight of the 149 villages have electricity, and few can be reached by paved roads. There is one district hospital, staffed by a clinical officer, and seven health posts staffed by nurses.
Data were collected through semi-structured interviews and focus group discussions (FGDs) followed by a pilot capsule trial. All data collection instruments were pre-tested and revised as needed. Informed consent was sought from all respondents. The semi-structured interviews and FGDs were conducted in Kintampo town plus four villages chosen purposively to reflect the district's range of health service accessibility, ethnicity, size of village and location (Table 1).
Community characteristics of the study villages
Several respondent groups were included, namely the potential target recipients of the VAS programme (women of reproductive age), gatekeepers (husbands and village leaders) and information providers (drug sellers, midwives, traditional birth attendants and health workers). A wide range of respondents were included in order to get a broad picture of potential barriers and facilitators for VAS adherence. The sample size for each respondent group was determined using saturation sampling (i.e. respondents were interviewed until no new information was learnt).
Semi-structured interviews were conducted with 50 women aged 15–45 years and with 30 husbands. The women were selected purposively to reflect differences in age ( ≤ 20, 21–35, 36–45 years), parity (0, 1–3, ≥ 4 children) and ethnicity (Northern tribe, Akan, Mo). Thirteen in-depth interviews were conducted with drug sellers, midwives, traditional birth attendants and health workers, all of whom were identified from the semi-structured interviews as being the most utilised in each village. Fieldworkers conducted an average of two interviews a day; they interviewed early in the morning when most people were at home. Fieldworkers entered compounds in different sections of the village and interviewed respondents who met their age, parity and ethnicity quota. During the interviews fieldworkers took field notes, which they converted to detailed English fairnotes on the day of the interview. Fairnotes were then transcribed into Microsoft Word®.
Six focus groups were conducted with six to 12 women aged 15–45 years, stratified by age and ethnicity. Women were selected by asking a community leader to select talkative women who matched the specified age and ethnicity requirements. FGDs were recorded and then transcribed into English.
The information collected in the semi-structured interviews and the FGDs included perceptions of illness and prevention, knowledge and experiences of vitamins and vitamin A, adherence experiences, perceived barriers to adherence and community solutions, desirable and undesirable capsule characteristics, acceptability of distribution strategies, and exposure and experience with communication channels. The data were collected by five trained fieldworkers in the local languages. The transcripts were explored through multiple readings to ensure familiarity with the data. Key analytic categories (issues, concepts and themes) were then identified and the interviews and FGDs were indexed into the categories and interpreted.
The semi-structured interviews and FGDs were collected over 2 months; after they were completed, 60 women aged 15–45 years were randomly selected for the capsule trial. Women were selected in two of the villages using the district's demographic surveillance database. Women who had responded to the semi-structured interviews were excluded. The aim of the capsule trial was to quantify the findings of the semi-structured interviews and to uncover any missed barriers. Each woman was visited at home and informed consent obtained and any questions answered. They were informed that we were planning to start a study exploring whether vitamin A is good for women's health and that we would like to understand the problems women are likely to have taking capsules. They were invited to help us identify any problems by taking a capsule each week so we could learn from their experiences. Each woman was given a small bag of low-dose cod-liver oil capsules (vitamin A capsules were unavailable), was instructed to take a capsule on the same day every week and told that the fieldworker would return in a month. Each month the fieldworker returned, distributed more capsules and administered a questionnaire about adherence, storage, reminding strategies and respondents' opinions of the capsules. The trial lasted 4 months and the response rate was 100%.
Data on availability and use of communication channels in the district were available from a Health Seeking Behaviour study conducted in the district in 1999Reference Hill, Kendall, Arthur, Kirkwood and Adjei16. The study included a health communication channel survey of 332 randomly selected women with children < 5 years of age in eight villages. Some of the information was dated by the time this adherence study was conducted, for example a new popular local radio station began operation in 2000; care was therefore taken with the interpretation of the original results.
Results
Knowledge of vitamins
In the semi-structured interviews and FGDs, few respondents, except the drug sellers and health workers, had heard of the word ‘vitamin’ or the conventional local language equivalent ‘medicine from food’. Of those who had heard of vitamins, few could describe their function or purpose, despite the widespread use of vitamins and vitamin-based tonics as a treatment for common illnesses. Druggists reported that vitamins and tonics make up a large proportion of their business (B complex and multivitamins being their best sellers) and were sold to treat specific symptoms and illnesses. There was a high level of knowledge in all respondent groups that ‘good-quality’ foods such as beans, palm oil, meat and fish are important for health, but methods of preventing adult illness centred round avoiding behaviours linked to the perceived cause of an illness, such as working in the sun.
Adopting a new behaviour
Taking ‘medicine’ such as VAS to prevent illness or maintain health constitutes a new behaviour and some of the women and the husbands said they could not see the importance of taking supplements when healthy. Taking a ‘medicine’ for long periods of time also constitutes a new behaviour as few women reported that they had ever taken ‘medicine’, even prenatal supplements, for longer than a month. Women reported that this new behaviour would be acceptable if they saw other women taking the supplement and were given detailed information before starting to take the supplement. This desire for information was in contrast to most women's experiences, as they were rarely told anything about the medicines or supplements they had taken. The information women said they would require (and in fact was frequently requested during capsule trial informed consent) included an explanation of why a healthy woman should take ‘medicine’ for a long time, what the supplement does, what illness it cures, whether it has side-effects, affects menstruation, has a ‘scent’ (bad taste) and costs anything, and what are the motives of the people supplying it.
Maintaining the behaviour
Despite the acceptability of taking ‘medicine’ for long periods and when not sick, a lack of motivation was identified by women in the semi-structured interviews and FGDs as a potential reason for non-adherence to VAS, as women would get weary of taking ‘medicine’ after a long period of time. The women also reported that forgetfulness would be an important reason for non-adherence; in the capsule trial 19% of women reported missing at least one capsule, the majority because they forgot due to being busy or having an unusual schedule that day. When women in the capsule trial missed one capsule they waited until the next dosing day or stopped completely due to concerns that the capsules must be taken exactly as prescribed, i.e. consistently every week. Women in the capsule trial used a variety of techniques to help them remember to take the capsules, such as storing the capsule with commonly used things (20% of women) and having someone remind them (47%).
It became evident that women in the capsule trial required information to be provided on an ongoing basis, as information given during the consent procedure was often forgotten or needed reinforcing if a woman's situation changed. For example, when women in the capsule trial became pregnant, delivered a baby or became ill they often stopped taking their capsules completely as they needed reassurance that the capsules were safe for their new condition. Other concerns about taking supplements during pregnancy, reported by women in the semi-structured interviews, FGDs and the capsule trial, were whether taking ‘medicine’ during pregnancy could make pregnancy-related nausea worse or make the baby grow big in the womb and cause a painful and difficult delivery. No concerns were raised about taking ‘medicine’ when breast-feeding. Husbands also reported concerns surrounding VAS and pregnancy, as they were worried that the supplement may affect fertility or be a secret form of family planning. Husbands also wanted to know why the supplement is only for women.
Women in the semi-structured interviews and the FGDs reported that they would stop taking VAS if they had side-effects. They would also stop if they developed an illness whilst taking the supplement, as they would assume that the supplement either caused the illness or was ineffective in preventing it, or they would think that the illness was a side-effect of the supplement. They also reported that they would not want to mix the supplement with ‘doctor’ or herbal medicine. In fact, 16% of women in the capsule trial reported that the capsule had caused a bad effect (weakness, tiredness, sleepiness, dizziness) and 12% reported being worried about taking the capsules. Capsule sharing was not a problem, as ‘medicines’ were only considered safe for those to whom they were prescribed. No sharing was reported in the capsule trial, and women and husbands in the semi-structured interviews reported that they would not share medicine prescribed for someone else.
Logistical barriers
Logistical barriers were identified by all of the respondent groups in the semi-structured interviews and the focus groups. Respondents felt that women who travelled may forget to take their VAS with them or may run out of the supplement whilst away. Several women in the capsule trial missed doses because of travel.
Eighteen per cent of women in the capsule trial reported losing some of their capsules: children's curiosity and inadequate packaging were often implicated in the loss. However, only 12% of women reported storage as a problem and 91% kept their capsules in good condition (as observed by the fieldworker).
Distribution strategies
The husbands and the women in the semi-structured interviews and the focus groups reported that the way VAS was distributed would affect adherence. Women had a strong preference for the supplement to be distributed through monthly home visits by a non-community member. The reasons given were that they felt that it would be difficult to find time to go and collect the capsules from a central place, that it would be difficult to ensure they could be found at home on a weekly basis, and that community members may be too busy to act as distributors and would not be able to answer questions or give advice. There was also a unanimous preference for taking the capsule on the same day and at the same time, as it was felt this would foster social support, reduce forgetfulness and would also allow reminding activities to be focused on a single day. Sunday morning emerged as the preferred time and day since women are mostly at home.
Capsule formulation
The size, shape, form and colour of the proposed supplement (a small, transparent, honey-coloured gelatine capsule) was acceptable to women in the semi-structured interviews and FGDs because it was slippery, round, and looked easy to swallow without chewing. However, most women said they preferred taking red medicines as they associated red with blood. The perception of blood as a life force that is used up through work and illness, and restored through diet, rest and certain medicines, is widely held.
Communication channels
A variety of potential communication channels were explored in the interviews with all the respondents. The channels included community groups, churches and mosques, schools, print materials (pamphlets, calendars, posters, newspapers, billboards), mass media (radio, television, videos, movies, loudspeaker vans, market announcements), traditional channels (town criers, songs, theatre) and health channels (maternal and child outreach workers, antenatal care). Of these, the options respondents felt would be most effective were town criers, radio, posters, church, mosque and market announcements, loudspeaker vans and a song.
The coverage of potential channels was explored using the communication channel survey. Most villages had radio reception but only 54% of households owned a radio and only 36% of women reported listening to the radio in the last month. Women listened at specific times of the day and to specific programmes. The radio was considered a very credible source of information and respondents felt that radio messages would be passed around the community. The fact that radio stations often broadcast in only one local language was mentioned as a disadvantage. Literacy was low with only 9% of women able to read a sentence when tested, but respondents felt they could find somebody who could read for them (57% of women reported that somebody in their household could read). Women liked the idea of having a calendar or poster visible at home to help them remember to take the capsule. Respondents felt that the poster should be mostly pictorial with some writing to add credibility. Most women felt that involving the churches and mosques was a good idea, although they acknowledged that not all women attend regularly.
Discussion
To date, intervention studies that have examined how to enhance adherence in developing countries have focused on the treatment of acute illness, and have had varying success. They explored improving instructions, packaging, dosing or treatment formulationReference Helitzer-Allen, Macheso, Wirima and Kendall17–Reference Hyder, Persson, Chowdhury and Ekstrom25, but IEC activities were absent. Evidence from programme evaluations and observational studies of prenatal iron supplementation indicates that there are many barriers to adherence that could be addressed through IECReference Galloway and McGuire11–Reference Griffiths15. These include: a lack of knowledge of the supplement's purpose; low prioritisation of nutrient deficiencies; misunderstanding instructions; side-effects; frustration at having to take the supplement for a long time; forgetfulness; migration; concerns that the supplement causes excessive bleeding during birth; concerns around taking the supplement when pregnant; and concerns that the supplement will make the foetus big and delivery difficultReference Galloway and McGuire11, Reference Ejidokun14, Reference Galloway, Dusch, Elder, Achadi, Grajeda and Hurtado26, Reference Winichagoon27. Most of these were also proposed by our respondents as being likely barriers to adherence to VAS, which suggests that prior to implementing any supplementation programme for women it would be important to explore these aspects through formative research. Respondents' concerns about taking the supplement with other ‘doctor’ or herbal medicine, and what to do if a dose is not taken at the prescribed time, may also be relevant for adherence in other supplementation programmes. Formative research should also address appropriate packaging to reduce capsule loss and favourable formulation such as capsule colour and size. For programmes that supplement non-pregnant women, also relevant for adherence are respondents' concerns that the supplement may affect menstruation, or is really family planning in disguise.
The study also identified the importance of an appropriate delivery system for the capsules, including special arrangements for women when they travel away from home. The experience of giving iron at antenatal care visits suggests that centralised distribution can be difficultReference Galloway and McGuire11, Reference Galloway, Dusch, Elder, Achadi, Grajeda and Hurtado26 and it is recognised that alternative delivery systems may be neededReference Yip13. Women in the present study stated a preference for monthly home delivery of the supplement, but this may not be feasible in large-scale settings. Appropriate distribution systems provide an accessible, constant supply of the supplement, and development of suitable delivery systems needs to be a key concern in any future vitamin A maternal mortality programme. Intermittent supply has been a major barrier to adherence in many supplementation programmesReference Galloway and McGuire11, Reference Mora12, Reference Galloway, Dusch, Elder, Achadi, Grajeda and Hurtado26.
The importance of formative research has been emphasised by othersReference Griffiths15, Reference Blum, Pelto, Kuhnlein and Pelto28, Reference Ghanekar, Kanani and Patel29, and the present study benefited from the multiple methods used. The semi-structured interviews and the FGDs identified issues to be explored in the capsule trial and also allowed issues that could not be piloted, such as preferred communication channels, to be explored in depth. The capsule trial quantified the findings of the semi-structured interviews and FGDs and thus provided information about the relative importance of the barriers. The capsule trial also uncovered barriers to adherence that the community could not identify or predict, such as women stopping taking the supplement if they missed a dose. Formative research is, however, only the first step in developing an IEC strategy. Programme planners need to ensure that knowledge gained from formative research is translated into appropriate actions and that communication experts are utilised in designing the IEC strategy.
For example, the IEC strategy for the subsequent vitamin A and maternal mortality trial in Ghana was developed in five steps. First, information from the formative research was presented at a workshop attended by project staff, communications experts and staff from the national vitamin A programme. The participants used the information to identify the aims of the IEC strategy: to allay fears, inform, motivate, remind, foster social support and reduce logistical constraints. Second, the workshop participants identified simple key messages (Table 2) to help achieve the IEC aims. Third, the participants used the communication information from the formative research to develop a multi-pronged IEC strategy (Table 3) that included appropriate communication channels for each message. Fourth, a creative brief detailing each proposed communication channel was developed and presented to advertising companies, communication experts, musicians and theatre directors. These experts developed a communication plan and put in tenders for material production. Fifth, the IEC materials were then developed with the input of the community and extensively pre-tested with the target populations.
Key messages for information, education and communication (IEC) activities, identified from the study findings
Information, education and communication (IEC) activities designed to support the Ghana vitamin A and maternal mortality trial
The present paper describes research to uncover potential barriers and facilitators to weekly VAS of women. The information collected was used to design an IEC strategy to enhance adherence to the supplement. Although other studies have identified the need for IEC for effective prenatal iron supplementationReference Galloway and McGuire11–Reference Yip13, no rigorous evaluations were located. The findings of the present work, that formative research can identify potential supplementation barriers and can be used to inform an IEC strategy, must be followed by monitoring and evaluation of IEC strategies to show whether they are effective or not.
Acknowledgements
Sources of funding: This study was supported by the United States Agency for International Development (USAID) and the UK Department for International Development (DfID).
Conflict of interest declaration: None.
Authorship responsibilities: B. K., P.A., C.K. and Z.H. designed the study. Z.H., E.A. and C.T.A. were responsible for data collection and analysis. Z.H. wrote the first draft of the paper and all authors, except P.A. who sadly died before the paper was completed, commented on and revised the paper.
Acknowledgements: The authors would like to thank Dr Anne Roberts, Ms Esi Amoaful, Ms Nana Gabrah-Aidoo, Dr Kojo Yeboah-Antwi, Mrs Rosanna Agble, Dr Henrietta Odoi Agyarko, Mrs Mary Arday-Kotei and Dr Sam Adjei for their valuable contributions at the IEC workshop. We would also like to thank all the staff of Kintampo Health Research Centre and all of the community members who participated for their support during this research.
