Women’s attitudes and beliefs towards specific contraceptive methods in Bangladesh and Kenya

In low- and middle-income countries, women’s beliefs about specific contraceptive methods are not well understood. This study documents the beliefs about eight contraceptive methods among women living in Matlab (rural Bangladesh), Nairobi slums and Homa-Bay (rural Kenya) and compares the opinions of current, past and never users of the most commonly used methods (oral contraceptives, injectables and implants). In each site, we interviewed 2424 to 2812 married women aged 15–39 years. As expected, we found that current contraceptive users were typically satisfied and had more positive beliefs about their method than past or never users. Nevertheless, large minorities of current users thought that their method might cause serious health problems, impair future childbearing and was unsafe to use for a long time; higher proportions (25–50%) reported that their method use caused unpleasant side effects. Past users of pills and injectables outnumbered current users and their beliefs were similar to never users. In all three sites, about half of past injectable users reported satisfaction with the method but the satisfaction of past implant users was lower. Despite high contraceptive use in these populations, adverse and inaccurate beliefs about the major methods persist, particularly in Kenya. This study provides useful guidance for counselling and informational campaigns. The generally negative views of past users imply that programmes may need not only to improve individual counselling but also strengthen community information campaign to change the overall climate of opinion which may have been influenced by dissatisfaction among past users.

Over the past half century, an extensive literature has been accumulated on women’s attitudes and beliefs concerning contraception in general and specific methods. The extensive evidence for low- and middle-income countries (LMICs) falls into four main categories. The most common category comprises studies that examine attitudes towards and beliefs about contraception in general, with little or no distinction between specific methods [1‐6]. The emphasis in many papers is on negative perceptions, often labelled as myths and misinformation. Commonly reported themes include the belief that use of modern methods will cause long term infertility, serious health damage, such as cancer and foetal abnormalities. In some studies, respondents associated contraception with promiscuity [2].

The second category concerns specific methods. Owing to the HIV pandemic, a huge literature on beliefs about condoms has been generated, much of which has been summarised by Maticka-Tyndale [7]. A review of perspectives on intra-uterine devices (IUDs) identified 14 studies from Africa, Asia and Latin America [8]. Positive features of this method included its high effectiveness and long-acting nature while commonly expressed concerns were health risks including cancer, ectopic pregnancy, infertility and harm to the husband during intercourse. A study in Kenya found that postpartum women preferred implants over IUDs because of less pain, less infringement of modesty and preference for a superficial insertion in the upper arm than in the uterus [9]. The literature on hormonal methods shows the dominant concerns to be side effects such as nausea, dizziness and weight change, menstrual disruption and infertility [10‐12]. In both Mali and Kenya, many women believed that oral contraceptives accumulate inside the body, causing infertility or a variety of diseases [13, 14].

The most influential, though indirect, body of evidence on views about contraceptive methods is based on self-reported reasons for non-use or discontinued use of contraception. The dominant source of data is the Demographic and Health Surveys (DHSs). The most recent analysis of reasons for non-use among married women not wishing to get pregnant found that, in most of the 52 countries studied, 20–33% cited side effects or health concerns. In 21 countries, this category of reason was the most common [15]. Moreover, in many countries, the majority of women giving this reason for contraceptive avoidance had previously used a modern method. Thus fear of side effects (and perhaps concerns about health) is partly based on personal experience and partly on hearsay evidence from friends or from media [16].

Reasons for discontinued use is more useful than reasons for non-use in distinguishing method-specific concerns. For oral contraceptives, injectables and IUDs, side effects or health concerns are the most common reason for stopping use. For condoms, objections from the husband or desire for a more effective method are the key reasons while for withdrawal and periodic abstinence accidental pregnancy dominates [17]. One limitation of this type of evidence is lack of comparable data for continuing users who perhaps are equally concerned about side effects or health risks as discontinued users but persist with use because of stronger motivation to avoid pregnancy or for some other reason.

The final main type of studies addresses the characteristics of contraceptive technology that women judge to be important. The single largest study of this type involved focus group discussions with 576 women in seven countries, six of which were LMICs [18]. Widespread agreement was found on the key importance of effectiveness and reversibility. Opinions varied by site on the desirability of a quick return to fertility after stopping the method and of the possibility of clandestine use. Most women found amenorrhea as a result of hormonal method use to be disturbing and similar results were found in Ghana and Nigeria [19, 20]. In a systematic review of contraceptive attributes that women take into account when choosing a method drawing on studies in USA and Europe as well as LMICs, ease of use, frequency of use, return to fertility, and side effects and health concerns were the most commonly mentioned [21].

Conspicuously and surprisingly lacking from the evidence-base for LMICs is detailed, quantitative documentation of the beliefs of women concerning all major methods of contraception. Our study addresses the omission by presenting descriptive survey data on method-specific beliefs in three populations. The data presented here form part of a wider prospective cohort study whose aims are to advance the understanding of reasons for unmet need for contraception and measurement of fertility preferences. The detailed rationale and study protocol may be found elsewhere [22]. In brief, its conceptual underpinnings posit that adverse method-specific beliefs constitute one cause of unmet need and unintended pregnancy, alongside four other factors (generic hostility to contraception, partner-related factors, weak or inconsistent fertility preferences and low perceived risk of pregnancy). The prospective design of the study will permit assessment of the relative power of these factors to predict method-specific adoption, continuation and pregnancy-incidence. Here, however, we present data from round one with the aim of documenting method-specific beliefs and of comparing the beliefs concerning oral contraceptives, injectables and implants of current, past and never users.

The survey was conducted in the icddr,b service area and the government service area of the Matlab Health and Demographic Surveillance System (HDSS), Bangladesh, the Nairobi Urban Health and Demographic Surveillance System (NUHDSS) in two slums in Nairobi, Kenya, and Homa Bay County in rural Western Kenya. Details of the three study population have been published [22], but, in summary, contraception has been well established in Matlab for decades and the total fertility rate (TFR) is 2.7 [23]. The population is predominantly Muslim. In Nairobi slums, the TFR is 3.5 and HIV prevalence is 8% in the adult population [24, 25]. In the two slums covered by the NUHDSS, the population is ethnically diverse, the majority belong to a Christian faith and mobility is high. Homa Bay’s population is predominantly rural, Christian and of Luo ethnicity. It has a high TFR of 5.2, high unmet need for contraception and a severe HIV epidemic with an infection level in the adult population of 26% [26]. All married or co-habiting women aged between 15 and 39 years residing in the three sites were eligible for the study. We set a target sample size for each site of 2600. The sample size was determined to detect a 30% difference in reproductive outcomes (pregnancy, use and non-use of contraceptives) at 95% confidence level and 80% power with an assumption of 10% non-response rate [27].

Data collection was carried out between August and December 2016. Respondents were randomly sampled from eligible female residents from the HDSS databases in the Matlab and Nairobi sites. In Homa Bay two-stage cluster sampling was used. A total of 34,308 women were identified as being married and aged between 15 and 39 years in the Matlab HDSS database. Out of these, 3109 were randomly selected and 2605 women completed the interviews. The Nairobi HDSS database identified a total of 5905 married or cohabiting women who were eligible for inclusion. Out of these, 3093 women were randomly sampled and interviews were completed with 2812 of the eligible women.

In Homa-Bay, we randomly selected 12 sub-locations (the smallest administrative unit in Kenya) in each three purposely identified sub-counties in Homa Bay County. All households with currently married or co-habiting women aged 15–49 years in the selected sub-locations were identified with the help of the local administration. Subsequently, the sampling frame was generated by listing all individuals in these households. In the second stage, 3118 were randomly sampled out of 5424 eligible women. A total of 2424 women completed the interviews.

All respondents provided written consents after having been informed about the objectives, procedures, benefits and risks of the study using an information sheet and informed consent form. All married and co-habiting adolescents aged 15–17 years were considered emancipated minors for whom parental permission is not required.

The questionnaire was developed through literature review and consultation with experts. We reviewed existing literature and more than 30 questionnaires on fertility preferences and reasons for non-use of family planning fielded in high-, middle- and low-income countries including instruments from the DHS, the Determinants of Unintended Pregnancy Risk study in New Orleans; the US- based National Survey of Family Growth (NSFG); and the Fog Zone study by the Guttmacher Institute. A questionnaire was developed and reviewed through consultation with dozens of experts in the field. The questionnaire was further refined after pre-testing.

Structured interviews, lasting on average for 45–60 min, were conducted using the local languages following a 7 to 10-day training. Very similar questionnaires were used in all three sites covering the following topics: background, reproduction, contraceptive knowledge and use, beliefs and attitudes towards contraception in general and specific methods and fertility preferences of the women. Two summary measures of method-specific beliefs were generated. A net positive score is the mean percent difference between positive and negative responses across nine women’s attributes, ignoring “don’t know” and perceived husband’s approval. Positive responses refer to perceptions that a method is easy to obtain and to use, very effective in preventing pregnancy and safe to use for a long time without a break, does not cause health problems, unpleasant side effects, menstrual disruption or infertility, and that half or more of friends, relatives, neighbours (i.e. social network) had used the method and their experiences were satisfactory. A familiarity measure which is simply the mean percent responding “don’t know” across eight attributes excluding experiences of method use among women’s social network and perceived husband’s approval. In addition, past and current users of a method were asked whether they were satisfied or dissatisfied with use.

Descriptive analysis and chi-square tests were carried out to assess differences in beliefs and attitudes among current, past and never users of pills, injectables and implants. Clustering was taken into account for the statistical analyses when analysing the data from Homa-Bay. We used STATA SE version 15.0 for the data analysis.

Ethical approvals for this study were obtained from the Institutional Review Boards of the countries and three participating institutions.

The number of women successfully interviewed was 2424 in Homa Bay, 2605 in Matlab, and 2812 in Nairobi (Additional file 1: Table S1). The average ages of women in the three sites were closely similar, 28 or 29 years (Table 1). About 70% in Matlab had secondary or higher schooling, compared with 40% in Nairobi and 23% in Homa Bay. The mean number of surviving children was lowest in Matlab, highest in Homa Bay and intermediate in Nairobi. Women in Matlab were much more likely to express a desire to stop childbearing than those in the Kenyan sites. In Nairobi, 75% were currently using a method of contraception, with injectables followed by implants, as the dominant methods. In Homa Bay, the level of current use was 65%, with a method-mix similar to that in Nairobi. Current use was lowest in Matlab (59%), where oral contraception pills and injectables were the favoured methods. But this unexpected result, in view of the far lower fertility rate in Matlab, may largely reflect the fact that 34% of husbands had been away for three months or longer preceding the date of interview, mainly as migrant workers, thus reducing the need for pregnancy-protection.

In view of the principle of cognitive dissonance—the tendency to align beliefs and attitudes to behaviour—it is no surprise that most current users of a method consider it to be satisfactory and express more positive views than past or never users. In apparent contradiction to high overall satisfaction is the finding that appreciable proportions of these current users appear to be concerned about serious health effects and acknowledge that the method causes unpleasant side effects. The view that it is unsafe to use a method for a long time without taking a break is widespread among current users, particularly in the Kenyan sites. Follow-up surveys will establish the impact of these perceptions on discontinuation.

Dissatisfaction among past users, who tend to outnumber current users, may be a powerful influence on the overall climate of opinion; indeed the views of past and never users tend to be similar. The low level of overall satisfaction among past users of injectables, pills and implants in Kenyan sites, associated with health worries and side effects, is a concern, particularly with implant users of whom only about 30% in Nairobi reported satisfaction. A continuation of the steep rise in the use of implants in Kenya, therefore, could be in jeopardy. Similarly in Matlab, the results imply considerable potential resistance to long-acting reversible methods (IUDs and implants), the promotion of which is a national priority.

The study is not without limitations. Information on perceived partner’s approval of specific methods is difficult to interpret. While close to 90% reported partner’s support of family planning in general, perceived approval is low in Matlab for use of implants, IUDs and female sterilisation, and also low in the Kenyan sites for use of IUDs and sterilisation. Whether or not, these results reflect women’s own misgivings about these methods remains unclear.

The value of this detailed information in explaining reproductive and contraceptive behaviour is uncertain and a verdict must await the results of follow-up surveys. Nevertheless, the results are of considerable interest because this study, to our knowledge, is the first to have obtained comparable quantitative information in LMICs on attitudes and beliefs on each of eight major family planning methods. Many of the findings from this study are consistent with earlier studies that have documented fears about side effects and damage to health to be widespread. The result regarding the belief that it is unsafe to use a method for a long term without taking a break is simiar to the common behaviours observed in the Philippines [28]. Other findings are new and have considerable possible implications for behaviour.

Moreover, the ability of this study to contrast the beliefs of current, past and never users of the three more prevalent methods represents a major contribution. In 2015 an expert meeting on misperceptions about contraceptives called for a revitalization of discussion on this issue. Participants identified the need for quantitative data that distinguish concerns stemming from documented side effects from those resulting from rumours or myths and the need for evidence about links between misperceptions and method choice [29]. This paper starts to address these gaps in evidence and further progress is expected when follow-up survey data become available.

In conclusion, high levels of contraceptive use can clearly co-exist with widespread misgivings about methods, even those that are widely used. Serious concerns about damage to health, long term fertility impairment, and dangers of prolonged use without taking a break were particularly common in the Kenyan sites and these beliefs may explain the high levels of discontinuation observed in Kenya and elsewhere in Africa. This documentation of beliefs provides useful guidance for counselling and informational campaigns. The generally negative views of past users imply that programmes may need not only to improve individual counselling but also strengthen community information campaign to change the overall climate of opinion which may have been influenced by dissatisfaction among past users.