The use of traditional medicine in maternity care among African women in Africa and the diaspora: a systematic review

The search included the following eight databases: CINAHL, EMBASE, Infomit, Ovid Medline, ProQuest, PsychINFO, PubMed and SCOPUS and was conducted from the inception of each database until 31 December 2016. The search terms employed were the same for all databases. A detailed search strategy was developed by author (ZS) with input from authors (TD) and (CS). The search strategy combined terms for: (i) women or females, (ii) African or African-born migrants, (iii) traditional medicine, and (iv) maternal health conditions. All possible synonyms of these terms were listed and combined using Boolean operators (see Additional file 1 for more details). The reference lists of all included articles were also checked for other relevant studies.

Studies were included if they reported use of TM by African women or African-born migrant women for reasons related to the preparation for pregnancy, promoting fertility, treating pregnancy related symptoms, maintenance of general wellbeing during pregnancy, inducing or assisting labour, terminating pregnancy (abortion), and enhancing milk secretion or postnatal wellbeing. Studies were also included if describing the views, attitudes or beliefs of women towards TM. The search strategy included primary research (quantitative, qualitative and mixed methods), grey literature and unpublished reports.

Studies were excluded if they focused on women’s use of TM for general purposes and other conditions which were not directly related to maternal health care (e.g. postmenopausal symptoms, breast cancer and prevention of mother to child transmission of HIV). Studies that reported combined use of TM and pharmaceuticals were excluded if the data on TM could not be separated. Ethno-botanical surveys were also excluded.

Author ZS conducted the search from November to December 2016. A step-by-step review strategy was implemented to identify all relevant studies. Studies retrieved by the search were assessed first by title and then by abstracts by author ZS. This was followed by reading full texts to identify the final studies for inclusion. Data was extracted according to a predefined reference by all authors (ZS, TD and CS) with disagreement resolved through discussion. The data extracted covered the country of studies, participants’ demographics, prevalence of TM use, details of TMs used, characteristics of TM users, maternal conditions treated by TM, reasons of use, source of information, disclosure of TM use to health professionals, and the method of data collection. All search results were imported into Endnote, a bibliographic management software system and analysed.

There is no agreed set of methods for assessing the quality of observational studies describing CAM use [23‐25]. Bishop and colleagues recently developed a Quality Assessment Tool (QAT) for a systematic review of the prevalence of complementary medicine use in pediatric cancer [23]. The QAT has also been modified by Grant and colleagues for a systematic review on use of CAM by people with cardiovascular diseases [25]. We have used the modified version of the QAT to undertake quality appraisal of the 18 quantitative studies (two qualitative studies were excluded from appraisal as QAT is not designed to assess quality of such studies). We were also unable to carry out a separate quality analysis for these articles because the number of studies was too small to allow us to draw firm conclusions. Author ZS and CS assessed the quality of studies with disagreements resolved by discussion.

A narrative synthesis of studies was undertaken. Data such as prevalence rates were analysed and grouped together for comparison between studies and/or countries. Quality assessment scores were calculated.

The database search identified 1949 potential references, from which 488 duplicates were removed. A total of 92 references were reviewed by abstract and 59 were not directly related to the aim of the review. The remaining 33 articles were examined by full text and a total of 20 articles were included [7, 8, 26‐43]. Thirteen articles which examined an ethno-botanical survey of herbs used in pregnancy, focused on use of both TM and pharmaceuticals, reported the use of TM for treatment of HIV and cancer were excluded (Fig. 1).

Of the 20 studies included, four were conducted in Nigeria [28, 30, 40, 41], three in South Africa [8, 33, 43], and two each in Ethiopia [27, 32], Uganda [31, 37] and Tanzania [39, 42]. The remaining seven studies were conducted in Zimbabwe [35], Zambia [26], Mali [36], Lesotho [34], Kenya [7], Morocco [29], and Egypt [38]. No study was found related to maternal use of TM or CAM among African women of the diaspora.

Studies were published between 1985 and 2016, from which 13 (65%) were published between 2014 and 2016. Of the selected articles, 17 studies were quantitative while two were qualitative. One study utilised a mixed research method which included a structured questionnaire survey with pregnant women and focus group discussions involving TM providers. From this study, the data that concerned only the women was extracted in keeping with the aim of this review.

Twenty studies investigated the use and/or the perception of TM for various maternal health conditions among 11, 858 African women. The sample size of the quantitative studies ranged from 72 to 5686 participants. Five studies included a sample of 500 or more participants [26, 30, 33, 40, 41]. Thirteen (72.2%) studies included pregnant women who were currently attending health facilities [26‐28, 30, 32‐38, 41, 42]. Five studies focused on surveying women who were mothers or had been pregnant in the past 2–5 years [7, 8, 29, 40, 43] preceding their data collection. Other studies sampled nursing mothers [28], women attending a fertility clinic [31], and women who underwent unsafe abortion [39]. Most studies reported on women of childbearing age (18–45 years).

As shown in Table 1, the modified QAT included four major assessment criteria: i) study methodology, ii) sampling, iii) participant characteristics, and iv) TM use. A total of 15 specific quality assessment criteria were weighted according to their relative importance as described by Bishop and colleagues [23]. Three items scored a maximum of 2 points, 8 items scored 1 point, and 4 items scored 0.5 points. The maximum score was 16.

Study quality varied significantly between studies with QAT percentage scores ranging from 25% to 59.4%. Fifteen studies attained less than 50% of the maximum score [7, 26‐29, 32‐34, 36‐42] (Table 2). All quantitative studies used retrospective data collection methods in which there was a potential for risk of recall bias. Only five studies used a retrospective data collection within the 12 months to minimise the risk of recall bias [7, 26‐28, 31]. Eleven studies (61.1%) collected data with a piloted questionnaire, while five (27.8%) studies adjusted for potential confounders in their analysis. Seven studies reported a response rate which ranged between 74% and 100%. Only three studies recruited a multicenter sample in an attempt to achieve a representative sample of participants to the larger population from which they were drawn. Fifteen studies collected socio-economic data and 16 studies reported the age of participants. Only five studies reported the ethnicity of the study participants, most studies were from regions or countries of homogenous populations.

Three studies reported they provided a definition of TM to study participants, although the definition was not stated in the manuscripts (8, 12, 14). Two articles provided an operational definition of TM for their study (13, 20). While all studies reported the prevalence of use of TM, only three studies reported the frequency or duration TM use.

In 18 quantitative studies, prevalence of TM use for maternal and reproductive health issues ranged from 12% [7] to 79.9% [36]. In 15 studies, the TM modality investigated was limited to herbal medicine only, 11 of these studies focused on the use of herbal medicine during pregnancy [27, 28, 30, 32‐34, 36‐38, 41, 42]. The prevalence of herbal medicine use among pregnant women in these 11 studies ranged from 20% [37] to 79.9% [36] (Fig. 2). A study in rural Tanzania reported a 43% prevalence rate of medicinal herb usage to induce abortion [39]. Another study from Uganda that examined herbal medicine use among women attending the infertility clinic reported a 76.2% prevalence rate.

Herbal medicines were mostly investigated (Table 3). Ten studies reported the specific types of herbal preparations commonly used by respondents [8, 27‐29, 32, 36, 38, 39, 41, 43]. Some of these studies used names of herbs in local languages while others used either English or Latin names of specific medicinal herbs. Frequently reported herbal medicines included ginger, garlic, aniseed, fenugreek, green tea, peppermint, eucalyptus, rue, garden cress, madder, cinnamon, bitter leaf, palm kernel, bitter kola, neem leaves, and jute leaves. Some of the medicinal herbs reported by their scientific names were Bidens pilosa, Commelina africana, Desmodium barbatum, Manihot esculenta, Ocimum suave and Sphaerogyne latifolia., Obetia radula, Rubia cordifolia, and Triumfetta microphylla, Lippia Chevalieri, Combretum micranthum, Parkia biglobosa, Peris heterophylla, Stylosanthes erecta, Ximenia Americana, Mitragyana inermis,and Combretum glutinosum [36, 39]. In one study, the use of different types of TM and other substances by pregnant women was reported [35]. These included holy water, soil from a burrowing mole, pouzolzia mixta, elephant dung, manual exercises, cocktails of unknown herbs, Cannabis stivum, castor oil, rooibos tea, and hot water/steam baths.

Lack of access to western medicine was reported to be the major determining factor that influenced women to use TM [7, 30, 36, 42]. Other determinants included user’s belief that TM was more effective than western medicine, as well as better accessibility and lower cost of TM.

A study from Nigeria reported significantly increased use of TM among pregnant women aged 20–30 years, with no formal education, a good income (those earning above 250 dollars monthly) and who were married [28]. In this study, gestational age, parity, ethnicity and occupation did not affect the use of TM. Another study identified that married women, women within the first three years of infertility diagnosis, and those never conceived were more likely to use TM [31]. The authors also noted that the majority of the participants sought biomedical care for infertility after three years of being unable to achieve a pregnancy. In most of the African culture, the main purpose of marriage is procreation, and if a woman fails to conceive soon after marriage, they seek help from traditional healers [31, 44].

A study form Ethiopia reported that during the first trimester of pregnancy, being less educated and having less knowledge about TM were associated with higher use of TM [32]. A Ugandan study found that women who used herbal medicine in the past were eight times more likely to use during the current pregnancy [37]. In two studies, a long distance to health facilities was found to be associated with increased herbal medicine usage during pregnancy [7, 37]. In one study which investigated the use of TM during childbirth, socioeconomic factors were not significantly associated with use of TM [40]. Two studies found multiparous women were more likely to use TM than first time mothers [29, 34]. In a Zambian study, users were not different from non-users in terms of age, education, ethnicity or income [26]. In this study, women who knew someone who had used TM during pregnancy were more likely to use TM, and women who thought that the use of TM may hurt their baby were less likely to use TM.

Fourteen studies reported reasons for TM use [7, 27, 29‐32, 34‐40, 42]. Some of the reasons were related to treating pregnancy related conditions including nausea, vomiting, nutritional deficiency, swollen feet, back pain, digestive problems, fever, cold, abdominal pain, edema, urinary tract infection, tiredness, headache, and waist pain [7, 27, 32, 36, 37] (Table 3). One study reported that TM was frequently used to get back in shape after delivery, facilitate child birth, increase breast milk secretion and reduce pain during pregnancy [29].

Two studies reported that the main reasons for TM use were for prevention of perineal tearing, placenta retention, breech presentation, postpartum hemorrhage, prolonged labor, preeclampsia, abortion and pains [34, 35]. Other reasons for TM use included general wellbeing during pregnancy, promotion of fetal growth, spiritual cleansing, to protect the pregnancy against evil influence, to have a male baby, for induction of labour, assisting childbirth, and as dietary supplements [34‐37, 40, 42].

Two qualitative studies from South Africa reported the constructed meaning, knowledge, and beliefs of pregnant women and mothers towards specific types of traditional health practices. Kooi and Theobland examined the women’s belief about kgaba remedy, a traditional therapy based on a mixture of plants and minerals that can vary among traditional healers and has not been officially documented [8]. Authors noted that Kgaba may contain mixture of up to 18 different medicinal herbs and can be prepared by combining these herbal remedies with ostrich eggshell, baboon urine, mud, and ashes of burnt herbs. They concluded that kgaba was an important component in the experience of pregnancy and labour, mainly to protect pregnant woman from evil and harm, and to induce labour. Similarly, Naidu examined the constructed meaning of isihlambezo, a herbal mixture used by many Zulu women in South Africa as a preventative health tonic during pregnancy [43]. The author reported that pregnant women perceive isihlambezo as a powerful medicine given by God and the ancestors to protect the mother and her unborn baby. The strong belief in the efficacy of isihlambezo was echoed in the understanding of many of the participants.

The source of information on TM was not addressed in the majority of studies. Four studies identified the sources of information on TM were from families, relatives, friends, traditional birth attendants, and local herb sellers/herbalists [27, 30, 32, 41].

Three studies reported the concurrent use of TM with prescription drugs [7, 30, 41]. In one study, although respondents used prescribed pharmaceuticals and herbal medicine concomitantly, few users disclosed their use to a healthcare professional [7]. In this study, when use of herbal medicine was disclosed, participants reported that the healthcare providers provided advice about side effects, or discouraged the use. One study found 40% of those using herbal medicine combined herbs with pharmaceutical drugs during pregnancy [41]. In another study, 13.7% of herbal medicine users reported that they support the use of TM combined with prescription medications during pregnancy [30].

Four studies reported data on disclosure of TM use to health care providers [7, 26, 31, 37]. In one study, 64% of users did not want to share their use of TM to clinicians. Reasons given for nondisclosure to physicians about TM use included fear that it might negatively impact their antenatal care [26]. Another study also reported 63.8% of users did not disclose TM use to attending physicians although their reason was not examined [31], whereas another study indicated 89.7% of respondents were not interested in disclosing use of herbal medicine to the healthcare providers [37].

This is the first comprehensive review to date of studies reporting the use of TM for maternal health among African women, and it increases our understanding of African women’s traditional health practices and highlights an extensive use of TM. A rigorous search strategy was conducted of the English language literature identifying published articles, grey literature and unpublished studies to minimise the potential for publication bias. However, the studies included focused primarily on use of herbal medicine during pregnancy, and therefore limit our understanding about other African traditional and cultural health practices. Lack of studies among the African diaspora also limits the generalizability of this finding to migrant African women.

Findings of this review may help policy makers as well as conventional and traditional health providers understand the traditional health practices and beliefs of African women. Identification of these traditional practices and the reasons why women choose to use them may support the development of strategies to improve maternal and reproductive health services for African women. It may also lead to better regulation of TM practices and the identification of gaps for future research.

Although TM is a primary source of health care for more than 80% of the African population, there is little information regarding the types of African TM, users’ profile, and reasons for and determinants of use. There is also an absence of research examining TM use among African-born migrant women and how their previous cultural health practices and beliefs may influence their health seeking behavior. Research that includes a broader operational definition of TM is needed to fully understand the traditional and cultural health practices and beliefs of African women in Africa and the diaspora. The reasons and determinants for higher reliance on TM among Africa women also need further investigation.