About Bioline  All Journals  Testimonials  Membership  News  Donations

The Journal of Health, Population and Nutrition
ISSN: 1606-0997 EISSN: 2072-1315
Vol. 24, Num. 4, 2006, pp. 472 - 478

Journal of Health, Population and Nutrition, Vol. 24, No. 4, Dec, 2006, pp. 472 - 478

Transition to Skilled Birth Attendance: Is There a Future Role for Trained Traditional Birth Attendants?

Lynn M. Sibley, Theresa Ann Sipe

Center for Research on Maternal and Newborn Survival, Emory University, 1520 Clifton Road, Suite 436, Atlanta, GA 30322, USA and Rollins School of Public Health, Behavioral Science and Health Education, Emory University,1518 Clifton Road, Atlanta, GA 30322
Correspondence and reprint requests should be addressed to: Dr. Lynn M. Sibley Associate Professor Center for Research on Maternal and Newborn Survival Emory University 1520 Clifton Road, Suite 436 Atlanta, Georgia 30322 USA Email: Fax:404-727-9676

Code Number: hn06056


A brief history of training of traditional birth attendants (TBAs), summary of evidence for effectiveness of TBA training, and consideration of the future role of trained TBAs in an environment that empha­sizes transition to skilled birth attendance are provided. Evidence of the effectiveness of TBA training, based on 60 studies and standard meta-analytic procedures, includes moderate-to-large improvements in behaviours of TBAs relating to selected intrapartum and postnatal care practices, small significant increases in women’s use of antenatal care and emergency obstetric care, and small significant decreas­es in perinatal mortality and neonatal mortality due to birth asphyxia and pneumonia. Such findings are consistent with the historical focus of TBA training on extending the reach of primary healthcare and a few programmes that have included home-based management of complications of births and the new­borns, such as birth asphyxia and pneumonia. Evidence suggests that, in settings characterized by high mortality and weak health systems, trained TBAs can contribute to the Millennium Development Goal 4—a two-thirds reduction in the rate of mortality of children aged less than 14 years by 2015—through participation in key evidence-based interventions. 

Key words: Traditional birth attendants; Skilled birth attendants; Meta-analysis;  Maternal mortality; Perinatal mortality; Training


Every year, an estimated four million babies die in the first four weeks of life, and a similar number of babies are stillborn. Moreover, half a million women die from pregnancy-related causes. Most of these deaths occur at home during the first postnatal or postpartum week, especially within the first 24 hours of birth. Most oc­cur in developing countries with weak or failing health systems, with South Asia and sub-Saharan Africa being especially hard-hit (1-3). For more than three decades, the World Health Organization (WHO) and other agen­cies of the United Nations promoted training of tradi­tional birth attendants (TBAs) as a global public-health strategy to reduce this tragic loss of life (4). Lack of evidence to demonstrate that trained TBAs can reduce maternal mortality led to controversy over their train­ing in relation to safe motherhood and a policy shift to skilled birth attendance (5-7). In this paper, we provid­ed a brief history of TBA training, reviewed evidence of the effectiveness of TBA training, and considered the future role of trained TBAs in an environment that emphasizes the transition to skilled birth attendance.            

Although TBAs have been trained since the late 1800s, important milestones over the last century illus­trate the shifting policies on TBA training as a global public-health strategy. Credit for the first formal train­ing programme is usually given to a British missionary midwife, Miss M.E. Wolfe, working in Sudan in 1921 (8,9). The Inter-Governmental Conference of Far-East­ern countries, held in Bangkok in 1937, called for the integration of TBAs into rural health programmes. By 1952, the United Nations Children’s Fund (UNICEF) began to supply trained TBAs with delivery-kits. The goal of these early programmes was to improve perina­tal healthcare (8). Nearly 20 years later, interest in pri­mary healthcare and in traditional medicine in relation to primary healthcare had grown to the extent that the UNICEF and WHO sponsored a technical consultation on TBA training. By the time of the 1978 Alma Ata Declaration, the WHO was fully in support of training TBAs to extend the reach of primary healthcare serv­ices. At that time, the WHO recommended that trained TBAs work side-by-side in ‘articulation’ with the mod­ern health system, so that the informal traditional and formal modern health systems could presumably co-exist without conflict. The success of the WHO’s en­couragement can be measured by the rapid increase in the number of countries undertaking TBA training. For example, in 1972, only 20 countries had TBA training programmes. It is now estimated that 85% of developing countries have some form of TBA training (8). With the advent of the safe motherhood initiative and without evidence to show that the risk approach and trained TBAs can reduce maternal mortality, there has been a gradual waning of enthusiasm for TBAs. In 1992, the WHO emphasized that, if TBAs were going to contribute to safe motherhood, they must be ‘inte­grated’ into the modern health system through training, supervision, and technical support (4). However, by 1997, the WHO and many safe motherhood advocates turned from TBA training to promote skilled birth at­tendance for all (8), most recently calling for a ‘new’ and ‘expanded role’ for TBAs, where TBAs act as ‘link workers’ to skilled birth attendants rather than as pri­mary care providers (10,11).

The broad goals of TBA training programmes are to reduce maternal and child mortality and morbidity and to improve the reproductive health of women. The objectives include: enhancing the linkages between the modern health system and community, increasing the number of TBA-attended births, and improving the skills and stature of TBAs (4). Training programmes vary, however, in addressing these objectives. For ex­ample, individuals, non-governmental organizations, and missions have trained TBAs through the private sector and also through local, state and national gov­ernment, and international agencies have trained them through the public sector. Training programmes may last from several days to several months and may in­clude clinical practice at a health facility, follow-up su­pervision, and continuing education (12).

The content of curricula of TBA training also var­ies. Most TBAs have been trained to upgrade their skills so as to be able to perform safe deliveries. Con­sistent with the emphasis on extending the reach of pri­mary healthcare, many TBAs have also been trained to take on the expanded functions of prevention, screen­ing, and referral (12,13). Very few programmes have included content on initial response and stabilization of maternal and newborn complications (e.g. resuscitation of newborns or detection and management of sepsis), and content that would be very useful to those who are confronted with emergencies and that might save lives (13).

The answer to the question: is TBA training effec­tive? depends on the objectives and content of a particular training programme and the outcomes measured. From 1997 to 2002, we conducted a meta-analysis of the available published and unpublished studies on the effectiveness of TBA training. We updated this review for selected topics in 2003 and 2004 (13-15). The goal of the meta-analysis was to provide information that would inform policy decisions about future TBA train­ing and about the needs of evaluation research. Specific objectives were (a) to describe the effect of training on TBA and maternal attributes, such as knowledge and behaviour, (b) to determine the impact on pregnancy outcomes, and (c) to describe the quality of the litera­ture on evaluation.

Materials and Methods

Details of materials, methods, and statistical analy­ses have been described elsewhere (13-15). Briefly, a broad search yielded nearly 1,200 documents, of which 60 studies met the following eligibility criteria: the intervention was TBA training; intervention group data were derived from trained TBAs (reference to intervention), and mothers and neonates whose care was provided by trained TBAs or who were living in areas where more than 50% of births were attended by trained TBAs (a proxy for exposure); comparison group data were available; dependent measures were related to knowledge, attitudes, behaviour, or maternal and perineonatal health outcomes; documents were in English and published during January 1970–June 1999; research design was either experimental or qua­si-experimental; and data were sufficient to calculate an effect size. The sample of studies, representing three regions of the world—Asia, Africa, and Latin America/Caribbean—contained 1,695 outcomes. TBA and ma­ternal attributes comprised 95% of the outcomes, and pregnancy outcomes comprised the remainder. The attribute outcomes were independently sorted into 24 maternal and child health content areas for sub-group analysis. In 2002, we repeated the search for studies having outcomes relating to antenatal care and refer­ral for obstetric complications. In 2004, we repeated the search for studies having outcomes relating to safe delivery and newborn-care practices. Altogether, these update searches yielded two new studies plus three documents that were more recent versions of studies already included in the original sample.

Most outcomes measured were reported as percent­ages. The percent difference associated with each out­come was converted to a common scale using the effect size index Cohen’s h, representing the standardized difference between the trained and the untrained TBA group on the outcome of interest. The variance-weight­ed mean effect size for each sub-group of outcomes was then calculated, and homogeneity tests were per­formed on the distributions of the weighted mean effect sizes. With a few exceptions, homogeneity of variance was rejected (α=<0.05), and the weighted mean effect size and 95% confidence interval were calculated us­ing formulae based on a random effects model. Sen­sitivity analysis was conducted to detect the presence of publication bias. In addition, stratified analyses, by study design and sampling procedure, were performed to examine the influence of these potential moderating variables on the weighted mean effect sizes. Lastly, sub-group analyses conducted to date include antenatal care, referral for emergency obstetric care, birthing and newborn-care practices, perinatal and newborn health outcomes.


The findings of the meta-analysis were positive for the sub-groups of outcomes measured (Table 1). Strati­fying each sub-group analysis by the study design and sampling procedure did not change the overall conclu­sions. The medium-to-large effect size values for be­haviour of TBAs relating to intrapartum, including safe delivery, clean delivery, and cord-care practices, represented significant increases of 44%, 103%, and 53% respectively for the trained TBA group over the untrained TBA group at baseline. Likewise, the me­dium-to-large effect size values for behaviour of TBAs relating to counselling on maternal nutrition, early exclusive breastfeeding, and immunization (primarily TT) represented significant increases (15%). The mod­erate-to-large effect sizes for knowledge and behaviour of TBAs relating to antenatal care—positive medium-to-large values—were also significant and represented 177% and 47% increases respectively for the trained TBA group over the untrained TBA group at baseline, as was the small effect size for maternal behaviour relating to use of antenatal care services—a 38% in­crease for women cared for by trained TBAs (14). The small effect sizes for both behaviour of TBAs and ma­ternal behaviour relating to referral of obstetric compli­cations were significant and represented 36% and 22% increases respectively for the trained TBA group over the untrained TBA group at baseline (15). However, the large standard errors and wide confidence intervals revealed the variability in these TBA and maternal attribute data. Lastly, the effect size value for perina­tal mortality represented a significant 8% decrease in mortality associated with TBA training (13). Although the number of studies included in the latter sub-analy­sis was small, the effect size values for birth asphyxia and mortality due to pneumonia represented 11% and 8% decreases respectively in cause-specific newborn mortality (13). Small standard errors and narrow con­fidence intervals highlighted the homogeneity of these outcome data.


TBA training is associated with moderate-to-large improvements in behaviours relating to selected in­trapartum and postnatal care practices, and small but significant decreases in perinatal mortality and neona­tal mortality due to birth asphyxia and pneumonia. The findings are consistent with the historical focus of TBA training on extending the reach of primary healthcare and a few TBA training programmes that have includ­ed management of birth asphyxia and pneumonia care. TBA training is also associated with small but signifi­cant increases in women’s use of antenatal care and emergency obstetric care, suggesting that referral is a multi-faceted process, and influence of TBAs on care-seeking behaviour may be limited.

Unfortunately, incomplete reporting by authors re­sulted in loss of studies and datasets and limited our ability to examine which intervention characteristics were associated with better outcomes, cost-effective­ness, and the association between training and maternal mortality—the rationale for the policy shift away from TBA training. Moreover, due to the variable quality of studies included in our sample, we could not address causality and only the magnitude and direction of the association between training and the outcomes could be measured. More rigorous studies are now being con­ducted. In a recent cluster-randomized controlled trial from Pakistan, Jokhio and colleagues showed that TBA training, linked to outreach and facility-based care, re­sulted in a statistically significant reduction of 30% in perinatal mortality (16). The estimated percent reduc­tion was similar for maternal mortality, but it was not significant.

Millennium Development Goal (MDG) 4 aims to reduce mortality of children, aged less than five years, by two-thirds between 1990 and 2015. MDG 5 aims to reduce maternal mortality by three quarters during this timeframe. Policy regarding the best strategies to meet these goals, however, swings between com­munity-based care and facility-based care, while safe motherhood and child-survival advocates compete for limited resources (2-3). The findings of the meta-analysis suggest that trained TBAs can contribute to MDG 4 through participation in key evidence-based interven­tions (Fig. 1). For example, at 90% coverage, universal, extra, and situational interventions would result in a re­duction of 15-32% in newborn mortality while improv­ing maternal health (1). There are no comparable data for maternal mortality. Empirical evidence nonetheless suggests that hydration, keeping the bladder empty dur­ing labour, and stimulating the uterus to contract after delivery reduces postpartum haemorrhage due to uter­ine atony. Thus, there is an urgent need for research on these and on other community-based interventions that may reduce postpartum haemorrhage, a leading cause of maternal death (Fig. 2).

In settings characterized by high mortality and weak healthcare systems, family and community care providers often include a mix of women, family mem­bers, TBAs, and other semi-skilled or skilled health workers. In these circumstances, “Are TBAs the best choice, among others, for training in health promotion and delivery care?” There are a number of considera­tions. One important consideration is the proportion of deliveries assisted by TBAs since this reflects, to a large extent, their number, distribution, role, status, presence of other providers, and preference of women and fami­lies. Our analysis of the 1994-2004 Demographic and Health Survey data from 44 countries for the period of three years preceding the survey (17) showed that the proportion of deliveries assisted by TBAs is extremely variable within and across countries, being highest for the rural areas. The proportion, however, is relative­ly low on average and is comparable to that of deliver­ies assisted by family members and no one combined (Table 2). It is, thus, essential to assess healthcare use-patterns, existing infrastructure, and availability of providers in local, regional and national settings for inputs into policy decision-making. We suggest that, where coverage is high, training TBAs to provide key evidence-based interventions and first-aid care for se­lected complications is a viable short-term strategy.

The authors of “Newborn health: a key to child survival” in a 2005 special issue of The Lancet—eloquently argued that women, newborns, and children all have rights, and all would greatly benefit from a health sys­tem that delivers proven interventions through a con­tinuum-of-care approach, i.e. from pregnancy through labour, birth, and the postpartum and postnatal periods, and into early childhood; and including home, periph­eral health facility, and hospital (1,18). They contended that, with sufficient coverage, affordable low-tech ap­proaches could reduce the number of newborn deaths by up to one-third and prevent over half of child deaths. Thus, they urge a balanced, phased approach to policy and programming such that service modalities of family-community care and outreach save lives now, especially among the poor, while the health system is simultaneously strengthened, and facility-based clini­cal care is made available and equitable, in time. To achieve even greater reductions in newborn, and early child mortality and to reduce maternal mortality, they emphasized that facility-based clinical care, including skilled birth attendance and an intact referral system, i.e. a functioning health system, is required. 

Given an environment of scarce resources, there is an understandable concern among safe motherhood advocates that investment in family-community care approaches will set-back advances made with respect to increasing skilled birth attendance and emergency obstetric care. We agree with the Millennium Task Force on Child Health and Maternal Health that, “A strategy designed to address maternal mortality as its true aim—and not just some welcome, but coincidental by product of a health intervention designed primarily for another purpose (averting newborn death, for ex­ample)—must include interventions that prevent and treat complications that kill women” (2). We believe that a balanced, phased approach encompasses such interventions and that, realistically, there is no ethical alternative. To improve pregnancy outcomes, there is a critical need, not only to upgrade health facilities and train, strategically deploy, and retain professional care providers, but also to evaluate, refine, and disseminate promising community-based approaches to care during birth and the immediate postpartum and postnatal peri­ods.


The meta-analysis of effectiveness of TBA training was made possible through funding support from the World Bank (Small Grants Program for Safe Mother­hood), USAID (Intra Health/PRIME I Project, Acade­my for Educational Development/SARA Project), US­AID, DFID and the Bill and Melinda Gates Foundation (University of Aberdeen/IMMPACT Project), and the Bill and Melinda Gates Foundation (Save the Children/Saving Newborn Lives Project). The views presented here represent those of the authors.


  1. Darmstadt GL, Bhutta ZA, Cousens S, Adam T, Walker N, de Bernis L (Lancet Neonatal Survival Steering Team). Evidence-based, cost-effective interventions: how many newborn babies can we save? Lancet 2005;365:977-88.
  2. Freedman LP, Waldman RJ, de Pinho H, Wirth ME, Chowdhury AMR, Rosenfield A. Who’s got the power? Transforming health systems for wom­en and children. London: Earthscan Publications, 2005. 185 p.
  3. World Health Organization. The world health re­port 2005: make every mother and child count. Ge­neva: World Health Organization, 2005. 243 p.
  4. World Health Organization. Traditional birth attendants: a joint WHO/UNICEF/UNFPA state­ment. Geneva: World Health Organization, 1992. 18 p.
  5. Maine D. Safe motherhood programs: options and issues. New York, NY: Center for Population and Family Health, Faculty of Medicine, School of Public Health, Prevention of Maternal Mortality, Columbia University, 1993. 36 p.
  6. Starrs A. The safe motherhood action agenda: priorities for the next decade. Report on the Safe Motherhood Technical Consultation, 18-23 October 1997, Colombo, Sri Lanka. New York, NY: Family Care International, 1998. 30 p.
  7. Bergstrom S, Goodburn E. The role of traditional birth attendants in the reduction of maternal mor­tality. In: De Brouwere V, Lerberghe VW, editors. Safe motherhood strategies: a review of the evi­dence. Antwerp: ITG Press, 2001:451. (Studies in health services organization and policy, no. 17).
  8. Fleming JR. What in the world is being done about TBAs? An overview of international and national attitudes to traditional birth attendants. Midwifery 1994;10:142-7.
  9. Genesis, developing access to women’s history sources in the British Isles. Durham University Li­brary. (, accessed on 16 December 2005).
  10. Tarnpol PW. Maternal survival: improving access to skilled care. A behavior change approach. The Change Project. Washington, DC: Academy for Edu-cational Development, 2005. 9 p.
  11. United Nations Fund for Population Activities. Consensus and concerns: Regional Workshop on Skilled Birth Attendants in South and West Asia. Is­lamabad: UNFPA Country Office, Pakistan, 2004. 38 p.
  12. Fortney JA, Smith JB. Training of traditional birth attendants: issues and controversies. Durham, NC: Family Health International, 1997. 3 p.
  13. Sibley L, Ann Sipe T. What can meta-analysis tell us about traditional birth attendant training and pregnancy outcomes? Midwifery 2004;20:51-60.
  14. Sibley LM, Sipe TA, Koblinsky M. Does tradition­al birth attendant training increase use of antenatal care? A review of the evidence. J Midwifery Wom­ens Health 2004;49:298-305.
  15. Sibley L, Sipe TA, Koblinsky M. Does traditional birth attendant training improve referral of women with obstetric complications: a review of the evi­dence. Soc Sci Med 2004;59:1757-68.
  16. Jokhio AH, Winter HR, Cheng KK. An Interven­tion involving traditional birth attendants and peri­natal and maternal mortality in Pakistan. N Engl J Med 2005;352:2091-9.
  17. MEASURE DHS+, demographic and health sur­veys, StatCompiler. Calverton, MD: Macro Interna­tional Inc., 2004. (, accessed on 16 December 2005).
  18. Tinker A, ten Hoope-Bender P, Azfar S, Bustreo F, Bell R. A continuum of care to save newborn lives. Lancet 2005;365:822-5.

© 2006 ICDDR,B: Centre for Health and Population Research

The following images related to this document are available:

Photo images

[hn06056t2.jpg] [hn06056f2.jpg] [hn06056t1.jpg] [hn06056f1.jpg]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2018, Site last up-dated on 05-May-2018.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Internet Data Center of Rede Nacional de Ensino e Pesquisa, RNP, Brazil