Counselling on breastfeeding: assessing knowledge and skills.

Counselling on breastfeeding: assessing knowledge and skills M.F. Rea,1 S.I. Venancio,2 J.C. Martines,3 & F. Savage3 Reported are the results of a randomized controlled trial to assess the effectiveness of the WHO/UNICEF 40-hour course ‘‘Breastfeeding counselling: a training course’’. The course was conducted in a maternity hospital which provides care to a low-income population in a metropolitan area in Sa˜o Paulo, Brazil. Health workers from 60 health units were randomly assigned to be either participants (20) or controls (40), and their breastfeeding knowledge and skills were assessed before and immediately after the course, as well as 3 months later. Immediately after the course the participants’ knowledge of breastfeeding had increased significantly compared to controls. Both their clinical and counselling skills also improved significantly. When assessed 3 months later, the scores remained high with only a small decrease. The implementation of the course was also evaluated. The methods used were participatory observation, key interviews and focus group discussion. In the 33 sessions of the course, the average score was 8.43 out of 10. Scores were highest for content and methodology of the theory sessions, and lowest for ‘‘use of time’’, ‘‘clinical management of lactation’’, and ‘‘discussion of clinical practice’’. ‘‘Breastfeeding counselling: a training course’’ therefore effectively increases health workers’ knowledge and their clinical and counselling skills for the support of breastfeeding. The course can be conducted adequately using the material and methodology proposed, but could be more satisfactory if the time allocated to exercises and clinical practice sessions were increased. Voir page 497 le re´sume´ en franc¸ais. En la pa´gina 497 figura un resumen en espan˜ol. Introduction Mothers often stop breastfeeding earlier than is desirable, even when they know that it is the ideal way to feed their infants. Low rates of breastfeeding are common in countries where average income is higher (1), and where women work outside the home, which can make breastfeeding difficult. In countries where income is sufficient for the purchase of breast-milk substitutes, they are marketed more intensely, with the result that many families come to regard them as indispensable. Health professionals have an extreme- ly important role to play in supporting women to enable them to breastfeed and in convincing them that breast-milk substitutes are unnecessary. Unfortunately, health professionals often have negative attitudes towards breastfeeding. Also during their training, few health professionals receive much up-to-date information on the subject and many are exposed to advertisements for infant formula. For generations, both health professionals and mothers have been instructed to follow fixed schedules for both breastfeeding and formula feeding, and to give water or teas between feeds. This can result in diminished breast-milk production and early cessa- tion of breastfeeding, and has frequently led to the development of a ‘‘bottle-feeding culture’’. The use of dummies to ‘‘pacify’’ a crying infant has also become an integral part of such cultures (2). In the past, schools of medicine, nutrition, nursing and public health failed to include breast- feeding in their curricula, though students were usually instructed on the preparation and handling of breast-milk substitutes. Textbooks were also inade- quate. In 1993, WHO and the International Baby Food Action Network (IBFAN) (3) evaluated the coverage of breastfeeding by 180 textbooks used in medical schools in over 90 countries, scoring the books on a scale of 0 to 1.00. The results ranged from 0.04 to 0.76, with only four books receiving a score of 0.5 or more. Practical management of breastfeeding received much less attention than theoretical aspects of lactation and its problems. Thus, although those who have read the books may be convinced of the value of human milk, they are unlikely to have learned the skills necessary to help mothers breastfeed, or to be able to teach others to do so. How can health professionals acquire appro- priate training on breastfeeding? The long-term strategy must be to make appropriate changes to their basic training; however, it has proved necessary to start with in-service training, to change existing practices and routines (4). This has led WHO and UNICEF to develop several courses on breastfeed- 1 Senior Researcher, Instituto de Saude, Rua Santo Antonio, 590, Sa˜o Paulo, 01314-000 Brazil. Requests for reprints should be sent to Dr Rea at this address. E-mail: marifrea@isaude.sp.gov.br. 2 Pediatrician, Instituto de Saude, Sa˜o Paulo, Brazil. 3 Medical Officer, Child and Adolescent Health and Development, World Health Organization, Geneva, Switzerland. Reprint No. 0007 Research 492 # World Health Organization 1999 Bulletin of the World Health Organization, 1999, 77 (6) ing, of different lengths and aimed at different target groups. Breastfeeding counselling: a training course (BFC) (5) differs from other courses because in addition to theoretical and clinical aspects of breastfeeding, it includes training on specific counselling skills. The word ‘‘counselling’’ is new to many people, and it signifies a different approach to giving advice. A counsellor does not tell a mother what to do: she/he gives appropriate information and helps the mother to decide what is best for her. A counsellor listens and tries to understand how a mother feels, and also tries to increase her confidence, so that the mother feels she is in control of the situation. Use of these skills has been effective in increasing the proportion of mothers who breastfeed exclusively (6). During a BFC course, there are four 2-hour practical sessions in which participants work with mothers and infants and practise clearly defined skills. These include six ‘‘listening and learning skills’’: using helpful non- verbal communication; asking open questions; using responses and gestures to show interest; reflecting back what the mother says; empathizing — showing understanding of how the mother feels; and avoiding words which imply judgement. There is another set of six skills for ‘‘building confidence and giving support’’: accepting what a mother thinks or feels; recognizing and praising what the mother and the baby are doing right; giving practical help; giving information that is of immediate relevance; using simple language; and making suggestions instead of giving commands. The rationale for this approach can be related to basic physiological principles. The flow of milk from the breast is stimulated by the oxytocin reflex, which can be affected by the woman’s emotions and can be conditioned. Seeing and hearing her infant can facilitate the reflex, while doubt and anxiety may temporarily inhibit it (7). One of the commonest reasons for early introduction of supplements or premature discontinuation of breastfeeding is the mother’s perception that her breast milk is insuffi- cient (8), which is often associated with lack of confidence and support. During the course, clinical skills are also practised. These include observation and assessment of breastfeeding, helping mothers to position and attach their infants at the breast, expression of breast milk, and clinical management of common difficul- ties such as sore nipples, mastitis and apparent insufficiency of milk. The course is provided to 20–24 mid-level health workers over a period of 5–6 days (40 hours) in a hospital which provides maternity and paediatric care. The training material consists of a Director’s guide, which provides detailed instruction on how to plan and conduct a course; a Trainer’s guide, which explains in detail how to teach the 33 sessions of the course; 50 overhead transparencies also available as a flipchart; 50 slides; and a Participants’ manual, which includes summaries of each session, written ex- ercises, and copies of all the checklists and forms needed in the course. In addition, there are printed answer sheets for the exercises, basic reference documents, and a video on the clinical management of breastfeeding. There have been few evaluations of materials and methods used in training courses on breastfeed- ing for health professionals. Armstrong admits that a single course may not have been the only factor leading to the changes in practice that she observed, but it may have acted as a catalyst (9). Short courses might be effective if given as part of a comprehensive educational programme (10). An evaluation of a 15-day course offered by the Santos Lactation Center (11) in Brazil found that the health units in which staff had attended a course had changed their practice to promote breastfeeding, but this had not happened in units where staff had not been so exposed. This article presents the results a study to evaluate the BFC course, including its methodology, and its effect on knowledge of breastfeeding and the clinical management and counselling skills of parti- cipants. Materials and methods A randomized controlled trial was conducted in a maternity hospital which provides care to a low- income population in a metropolitan area in Sa˜o Paulo, Brazil. A total of 60 potential participants, all health professionals, were randomly allocated to two groups of unequal size: 20 to an ‘‘exposed’’ group (participants) and 40 to a control group. Each potential participant belonged to a different health unit. Because it was likely that there might be a greater drop-out rate among the controls, we selected six extra professionals to compensate. Potential partici- pants and controls all had a minimum of 8 years’ schooling and currently worked in an area dealing with the care of mothers and children. The sample size was calculated to detect differences of 40% in knowledge scores between the ‘‘exposed’’ and control groups with a significance level of 95%. To ensure baseline comparability of the two groups, we matched potential participants in groups of three on the basis of their pre-test performance (i.e. previous knowledge of breastfeed- ing), schooling (mid-level or university level), and function in the health service (maternity/rooming-in ward or outpatient units). One member of each group of three was randomly allocated to participate in the course, the other two acting as controls. The implementation of the course, and the suitability of the materials and the methodologies that it proposes, were evaluated as outlined below. . By participatory observation conducted by a project team researcher and a trained research assistant: both filled in their forms independently, with the results being compared and discussed later. . By a semi-structured (recorded) interview with the course directors. Counselling on breastfeeding: assessing knowledge and skills 493Bulletin of the World Health Organization, 1999, 77 (6) . By focus groups together with trainers at the end of the course. . By focus group sessions with participants im- mediately after the course and three months later. We observed the sessions covering the theory elements of the course (i.e. everything except the clinical practice) in order to assess achievement of the objectives. Content, methodology, use of time, and trainers’ and participants’ performance were evalu- ated. Evaluation of the clinical practice included preparation of the class, management and discussion of practical activity, use of time, and trainers’ and participants’ performance. The content and methodology items were each worth a maximum of three points, according to the percentage fulfilment of objectives: <60% = 1, 60– 80% = 2, and >80% = 3. For the use of time, one point was awarded if the achievement for assessed items was >80%, allowing the time limit to be exceeded by up to 10 minutes. Classes where fulfilment was <80% were rated zero. A maximum score of 1.5 was given for the performance of the trainer and for the participants according to the following criteria: <60% = 0.5, 60–80% = 1, and >80% = 1.5. For the four sessions of clinical practice, the preparation, clinical management, and discussion were worth a maximum score of two points each. The following criteria were adopted, according to the percentage of each item fulfilled: <60% = 0, 60– 80% = 1, and >80% = 2. Use of time, and trainers’ and participants’ performance were scored in the same way as for the theory classes. Thus, both theory and practice classes could each score a maximum of 10 points if awarded a maximum score for each of the items assessed. Impact of the course on knowledge, skills and attitudes of participants The impact of the course on the participants’ knowledge of breastfeeding was evaluated by means of a test containing 13 multiple-choice questions on the topics covered in the course. All 60 professionals were tested before and immediately after the course. Changes in the participants’ skills and attitudes were evaluated by observing their behaviour in clinical consultations before and immediately after the course. All professionals were observed by two researchers (M.F.R., S.I.V.) during consultations with mothers in the rooming-in ward (to ‘‘standardize’’ the type of consultation, since all these mothers had a similar range of problems). The results were analysed after grouping the skills into blocks. Three months later, a further clinical evaluation and post-test assessment of knowledge on breast- feeding were made using the participants only, to determine whether the effects of the course persisted in the medium term. Results Implementation of the course Table 1 shows the mean scores obtained for each of the items analysed. Analysis of the theory sessions indicated that the average score for content was 2.95 (maximum = 3). For the application of the metho- dology, the average score for the 29 theory sessions was 2.60 (maximum = 3). Scores for use of time were low (0.23 out of a maximum of 1) mainly because sessions continued for longer than planned. Scores for the clinical practice items clinical management (1.38 out of 2) and discussion (1.00 out of 2) were also relatively low, apparently because trainers sometimes taught from their own experience rather than following the suggested procedure in the course manual. This may be why some sessions continued longer than recommended, and therefore may not have covered the planned topics adequately. Over the full 33 sessions the course as a whole received an average score of 8.43 (SD =0.98) out of a maximum of 10. The trainers’ performance received quite favourable evaluations (average: 1.8 out of a total 2 points), as did participants’ involvement (average: 1.43 out of a total 2 points). Assessment of breastfeeding knowledge of participants In the pre-test assessment, the exposed group (participants) averaged 6.23 out of 10 points, while the control group averaged 6.06 (P = 0.95). This is evidence that the process of dividing the potential participants into threes resulted in groups that were initially homogeneous in terms of their knowledge of breastfeeding. In the post-test assessment held immediately after completion of the course, the exposed group averaged 8.35 and the control group 5.54 (P>0.001). Three months after the course, the average score of the 20 participants was 7.80. Thus although performance had declined slightly in relation to that immediately post-test (Fig. 1), participants maintained a statistically significant increase in knowledge skills compared to the pre-test. Assessment of clinical and counselling skills Table 2 shows the average scores obtained for each skill item by the professionals in the exposed and control groups. Comparison of the average achieved by each of the two groups before and after the course reveals that the averages of all items improved for the exposed group, as assessed by both observers. In the control group, the scores for some items remained constant or even declined. Using the Kruskal–Wallis variance test, we found that there were no statistically significant differences between the average pre-course scores of the exposed or control groups for any of the items considered. The post-test assessment revealed Research 494 Bulletin of the World Health Organization, 1999, 77 (6) statistically significant differences between the exposed and control groups for all items analysed. Table 3 shows the percentage changes in counselling skills at pre- and post-testing, calculated by dividing the averages of the exposed and control groups for each of the variables by the total score that could be obtained for each item. Even taking into account the differences between the two researchers, the indicators show the course significantly improved breastfeeding clinical management and acquisition of counselling skills. Late post-test assessment of skills Table 4 shows participants’ skills 3 months after the course. Although average scores were slightly lower in the late post-test assessment than immediately after the course, the differences between pre-test and late post-test results were still statistically significant for all items analysed (P<0.05). Performance declined least for non-verbal communication and most for breastfeeding history. Discussion The assessment of the BFC course indicated that it was well received by the trainers. Most of the sessions covered 100% of the proposed content. However, the content was not entirely new to either coordina- tors or trainers. In fact, one of the criteria for their selection was prior knowledge and even experience of teaching courses on breastfeeding. Trainers placed particular emphasis on the development of counsel- ling skills, the area that seemed to be the most unfamiliar and challenging. None of them referred to the development of clinical skills, such as assessment of a breastfeed, as an essential feature of the course. This suggests that this area received less emphasis, which seems to be reflected in participants’ some- what lower scores for clinical management in the evaluation. Clearly more attention should be given to this in future courses. In all sessions the greatest problem in following the course guidelines was lack of time. This might have been because the course was held for 4 hours Table 1. Scores according to items analysed in WHO/UNICEF breastfeeding counselling course Item Mean score Content (3)a 2.95 Methodology (3) 2.60 Preparation of the practice (2) 1.75 Handling of the practice (2) 1.38 Discussion of the practice (2) 1.00 Trainers’ performance (1.5) 1.48 Participants’ performance (1.5) 1.43 Use of time (1.0) 0.23 a Figures in parentheses are the maximum values for each item. Fig. 1. Performance of exposed and control groups on knowledge of breastfeeding in the pre-test, immediate post-test and late post-test, Sa˜o Paulo, 1996 Table 2. Average scores (+ SD), before and after the course, for each item of the clinical consultation for professionals in the exposed and control groups Skill Pre-course Post-course Exposed Control Exposed Control group group group group Clinical history (10 points)a 3.70 + 1.03 3.63 + 1.62 5.23 + 1.47b 3.76 + 1.27 Assessment of a breastfeed (14 points) 6.75 + 2.77 6.74 + 3.36 9.90 + 1.37b 8.59 + 2.40 Non-verbal communication (25 points) 17.25 + 2.69 15.95 + 3.78 22.15 + 2.1b 17.70 + 3.70 Listening and learning (25 points) 12.80 + 3.16 12.91 + 3.36 19.35 + 3.69b 12.18 + 3.35 Building confidence and giving support (45 points) 26.55 + 4.92 25.43 + 5.26 36.00 + 4.87b 24.33 + 5.91 a Figures in parentheses are the maximum number of points for the skill concerned. b P-value < 0.05. Counselling on breastfeeding: assessing knowledge and skills 495Bulletin of the World Health Organization, 1999, 77 (6) per day for two weeks, instead of 8 hours per day for one week, or due to the number of sessions recommended. The lack of time was especially detrimental to the completion of exercises and to clinical practice, both of which are of fundamental importance to the course. Person-to-person skills are important for the support of breastfeeding (12), and their acquisition is particularly time consuming. The difficulty in conducting the clinical practice sessions seemed in some cases to be due to the more experienced trainers teaching in their usual manner, departing from the course methodology and sometimes from the technical content. Despite this, the analysis showed highly significant improvements in participants’ knowledge of breastfeeding. Clinical management and counselling skills also showed significant and sustained improvement following the course, but some further comments are necessary in this respect. Taking a breastfeeding history was the skill least well retained by participants. The content of the session was covered (100%), but there were serious time difficulties with the exercises and thus few opportunities for discussion. Assess- ment of a breastfeed was another skill that participants found difficult to retain. Participants began with a score of around 48%, which increased to 71% immediately after the course, but their perfor- mance fell to around 62% at 3 months, suggesting that although there was definite improvement some participants rapidly revert to old habits. Listening and learning skills were seen to improve from 68% to 88%, and this improvement was maintained in the late evaluation. Prior to the course, the greatest difficulty here was experienced with ‘‘appropriate touch’’. This practice is quite natural in Brazilian culture and low scores in the pre-test may reflect inhibition by the participants, who felt less inhibited after the course. ‘‘Non-verbal communication’’ was adequately practised by half the participants before the course, 77% immediately after training, and 72% three months later. The hardest skill to learn was ‘‘empathy’’: after the course, 2 out of 3 participants were able to demonstrate this skill. The ability to ‘‘show interest’’ increased from around 42% to 82%. The skills of building confidence and giving support were already practised by around 59% of participants before the course, by 80% immediately after the course, and by 78% in the late post-test. ‘‘Giving practical help’’, ‘‘giving praise’’ and ‘‘making sugges- tions, not commands’’ were the best learned components of these skills. These data are consistent with an evaluation of the performance of health workers following a similar course in Bangladesh, who were able to counsel Table 3. Average percentage achievement of counselling skills in clinical practice, by participants in the exposed and control groups at pre- and post-testing Item % achievement in Exposed group Control group Pre-test Post-test ~a Pre-test Post-test ~a Keep head at same level 70.4 93.0 22.6 65.0 82.2 17.2 Pay attention 80.0 94.4 14.4 72.2 83.0 10.8 Remove barriers 74.6 91.0 16.4 66.6 83.2 16.6 Take time 74.0 90.0 16.0 66.2 76.2 10.0 Touch appropriately 52.0 80.6 28.6 52.2 64.4 12.2 Ask open questions 61.6 81.6 20.0 55.2 58.6 3.4 Show interest 41.6 81.6 40.0 48.0 58.4 10.4 Reflect on what mother says 64.6 87.6 23.0 62.8 73.0 10.2 Empathize 42.0 66.0 24.0 45.4 47.0 1.6 Avoid judging 50.0 73.4 23.4 50.2 57.6 7.4 Accept what mother thinks and feels 58.4 80.6 22.2 55.0 64.2 9.2 Praise 59.0 86.6 27.6 51.6 59.2 7.6 Give practical help 47.6 77.0 29.4 47.8 59.8 12.0 Give little, relevant information 62.0 80.0 18.0 60.2 61.6 1.4 Use simple language 73.0 88.4 15.4 73.8 80.6 6.8 Suggest, do not order 62.6 89.4 26.8 57.6 65.4 7.8 a ~= % achievement of skill at post-testing – % achievement of skill at pre-testing. Table 4. Average percentage achievement by participants in counselling skills and breastfeeding clinical management before and immediately after the course and upon late evaluation % achievement Pre-course Post-course Late valuation Clinical history 37 52 47 Assessment of a breastfeed 48 71 62 Non-verbal communication 68 88 87 Listening and learning 51 77 72 Confidence and support 59 80 78 Research 496 Bulletin of the World Health Organization, 1999, 77 (6) mothers effectively and to enable them to breastfeed exclusively for longer periods (6). Thus the BFC course resulted in a substantial acquisition of both knowledge and skills. Three months afterwards, there was no substantial loss for any item evaluated, despite the difficulty of practising breastfeeding counselling in the health services of a large metropolitan area. We recommend that the time allocated for development of skills in the clinical practice and exercise sessions of the course be increased. Some theoretical sessions (such as health care practices) may be shorter for the target audience for this course. We also recommend that, for training to be most effective, it should be followed up and participants enabled to continue practising the skills that they have acquired. To ensure that this happens, it is necessary for the local health supervisor to be included in the course training. This requirement should be specified in the Director’s guide along with the recommendations for selecting trainers and participants. In conclusion, Breastfeeding counselling: a training course effectively increases health workers’ clinical and counselling skills for the support of breastfeeding. Adequate time must, however, be allocated for exercises and clinical practice, which are important for the learning of skills. Where health workers have already been exposed to courses covering the theoretical aspects of breastfeeding, the practical training is likely to be the most effective part of the BFC course. n Acknowledgements The article is based on the research project Assess- ment of the impact and implementation of the WHO/UNICEF Breastfeeding Counselling Train- ing Course, carried out by Instituto de Saude, Sa˜o Paulo, with the support of WHO. Re´sume´ Conseil en matie`re d’allaitement au sein : e´valuation des connaissances et des compe´tences L’article de´crit les re´sultats d’un essai controˆle´ randomise´ destine´ a` e´valuer l’efficacite´ d’un cours OMS/UNICEF de 40 heures sur le conseil en matie`re d’allaitement au sein. Le cours a e´te´ organise´ dans une maternite´ d’une grande agglome´ration qui accueille une population a` faible revenu. Les agents de sante´ de 60 unite´s de sante´ ont e´te´ de´signe´s au hasard, 20 comme participants et 40 comme te´moins, et leurs connaissances et leurs compe´tences concernant l’allaitement au sein ont e´te´ e´value´es avant et imme´diatement apre`s le cours, et une nouvelle fois 3 mois plus tard. Imme´diatement apre`s le cours, les connaissances des participants sur l’allaitement au sein avaient sensiblement progresse´ par rapport a` celles des te´moins. Leurs compe´tences tant au plan clinique qu’en matie`re de conseil s’e´taient aussi nettement ame´liore´es. Trois mois plus tard, les re´sultats demeuraient satisfai- sants, malgre´ un le´ger recul. La mise en œuvre du cours a aussi e´te´ e´value´e. Les me´thodes utilise´es e´taient l’observation active, les entretiens clefs et la discussion de groupe. Pour les 33 se´ances du cours, la note moyenne e´tait 8,43 sur 10. Les notes les plus e´leve´es ont e´te´ attribue´es au contenu et a` la me´thodologie des cours the´oriques, et les plus faibles a` l’utilisation du temps, a` la gestion clinique de la lactation et a` la discussion de la pratique clinique. Le cours sur le conseil en matie`re d’allaitement au sein permet donc d’ame´liorer effectivement les connais- sances des agents de sante´ et leurs compe´tences cliniques dans ce domaine. Le cours peut eˆtre dispense´ de fac¸on satisfaisante avec les mate´riels et la me´thodo- logie propose´s mais il serait pre´fe´rable de pre´voir plus de temps pour les exercices et la pratique clinique. Resumen Orientacio´n sobre la lactancia materna: evaluacio´n de los conocimientos y las aptitudes Se presentan los resultados de un ensayo controlado aleatorizado que se llevo´ a cabo para evaluar la eficacia del curso OMS/UNICEF de 40 horas «Orientacio´n sobre la lactancia materna: curso de formacio´n». El curso se impartio´ en una maternidad que atiende a personas de bajos ingresos de una zona metropolitana. Entre el personal sanitario de 60 unidades de salud, se escogio´ aleatoriamente a 20 participantes y 40 testigos. Sus conocimientos y aptitudes en materia de lactancia materna fueron evaluados antes del curso e inmediata- mente despue´s del mismo, ası´ como tres meses ma´s tarde. Inmediatamente despue´s del curso los conoci- mientos de los participantes sobre la lactancia materna habı´an aumentado significativamente en comparacio´n con los controles. Sus aptitudes, tanto clı´nicas como de orientacio´n, tambie´n mejoraron significativamente. En la evaluacio´n realizada tres meses ma´s tarde las puntua- ciones seguı´an siendo altas, observa´ndose so´lo una ligera disminucio´n. Se evaluo´ tambie´n la ejecucio´n del curso. Los me´todos empleados fueron la observacio´n participativa, la realizacio´n de entrevistas a los directores de los cursos y la discusio´n por grupos focales. En las 33 sesiones del curso la puntuacio´n promedio fue de 8,43 sobre 10. Las puntuaciones ma´s altas correspondieron al contenido y a la metodologı´a de las sesiones de teorı´a, y las ma´s bajas al «uso del tiempo», el «manejo clı´nico de la lactancia» y el «estudio de la pra´ctica clı´nica». Counselling on breastfeeding: assessing knowledge and skills 497Bulletin of the World Health Organization, 1999, 77 (6) Ası´ pues, el curso de formacio´n aquı´ considerado permite aumentar eficazmente los conocimientos de los agentes de salud y sus aptitudes clı´nicas y de orientacio´n en apoyo de la lactancia materna. El curso puede impartirse adecuadamente con el material y la metodo- logı´a propuestos, pero sus resultados podrı´an ser ma´s satisfactorios si se aumentara el tiempo asignado a los ejercicios y a las sesiones sobre la pra´ctica clı´nica. References 1. Contemporary patterns of breastfeeding. Report on the WHO Collaborative Study on Breastfeeding. Geneva, World Health Organization, 1981. 2. Victora CG et al. Pacifier use and short breastfeeding duration: causes, consequence, or coincidence. Pediatrics, 1997, 99: 445–453. 3. Courant GT et al. An evaluation of the breastfeeding content of selected medical textbooks. Washington, DC, Institute for Reproductive Health, Georgetown University Medical Center, 1993. 4. Naylor A. Professional education and training for trainers. International journal of gynaecology and obstetrics, 1990, 31: 25–27. 5. Breastfeeding counselling: a training course. Unpublished document WHO/CDR/93.3–6 (UNICEF/NUT/93.1–4). Available upon request from Health Systems and Community Health, World Health Organization, 1211 Geneva 27, Switzerland. 6. Haider R et al. Breast-feeding counselling in a diarrhoeal diseases hospital. Bulletin of the World Health Organization, 1996, 74: 173–179. 7. Mepham TB. Physiology of lactation. Milton Keynes, England, Open University Press, 1987. 8. Segura-Millan S, Dewey KG, Perez-Escamilla R. Factors associated with perceived insufficient milk in a low-income urban population in Mexico. Journal of nutrition, 1994, 124: 202–212. 9. Armstrong H.C. Breastfeeding promotion: training of mid-level and outreach health workers. International journal of gynaecology and obstetrics, 1990, 31: 91–104. 10. Valdes V et al. The effects on professional practices of a three- day course on breastfeeding. Journal of human lactation, 1995, 11:185–190. 11. Westphal MF et al. Breastfeeding training for health profes- sional and resultant institutional changes. Bulletin of the World Health Organization, 1995, 73: 461–468. 12. Kyenkya-Isabirye M, Magalha˜es R. The mothers’ support group role in the health care system. International journal of gynaecology and obstetrics, 1990, 31: 85–90. Research 498 Bulletin of the World Health Organization, 1999, 77 (6)