Pii: s0020-7292(01)00517-3

International Journal of Gynecology & Obstetrics 75 Ž The gap between evidence and practice in maternal J. Villar a,U, G. Carrolib, A.M. Gulmezoglua a UNDPrUNFPArWHOrWorld Bank Special Programme of Research, De¨elopment and Research Training in Human HRP , Department of Reproducti¨e Health and Research, WHO, Gene¨a, Switzerland b Centro Rosarino de Estudios Perinatales ( Abstract
The expected improvement in maternal and perinatal health in developing countries has not yet materialized. In addition to the factors related to socioeconomic conditions, we have identified areas where large gaps betweenevidence and practice are apparent. These gaps are in clinical care, implementation of effective practices and inselecting research priorities. We present examples from our own research and the literature to illustrate these points.
ᮊ 2001 International Federation of Gynecology and Obstetrics. All rights reserved.
Keywords: Evidence based medicine; Clinical practice; Maternal and perinatal health 1. Introduction
pointing situation. We think a major factor is thecontinued presence of glaring gaps between our The improvement of maternal and perinatal knowledge and practices. Because of these gaps, health in developing countries, and the preven- improved quality of care and expanded coverage tion of the leading pregnancy-related conditions of maternal health services to the large unde- worldwide, such as pre-eclampsia, preterm deliv- ery, intrauterine growth retardation and prelabor There are three major areas where these gaps rupture of membranes, has not materialized as can be identified: the gap in implementing evi- expected, as in other areas of medicine. There dence-based practices; the gap in the strategies could be many reasons for this current disap- for changing practices and the gap in selectingresearch priorities. These gaps need to be recog-nized and addressed according to two important ଝ This paper was partially presented as a Keynote address public health principles: only interventions that at the International Conference on the Humanization ofChildbirth, Fortaleza, Brazil, 2000.
have been shown to be effective by strong re- U Corresponding author. WHO, 1290 Geneve, 27, Switzerland.
search evidence Žnon-biased. should be promoted 0020-7292r01r$20.00 ᮊ 2001 International Federation of Gynecology and Obstetrics. All rights reserved.
PII: S 0 0 2 0 - 7 2 9 2 Ž 0 1 . 0 0 5 1 7 - 3 J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( and implemented, and, the same evidence of ef- ficacy and safety must be required for both drug Proportion of women receiving selected clinical activities inthe baseline survey of the WHO Antenatal Care Trial and non-drug forms of care, including health pol-icy, health systems and educational interventions.
Examples from research projects conducted by us as well as from the literature will be presented,with the objective of highlighting the obstacles to a better and more humane maternal healthcare.
2. The gap in implementing evidence-based
practices
pelling evidence that routine iron supplementa- sential element for the screening, primary or sec- tion reduces iron deficiency anemia during preg- ondary prevention and treatment of pregnancy nancy, particularly among women with borderline complications. Its format, including the timing iron stores. Iron supplements were routinely pro- and its components, has been recognized as areas vided in three of the four study sites in the WHO preparation phase of the WHO multicenter ran- was no such policy and only a small proportion of domized controlled trial to evaluate a new model routine provision of iron supplements was imple- tal practices and procedures was conducted w x mented in the trial in the intervention group, The survey included a review of medical records which resulted in an increase in the proportion of of 2913 women randomly selected at 53 antenatal women with iron supplementation from 20.6% in clinics and interviews with all staff providing ante- the control group to 85.5% in the intervention natal care in these clinics participating in the group. This higher proportion of iron intake was trial. The most striking result was the large dif- associated with a lower rate of severe postpartum ferences across sites in the use of several antena- tal care practices and procedures, such as routine intervention group vs. 13.3% in the control group vaginal examination or formal risk scoring, activi- 5 . It is evident from these data that when mod- ties for which there are no evidence of benefit w x est efforts are made to provide activities proven In two study sites, routine vaginal examination to be effective, the benefits on biological out- comes could be observed in a short period of approximately 42% and, in the fourth site, 99% of women had routine vaginal examinations. There Effective interventions during intrapartum and were similar but not consistent differences in immediate postpartum periods to reduce mater- several other practices. The practice of antenatal nal mortality and severe morbidity have also been care in these clinics seemed to be not-evidence- identified. These interventions include the avail- based and with large variations in locally adapted ability and capacity to provide parental antibi- otics, uterotonics and anticonvulsants Žmag- widely used form of healthcare can be left to such nesium sulfate., perform manual removal of pla- subjective judgement for its implementation? centa, conduct assisted vaginal delivery, perform Even if some practices may not be harmful, their cesarean section when indicated and provide implementation will result in inappropriate use of blood transfusions. Unfortunately recent survey scarce resources and waste of time and money.
found that only approximately half of the services The case of iron supplementation is another in 49 developing countries were able to provide example of the gap between the evidence and J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( Table 2Interventions introduced to medical practice without strong unbiased evidence Conversely, a survey conducted in a hospital in obstetrics risk factors, their age, obstetricians’ China in 1999 showed that intrapartum activities income and years of training, by method or by that should be eliminated from clinical practice amount of payment. We concluded that there is a Ži.e. pubic shaving, rectal examination during very strong influence of the obstetrician’s atti- tudes to clinical practice over the cesarean sec- and recommended not to be used routinely Želec- The ‘attitude’ of obstetricians to cesarean sec- tion has been studied in more detail in England This situation in Chinese hospitals is not unique.
and the USA. In London, 31% of female and 8% Table 2 presents a list of interventions that have of male obstetricians preferred cesarean section limited beneficial effects on substantive perinatal for themselves or their partners, as the mode of outcomes according to systematic reviews of ran- delivery of uncomplicated, singleton pregnancy in domized controlled trials, which are still largely implemented in both developed and developing another recent survey, 56% of male and 33% of countries. Table 3, as a dramatic contrast, pre- female North American obstetricians would per- sents examples of interventions of known effec- sonally prefer to be delivered by cesarean section tiveness that are implemented only on a limited These unnecessary surgical interventions are The attitudes of health service providers clearly applied to populations with low or medium risk of play a major role in the implementation of poli- pregnancy complications, while, at the same time, cies. For example, there is a consistent and un- large numbers of women in the poorest regions of necessary increase in the cesarean section rate in the world, who are at high risk of obstructed many developed and developing countries. When labor, do not have access to surgical facilities.
we explored associated factors in an urban area in The gap between evidence and practice is also Latin America, it was not possible to explain the evident among those who are traditionally con- differences in cesarean section rates in a group of sidered to be closer to the interests of women’s obstetritions by women’s socioeconomic and needs. A recent survey found that most women Table 3Effective interventions with limited implementation Maternal corticosteroids in preterm labor Unavailable where neededroverused in low risk populations J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( chose to be described by staff in antenatal care CME , guidelines and specific interventions clinics in a rural area in the UK as ‘patient’ aimed at changing practices found that passive 39% , the second choice was ‘pregnant women’ dissemination of written educational materials and the less favored were ‘client’, ‘consumer’ or and didactic educational sessions, the most widely spread forms of CME, are largely ineffectivestrategies to improve medical practices w ventions considered more effective in these re- 3. The gap in strategies for changing practices
views were reminders of a specific procedure orpractice, educational outreach for prescribing, in- Clinicians need information to update their teractive educational workshops and multifaceted practice with effective forms of care, but most interventions, mostly not used in standard medi- textbooks, expecially those translated, are out of date for treatments and journals are in excessive A study reporting the time doctors spend read- numbers to be followed by individuals. As a re- ing per week in an U.K. academic institution sult, the knowledge and performance of physi- found that the proportion of physicians not read- cians deteriorate and new practices are intro- ing ‘at all’ ranged from 70% for house officers to duced without proper knowledge ŽFig. .
15% for senior registrars. The median number of Two strategies are commonly used to influence minutes per week among those that did read practice with new evidence: continued medical ranged from 90 min among medical students to education and printed materials of extracts or 20 min among house officers and senior house reviews of the medical literature. An overview of officers ŽCentre for Evidence Based Medicine web 51 reviews of Continuing Medical Education Fig. 1. Cartoon published in a newspaper of wide circulation illustrating the public concern about the introduction of newtechnologies without proper evaluation or training.
J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( sources.html. ᎏ last accessed on 4 April improving knowledge, using a proactive dissemi- There is a pattern of inverse correlation between nation strategy?’ It is possible that institutional years since graduation and medical knowledge of barriers and provider attitudes could be difficult the current best care which is called the ‘slippery to overcome. Nevertheless, it is important and slope’ Žthe more years since graduation, the less necessary to rigorously evaluate promising strate- gies to implement effective practices.
International agencies and professional organi- zations can play an important role in providing upto date access and reliable information to health 4. The gap in selecting research priorities
workers by using strategies to systematically sum-marize the large body of information. Although The selection of relevant research topics and several initiatives are currently in place, most are their implementation, including the availability of aimed at developed country practitioners. One resources and funding, are key elements in the initiative to address this gap, focusing on develop- production of research results that are relevant to ing countries, is The WHO Reproducti¨e Health practice. We distinguish three dichotomies in re- search on maternal health, relevant to the condi- World Health Organization and the Cochrane tions most prevalent in developing countries: Ž .
Collaboration. RHL provides access to Cochrane research to advance curative vs. preventive care; systematic reviews in some of the most important 2 priorities in the ‘‘north’’ vs. the priorities in reproductive health areas on a free-subscription the ‘‘south’’; and Ž3. potentially biased vs. unbi- basis for developing countries. Each of the re- views is accompanied by a commentary bringingin an under-resourced setting perspective and a 4.1. Research to ad¨ance curati¨e ¨s. pre¨enti¨e care practical guide on how to implement the recom-mended practices. Twenty thousand copies of It is now well-recognized that most of the health RHL are published in English and Spanish and gains in the industrialized world in the last cen- updated annually. Currently, we are evaluating a tury are attributable to public health efforts to multifaceted strategy to implement RHL recom- prevent illness and not just to heroic measures to mendations in 40 hospitals in Mexico and Thai- land, in a cluster randomized controlled trial. The strategies for the leading pregnancy specific causes main question is: ‘can we change practices by of maternal and perinatal morbidity and mortality Table 4Current state of knowledge about pregnancy specific conditions U Even using recommended preventive strategies about 4% of women will have postpartum haemorrhage of G1000 ml of blood.
J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( are mostly lacking Žpre-eclampsiareclampsia, 4.3. Potentially biased ¨s. unbiased research method postpartum hemorrhage, prelabor rupture ofmembranes, preterm delivery, small for gestation-al age. The biology, prevention and treatment for these conditions are largely unknown ŽTable less biased study design option to select the most There are other pregnancy-specific conditions effective interventions and to stop the transfer of anemia for which the biology, prevention, and knowledged by developed country health authori- treatment are known, but the challenge to make ties, and implemented decades ago when major them available to all women remains ŽTable public health achievements were made Ži.e. polio Conversely, for most of the infectious diseases affecting pregnant women, such knowledge is However, strong resistance is often encoun- mostly available, but the implementation of pre- tered among funding agencies, health workers ventive programs has proven to be a difficult task and academic circles in the implementation of such randomized controlled trials. For example, itis often said that we cannot wait for the results of 4.2. Priorities in the north ¨s. priorities of the south large trials, and action must be taken ‘now.’ How-ever, the introduction of forms of care that are The area of maternal health research has been not supported by evidence of their effectiveness is strongly influenced, as in others areas of medicine, more detrimental to healthcare, as it is almost by the priorities of funding agencies and academicinstitutions in developed countries. To evaluate impossible to abandon an ineffective and gener- the extent of this bias we reviewed all 9014 trials ally costly treatment, once it has been imple- included in the Pregnancy and Childbirth Module of the Cochrane Library 2000. Only 45 trials trials can be completed in short period of time in 0.5% concerned postpartum hemorrhage, 156 developing countries, as has been demonstrated recently for major health conditions ŽTable 1.2% intrauterine growth retardationrsmall for The WHO evaluation of misoprostol for the pre- gestational age, three of the leading maternal and vention of severe postpartum hemorrhage re- perinatal conditions in developing countries. Con- cruited 18 500 women in less than 2.5 years w versely, preterm delivery, the component of low Another argument against large RCTs is that birthweight most prevalent in developed popula- they are too costly for developing countries and 14 was studied as an outcome in 1203 trials that a simpler, ‘quick and dirty research’ is more 13.3% . A recent evaluation of funding patterns appropriate. In contrast, large resources are used of National Institute of Health, USA, also showed in developing countries in the implementation of the lower priority given to perinatal conditions for ineffective and sometimes harmful forms of care, while large trials conducted in developing coun- Table 5Interventions evaluated during 2000 by international collaborative efforts J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( tries are of low cost by any developed country peer-reviewed journal, which is not aimed at dis- standard. Conducting large-scale, cutting-edge semination for changing practice, but rather to health service research in developing countries, allow the scientific community to scrutinize the 5 particularly the trials evaluating non- work and challenge or accept the results. This is a central point and should never be neglected.
appropriate research methodology in parallel, is a Therefore, further planned strategies are neces- sary for dissemination of research results at ser- Furthermore, it is usually argued that for many vice level and to promote their implementation.
health services practices, trials are not needed There is a big gap between research funding because ‘common sense’ indicates that they are and magnitude of disease burden in maternal ‘logically’ effective and harmless. A relevant ex- healthcare in developing countries. Most preg- ample of such misconception is the case of bed nancy-specific conditions have poorly understood rest during pregnancy for prevention of adverse etiology, pathogenesis and consequently, there are outcomes such as preterm birth, impaired fetal no readily available specific prevention, the back- growth and preeclampsia. A review of 15 trials bone of public health. A significant change in investigating bed rest as a primary treatment for research funding to improve our understanding of different medical conditions did not find an im- the etiology and pathogenesis of these disorders provement in any outcome, and additionally in and to evaluate promising preventive strategies is nine trials there was evidence of significant wors- ening of some conditions such as preeclampsia Finally, it could be argued that all of these 20 . There is also the misconception that RCTs considerations do not apply to the large section of are inappropriate for evaluating social interven- the pregnant population who do not have access tions. Such study design was popular among social to services. We think that these underserved scientists during the 1980s, but became less popu- populations comprise two groups: those who make lar as policy makers reacted negatively to the an initial contact, but because of the poor quality evidence of ‘near zero’ effects from several trials of the services do not come back; and those who genuinely have no access to services. We believethat the first step in the humanization of mater-nal healthcare is to make services ‘effective’ and 5. Conclusion
We have presented examples to support the References
concept there are systematic gaps between evi-dence of effectiveness, patients’ preferences, re-search priorities and what is actually practiced.
Villar J, Khan-Neelofur D. Patterns of routine antenatalcare for low-risk pregnancy ŽCochrane Changing or updating clinical practice is not a Cochrane Library, Issue 4. Oxford: Update Software, new challenge. W. Osler in 1906 recognized the risks of not updating our practices by saying ‘from Villar J, Bakketeig L, Donner A, Al-Mazrou Y, Ba’aqeel our teachers and associates, from our reading, from the social atmosphere about us, we catch domized controlled trial: rationale and study design.
Pediatr Perinat Epidemiol Piaggio G, Ba’aqeel H, Bergsjø P, Carroli G, Farnot U, Lumbiganon P et al. The practice of antenatal care: concentrate at earlier stages of the clinical train- comparing four study sites in different parts of the world ing to prepare the practitioners to a continuous participating in the WHO Antenatal Care Randomised process of learning of how to break the ‘force of habit.’ Researchers are usually not effective in Carroli G, Rooney C, Villar J. How effective is antena- the dissemination of their findings to practition- tal care in preventing maternal mortality and serious ers. We are content with publishing in a scientific, morbidity? Pediatr Perinat Epidemiol 2001;15:1᎐42.
J. Villar et al. r International Journal of Gynecology & Obstetrics 75 ( Villar J, Ba’aqueel H, Piaggio G, Lumbiganon P, Beli- Belizan JM. A randomized trial of psychosocial support zon JM, Farmot V et al. The WHO Antenatal Care Randomized Controlled Trial: the evaluation of a new model of routine antenatal care. Lancet 2001;357: JM. Randomized trials in maternal and perinatal Bulatao R, Ross J. Rating developing ᎏ country efforts medicine: global partnerships are the way forward. Br J to improve maternal and neonatal health. Glastonbury, Obstet Gynaecol 1998;105:1244᎐1247.
CT: Futures Group International, 2001.
Allen C, Glazsiou P, del Mar C. Bed rest: a potentially Belizan JM, Quaranta P, Paquez E, Villar J. Caesarean harmful treatment needing more careful evaluation’.
section and fear of litigation. Lancet 1991;338:1462.
Al-Muffti R, McCarthy A, Fisk N. Obstetricians’ perso- Oakley A. Experimentation and social interventions: a nal choice and mode of delivery. Lancet 1996;347:544.
forgotten but important history. Br Med J 1998;317: Gabbe S, Holzman G. Obstetricians choice of delivery.
Osler W. Harvey and his discovery. The Harveian Ora- Byrne DL, Asmussen T, Freeman JM. Descriptive terms tion delivered at the Royal College of Physicians, Lon- for women attending antenatal clinics: mother knows don 1906. The Collected Writings of Sir William Osler, best? Br J Obstet Gynaecol 2000;107:1233᎐1236.
Freemantle N, Harvey EL, Wolf F, Grimshaw JM, Grilli Thacker S, Stroup D. Continuous electronic heart rate R, Bero LA. Printed educational materials: effects on monitoring for fetal assessment during labour ŽCoch- professional practice and healthcare outcomes ŽCoch- Reviews . The Cochrane Library, June 4. Oxford: Review . The Cochrane Library, Issue 4. Oxford: Bricker L, Neilson JP. Routine ultrasound in late preg- nancy Žafter 24 weeks gestation. ŽCochrane The Cochrane Library, Issue 4. Oxford: Update Soft- Bloom B. Letter to the alumni. Boston, MA: Harvard Carroli G. Episiotomy for vaginal delivery ŽCochrane Review . The Cochrane Library, Issue 4. Oxford: Up- Villar J, Belizan JM. The relative contribution of prema- turity and fetal growth retardation to low birthweight in developing and developed societies. Am J Obstet Gy- Crowley P. Prophylactic corticosteroids for preterm birth Review . The Cochrane Library, Issue 1. Ox- Gross CP, Anderson GF, Powe NR. The relation between funding by the National Institute of Health and Smaill F. Antibiotics for asymptomatic bacteriuria in the burden of disease. N Engl J Med 1999;340: Issue 1. Oxford: Update Software, 2001.
Villar J, Carroli G. Methodological issues of randomised Duley L, Henderson-Smart D. Magnesium sulphate ver- controlled trials for the evaluation of reproductive health interventions. Prev Med 1996;25:365᎐375.
Cochrane Library, Issue 1. Oxford: Update Software, AM, Villar J, Ngoc NN, Piaggio G, Carroli G, Adetoro L et al. for the WHO Collaborative Group Hodnett ED. Caregiver support for women during child- to Evaluate Misoprostol in the Management of the Third Stage of Labour. The WHO multicentre double- blind randomized controlled trial to evaluate the use of Hofmeyr GJ, Kulier R. External cephalic version for misoprostol in the management of the third stage of Cochrane Library, Issue 1. Oxford: Update Software, Villar J, Farnot U, Barros F, Victoria C, Langer A,

Source: http://bhpelopartonormal.pbh.gov.br/estudos_cientificos/arquivos/the_gap_between_evidence_and_practice_in_maternal_healthcare.pdf