BIRTH 33:4 December 2006
Beyond Evidence: The Complexity of Maternity Care Preface: This paper has been gestating for a long time–the sixty years that I have been associated
with maternity care. It was conceived during an era of medical authoritarianism, born in a time of nascent childbirth education and family-centered maternity care, matured in a period of enthusiastic (but not unquestioning) homage to evidence-based obstetrics, and culminated in a reluctant but comforting acceptance of uncertainty. It has passed through stages of enthusiasm, of disillusionment, and of reevaluation. It is, to use an ancient word I only recently learned, a clinamen, a swerve, a point of intellectual revision. Others in the Clinamen Collaboration share the authorship of this paper with me. Murray Enkin
Lest anyone yearn for the good old days, the halcyon days of yore, they weren’t. My father’s mother was one of the many women who died in childbirth. Although maternal mortality, which in the 1930s was still as high as it had been a century before (1), began to fall precipitously in the 1940s (at least in the West), it left its residue of fear. When I began practice in the late 1940s, pregnancy was still seen as a time of danger, and maternity care was enmeshed in a tangle of enshrined do’s and don’ts, prescriptions and proscriptions, dictated by medical tradition. Companions (witnesses?) were excluded from accompanying, helping, or even seeing laboring women. The pain of labor was controlled by massive narcosis, and that of childbirth by general anesthesia. But growing obstetrical expertise also brought great benefits. Care became more effective, childbirth both more comfortable and safer. It was a win-win situation. Women were pleased, and doctors gratified. It was a time of great, and justified, optimism. Nevertheless some disturbing problems remained. Experts did not agree either with women’s demands or with each other. Different interventions were used,
Address correspondence to Dr. Murray W. Enkin, 1001 Bay Street, Apt. 2404, Toronto, ON M5S 3A6, Canada. Ó 2006, Copyright the Authors Journal compilation Ó 2006, Blackwell Publishing, Inc.
in different ways, and often with differing results. Expectations grew; as more was achieved, still more was demanded. Something new was needed. A New Approach to Evidence Something new came along. At first largely unknown and ignored, randomized clinical trials were introduced to medicine in the 1950s as an unbiased way to determine the real effects of medical interventions. The medical profession was slow to recognize their value, and obstetrics was one of the slowest specialties to make use of this new approach to evidence. In his 1979 review of the medical profession, Archie Cochrane, after first considering the abysmal record of psychiatrists, surgeons, and cardiologists, decided to award the ‘‘wooden spoon’’ to the obstetricians for the poorest record in evaluating their practices. He justified his ranking by pointing out that The specialty missed its first opportunity in the sixties, when it failed to randomise the confinement of low risk pregnant women at home or hospital. . . . Then, having filled the emptying beds by getting nearly all pregnant women into hospital, the obstetricians started to introduce a whole series of expensive innovations into the routines of pre and postnatal care and delivery, without any rigorous evaluation. The list is long, but the most important were induction, ultra-sound, fetal monitoring, and placental function tests. The specialty reached its apogee in 1976 when they produced 20 per cent fewer babies at 20 per cent more cost (2).
266 Cochrane was, however, cautiously optimistic. He presciently went on to ‘‘admit that my spies tell me there will soon be a torrent of evaluations in this field.’’ His spies were right; obstetrics indeed rose to his challenge. Iain Chalmers, director of the fledgling National Perinatal Epidemiology Unit in Oxford, set about the monumental task of finding, analyzing, systematically reviewing, and synthesizing the data from all controlled trials in perinatal medicine. One of us (MWE) was able to join him early in his endeavors, when he first embarked on his massive international search, at first without the benefit of computers. The result, 10 years later, was the electronic Oxford Database of Perinatal Trials (3), and then an encyclopedic, multi-authored, two-volume text (4), the first systematic review of controlled trials in any entire medical specialty. This was shortly followed by a paperback synopsis of the larger work (5), which could make the findings more generally available to a wider audience. At first all went well. Most of the time the results came out as we had predicted, and they provided good rationale for the changes we wanted to see in existing practices. When controlled trials failed to show any benefits from traditional practices like predelivery shaving (6,7) and enemas (8), we could convincingly recommend that such uncomfortable practices be discontinued. We were happy when trials of labor support from either companions or professionals showed clear improvements in outcome (9) for the mothers, as we knew it would. We were excited when randomized trials demonstrated effective ways to prevent serious disorders (10), or clearly resolved controversies about the best treatment for dangerous pregnancy complications (11). We believed we had the answer. This pioneering approach received academic accolades (12), but seemed to have little influence on obstetrical practice. It took what seemed to us to be ages before the profession and the public began to appreciate how effective randomized trials could be as a way to choose between alternative forms of care. We should have been more patient; shifts in paradigms do not occur quickly. The age of Galenic medicine lasted 1,200 years, the preeminence of pathology in medical thinking continued for over 300 years, and the deference to obstetrical expertise predominated for most of the last century. But the status quo shifted, in what seemed to be overnight. Only history can tell us who or what will receive the credit (or blame) for the meteoric rise of what came to be called evidence-based medicine. The term was coined by Gordon Guyatt, of McMaster University in Hamilton, Ontario, only in 1991 (13), and achieved widespread acceptance the following year (14). Over a period of years rather than decades or centuries the paradigm changed. The Cochrane
BIRTH 33:4 December 2006
Collaboration began as a single center in the United Kingdom, then became an immense multinational organization (15). The Oxford Database of Perinatal Trials became only a tiny part of the Cochrane Library, which systematically reviewed randomized trials of health care in almost all health disciplines. ‘‘Evidence-based medicine’’ became the new mantra, the new authority.
Authority Can Be Wrong And therein lies the rub. Any authority can be wrong, and can mislead even when it is right. Just as clinical expertise had contributed so tremendously to the improvement in health care and in health but was sometimes mistaken, so evidence-based obstetrics has answered many important questions about ‘‘best’’ care in pregnancy and childbirth, but by no means all. When the results of trials do not conform to what we believe from other evidence, we had to devise ad hoc rationalizations to explain the discrepancies. A systematic review of trials of antenatal education for labor or parenthood failed to demonstrate any clearcut benefits from our intervention. The authors concluded only that ‘‘No recommendations for practice changes can be made at this time since there exists insufficient evidence to determine the effects of person-to-person antenatal education for childbirth and/or parenthood’’ (16). Yet women obviously get something from prenatal classes, because they continue to attend them. A Cochrane review of programs offering comprehensive social support for pregnant women at high risk of having an overly small baby failed to demonstrate improvements in any measurable perinatal outcomes (17). The authors of this review explained their unexpected results with the comment, ‘‘Pregnant women need and deserve to have the help and support of caring family members, friends, and health professionals. However, such support is unlikely to be powerful enough to overcome the effects of a lifetime of poverty and disadvantage, or a longstanding pregnancy complication, and thereby influence the remaining course of a pregnancy’’ (17). Electronic fetal heart rate monitoring has been extensively studied, and the only clinically significant benefit demonstrated from its use is a reduction of neonatal seizures (18), which was subsequently shown to have no long-term adverse effects (19). The counter-intuitive results of these trials were largely ignored. Electronic fetal monitoring is still almost routinely practiced in many hospitals. Sometimes, of course, the way a trial is conceived or carried out is not perfect. Questions can be, and have
BIRTH 33:4 December 2006
been, raised about the methodology of even some recent, major, large-scale, highly influential trials (20-23). More fundamentally, however, questions have been, and increasingly are being, raised about the types of question that trials can address, and the extent to which they should influence our care (24). Even good evidence can lead to bad practice if applied in an unthinking way. Moreover, there are biases inherent in even the best trials. One such bias is in the choice of outcomes to be studied. These must be specified in advance, and hence reflect the values (and biases) of the investigators. They must be limited in number; if we choose too many outcomes to study, some will be statistically significant by chance; if we choose too few, how can we tell if we really chose the right ones? We are limited by cost and feasibility. We have to study the outcomes that we can study, rather than the outcomes we would like to study. We tend to use surrogate rather than clinical outcomes, short-term outcomes rather than long term, in the necessary interest of feasibility. Perhaps the most significant and ubiquitous bias is in the choice of what to study. Evidence-based obstetrics pays attention to the interventions that have been studied in randomized trials. Other, potentially more useful interventions that have not been studied by randomized trials, tend to be ignored. Sometimes the determining factor is the commercial importance of the intervention studied, or a vested interest of either the researcher or the funding agency. And some things just cannot be studied by randomized trials. Randomized trials, currently at the top of the evidence-based hierarchy, are perfectly suited to evaluate the (average) relative effects of alternative forms of care, for both simple and complicated problems, where the form of care used is the principal cause of the outcome found. They are less suitable, and often seriously misleading, for complex problems, where the outcomes depend more on the web of interactions between the care, the individuals concerned, and the context in which they occur.
From Complicated to Complex It is well, at this point, to try to clarify the differences among simple, complicated, and complex issues (25) (Table 1). Some problems, such as how to bake a cake, or how to suture an episiotomy, are pretty straightforward. You get a good recipe, and you follow it. We have lots of good recipes in maternity care: corticosteroids for lung maturation with preterm birth; cesarean section for placenta previa; transfusions for women who have lost excessive blood; magnesium sulphate for eclampsia. There is a direct, linear rela-
tionship between what is done and what results. Properly used, the results will be as expected. A recipe must be tested. When an effect is sufficiently large, it can be tested by simple observation. For smaller, but still important effects, these recipes can only be adequately tested by randomized trials. Other basically linear problems are more complicated. You need much more than a simple recipe to put a space ship in orbit, or manage a brittle diabetes. You need sophisticated equipment and a highly trained team. But if you make your preparations carefully, have everything in place, and take meticulous care at every step, you can be reasonably confident that you will likely succeed. With specialized skills, advanced technology, and coordinated teamwork, many complicated problems in maternity care can now be confidently addressed. Severe diabetes can be managed successfully, genetic deviations from normal can be anticipated, fetal growth can be monitored. Randomized trials play a major role in the evaluation of these complicated recipes But when a problem is complex, rather than just complicated, we can never be entirely sure of what is going to happen. There is no direct, linear effect between what we do, and what results. Perhaps the best, or the most familiar, example of a complex problem is how to raise one’s child. Formulas, recipes, guidelines, have a limited application. Expertise in parenting can help, but is not enough. Every child is unique, and an approach that results in a successful outcome for one child may be disastrous for another. Many of the influences that determine how your child will grow up are beyond your control. Chance events can have far-reaching consequences: a meeting with a friend, or with a bully; a teacher who fosters your child’s talents, or one who blunts her curiosity. Despite all this, the birth of a child is a time for optimism. We look forward to raising that child as both a challenge and an opportunity. We have solved, or are well on the way to solving, many of our complicated problems in maternity care. These respond to the well-established toolkit of the evidence-based paradigm. How can we dare to even think of challenging this paradigm that has served us so well, that has proved so successful, that has helped us to cure some of our most serious diseases, to solve so many ‘‘insoluble’’ problems, to answer so many ‘‘unanswerable’’ questions?
Breaking Through the Tangle Because we must dare, we must think. Many, if not most, of our remaining problems are complex ones, rather than merely complicated. They have multiple,
BIRTH 33:4 December 2006
Table 1. Differences Among Simple, Complicated, and Complex Issues*
Linear Simple Baking a Cake or Suturing an Episiotomy
Sending a Rocket into Orbit or Managing a Brittle Diabetes
Raising a Child or Delivering a Baby
Recipe is essential Recipe is tested to assure easy replication
Recipe is critical Doing once increases assurance that the next will be OK
No particular expertise is required, but expertise increases success rate
High levels of expertise in a variety of ﬁelds are necessary for success
Recipes produce standard products
Rockets, like surgical procedures, are similar in critical ways
Direct causal relationship between what is done and what occurs Best recipes give good results every time Test with randomized trial
Direct causal relationship between what is done and what occurs There is a high degree of certainty about the outcome Test with randomized trial
An optimistic approach to the problem is possible
An optimistic approach to the problem is possible
Recipes have limited application Raising one child provides experience, but no assurance of success with the next Expertise can contribute, but is neither necessary nor sufﬁcient to assure success Every child, like every birth, is unique, and must be understood as an individual Results are contingent on factors beyond the control of the operator Uncertainty of the outcome remains Test with appropriate methodology; qualitative, quantitative, narrative, innovative An optimistic approach to the problem is possible
*Modified from Glouberman and Zimmerman (25).
interrelated, interconnected, interwoven, hopelessly tangled causes. They respond in unexpected ways to well-intentioned interventions, even ones based on good evidence. They are not tidy, like respectable problems should be. They demand a new approach. The fundamental mistake of evidence-based medicine, evidence-based obstetrics, is to treat complex problems as if they were merely complicated. We have hoped, even expected, that medical research would reduce our apparently complex problems to their simple components, so that we could address them with our current paradigm. But clinical problems, like preterm labor or small-for-gestational-age babies, do not fit into our predefined slots. Neither do more broadly based problems like overextended facilities, shortage of maternity care practitioners, the current epidemic of cesarean sections, or the specter of medicolegal liability that frightens and constrains us. Our growing understanding of the nature of complex systems, from the physical to the social, can help us understand the changes that are taking place in maternity care. The strength and acceptance of family-centered over practitioner-centered care, the resurgence of midwifery as an honored profession, all speak to the power of self-organization within complex systems. Naı¨ ve efforts to simplify the management of pregnancy and childbirth through standardized formulas, evidence-based protocols, are failing, and we are
beginning to recognize anew the complexity of pregnancy and birth as life events to be experienced, rather than diseases to be managed. The present evidencebased paradigm, while giving lip service to women and their partners, and the context of each pregnancy and birth, fails to fully appreciate that it is not simply the woman or the setting, the attendant or the policies, that influence the outcome. Rather, it is the complex interrelationships among these separate elements. Although not new, this renewed understanding can guide us in our search to constantly improve care, at all levels. It cannot provide us with the comforting protocols that we have come to expect, but it can point us to the steps we can take to move forward. First and foremost, we need to accept the uncomfortable reality that there are no comprehensive formulas. A cookbook for maternity care is not in the cards. The fruitless search for the magic bullet can only lead to frustration. Second, we must learn to think of the relationships among the disparate factors that influence each birth, each setting, each situation, rather than of the factors in isolation. We must allow new forms of research to evolve, to produce new kinds of evidence (26), and to accept the value of this new evidence. Third, we must advance quickly by moving slowly. A revolution is neither needed nor desirable. Many aspects of current understanding, current approaches, and many current practices work very well. We must be
BIRTH 33:4 December 2006
careful not to discard them, but to nourish them, adapt them, and build on them. We can, we have, and we will. Murray W. Enkin, MD, FRCSC, LLD(Hon), DSc(Hon) Sholom Glouberman, PhD Philip Groff, PhD Alejandro R. Jadad, MD, DPhil, FRCPC Anita Stern, RN* Murray Enkin is Professor Emeritus, McMaster University, Hamilton, Ontario; Sholom Glouberman is Philosopher in Residence, Baycrest Centre for Geriatric Care, Toronto; Philip Groff is Director, Research and Evaluation, Smartrisk, Toronto; Alejandro R. Jadad is Professor, Chair, Chief Innovator, Centre for Global eHealth Innovation, Toronto; and Anita Stern is a Doctoral candidate, McMaster University, Hamilton, Ontario, Canada. *For the Clinamen Collaboration: The Clinamen Collaboration is a small study group from various parts of the health care field, who are trying to understand more about the complex nature of health. It includes a philosopher, a psychologist, a nurse, and several physicians.
References 1. Porges RE. The response of the New York Obstetrical Society to the report by the New York Academy of Medicine on maternal mortality, 1933-1934. Am J Obstet Gynecol 1985; 62:642–649. 2. Cochrane AL. 1931-1971: A critical review with particular reference to the medical profession. In: Teeling-Smith G, Wells N, eds. Medicines for the Year 2000. London: Office of the Health Economics, 1979. 3. Chalmers I, Hetherington J, Newdick M, Mutch L, Grant A, Enkin M, Enkin E, Dickersin K. The Oxford Database of Perinatal Trials: Developing a register of published reports of controlled trials. Control Clin Trials 1986;7:306–324. 4. Chalmers I, Enkin M, Keirse, MJNC., eds. Effective Care in Pregnancy and Childbirth. Oxford: Oxford University Press, 1989. 5. Enkin M, Keirse MJNC, Chalmers I. A Guide to Effective Care in Pregnancy and Childbirth. Oxford: Oxford University Press, 1989. 6. Johnston RA, Sidall RS. Is the usual method of preparing patients for delivery beneficial or necessary? Am J Obstet Gynecol 1922;4:645–650. 7. Kantor HI, Rember R, Tabio P, Buchanon R. Value of shaving the pudendal-perineal area in delivery preparation. Obstet Gynecol 1965;25:509–512. 8. Romney ML, Gordon H. Is your enema really necessary? BMJ 1981;282(6272):1269–1271.
269 9. Hodnett ED, Gates S, Hofmeyr GJ, Sakala C. Continuous support for women during childbirth. The Cochrane Database of Systematic Reviews 2003, Issue 3. Art. No.: CD003766. DOI: 10.1002/14651858. 10. Liggins GC, Howie RN. A controlled trial of antepartum glucocorticoid treatment for prevention of the respiratory distress syndrome in premature infants. Pediatrics 1972;50:515–525. 11. Duley L, Henderson-Smart D. Magnesium sulphate versus diazepam for eclampsia. The Cochrane Database of Systematic Reviews 2003, Issue 4. Art. No.: CD000127. DOI: 10.1002/ 14651858. 12. Chalmers I, Enkin MW, Keirse MJNC. Establishing systems for preparing and updating systematic reviews of randomized controlled trials of health care. Milbank Q 1993;71:1–25. 13. Guyatt G. Evidence-based medicine. ACP J Club 1991;114-A-16. 14. Guyatt G, Cairns J, Churchill D, et al. [‘Evidence-Based Medicine Working Group’] Evidence-based medicine. A new approach to teaching the practice of medicine. JAMA 1992;268: 2420–2425. 15. Chronology of the Cochrane Collaboration. Available at: http://www.cochrane.org/docs/cchronol.htm. Accessed March 3, 2006. 16. Gagnon AJ. Individual or group antenatal education for childbirth/parenthood. The Cochrane Database of Systematic Reviews 2000, Issue 4. Art. No.: CD002869. DOI: 10.1002/ 14651858. 17. Hodnett ED, Fredericks S. Support during pregnancy for women at increased risk of low birthweight babies. The Cochrane Database of Systematic Reviews 2003, Issue 3. Art. No.: CD000198. DOI: 10.1002/14651858. 18. Thacker SB, Stroup D, Chang M. Continuous electronic heart rate monitoring for fetal assessment during labor. The Cochrane Database of Systematic Reviews 2001, Issue 2. Art. No.: CD000063. DOI: 10.1002/14651858. 19. Grant A, O’Brien N, Joy MT, et al. Cerebral palsy among children born during the Dublin randomized trial of intrapartum monitoring. Lancet 1989:8674:1233–1236. 20. Menticoglou SM, Hall PF. Routine induction of labour at 41 weeks gestation: Nonsensus consensus. BJOG 2002;109: 485–491. 21. Kotaska A. Inappropriate use of randomised trials to evaluate complex phenomena: Case study of vaginal breech delivery. BMJ 2004;329:1039–1042. 22. Keirse, MJNC. Evidence-based childbirth only for breech babies? Birth 2002;29:55–59. 23. Glezerman M. Five years to the term breech trial: The rise and fall of a randomized controlled trial. Am J Obstet Gynecol 2006;194:20–25. 24. Grossman J, Mackenzie FJ. The randomized controlled trial: Gold standard, or merely standard? Perspect Biol Med 2005; 48:516–534. 25. Glouberman S, Zimmerman B. Complicated and Complex Systems: What Would Successful Reform of Medicare Look Like? Ottawa: Commission on the Future of Health Care in Canada, 2002. 26. Enkin MW, Jadad AR. Using anecdotal information in evidence based health care: Heresy or necessity? Ann Oncol 1998; 9:963–966.