Evidence And Non-Evidence Based Treatment Options
Evidence And Non-Evidence Based Treatment Options
Limitations to Evidence Based Practice
THOMAS MAIER
The promotion of evidence-based medicine (EBM) or, more generally, of evidence-based practice (EBP) has strongly characterized most medical disciplines over the past 15 to 20 years. Evidence-based medicine has become a highly influential concept in clinical practice, medical education, research, and health policy. Although the evidence-based approach has also been increasingly applied in related fields such as psychology, education, social work, or economics, it was and still is predominantly used in medicine and nursing. Evidence-based practice is a general and nonspecific concept that aims to improve and specify the way decision makers should make decisions. For this purpose it delineates methods of how professionals should retrieve, summarize, and evaluate the available empirical evidence in order to identify the best possible decision to be taken in a specific situation. So EBP is, in a broader perspective, a method to analyze and evaluate large amounts of statistical and empirical information to understand a particular case. It is therefore not limited to specific areas of science and is potentially applicable in any field of science using statistical and empirical data. Many authors often cite Sackett, Rosenberg, Muir Gray, Haynes, and Richardson’s (1996) article entitled “Evidence-based medicine:What it is and what it isn’t” as the founding deed of evidence-based practice. David L. Sackett (born 1934), an American-born Canadian clinical epidemiologist, was professor at the Department of Clinical Epidemiology and Biostatistics of McMaster University Medical School of Hamilton, Ontario, from 1967 to 1994. During that time, he and his team developed and propagated modern concepts of clinical epidemiology. Sackett later moved to England, and from 1994 to 1999, he headed the National Health Services’ newly founded Centre for Evidence-Based Medicine at Oxford University. During that time, he largely promoted EBM in Europe by publishing articles and textbooks as well as by giving numerous lectures and training courses. David Sackett is seen by many as the founding father of EBM as a proper discipline, although he would not at all claim this position for himself. In fact, Sackett promoted and elaborated concepts that have been described and used by others before; the origins of EBM are rooted back in much earlier times. The foundations of clinical epidemiology were already laid in the 19th century mainly by French, German, and English physicians systematically studying the prevalence and course of diseases and the effects of therapies. As important foundations of the EBMmovement, certainly the works and insights of the Scottish epidemiologist Archibald (Archie) L. Cochrane (1909–1988) have to be c04 18 April 2012; 19:44:27 55 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. mentioned. Cochrane, probably the true founding father of modern clinical epidemiology, had long before insisted on sound epidemiological data, especially from RCTs, as the gold standard to improve medical practice (Cochrane, 1972). In fact, the evaluation of epidemiological data has always been one of the main sources of information in modern academic medicine, and many of the most spectacular advances of medicine are direct consequences of the application of basic epidemiological principles such as hygiene, aseptic surgery, vaccination, antibiotics, and the identification of cardiovasular and carcinogenic risk factors. One of the most frequent objections against the propagation of EBM is, “It’s nothing new, doctors have done it all the time.” Rangachari, for example, apostrophized EBM as “old French wine with a new Canadian label” (Rangachari, 1997, p. 280) alluding to the French 19th century epidemiology pioneer Pierre Louis, who was an influencing medical teacher in Europe and North America, and to David L. Sackett, the Canadian epidemiologist. Even though the “conscientious, explicit and judicious use of the current best evidence in making decisions about the care of individual patients” (Sackett et al., 1996, p. 71) seemsto be a perfectly reasonable and unassailable goal, EBM has been harshly criticized from the very beginning of its promotion (Berk &Miles Leigh, 1999; B. Cooper, 2003; Miles, Bentley, Polychronis, Grey, and Price, 1999; Norman, 1999; Williams & Garner, 2002). In 1995, for example, the editors of The Lancet chose to publish a rebuking editorial against EBM entitled “Evidence-based medicine, in its place” (The Lancet, 1995): The voice of evidence-based medicine has grown over the past 25 years or so from a subversive whisper to a strident insistence that it is improper to practise medicine of any other kind. Revolutionaries notoriously exaggerate their claims; nonetheless, demands to have evidence-based medicine hallowed as the new orthodoxy have sometimes lacked finesse and balance, and risked antagonising doctors who would otherwise have taken many of its principles to heart. The Lancet applauds practice based on the best available evidence–bringing critically appraised news of such advances to the attention of clinicians is part of what peer-reviewed medical journals do–but we deplore attempts to foist evidencebased medicine on the profession as a discipline in itself. (p. 785) This editorial elicited a fervid debate carried on for months in the letter columns of The Lancet. Indeed, there was a certain doggedness on both sides at that time, astonishing neutral observers and rendering the numerous critics even more suspicious. The advocates of EBM on their part acted with great self-confidence and claimed no less than to establish a new discipline and to put clinical medicine on new fundaments; journals, societies, conferences, and EBM training courses sprang up like mushrooms; soon academic lectures and chairs emerged; however, this clamorous and pert appearance of EBM repelled many. A somehow dogmatic, almost sectarian, tendency of the movement was noticed with discontent, and even the deceased patron saint of EBM, Archie Cochrane, had to be invoked in order to push the zealots back: How would Archie Cochrane view the emerging scene? His contributions are impressive, particularly to the development of epidemiology as a medical science, but would he be happy about all the activities linked with his name? He was a freethinking, iconoclastic individual with a healthy cynicism, who would not accept dogma. He brought an open sceptical approach to medical problems and we think that he would be saddened to find that his name now embodies a new rigid medical orthodoxy while the real impact of his many achievments might be overlooked. (Williams & Garner 2002, p. 10) THE DEMOCRATIZATION OF KNOWLEDGE How could such an emotional controversy arise about the introduction of a scientific 56 Overview and Foundational Issues c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. method (Ghali, Saitz, Sargious, & Hershman, 1999)? Obviously, the propagation and refusal of EBM have to be seen not only from a rational scientific standpoint but also from a sociological perspective (Miettinen, 1999; Norman, 1999): The rise of the EBM movement fundamentally reflects current developments in contemporary health care concerning the allocation of information, knowledge, authority, power, and finance (Berk & Miles Leigh, 1999), a process becoming more and more critical during the late 1980s and the 1990s. Medicine has, for quite some time, been losing its prestige as an intangible, moral institution. Its cost-value ratio is questioned more and more and doctors are no longer infallible authorities. We do not trust doctors anymore to know the solution for any problem; they are supposed to prove and to justify what they do and why they do it. These developments in medicine parallel similar tendencies in other social domains and indicate general changes in Western societies’ self-conception. Today we are living in a knowledge society, where knowledge and information is democratized, available and accessible to all. There is no retreat anymore for secret expert knowledge and for hidden esoteric wisdom. The hallmarks of our time are free encyclopedic databases, open access, the World Wide Web, and Google©. In the age of information, there are no limitations for filing, storage, browsing, and scanning of huge amounts of data; however, this requires more and more expert knowledge to handle it. So, paradoxically, EBM represents a new specialized expertise that aims to democratize or even to abolish detached expert knowledge. The democratization of knowledge increasingly questions the authority and selfsufficiency of medical experts and has deeply unsettled many doctors and medical scientists. Of course, this struggle is not simply about authority and truth; it is also about influence, power, and money. For all the unsettled doctors, EBM must have appeared like a guide for the perplexed leading them out of insecurity and doubt. Owing to its paradoxical nature, EBM offers them a new spiritual home of secluded expertise allowing doctors to regain control over the debate and to reclaim authority of interpretation from bold laymen. For this purpose, EBM features and emphasizes the most valuable label of our time that is so believable in science: science- or evidencebased. In many areas of contention, terms like evidence-based or scientifically proven are used for the purpose of putting opponents on the defensive. Nobody is entitled to question a fact, which is declared evidence-based or scientifically proven. By definition, these labels are supposed to convey unquestioned and axiomatic truth. It requires rather complex and elaborate epistemological reasoning to demonstrate how even true evidence-based findings can at the same time be wrong, misleading, and/or useless. All these accounts and arguments apply in particular to the disciplines of psychiatry and clinical psychology, which have always had a marginal position among the apparently respectable disciplines of academic medicine. Psychiatrists and psychologists always felt particularly pressured to justify their actions and are constantly suspected to practice quackery rather than rational science. It is therefore not surprising that among other marginalized professionals, such as the general practitioners, psychiatrists and psychotherapists made particularly great efforts over the last years to establish their disciplines as serious matters of scholarly medicine by diligently adopting the methods of EBM (Geddes & Harrison, 1997; Gray & Pinson, 2003; OakleyBrowne, 2001; Sharpe, Gill, Strain, & Mayou, 1996). Yet, there are also specific problems limiting the applicability of EBP in these disciplines.
EMPIRICISM AND REDUCTIONISM In order to understand the role and function of EBP within the scientific context, it may be Limitations to Evidence-Based Practice 57 c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. helpful to give a brief overview of the theoretical backgrounds of science in general. What is science and how does it proceed? Science can be seen as a potentially endless human endeavour that aims to understand and determine reality. Not only are physical objects matters of science, but also immaterial phenomena like language, history, society, politics, economics, human behavior, thoughts, or emotions. Starting with the Greek scientists in the ancient world, but progressing more rapidly with the philosophers of the Enlightenment, modern science adopted defined rules of action and standards of reasoning that delineate science from nonscientific knowledge such as pragmatics, art, or religion. Unfortunately, notions like science, scientific, or evidence are often wrongly used in basically nonscientific contexts causing unnecessary confusion. The heart and the starting point of any positive science is empiricism, meaning the systematic observation of phenomena. Scientists of any kind must start their reasoning with observations, possibly refined through supportive devices or experimental arrangements. Although positive science fundamentally believes in the possibility of objective perception, it also knows the inherent weaknesses of reliability and potential sources of errors. Rather than have confidence in single observations, science trusts repeated and numerous observations and statistical data. This approach rules out idiosyncratic particularities of single cases to gain the benefit of identifying the common characteristics of general phenomena (i.e., reductionism). This approach of comprehending phenomena by analytically observing and describing them has in fact produced enormous advancements in many fields of science, especially in technical disciplines; however, contrasting and confusing gaps of knowledge prevail in other areas such as causes of human behavior, mind–body problems, or genome–environment interaction. Some areas of science are apparently happier and more successful using the classical approach of positive science, while other disciplines feel less comfortable with the reductionist way of analyzing problems. The less successful areas of science are those studying complex phenomena where idiosyncratic features of single cases can make a difference, in spite of perfect empirical evidence. This applies clearly to medicine, but even more to psychology, sociology, or economics. Medicine, at least in its academic version, usually places itself among respectable sciences, meeting with and observing rules of scientific reasoning; however, this claim may be wishful thinking and medicine is in fact a classical example of a basically atheoretical, mainly pragmatic undertaking pretending to be based on sound science. Inevitably, it leads to contradictions when trying to bring together common medical practice and pure science. COMPLEXITY Maybe the deeper reasons for these contradictions are not understood well enough. Maybe they still give reason for unrealistic ideas to some scientists. A major source of misconception appears to be the confused ontological perception of some objects of scientific investigation. What is a disease, a disorder, a diagnosis? What is human behavior? What are emotions? Answering these questions in a manner to provide a basis for scientific reasoning in a Popperian sense (see later) is far from trivial. Complex objects of science, like human behavior, medical diseases, or emotions, are in fact not concrete, tangible things easily accessible to experimental investigation. They are emergent phenomena, hence they are not stable material objects, but exist only as transitory, nonlocal appearances fluctuating in time. They continuously emerge out of indeterminable complexity through repeated self-referencing operations in complex systems (i.e., autopoietic systems). Indeterminable complexity or deterministic chaos means that a huge number of mutually interacting parameters autopoietically 58 Overview and Foundational Issues c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. form a system, rendering any precise calculation of the system’s future conditions impossible. Each single element of the system perfectly follows the physical rules of causality; however, the system as a whole is nevertheless unpredictable. Its fluctuations and oscillations can be described only probabilistically. In order to obtain reasonable and useful information about a system, many scientific disciplines have elaborated probabilistic methods of approaching their objects of interest. Thermodynamics, meteorology, electroencephalography, epidemiology, and macroeconomics are only a few such examples. Most structures in biological, social, and psychological reality can be conceived as emergent phenomena in this sense. Just as the temperature of an object is not a quality of the single molecules forming the object—a single molecule has no temperature— but a statistic description of a huge number of molecules, human behavior cannot be determined through the description of composing elements producing the phenomenon—for example, neurons—even if these elements are necessary and indispensable preconditions for the emergence of the phenomenon. The characteristics of the whole cannot be determined by the description of its parts. When the precise conditions of complex systems turn out to be incalculable, the traditional reaction of positive science is to intensify analytical efforts and to compile more information about the components forming the system. This approach allows scientists to constantly increase their knowledge about the system in question without ever reaching a final understanding and a complete determination ofthe function ofthe system. This is exactly what happens currently in neurosciences. Reductionist approaches have their inherent limitations when it comes to the understanding of complex systems. A similar problem linked to complexity that is particularly important is the assumed comparability of similar cases. In order to understand an individual situation, science routinely compares defined situations to similar situations or, even better, to a large number of similar situations. Through the pooling of large numbers of comparable cases, interfering individual differences are statistically eliminated, and only the common ground appears. The conceptual assumption behind this procedure is that similar—but still not identical— cases will evolve similarly under identical conditions. One of the most important insights from the study of complex phenomena is that in complex systems very small differences in initial conditions may lead to completely different outcomes after a short time—the socalled butterfly effect. This insight is well known to natural scientists; however, clinical epidemiologists do not seem to be completely aware of the consequences of the butterfly effect to their area of research. FROM KARL POPPER TO THOMAS S. KUHN Based on epistemological considerations, the Anglo-Austrian philosopher Karl Popper (1902–1994) demonstrated in the 1930s the limitations of logical empiricism. He reasoned that general theories drawn from empirical observations can never be proven to be true. So, all theories must remain tentative knowledge, waiting to be falsified by contrary observations. In fact, Popper conceived the project of science as a succession of theories to be falsified sooner or later and to be replaced by new theories. This continuous succession of new scientific theories is the result of natural selection of ideas through the advancement of science. According to Popper, any scientific theory must be formulated in a way to render it potentially falsifiable through empirical testing. Otherwise, the theory is not scientific: It may be metaphysical, religious, or spiritual instead. This requires that a theory must be formulated in terms of clearly defined notions and measurable elements. Popper’s assertions were later qualified as being less absolute by the American philosopher of science Thomas S. Kuhn (1922–1996). Limitations to Evidence-Based Practice 59 c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. Kuhn, originally a physicist, pointed out that in real science any propagated theory could be falsified immediately by contrary observations because contradicting observations are always present; however, science usually ignores or even suppresses observations dissenting with the prevailing theory in order to maintain the accepted theory. Kuhn calls the dissenting observations anomalies, which are—according to him—always obvious and visible to all, but nevertheless blinded out of perception in order to maintain the ruling paradigm. In Kuhn’s view, science will never come to an end and there will never be a final understanding of nature. No theory will ever be able to integrate and explain consistently all the observations drawn from nature. At this point, even the fundamental limitations to logical scientific reasoning demonstrated by Go¨del’s incompleteness theorems become recognizable (cf. also Sleigh, 1995). Based on his considerations, Kuhn clear-sightedly identified science to be a social system, rather than a strictly logical and rational undertaking. Science, as a social phenomenon, functions according to principles of Gestalt psychology. It sees the things it wants to see and overlooks the things that do not fit. In his chief work The Structure of Scientific Revolutions, Kuhn (1962) gives several examples from the history of science supporting this interpretation. It is in fact amazing to see how difficult it was for most important scientific breakthroughs to become acknowledged by the contemporary academic establishment. Kuhn uses the notion normal science to characterize the established academic science and emphasizes the self-referencing nature of its operating mode. Academic teachers teach students what the teachers believe is true. Students have to learn what they are taught by their teachers if they want to pass their exams and get their degrees. Research is mainly repeating and retesting what is already known and accepted. Journals, edited and peerreviewed by academic teachers, publish what conforms with academic teachers’ ideas. Societies and associations—headed by the same academic teachers—ensure the purity of doctrine by sponsoring those who confirm the prevailing paradigms. Dissenting opinions are unwelcome. Based on Kuhn’s view of normal science, EBP and EBM can be identified as classical manifestations of normal science. The EBP helps to ensure the implementation of mainstream knowledge by declaring to be most valid what is best evaluated. Usually the currently established practices are endorsed by the best and most complete empirical evidence; dissenting ideas will hardly be supported by good evidence, even if these ideas are right. Since EBP instructs its adherers to evaluate the available evidence on the basis of numerical rules of epidemiology, arguments like plausibility, logic consistency, or novelty are of little relevance. AN EXAMPLE FROM RECENT HISTORY OF CLINICAL MEDICINE When in 1982 the Australian physicians Barry Marshall and Robin Warren discovered Helicobacter pylori in the stomachs of patients with peptic ulcers, their findings were completely ignored and neglected by the medical establishment of that time. The idea that peptic ulcers are provoked by an infectious agent conflicted with the prevailing paradigm of academic gastroenterology, which conceptualized peptic ulcers as a consequence of stress and lifestyle. Although there had been numerous previous reports of helicobacteria in gastric mucosa, all these findings were completely ignored because they conflicted with the prevailing paradigm. As a consequence Marshall and Warren’s discovery was ignored for years because it fundamentally challenged current scientific opinion. They were outcast by the scientific community, and only 10 years later their ideas slowly started to convince more and more clinicians. Now, 25 years later, it is common basic clinical knowledge that 60 Overview and Foundational Issues c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. Helicobacter pylori is one of the major causes of peptic ulcers, and eradication therapy is the accepted and rational therapy for gastric ulcers. Finally, in 2005 Barry Marshall and Robin Warren gained the Nobel Price for their discovery (Parsonnet, 2005). BENEFITS AND RISKS OF EVIDENCE-BASED PRACTICE The true benefits of EBP for patients and society in terms of outcomes and costs have not been proven yet—at least not through sound empirical evidence (B. Cooper, 2003; Geddes & Harrison, 1997). Nevertheless, there is no doubt that the method has a beneficial and useful potential. Many achievements of EBP are undisputable and undisputed, hence they are evident. Owing to the spread of methodical skills in retrieving and evaluating the available epidemiological evidence, it has become much harder to apply any kind of obscure or idiosyncratic practices. The experts’ community, as well as the customers and the general public, are much more critical toward pretended effects of treatments and ask for sound empirical evidence of effectiveness and safety. It is increasingly important not only to know the best available treatment, but also to prove it. The EBP is therefore a helpful instrument for doctors and therapists to justify and legitimate their practices to insurance, judiciary, politics, and society. Furthermore, individual patients might be less at risk to wrong or harmful treatment due to scientific misapprehension. Of course, common malpractice owing to inanity, negligence, or viciousness will never be eliminated, not even by the total implementation of EBP; however, treatment errors committed by diligent and virtuous doctors are minimized through careful adherence to rational guidelines. In general, clinical decision-making paths have become more comprehensible and rational, probably also due to the spread of EBP. As medicine is in fact not a thoroughly scientific matter (Ghali et al. 1999), continuous efforts are needed to enhance and renew rationality. The EBP contributes to this task and helps clinicians to maintain rationality in a job where inscrutable complexity is daily business. In current medical education, the algorithms of EBP are now instilled into students as a matter of course. Seen from that perspective, EBP is also an instrument of discipline and education, for it compels medical students and doctors to reflect continuously all their opinions and decisions scientifically (Norman, 1999). Today EBP has a great impact on the education and training of future doctors, and it thereby enhances the uniformity and transparency of medical doctrine. This international alignment of medical education with the principles of EBP will, in the long run, allow for better comparability of medical practice all over the world. This is an important precondition for the planning and coordination of research activities. Thus, the circle of normal science is perfectly closed through the widespread implementation of EBP. GENERAL LIMITATIONS TO EVIDENCE-BASED PRACTICE It has been remarked, not without reason, that the EBP movement itself has adopted features of dogmatic authority (B. Cooper, 2003; Geddes et al., 1996; Miles et al., 1999). This appears particularly ironic, because EBP explicitly aims to fight any kind of orthodox doctrine. The ferocity of some EBP adherents may not necessarily hint at conceptual weaknesses of the method; rather, it is more likely a sign of an iconoclastic or even patricidal tendency inherent to EBP. Young, diligent scholars, even students, possibly without any practical experience, are now entitled to criticize and rectify clinical authorities (Norman, 1999). This kind of insurgence must evoke resistance from authorities. If the acceptance of EBP among clinicians should be enhanced, Limitations to Evidence-Based Practice 61 c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. it is advisable that the method is not only propagated by diligent theoreticians, but mainly by experienced practitioners. One of the first and most important arguments against EBP is reductionism (see earlier, Welsby, 1999). Complex and maybe fundamentally diverse clinical situations of individual patients have to be condensed and aggregated to generalized questions in order to retrieve empirical statistical evidence. Important specific information about the individual cases is inevitably lost owing to this generalization. The usefulness of the retrieved evidence is therefore inevitably diluted to a very general and dim level. Of course, there are some frequently used standard interventions, which are really based upon good empirical evidence (Geddes et al., 1996). EXAMPLES FROM CLINICAL MEDICINE Scabies, a parasitic infection of the skin, is an important public health problem, mainly in resource-poor countries. For the treatment of the disease, two treatment options are recommended: topical permethrin and oral ivermectin. Both treatments are known to be effective and are usually well tolerated. The Cochrane Review concluded from the available empirical evidence that topical permethrin appears to be the most effective treatment of scabies (Strong & Johnstone, 2007). This recommendation can be found in up-to-date medical textbooks and is familiar to any well-trained doctor. Acute otitis media in children is one of the most common diseases, one ofthe main causes for parents to consult a pediatrician, and a frequent motive for the prescription of antibiotics, even though spontaneous recovery is the usual outcome. Systematic reviews have shown that the role of antibiotic drugs for the course of the disease is marginal, and there is no consensus among experts about the identification of subgroups who would potentially profit from antibiotics. In clinical practice, in spite of lacking evidence of its benefit, the frequent prescription of antibiotic drugs is mainly the consequence of parents’ pressure and doctors’ insecurity. A recent meta-analysis (Rovers et al., 2006) found that children youngerthan 2 years of age with bilateral acute otitis media and those with otorrhea benefited to some extent from antibiotic treatment; however, even for these two particular conditions, differences were moderate: After 3–7 days, 30% of the children treated with antibiotics still had pain, fever, or both, while in the control group the corresponding proportion was 55%. So, the available evidence to guide a clinician when treating a child with acute otitis media is not really significant and the decision will mostly depend on soft factors like parents’ preferences or practical and economical considerations. Evidently, clinicians choosing these interventions do not really need to apply the algorithms of EBP to make their decisions. They simply administer what they had learned in their regular clinical training. The opponents of EBP rightly argue that the real problems in clinical practice arise from complex, multimorbid patients presenting with several illnesses and other factors that have to be taken into account by the treating clinician. In order to manage such cases successfully there is usually no specific statistical evidence available to rely on. Instead, clinicians have to put together evidence covering some aspects of the actual case and hope that the resulting treatment will still work even if it is not really designed and tested for that particular situation. Good statistical evidence meeting the highest standards of EBP is almost exclusively derived from ideal monomorbid patients, who are rarely seen in real, everyday practice (Williams & Garner, 2002). It is not clear at all—and far from evidence-based—whether evidence from ideal cases can be transferred to 62 Overview and Foundational Issues c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. more complex cases without substantial loss of validity. Another argument criticizing EBP points at an epistemological problem. Because the EBP operates retrospectively by evaluating what was done in the past, it cannot directly contribute to developing new strategies and to finding new therapies. The EBP helps to consolidate well-known therapies, but cannot guide researchers toward scientific innovations. No scientific breakthrough will ever be made owing to EBP. On the contrary, if all clinicians strictly followed recommendations drawn from available retrospective evidence and never dared to try something different, science would stagnate in fruitless selfreference. There is a basically conservative and backward tendency inherent to the method. Although it cannot exactly be called antiscientific on that account (B. Cooper, 2003; Miles et al., 1999), EBP is a classical phenomenon of normal science (Kuhn, 1962). It will not itself be the source of fundamental new insights. Finally, there is an external problem with EBP, which is probably most disturbing of all: Production and compilation of evidence available to clinicians is highly critical and exposed to different nonscientific influences (Miettinen, 1999). Selection of areas of research is based more and more on economic interests. Large, sound, and therefore scientifically significant epidemiologic studies are extremely complex and expensive. They can be accomplished only with the support of financially potent sponsors. Compared with public bodies or institutions, private companies are usually faster and more flexible in investing important amounts of money into medical research. So, for many ambitious scientists keen on collecting publishable findings, it is highly appealing to collaborate with commercial sponsors. This has a significant influence on the selection of diseases and treatments being evaluated. The resulting body of evidence is necessarily highly unbalanced because mainly diseases and interventions promising important profits are well evaluated. For this reason, more money is probably put into trials on erectile dysfunction, baldness, or dysmenorrhea than on malaria or on typhoid fever. So, even guidelines based on empirical evidence—considered to be the ultimate gold standard of clinical medicine—turn out to be arbitrary and susceptible to economical, political, and dogmatic arguments (Berk & Miles Leigh, 1999). So, EBP’s goals to replace opinion and tendency by knowledge are in danger of being missed, if the relativity of available evidence is unrecognized. The uncritical promotion of EBP opens a clandestine gateway to those who have interests in controlling the contents of medical debates and have the financial means to do so. Biasing clinical decisions in times of EBP is probably no longer possible by false or absent evidence; however, the selection of what is researched in an EBP-compatible manner and what is published may result in biased clinical decisions (Miettinen, 1999). One of the most effective treatment options in many clinical situations—watchful waiting—is notoriously under-researched because there is no commercial or academic interest linked to that treatment option. Unfortunately, there will never be enough time, money, and workforce to produce perfect statistical evidence for all useful clinical procedures. So, even in the very distant future, clinicians will still apply many of their probably effective interventions without having evidence about their efficacy and effectiveness; thus, EBP is a technique of significant but limited utility (Green & Britten, 1998; The Lancet, 1995; Sackett et al., 1996). EXAMPLE FROM CLINICAL MEDICINE Lumbar back pain is one of the most frequent health problems in Western countries. About 5% of all low back problems are caused by prolapsed lumbar discs. The treatment is Limitations to Evidence-Based Practice 63 c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. mainly nonsurgical and 90% of acute attacks of nerve root pain (sciatica) settle without surgical intervention; however, different forms of surgical treatments have been developed and disseminated. Usually these methods are considered for more rapid relief in patients whose recovery is unacceptably slow. The Cochrane reviewers criticize that “despite the critical importance of knowing whether surgery is beneficial for disc prolapse, only four trials have directly compared discectomy with conservative management and these give suggestive rather than conclusive results” (Gibson & Waddell, 2007, p. 1). They concluded: Surgical discectomy for carefully selected patients with sciatica due to lumbar disc prolapse provides faster relief from the acute attack than conservative management, although any positive or negative effects on the lifetime natural history of the underlying disc disease are still unclear. (p. 2) Surgical treatments of low back pain hold an enormous commercial potential due to the worldwide frequency of the problem. It appears obvious that there are only a few trials comparing conservative treatment with surgery. SPECIFIC LIMITATIONS TO EBP IN PSYCHIATRY, PSYCHOTHERAPY, AND CLINICAL PSYCHOLOGY In psychiatry and psychotherapy, there is an ambivalent attitude toward EBP. Attempting to increase their scientific respectability, some psychiatrists and clinical psychologists zealously adopted EBP algorithms (Geddes & Harrison, 1997; Gray & Pinson, 2003; OakleyBrowne, 2001; Sharpe et al., 1996) and started evidence-based psychiatry. Others remain hesitant or doubtful about the usefulness of EBP in their field, and several authors have addressed different critical aspects of evidence-based psychiatry (Berk & Miles Leigh, 1999; Bilsker, 1996; Brendel, 2003; Geddes & Harrison, 1997; Goldner & Bilsker, 1995; Harari, 2001; Hotopf, Churchill, & Lewis, 1999; Lawrie, Scott, & Sharpe, 2000; Seeman, 2001; Welsby, 1999; Williams & Garner, 2002) with all of them fundamentally concerning practical and scientific particularities of psychiatry and clinical psychology. Next, we shall try to clarify these arguments. The evidence-based approach to individual cases is critically dependent on the validity of diagnoses. This is an axiomatic assumption of EBP, which is rarely analysed or scrutinized in detail. If in a concrete case no diagnosis could be attributed, the case would not be amenable to EBP, and no evidence could support decisions in such a case. If the diagnosis is wrong, or—even more intricate—if cases labeled with a specific diagnosis are still not homogenous enough to be comparable in relevant aspects, EBP will provide useless results. EXAMPLE FROM PSYCHIATRY According to DSM-IV, eating disorders are classified in different categories: anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder (BED), and eating disorder not otherwise specified (EDNOS). These categories are clinically quite distinct and diagnostic criteria are clear and easily applicable. In spite of the phenomenological diversity of the disease patterns, there is a close relationship between the different forms of eating disorders. In clinical practice, switches between different diagnoses and temporary remissions and relapses are frequent. In the course of time, patients may change their disease pattern several times: At times they may not meet the criteria for a diagnosis anymore, although they are not completely symptom free, and later they may relapse to a full-blown eating disorder again or may be classified as having EDNOS. 64 Overview and Foundational Issues c04 18 April 2012; 19:44:28 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. Corresponding to these clinical impressions, longitudinal studies demonstrate that the stability of eating disorder diagnoses over time is low ( Fichter & Quadflieg, 2007; Grilo et al., 2007; Milos, Spindler, Schnyder, & Fairburn, 2005). Based on systematic evaluation of the available evidence, however, treatment guidelines give specific recommendations for the different conditions (National Institute for Clinical Excellence [NICE], 2004). For patients with AN, psychological treatment on an outpatient basis is recommended. The treatment should be offered by “a service that is competent in giving that treatment and in assessing the physical risk of people with eating disorders” (p. 60). For patients with BN, the NICE guideline proposes as a possible first step to follow an evidence-based self-help program. As an alternative, a trial with an antidepressant drug is recommended, followed by cognitive behavior therapy for bulimia nervosa. In the absence of evidence to guide the treatment of EDNOS, the NICE guideline recommends pragmatically that “the clinician considers following the guidance on the treatment of the eating problem that most closely resembles the individual patient’s eating disorder” (p. 60). So even though specific diagnoses of eating disorders are not stable and a patient with AN might be diagnosed with BN a few months later, treatment recommendations vary considerably for the two conditions. It becomes obvious that different treatment recommendations for seemingly different conditions reflect rather accidental differences in the availability of empirical evidence than real differences in the response of certain conditions to specific treatments. Hence, the guidance offered by the guideline is basically a rather unstable crutch, and of course, cognitive behavior therapy or an evidence-based self-help program might be just as beneficial in AN or in EDNOS than it is in BN, even though nobody has yet compiled the statistical evidence to prove this. What does the validity of a diagnosis mean? The question concerns epistemological issues and requires a closer look to the nature of medical diagnoses with special regard to psychiatric diagnoses. R. Cooper (2004) questioned if mental disorders as defined in diagnostic manuals are natural kinds. In her thoughtful paper, the author concluded that diagnostic entities are in fact theoretical conceptions, describing complex cognitive, behavioral, and emotional processes (R. Cooper, 2004; Harari, 2001). Diagnostic categories are based upon observations, still they are strongly influenced by theoretical, social, and even economical factors. The ontological structure of psychiatric diagnoses is therefore not one of natural kinds. They are not something absolutely existing that can be observed independently. Rather they are comprehensive theoretical definitions serving as tools for communication and scientific observation. Kendell and Jablensky (2003) have also recently addressed the issue of diagnostic entities and concluded that the validity of psychiatric diagnoses is limited. They analysed whether diagnostic entities are sufficiently separable from each other and from normality by zones of rarity. They concluded that this was not the case; rather, they concluded that psychiatric diagnoses often overlap (R. Cooper, 2004; Welsby, 1999), shift over time within the same patient, and several similar diagnoses can be present in the same patient at the same time (comorbidity). Not surprisingly, diagnosis alone is a poor predictor of outcome (Williams & Garner, 2002). Acknowledging this haziness of diagnoses, one realizes these problems when trying to match individual cases to empirical evidence. When even the presence of a correctly assessed diagnosis does not assure comparability to other cases with the same diagnosis, empirical evidence about mental disorders is highly questionable (Harari, 2001). Of course, limited validity does not imply complete absence of validity, and empirical evidence on mental disorders is still useful to some extent; however, insight Limitations to Evidence-Based Practice 65 c04 18 April 2012; 19:44:29 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. into the limitations is important and that insight points out that psychiatric diagnoses represent phenomenological descriptions rather than natural kinds. Several authors have treated the same issue when writing about the complexity of cases, the problem of subsyndromal cases, and of single cases versus statistical evidence (Harari, 2001; Welsby, 1999; Williams & Garner, 2002). NONLINEAR DYNAMICS IN THE COURSE OF DISEASES It might be fruitful to look at evidence-based psychiatry from another perspective and to address the issues of complexity and nonlinear dynamics. With regard to their physical and mental functioning, humans can be conceptualized as systems of high complexity (Luhmann, 1995). This means that they cannot be determined precisely, but only in a probabilistic manner; however, probabilistic determination is sufficient for most purposes in observable reality. Human life consists fundamentally in dealing with probabilities. Social systems and human communication are naturally designed to manage complexity more or less successfully. Medicine itself is a social system (Luhmann, 1995) trying to handle the effects of complexity (Harari, 2001), for example, by providing probabilistic algorithms for treatments of diseases. In most situations, medicine can ignore the particular effects emerging from the complex nonlinear structure of its objects, although such effects are always present. Only sometimes do these effects become obvious and irritating, as for example in fluctuations of symptoms in chronic diseases, variations in response to treatment, unexpected courses in chronic diseases, and so on. Such phenomena can be seen as manifestations of the butterfly effect (see earlier). This insight questions deeply the core principle of EBP that assumes that it is rational to treat similar cases in the same manner because similarity in the initial conditions will predict similar outcomes under the identical treatment. The uncertainty of this assumption is particularly critical in psychiatry and psychotherapy. In these fields similar appearance is just a palliation for untraceable difference, and this exact difference may crucially influence the outcome. Addressing such problems is daily business for psychiatrists and psychotherapists, so their disciplines have developed special approaches. Diagnostic and therapeutic procedures in these disciplines are much less focused on critical momentary decisions, but more on gradual, iterative procedures. Psychiatric treatments and even more psychotherapy are self-referencing processes, where assessments and decisions are constantly reevaluated. Instead, EBP focuses primarily on decision making as the crucial moment of good medical practice. One gets the impression that EBM clinicians are constantly making critical decisions, and after having made the right decision, the case is solved. Maybe it is because of this misfit between the proposals of the method and real daily practice that many psychiatrists are not too attracted by EBP. EXAMPLE FROM PSYCHIATRY The diagnosis of posttraumatic stress disorder (PTSD) was first introduced in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) in 1980. Before that time, traumatized individuals were either diagnosed with different nonspecific diagnoses (e.g., anxiety disorders, depression, neurasthenia) or not declared ill at all. Astonishingly, the newly discovered entity appeared to be a clinically distinct disorder and the corresponding symptoms (re-experiencing, avoidance, hyperarousal) were quite characteristic and easily identifiable. Within a short time after its invention (Summerfield, 2001), PTSD became a very popular disorder; 66 Overview and Foundational Issues c04 18 April 2012; 19:44:29 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. clinicians and even patients loved the new diagnosis (Andreasen, 1995). The key point for the success of the new diagnosis was that it is explicitly based on the assumption of an external etiology; that is, the traumatic experience. This conception makes PTSD so appealing for the attribution of cause, responsibility, and guilt is neatly separated from the affected individual. PTSD allows for the exculpation of the victim, a feature that was particularly important when caring for Holocaust survivors and Vietnam War veterans. But what was almost proscribed for some time after the introduction of PTSD is now evidence-based: Preexisting individual factors play an important role in the shaping of posttraumatic response. Whether or not an individual develops PTSD after a traumatic experience is not only determined by the nature and the intensity of the traumatic impact, but also by various pretraumatic characteristics of the affected individual. Furthermore, PTSD is not the only posttraumatic mental disorder. A whole spectrum of mental disorders is closely linked to traumatic experiences, although they lack the monocausal appearance of PTSD. Anyway, the most frequent outcome after traumatic experiences is recovery. In the second rank of frequency comes major depression. Borderline personality disorder is fully recognized now as a disorder provoked by traumatic experiences in early childhood. Dissociative disorders, chronic somatoform pain, anxiety disorders, substance abuse, and eating disorders are equally related to traumatic experiences. Not surprisingly, PTSD is often occurring as a comorbid condition with one or more additional disorder or vice versa. In clinical practice, traumatized patients usually present more complex than expected. This may explain to some extent why PTSD was virtually overlooked by clinicians for many decades before its introduction, a fact that is sometimes hard to understand by younger therapists who are so familiar with the PTSD diagnosis. At any rate, the high-functioning, intelligent, monomorbid PTSD patient is indeed best evaluated in clinical trials, but rarely seen in everyday practice. PTSD was right in the focus of research since its introduction. Also from a scientific point of view, the disorder is appealing because it is provoked by an external event. PTSD allows ideally for the investigation of the human-environmentinteraction, whichis a crucial issue for psychiatry and psychology in general. The number of trials on diagnosis and treatment of PTSD is huge, and the disorder is now probably the best evaluated mental disorder. What is the benefit of the accumulated large body of evidence on PTSD for clinicians? There are several soundly elaborated guidelines on the treatment of PTSD (American Psychiatric Association, 2004; Australian Centre for PosttraumaticMental Health, 2007; NICE, 2005), meta-analyses, and Cochrane Reviews providing guidance for the assessment and treatment of the disorder. When we look at the existing conclusions and recommendations, we learn that: Debriefing is not recommended as routine practice for individuals who have experienced a traumatic event. When symptoms are mild and have been present for less than 4 weeks after the trauma, watchful waiting should be considered. Trauma-focused cognitive behavior therapy on an individual outpatient basis should be offered to people with severe posttraumatic symptoms. Eye movement desensitization and reprocessing is an alternative treatment option. Drug treatment should not be used as a routine first-line treatment in preference to a trauma-focused psychological therapy. Drug treatment (Specific Serotonin Reuptake Inhibitors) should be considered Limitations to Evidence-Based Practice 67 c04 18 April 2012; 19:44:29 Hersen, Michel, and Peter Sturmey. Handbook of Evidence-Based Practice in Clinical Psychology : Adult Disorders, John Wiley & Sons, Incorporated, 2011. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/ashford-ebooks/detail.action?docID=817356. Created from ashford-ebooks on 2017-11-07 11:26:43. Copyright © 2011. John Wiley & Sons, Incorporated. All rights reserved. for the treatment of PTSD in adults who express a preference not to engage in trauma-focused psychological treatment. In the context of comorbid PTSD and depression, PTSD should be treated first. In the context of comorbid PTSD and substance abuse, both conditions should be treated simultaneously. These recommendations are obviously clear, useful, and practical. They give real guidance to therapists and do not leave much room for doubts or insecurity. On the other hand, they are basically very simple, almost trivial. For trauma therapists, these recommendations are commonplace and serve mainly to endorse what they are practicing anyway. The main points of the guidelines for the treatment of PTSD could be taught in a 1-hour workshop. The key messages of the guidelines represent basic clinical knowledge on a specific disorder as it has been instructed in times before EBP. Through their standardizing impact on the therapeutic community, guidelines may in fact align and improve the general service quality offered to traumatized individuals, although this effect has not yet been demonstrated by empirical evidence. The treatment of an individual patient remains a unique endeavor where interpersonal relationship, flexibility, openness, and cleverness are crucial factors. This challenge is not lessened by evidence or guidelines.
