Focus on Alternative and Complementary Therapies
www.pharmpress.com/fact
Focus Alternat Complement Ther©2005 Pharmaceutical Press
Focus Altern Complement Ther 2007; 12: 155–7
Making decisions in clinical practice is not simple, especially if one aspires to use scientific evidence as a guide. When a patient presents with a clinical condition or illness, there are several types of information a practitioner may need in order to make the best decision for and with the patient. These types of information require various types of high-quality evidence. Good evidence-based medicine (EBM) requires at least five distinct types of information:
1. Processes: First, ideally one needs to know the cause of the disease and how it is progressing (its pathophysi-ology and disease processes). Sometimes knowing the cause does not help in finding a cure, however, and so additional information on what mechanisms the person has to counter or recover from the disease are needed (the salutophysiology and recovery processes). How potential diagnostic and therapeutic interventions influence these processes can be a critical piece of information when considering differing alternatives. This type of information usually comes from basic science research done in the laboratory, or detailed clinical information on one or both of the above ‘-ologies’ collected in either a research practice or clinical research centre.
2. Preferences: Before selecting interventions from these possibilities, however, especially ones that may be delivered over long periods or have long-term consequences, one needs to know how these interventions may interact, both with each other and with the person’s needs and desires. Therefore, the desires and perceptions of the patient and the patient’s family are another type of information needed. Patients often want a cure but sometimes other outcomes are more important. In chronic disease, they may be seeking hope (what is their prognosis?), or forgiveness (did they cause this?), or they may be more concerned with suffering (will it be painful?), or the impact on their family (will I be a burden?) or finances (how will I pay for it?). Unvoiced issues may determine the preferred outcomes, such as the trade-off between the side-effects of treatment and the consequences of an untreated condition. Sometimes these various outcomes may collide, making decisions challenging even in the face of good evidence. This type of information usually comes from qualitative research (a form of which is done every day in the clinical interview), with open-ended queries, sensitivity for and skill in uncovering the ‘hidden agendas’ and triangulation of information from several avenues.
3. Probabilities: Before pulling an intervention out of their medical bag, physicians would like to know the probability of success for that patient. Once a treatment is implemented, is there a 20% or 80% chance of improvement in the main outcome of concern? What about the other outcomes and consequences of the intervention? Is probability data collected from a referral centre or in a clinical trial likely to occur in the local setting or with the patient at hand who is not ran-domised, blinded or informed in the same way? Is the healthcare team able to deliver, monitor and manage the treatment in an effective way? Is the proper training and system in place for risk management of the intervention to prevent errors and assure quality in its delivery? This type of information usually comes from observational data, preferably drawn at a local level, and timely quality assurance information with proper feedback to the healthcare team.
4. Specificities: After all of this, the physician, healthcare team and care delivery system, including those who set standards and pay for services, want to know if those services truly deliver anything unique or specific to the mix. That is, do they impact the patient’s outcome any differently than alternative services that the health system could deliver? In other words, of all the possible services the healthcare system could provide a patient, which should be implemented? Because personal desires and expectancies, including patient, practitioner and cultural beliefs are mixed into every diagnosis and treatment system, should we not see if the part of the intervention we think is producing the effect actually is contributing to the outcome? This, of course, requires comparative information in which the part we think is producing the effect is removed from one part of an experiment and left in another. If the part we think is producing the effect is a chemical, it becomes fairly easy to control for belief and expectancy (such as in placebo-controlled drug and herbal studies). If the intervention is not a chemical (such as surgery, acupuncture, psychotherapy, spiritual healing, rehabilitation, massage and manipulation, etc.), then properly controlling for belief and expectancy and knowing which part of the mix to ‘pull out’ and compare becomes a challenge.
5. Practicalities: Can this treatment be used in this practice on this patient? This is a decision level that deals with effectiveness not efficacy. Given the skills of the doctor, given the training and education of the doctor, given the type of practice and its resources, given the nature of the patient such as other morbidities, willingness and ability to adhere to the treatment, can this treatment be feasible in this setting? Most experimental evidence (trials) and process evidence (biological mechanisms) cannot answer these types of questions. A patient may have a very strong preference for antiretroviral treatment, they may have an intact immune system, but may not have the ability to carry out the very complicated and demanding regimen required to monitor and take the drugs. The physician must take into account not only the variability in the patient but also in the provider. A therapy that has been trialled in the National Cancer Research Centers with extensive equipment, infrastructure, resources and highly trained staff including physicians, nurses, technicians, pharmacists and other adjunctive professions is a long way removed from an average general practice. This type of evidence also comes from observational studies but of the kind anthropologists and sociologists provide in ethnographic observations. These attempt to look at the impact of the social and cultural context and the impact of the health encounter on the outcomes. In a sense they put the practice back into EBM.
Evidence-based medicine. When the idea of EBM first took hold as an important approach to the use of science in medicine, it was ‘comparative’ information that dominated the approach and was more preferred.1 It formed the basis for formulating EBM guidelines, is at the heart of the ‘best evidence’ summaries of the British and American Medical Associations, the guidelines approaches from various medical societies and government agencies and the Cochrane Collaboration methodologies for systematic reviews. Because the RCT is clearly the best way of obtaining comparative information, most guidelines call for it and EBM teaching procedures still start with it and often focus exclusively on it. Given the complexities of clinical decision-making and the various types of evidence just described, however, it is now becoming clear that EBM approaches that focus exclusively on this type of evidence are inadequate.2,3 They are too simplistic and there is currently not enough evidence to place them at the top of any evidence hierarchy. It is not too surprising then that most physicians don’t rely on these approaches for making clinical decisions.4
Evolving EBM approaches. Recognising that new, more sophisticated approaches to EBM that incorporate the full spectrum of information needed for making clinical decisions, a new type of EBM is emerging. Examples are the RAND ‘appropriateness’ approaches,5 the US Agency for Healthcare Research and Quality’s (AHRQ) efforts on consumer or patient-centred evidence evaluations,6,7 and ‘care pathway’ applications of EBM,8 among others. In a recent meeting on EBM convened by the US Institute of Medicine, a number of new and innovative approaches to EBM were discussed. Most of these approaches call for systematic incorporation of concepts such as goals, problem formulation, values, sufficiency and other factors into the development of consistent and customised EBM decisions that account for the complexity of information needed in the clinical setting. For a discussion of these topics see the special issue on EBM in Health Affairs, published in 2005.9 Taken together, it might be worth while to distinguish the more academic debate around EBM from the practical application by talking about evidence-based practice.
Keeping up with EBM in CAM research. CAM needs to keep pace with developments in EBM. Recycled debates about the value and exclusivity of RCT data for making black and white decisions about the use of CAM in health care are of little value. This has often led to critiques of the relevance of EBM to CAM. But these frequently lack any acknowledgement that within EBM many of these challenges are being dealt with. A clear example is in observational data. The question is no longer should it be included, but can we distinguish good from bad observational data, can we have confidence that the probabilities that come from such data are trustworthy? This is a huge advance from the previous position of if you can only speak observationally you should not speak at all.
Improved approaches to EBM, such as those being developed in mainstream medicine, need to be taken up and applied to CAM. Frameworks for such an advance have been produced such as the ‘Evidence House’10,11 and the ‘Circular’ method12 for evaluation and application of science in medicine. It is time for the field to go beyond simplistic approaches to EBM and develop improved standards that use all essential types of scientific information in a rational and systematic way and have sufficient quality criteria and granularity for use in clinical settings. Finally, such models need to be developed and evaluated in care settings for feasibility, and impact on satisfaction, quality, costs and patient value.13
CAM has an opportunity here. EBM is far less advanced in CAM and has not yet come to dominate CAM ideologically. CAM can learn from the errors of biomedicine. Furthermore, because CAM places more emphasis on the health encounter and the mind/body/spirit trilogy of health, there has traditionally been a much greater respect for non-hierarchical forms of evidence in CAM. The RCT has never dominated CAM in the way it has in biomedicine. This has partly to do with philosophical differences but it also has partly to do with coming out the gate a little slower, being some distance behind at the first bend, not being the first horse to take the fence and being the last one to encounter the water jump. There are advantages in both horse racing and academia in not being in first place – at least until the finish line.
Recommendation. Toward this goal, we recommend that the international research community, perhaps through the auspices of the International Society for Complementary Medicine Research, move towards establishing improved standards of EBM for the application of CAM in health care.
