Evidence-based medicine and dentistry is not new but is not always well understood. It is a way of thinking that should permeate every aspect of clinical practice. This chapter describes this philosophy, provides an overview of its components and provides an approach on how to make best use of the scientific literature and the benefits of evidence-based medicine.
Evidence-based medicine is the explicit and judicious use of current best evidence to guide health care decisions. It integrates this best research evidence with clinical expertise and patient values. The aim of evidence-based medicine is to optimise clinical outcomes and quality of life for patients.
This approach may be used for individual patients, or for planning and purchasing care for groups of patients. Patients will benefit if their clinician is abreast of the latest data but he or she also needs to be able to take a good history, carry out a thorough examination and have an understanding of the patients’ values and preferences.
When working with patients, there is a constant need to seek information before making a clinical decision and professionals need to develop the habit of learning by inquiry, so when confronted with a clinical question they can look for the current best answer as efficiently as possible. It can be difficult to find the current answer in a large database such as MEDLINE with over ten million references. A specialised database such as the Cochrane Library or Best Evidence can be a better place to start. Best-evidence resources are growing in number and are accessible as never before.
Best research evidence is clinically relevant research from basic science and clinical research. It either validates previously accepted diagnostic tests, preventive regimens and treatments, or replaces them with new ones that are more powerful, more accurate, more effective and safer. The strength of evidence from various study designs is shown in Fig. 1.1.
Do not look at promotional brochures, which often contain unpublished material. Ignore anecdotal ‘evidence’, such as the fact that a dental celebrity is using a particular product. Do not accept the newness of a product as an argument for changing to it as the opposite might have a good scientific argument.
Clinical expertise is the ability to use clinical skills and past experience to rapidly identify each patient’s unique oral health state and diagnosis, their individual risks and benefits of potential interventions and their personal values and expectations.
Patient values are the unique preferences, concerns and expectations each patient brings to a clinical encounter and which must be integrated into clinical decisions if they are to serve the patient. It is usual practice for the clinician to describe the diagnosed condition or disease to the patient and then describe the treatment available together with the harms that the treatment may potentially cause. To determine the patient values, the clinician could go on to ask the patient to make a value judgement about these two, that is, which is worse and by how much. The patient may need to think about this or discuss it with family members. The clinician may also describe the outcomes of forgoing or accepting treatment. For example, when the consultation concerns the removal of a lower wisdom tooth, the clinician may ask the patient to compare the distress caused by the pericoronitis with the anticipated distress of temporary pain and swelling and possible altered sensation. The patient should also take into account the likelihood of future episodes of pericoronitis if they forgo surgery.
The aim of evidence-based medicine is to improve clinical outcomes for patients and there is plenty of evidence that this is the case. One example is that myocardial infarction survivors, who are prescribed aspirin or beta-blockers, have lower mortality rates than those who aren’t prescribed these drugs. Another example would be the benefit of using streptomycin for pulmonary tuberculosis as demonstrated by the historic Medical Research Council trials. These are generally regarded as the first of the modern randomised controlled trials.
The randomised controlled trial provides the underlying basis for evidence-based medicine and the number of trials is growing exponentially with more than 150 000 listed by the Cochrane Library. However, there are limitations to evidence-based medicine. There is a shortage of consistent scientific evidence, difficulties in application of research evidence to individual patients and barriers to the practice of high-quality care. Some clinicians misunderstand the philosophy of evidence-based medicine and incorrectly believe that it means a loss of clinical freedom, or that it ignores the importance of clinical experience and of individual values, which is not the case.
Randomised controlled trials may be used to compare health screening, diagnostic and preventative strategies, in addition to different treatments. They are recognised as one of the simplest, yet most powerful and revolutionary clinical research tools that we have. People are allocated at random to receive one of several clinical interventions, and comparisons are made (Fig. 1.2).
• The patients who take part in the trial are referred to as ‘participants’ or the study population. Participants don’t have to be ill as the study can be conducted in healthy volunteers or members of the general public.
• One of the interventions is usually regarded as the standard of comparison or ‘control’, hence the name randomised controlled trial. The group of participants who receive the control are known as the ‘control group’. The control may be conventional treatment, placebo or no treatment.
• Outcomes are measures, so randomised controlled trials are regarded as quantitative studies. They compare two or more interventions and so are regarded as comparative studies. Case-series studies may also be quantitative but do not include comparisons among groups.
Random allocation means that all participants have the same chance of being assigned to each of the study groups. This ensures that the groups are balanced for the disease severity or other predictors of prognosis and are not biased. The randomisation should be concealed from the clinicians who entered patients into the trial so they don’t know which treatment the patient will receive, otherwise they may consciously or unconsciously distort the balance of the groups being compared.
The best method for allocation to study group is to use random-number tables or computer-generated sequences. Some investigators report using ‘odd or even’ birth year or hospital number but there may be problems with these ‘quasi-randomisation’ methods. The investigator may subvert the allocation because he or she knows which group the patient will be in and the study results could be biased as the groups are not properly balanced. For example, if comparing different surgical techniques for the removal of wisdom teeth, it would be important to have an equal mix of simple and difficult cases in the different groups and not all the simple cases in one group and all the difficult cases in another. If the groups are kept as similar as possible at the start of the study then it will be easier to isolate and quantify the impact of the intervention.
Ideally all patients and clinicians involved in the trial should be blind to the intervention so that all groups are treated equally, apart from the experimental treatments that are being compared. If this isn’t the case then the study may be biased by patients who report symptoms, and clinicians who interpret them, influenced by their hunches and opinions about the anticipated treatment effectiveness. It is, however, not always possible to blind all trials. In surgical trials, for example, the surgeon will be aware of which technique of the alternatives he or she is using, but it may be feasible to have clinicians other than the operating surgeon who are blind to the study group, carrying out the postoperative assessments. This would be described as a single-blind trial.
A trial is described as ‘double-blind’ when both the participants and the investigator are blind to the intervention. Some trials require a double-dummy. This may be the case, for example, in an oral medicine trial when two or more mouthrinse interventions need to look and taste the same. The double-blind, double-dummy, randomised controlled trial can also be useful when, for example, a drug in tablet form is to be compared with a drug in injection form. Participants in one of the study groups would receive a tablet containing the active drug together with an injection of placebo, and the other study group would receive a placebo tablet with an injection of the active group.
All patients entered to the study should be accounted for at its conclusion. Ideally no patients should be lost to follow-up because these patients could have had outcomes that would affect the conclusions of the study. They may have dropped out because of an adverse outcome. One way of dealing with the data where there are patients who have been lost is to assign the worst-case outcome to all of those lost to follow-up. However, some consider that a loss of more than 20% is unacceptable.