Consider a patient who has suffered from two episodes of major depressive disorder both of which have caused substantial functional impairment. On each occasion his symptoms have responded to treatment with a selective serotonin reuptake inhibitor. Following remission of symptoms, the patient has been advised to continue treatment for 6 months before gradually discontinuing the drugs. His psychiatrist is now considering whether or not to advise long-term treatment with antidepressant medication to reduce the risk of relapse. The patient wants to know the risk of relapse without treatment and how much this would be reduced by continuing the drugs. The process of rapidly finding the best answer begins by formulating a clinical question:

1 in patients with major depressive disorder who have responded to drug treatment (the problem)

2 how effective are antidepressants (the intervention)

3 compared with alternative treatments (including none) (the comparison intervention)

4 in preventing relapse (the outcome)?

The next step is to classify the question. This example clearly concerns a question about therapy. Most of the questions that arise in clinical practice concern therapy, diagnosis, prognosis, or aetiology. Once the question has been formulated and classified, this suggests the most reliable research architecture (Table 1.10.1)

A commonly used hierarchy of evidence for studies of treatments is as follows:

Ia Evidence from a systematic review of randomized controlled trials,

Ib Evidence from at least one randomized controlled trial,

IIa Evidence from at least one controlled study without randomization,

IIb Evidence from at least one other type of quasi-experimental study,

III Evidence from non-experimental descriptive studies, such as comparative studies, correlation studies, and case-control studies,

IV Evidence from expert committee reports or opinions and/or clinical experience of respected authorities.

In this hierarchy, randomized evidence is considered, on average, to be more reliable thsan non-randomized evidence, and a systematic review of randomized evidence is considered to be the best defence against systematic bias.

Hierarchies of evidence have also been formulated for non-therapeutic studies, such as studies of aetiology, diagnosis, and prognosis. Again, the fundamental feature of these hierarchies is that the study architectures with the least susceptibility to bias are considered most reliable. The study design considered most reliable for each type of clinical question is shown in Table 1.10.1.

The need for systematic reviews and the methodology used are described in Chapter 6.1.1.2. The recognition of the need for systematic reviews of randomized controlled trials, and the development of the scientific methodology of reviews, has been one of the most striking advances in health services research over the last decade. One key development was the founding of the Cochrane Collaboration, an international organization with the objective of
producing regularly updated systematic reviews of the effectiveness of all health care interventions.⁽¹²⁾

In some areas of health care there is sufficient evidence, coexisting with substantial clinical uncertainty, that it is worth developing clinical practice guidelines. Clinical practice guidelines have been defined as ‘systematically developed statements to assist practitioner decisions about appropriate health care for specific clinical circumstances’.⁽¹³⁾ Guidelines have been developed for several years, but there have been recent advances in the methodology of producing explicitly evidence-based guidelines. Evidence-based clinical practice guidelines are developed by a guideline development group consisting of key stakeholders who decide on the precise clinical questions to be answered. The evidence is then systematically reviewed and classified according to a hierarchy of evidence (see above) and presented to the guideline development group. The group then makes recommendations as appropriate. The degree to which the recommendations are directly based on the evidence is described using a second level of classification⁽¹⁴⁾:

Clinical practice guidelines are usually developed at a national level and need tailoring to suit local circumstances. Professional and scientific bodies such as the American Psychiatric Association and the British Association for Psychopharmacology often take the lead in developing national guidelines. Increasingly, health care providing organizations are developing clinical practice guidelines as a way of assuring quality, increasing standardization of care and controlling costs. For example, in the United Kingdom, the National Institute for Health and Clinical Excellence (NICE) is now the main body producing guidelines across all disease areas. These guidelines are extremely rigorous in terms of methodology and also routinely include economic analyses of the cost-effectiveness of health care technologies. At their best, these guidelines produce the most accurate syntheses of current knowledge available. NICE also produces Health Technology Appraisals (HTAs) of single interventions to determine the cost-effectiveness of new technologies (mainly medicines) prior to their introduction into the taxpayer-funded National Health Service. Cost-effectiveness analysis requires the translation of disease-specific estimates of clinical effectiveness into the common metric of Quality Adjusted Life Years (QALYs) using sophisticated modelling techniques. Decisions about whether to allow reimbursement of the treatment depends on the cost per QALY (incremental cost-effectiveness ratio, ICER): a treatment with an ICER of more than £30 000 is unlikely to be approved. NICEs HTA decisions are particularly likely to be controversial when there is some evidence that the treatment works, but that the ICER is found to be too high—for example, in the case of acetylcholinesterase inhibitors in Alzheimer’s disease.

There are several limitations to clinical practice guidelines. Firstly, evidence-based clinical practice guidelines are expensive and time consuming to produce and rapidly become out of date. Secondly, to influence practice, evidence-based clinical practice guidelines need to be actively disseminated and implemented. Guidelines that are developed nationally and passively sent out to doctors are often not used.⁽¹⁵⁾ A number of active approaches are effective in helping change clinicians’ behaviour:⁽¹⁶⁾

There is some evidence that guidelines can improve patient outcomes by their effect on clinical practice, especially when they are made relevant to local circumstances.⁽¹⁵⁾ In one study in the United States, 217 patients with depressive disorders were randomly assigned to usual care or a multifaceted intervention designed to achieve the Agency for Health Care Policy and Research guidelines on management of depression.⁽¹⁷⁾ Patients in the intervention group were much more likely to be treated in accordance with the guidelines, and this led to improved outcomes in patients with major depressive disorder (more than 50 per cent reduction on the Symptom Checklist-90 Depressive Symptom Scale at 4 months in 74 per cent of experimental patients compared with 44 per cent of control patients).

Understandably, clinicians also seek guidance in important clinical questions that are poorly served by high-quality, especially randomized, evidence. To assist in these clinical decisions, Frances and his colleagues have developed an innovative method of guideline development based on a systematic survey of the views of clinical experts.^(18,19)

The development of the Internet—or World Wide Web—during the 1990s facilitated the development of evidence-based practice. The Internet has now become a vast information resource for doctors and patients. Improved access to information afforded to patients means that they are often very well informed about their condition. This is one of the factors contributing to the need for doctors to improve their own access to information. The Internet has several drawbacks including the disorganization of the information and the lack of quality control.^(20,21) Web portals have been developed that provide organized and indexed access to critically appraised websites (e.g. www.nelh.nhs.uk). The Web has now become the main medium for transmitting and storing knowledge.