Article Text

Download PDFPDF

Guideline adherence
Guideline adherence rates and interprofessional variation
  1. R Peveler
  1. Correspondence to:
 Professor R Peveler, Professor of Psychiatry, University of Southampton, Royal South Hants Hospital, Southampton SO14 0YG, UK;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Variation in professional practice is a complex issue, but probably largely reflects differences in training, and there is evidence that guidelines and education alone have little impact on professional behaviour.

The increasing recognition of depressive illness as a major public health problem1 has intensified research efforts and also highlighted the extent of variation in professional practice. Because depressive illness is so prevalent, most cases are not managed by specialists. Although healthcare systems differ, general medical practitioners are the professional group most often involved. The training in mental health provided for such doctors is variable, but usually does not mirror the high prevalence of the common conditions—for example, in the UK less than half of general practitioners may receive specialist training experience in mental health. There is similar variation in the training of other professional groups. It is therefore no surprise that there is variation in professional practice both within and between such groups, as shown by Tiemeier et al2 in their vignette study published in this issue of QSHC.

This study creates an odd world of hypothetical patients. While the presentation of an apparently uniform stimulus to groups of health professionals has superficial scientific appeal, extrapolation of findings to clinical practice is extremely difficult. It is revealing that the “gold standard” for comparison purposes could not be derived from published evidence but had itself to be generated by panels of “experts”. Until the levels of agreement between such experts are known, surely it is premature to conclude that views of others are “inappropriate” simply because they disagree? Although carefully designed and conducted, the study is also compromised by the low response rates among some groups.

It seems questionable to “pool” the judgements of professional groups and then to compare each with the pooled scores. This implies that the context of the clinical encounter would have no impact on the treatment decision. In the UK most general practitioners have very limited access to psychotherapeutic treatments, so it would be expected that other professionals would recommend such approaches more often, simply because the patient would have been referred by a general practitioner (and usually would already have had a trial of pharmacological treatment). Professionals are more likely to choose treatments which they know are available to their patients, whatever guidelines may suggest.

The study highlights the important fact that overtreatment is as important a problem as undertreatment, an issue which has had too little attention in previous work. The observation that professionals may “undertreat” patients with dominating psychosocial problems is also valuable, and consistent with both clinical experience and newly emerging evidence.

One hope attached to “evidence-based” practice is that variation between professionals might be reduced. Clearly there is a tacit assumption that reduction in variation must necessarily represent an improvement in quality of care, and lead to better patient outcomes. In the early 1990s such hopes gave rise to the proliferation of guidelines, and publication of studies such as the Gotland study3 which suggested that educating general practitioners could lead to a measurable improvement in clinical outcomes. However, a large well designed randomised controlled trial of guideline based education in the UK4 was unable to demonstrate expected benefits in outcomes. Although it is possible that this study was not sufficiently powerful to detect benefit, any hypothetical benefit can at best be only modest in size. More likely, the failure to demonstrate benefit reflects either the ineffectiveness of education or the lack of validity of current guidelines.

We know that the “evidence base” itself suffers from a number of deficiencies. More evidence is available from secondary care than from primary care—even though many more patients are treated in the latter setting—because the infrastructure to support trials is less well developed there. As a result, there is far less evidence about the treatment of mild depression than of severe forms of the condition. This alone may explain why specialists' decisions are closer to guidelines than those of non-specialists. In addition, more evidence is available to guide drug treatment than psychotherapy because medication trials are more likely to be funded by the manufacturers of patentable products. There are also difficulties in the “conscientious, explicit and judicious” application of evidence to individual patients when systematic diagnostic and psychopathological evaluation does not form part of routine practice in primary care, simply because there is not time to do it.

The vignette study by Tiemeier et al leaves unanswered the question of where patients' preferences for treatment fit in. Difficulties arise from the fact that there is wide variation in patients' beliefs and expectations about the treatment of depression. Medical practice, like politics, is the art of the possible, and many will find the considerable time and effort needed to persuade a reluctant patient to accept antidepressant medication too demanding in the context of busy general practice.

Variation in professional practice is a complex issue, but probably largely reflects differences in training, and there is evidence that guidelines and education alone have little impact on professional behaviour.