Before deciding whether you need a guideline, it is vital to clarify your objectives, as Dr Mark Charny explains in the first of a new series


First, ask yourself what you want to achieve with the guideline. A guideline is a tool – a means to an end. If you establish the objectives, it will be much easier to decide whether a specific guideline is likely to help.

For example, you may want to

  • Improve outcomes
  • Control costs
  • Identify unmet need
  • Achieve greater efficiency
  • Ensure that your patients are more informed
  • Avoid complaints.

Some guidelines are very prescriptive: a drug formulary in a hospital may effectively prevent the doctor from prescribing anything other than the one agreed drug for a particular condition. Some are advisory, suggesting perhaps that in general it is good practice to observe a patient in certain circumstances rather than operate. Others are exhortatory: urging doctors, for example, to carry out audit in multidisciplinary teams.

No guideline should be less precise or complete than the evidence permits, and no more precise than the problem requires.

Guidelines may be very detailed, complete and specific: for example, they may suggest precisely when to test the urine of a patient with symptoms suggestive of cystitis, at what stage and on what basis to prescribe treatment (if at all), and what drug(s) to prescribe.

At the other extreme, a guideline may be general, incomplete and non-specific: for example, suggesting that when a middle-aged patient visits the surgery for some reason, the GP should take the opportunity to test blood pressure and enquire about his/her lifestyle and give appropriate advice. The level of detail will depend on the evidence available and the type of activity being considered.

It is inherently more difficult to be precise about counselling than about drug treatment and, in practical terms, more difficult to get evidence about treatments for rare conditions than common ones.

Guidelines may apply to all activities carried out by health professionals. Apart from clinical areas, guidelines have been produced on screening, counselling, diagnosis, test requests, medical treatment, surgical interventions, follow-up, record keeping, communication, audit, CME and CPD, patient information, health promotion, machine calibration and maintenance and teamwork.

A suitable guideline should be constructed to address the problem in the same terms in which it is understood by those who will use it.

A guideline concentrating on the treatment of angina or hypertension is not useful on its own if the goal is to reduce the burden of cardiovascular disease in the practice.

On the other hand, a guideline about promoting a healthier lifestyle is not the right tool to address concerns about blood pressure screening and the treatment of asymptomatic hypertensives.

Even guidelines addressing what appears to be the same clinical issue may differ greatly in the context within which they apply.

For example, a guideline on back pain intended to advise orthopaedic surgeons in an outpatient department is unlikely to help GPs manage back pain in the surgery.

In such cases, the problem is that superficially similar patient populations are not the same. Being clear about the clinical context will help you decide whether a particular guideline is appropriate to your patients.

Guidelines come from many sources: they may be based on systematic reviews of published randomised controlled trials, or they may simply express the opinion of an individual. In between, guidelines may be based on whatever evidence there is, with the gaps filled in by the consensus views of a panel of experts.

The important point is that the guideline should be fit for the purpose. Where there is good evidence, nothing less than evidence-based recommendations will do. Where the problem is of great significance to patients, such as what treatment offers the best prospect of survival to a patient with a certain stage and type of cancer, we will be wary of relying on the opinions of an individual, preferring an expert consensus view where evidence is not available.

In other circumstances, such as giving patients information, we might be happy to adopt suggestions from an individual in the absence of anything more structured.


It is essential to be clear about your interest in a guideline.

  • What is the issue?
  • Who will use the guideline?
  • What is the objective?

This will help you decide what you need, and whether any particular document meets that need.

In that way, you will avoid trouble with nomenclature: whether you or I call something a protocol, an algorithm, a standard, a recommendation, guidance, advice, or a good practice guide is irrelevant. What matters is whether it does the job you want it to do.

Guidelines in Practice, January/February 1999, Volume 2
© 1999 MGP Ltd
further information | subscribe