Process and methods

9 Developing and wording guideline recommendations

Many users of clinical guidelines do not have time to read the full document, and may want to focus only on the recommendations. It is therefore vital that recommendations are clear, can be understood by people who have not read the full guideline, and are based on the best available evidence of clinical and cost effectiveness. This chapter addresses key areas in developing guideline recommendations:

  • interpreting the evidence to make recommendations

  • wording the recommendations

  • prioritising recommendations for implementation

  • formulating research recommendations.

These processes are at the heart of the work of the Guideline Development Group (GDG). However, they are not straightforward and it may not be easy for the GDG to reach agreement. Consensus techniques may need to be used within the GDG (see section 3.5).

9.1 Interpreting the evidence to make recommendations

The GDG must decide what the evidence means in the context of the review questions and economic questions posed, and decide what recommendations can usefully be made to healthcare and other professionals.

In the full guideline, the aim should be to show clearly how the GDG moved from the evidence to the recommendation. This is done in a section called 'evidence to recommendations' so that it can be easily identified. This section may also be a useful way to integrate the findings from several evidence reviews that are related to the same recommendation(s).

Underpinning this section is the concept of the 'strength' of a recommendation (Schünemann et al. 2003). This takes into account the quality of the evidence but is conceptually different. Some recommendations are 'strong' in that the GDG believes that the vast majority of healthcare and other professionals and patients would choose a particular intervention if they considered the evidence in the same way that the GDG has. This is generally the case if the benefits clearly outweigh the harms for most people and the intervention is likely to be cost effective. However, there is often a closer balance between benefits and harms, and some patients would not choose an intervention whereas others would. This may happen, for example, if some patients are particularly averse to some side effect and others are not. In these circumstances the recommendation is generally weaker, although it may be possible to make stronger recommendations about specific groups of patients.

For all recommendations, a general principle of NICE clinical guidelines is that patients should be informed of their choices and be involved in decisions about their care. Patients may choose not to accept the advice to have the most cost-effective intervention, or they may opt for a treatment that has the same or lower long-term health and personal social service costs if, for example, they feel that its side effects are more tolerable. There might be little evidence of differences in cost effectiveness between drugs within a class, and the clinician and patient might choose between these drugs on the basis of side-effect profile. However, it is not usually possible to offer patients interventions that are above NICE's threshold for cost effectiveness (see section 7.3) because the opportunity cost of that course of action has been judged to be too great (see section 7.1.1).

The GRADE system (see section 6.2.1.1) allocates labels or symbols to represent the strength of a recommendation. NICE has chosen not to do this, but instead to reflect the concept of strength in the wording of the recommendation (see section 9.3.3). The GDG's view of the strength of a recommendation should be clear from its discussions, as reported in the full guideline.

The following points will need to be covered in the discussions and can also be used as a framework for reporting those discussions.

9.1.1 Relative value placed on the outcomes considered

Often more outcome data are available than are actually used in decision-making. It is therefore important to have explicit discussion of which outcomes are considered important for decision-making (including consideration of the perspective of the decision-makers) when developing review protocols (see section 4.4), and of what relative importance was given to them. This might be done informally (for example, 'death was considered the most important outcome') or formally (for example, by the use of utility weights).

This discussion should be clearly separated from discussion of how this will play out when the evidence is reviewed, because there is a potential to introduce bias if outcomes are selected on the basis of the results. An example of this would be choosing only outcomes for which there were statistically significant results.

It may be important to note outcomes that were not considered to be important for decision-making, and why (such as surrogate outcomes if longer-term, more relevant outcomes are available). If the same set of outcomes is used for a number of review questions, it might be more efficient to record this information once and then refer back to it.

9.1.2 Trade-off between clinical benefits and harms

A key stage in moving from evidence to recommendations is weighing up the magnitude and importance of the benefits and harms of an intervention. This may be done qualitatively (for example, 'the evidence of a reduction in mortality outweighed a small increase in side effects'), or quantitatively using a decision model.

9.1.3 Trade-off between net health benefits and resource use

If there are net health benefits from an intervention, there should be an explanation of how the implications of resource use were considered in determining cost effectiveness. Again, this may be informal, or may be more formal and include the use of economic modelling. If there is no clear evidence of net health benefit, cost and resource use could be discussed here.

9.1.4 Quality of the evidence

There should be discussion of how the presence, likely magnitude and direction of potential biases and uncertainty in the clinical and economic evidence have influenced the recommendation, and why. This should reflect the judgement on the quality of the evidence as described in the GRADE profile and the NICE economic profile. Lower-quality evidence makes it more difficult to justify a strong recommendation in general, although there may be exceptions to this. For example, evidence on the frequency of adverse effects is often of low quality, but a strong recommendation might be made not to use a particular drug thought to have teratogenic effects in women of child-bearing potential.

The discussion of uncertainty may include consideration of whether the uncertainty is sufficient to justify delaying making a recommendation to await further research, taking into account the potential harm of failing to make a clear recommendation.

9.1.5 Other considerations

If the 'evidence to recommendations' section combines consideration of several possible interventions, it may include discussion of the position of an intervention within a pathway of care.

This is also the appropriate place to note how the GDG's responsibilities under equalities legislation and NICE's equality scheme have been discharged in reaching the recommendation(s). This covers inequalities related to age, disability, gender reassignment, marriage and civil partnership, race, religion or belief, sex and sexual orientation and socioeconomic status. The GDG will need to consider whether:

  • the evidence review has addressed areas identified in the scope as needing specific attention with regard to equalities issues

  • criteria for access to an intervention might be discriminatory, for example through membership of a particular group, or by using a test that might discriminate unlawfully

  • people with disabilities might find it impossible or unreasonably difficult to receive an intervention

  • guidance can be formulated so as to promote equalities, for example by making access more likely for certain groups, or by tailoring the intervention to specific groups.

Before the guideline is signed off, an equality impact assessment (EIA) form is completed by the National Collaborating Centre (NCC) or the NICE Internal Clinical Guidelines Programme[14] and the GDG to demonstrate how equality issues have been identified and considered during development. The EIA form is signed by the NCC Director and GDG Chair, and countersigned by the Centre for Clinical Practice (CCP) lead for the guideline, before being posted on the NICE website. Further guidance on how to complete the EIA form is outlined in the document Positively equal: a guide to addressing equality issues in developing NICE clinical guidelines.

It may be useful to briefly discuss the extent of change in practice that will be needed to implement a recommendation, and the possible need for carefully controlled implementation with, for example, training programmes or demonstration projects.

9.1.6 Challenges in formulating recommendations

There are many reasons why it can be difficult for a GDG to reach a decision about a recommendation. The evidence base is always imperfect, and so there is always a degree of judgement by the GDG. There may be very little, or no, good-quality evidence that directly addresses the review question the GDG has posed. In this situation, there are several options to consider:

  • The GDG may wish to look at evidence that is likely to be more at risk of bias than the evidence they had hoped to find. For example, if the GDG had set out to collect only randomised trials for a question of effectiveness, but found none, they might consider looking for good-quality non-randomised studies. However, there is a risk that considerable time and effort is spent finding and reviewing studies that are likely to be biased and so are hard to interpret. This approach should be pursued only if there is reason to believe that it will help the GDG to formulate a recommendation.

  • The GDG may wish to extrapolate from high-quality evidence in a related area, for example in a largely similar patient group or for a closely related intervention. The GDG will need to make its approach explicit, stating the basis it has used for extrapolating from the data and the assumptions that have been made. This will need to include consideration of the plausibility of the assumptions. This approach is unlikely to be helpful if the evidence is derived from a question that is too different from the review question, or if the evidence is not of the highest quality.

  • The GDG may consider basing a recommendation on its view of current most cost-effective practice. Formal consensus techniques may be used to elicit opinions from the GDG, although NICE does not recommend a particular approach. Importantly, it is not usually appropriate to involve stakeholders from outside the GDG in this process, as they will be offering opinions on recommendations without having seen the evidence considered by the GDG; in addition, stakeholders will not have agreed to adhere to the principles underlying NICE's decisions on recommendations. This approach would also allow some stakeholders input to the decision-making process that other stakeholders will not have. GDGs should therefore be particularly cautious about using and interpreting the results of such exercises involving stakeholders outside the GDG, and should discuss any proposed use with NICE. The final decision on whether such work with external stakeholders is warranted will be made by NICE.

When formulating recommendations, there are likely to be instances when members of the GDG disagree about the content of the final guideline. Formal consensus methods can be used for agreeing the final recommendations (see section 3.5). Whatever the approach used, there should be a clear record of the proceedings and how areas of disagreement have been handled. This may be summarised in the full guideline.

9.2 'Only in research' recommendations

If evidence of effectiveness is either lacking or too weak for reasonable conclusions to be reached, the GDG may recommend that particular interventions are used within the NHS only in the context of research. Factors that will be considered before issuing such recommendations include the following:

  • The intervention should have a reasonable prospect of providing benefits to patients in a cost-effective way.

  • The necessary research can realistically be set up or is already planned, or patients are already being recruited.

  • There is a real prospect that the research will inform future NICE guidance.

9.3 Wording the guideline recommendations

Writing the recommendations is one of the most important steps in developing a clinical guideline. Many people read only the recommendations, so the wording must be concise, unambiguous and easy to translate into clinical practice. Each recommendation, or bullet point within a recommendation, should contain only one main action.

The wording of recommendations should be agreed by the GDG, and should:

  • focus on the action that needs to be taken

  • include what readers need to know

  • reflect the strength of the recommendation

  • emphasise the involvement of the patient (and/or their carers if needed) in decisions on treatment and care

  • use plain English where possible and avoid vague language

  • follow NICE's standard advice on recommendations about drugs, waiting times and ineffective interventions.

The rest of this section explains these points in more detail. The lead editor for the guideline from NICE can advise on the wording of recommendations.

9.3.1 Focus on the action

Recommendations should begin with what needs to be done. When writing recommendations, keep in mind a reader who is saying, 'What does this mean for me?'. Recommendations should be as specific as possible about the exact intervention being recommended and the group of people for whom it is recommended (see also section 9.3.2).

Use direct instructions because they are clearer and easier to follow. Most recommendations should be worded in this way. Assume you are talking to the healthcare professional who is working with the patient at the time.

Examples

  • Record the person's blood pressure every 6 months.

  • Ask people in high-risk groups whether they have symptoms.

  • Carry out and record a focused baseline assessment for people with faecal incontinence to identify the contributory factors.

Exceptions

  • Recommendations about service organisation, or if the audience is not the healthcare professional. For example:'Care should be provided by a multidisciplinary team.'

  • Recommendations that a specific type of healthcare professional should carry out an intervention. For example: 'An occupational therapist should assess the patient's needs.'

  • Recommendations that use 'must' or 'must not' (see section 9.3.3.1).

Start with a verb describing what the reader should do, such as 'offer', 'measure', 'advise', 'discuss', 'ask about' (see sections 9.3.3 and 9.3.4 for advice on the choice of verb).

Examples

  • Advise pregnant women to limit their intake of oily fish to two portions a week.

  • Perform surgery within 48 hours of symptom onset.

  • Offer relaxation techniques for managing pain, sleep problems and comorbid stress or anxiety.

Exceptions

  • Sometimes it is clearer to start with details of the patient group or other details, particularly if recommending different actions for slightly different circumstances or to make the sentence structure simpler. For example: 'If surgery is an option, refer the patient to a specialist surgeon to discuss the risks and benefits.'

9.3.2 Include what readers need to know

Recommendations should contain enough information to be understood without reference to the evidence or other supporting material. But do not add unnecessary details, because recommendations are more likely to be followed if they are clear and concise.

  • Define any specialised terminology that is used in the recommendations. Avoid using abbreviations unless your audience is likely to be more familiar with the abbreviation than with the term in full. If abbreviations are essential, define them at first mention and in a glossary.

  • Define the target population if it is not obvious from the context. Often it is necessary to define the population only in the first of a group of recommendations, if it is clear that the subsequent recommendations in that section relate to the same population.

  • Include cross-references to other recommendations in the guideline if necessary to avoid the need to repeat information such as treatment regimens.

  • Do not include reasons justifying the recommendation unless this will increase the likelihood that it will be followed – for example, if it involves a change in usual practice or needs particular emphasis.

  • Include only one main action in each recommendation or bullet point.

9.3.3 Reflect the strength of the recommendation

The description of the process of moving from evidence to recommendations in section 9.1 indicates that some recommendations can be made with more certainty than others. This concept of the 'strength' of a recommendation should be reflected in the consistent wording of recommendations within and across clinical guidelines. There are three levels of certainty:

  • recommendations for interventions that must (or must not) be used

  • recommendations for interventions that should (or should not) be used

  • recommendations for interventions that could be used.

The NICE guideline includes a standard section about how wording reflects the strength of recommendations.

9.3.3.1 Recommendations for interventions that must or must not be used

Recommendations that an intervention must or must not be used are usually included only if there is a legal duty to apply the recommendation, for example to comply with health and safety regulations. In these instances, give a reference to supporting documents. These recommendations apply to all patients.

However, occasionally the consequences of not following a recommendation are so serious (for example, there is a high risk that the patient could die) that using 'must' (or 'must not') is justified. Discuss this with the Guidelines Commissioning Manager at NICE, and explain in the recommendation the reason for the use of 'must'.

If using 'must', word the recommendation in the passive voice ('an intervention must be used') because the distinction between 'should' and 'must' is lost when the recommendation is turned into a direct instruction.

Examples

  • Ultra-rapid detoxification under general anaesthesia or heavy sedation (where the airway needs to be supported) must not be used. This is because of the risk of serious adverse events, including death.

  • Gloves used for direct patient care:

    • must conform to current EU legislation (CE marked as medical gloves for single use) and

    • should be appropriate for the task.

9.3.3.2 Recommendations for interventions that should or should not be used – 'strong' recommendations

For recommendations on interventions that 'should' be used, the GDG is confident that, for the vast majority of people, the intervention (or interventions) will do more good than harm, and will be cost effective.

Use direct instructions for recommendations of this type where possible (see section 9.3.1), rather than using the word 'should'. Use verbs such as 'offer', 'refer', 'advise' and 'discuss'.

Example

  • Offer bariatric surgery as a first-line option (instead of lifestyle interventions or drug treatment) to adults with a BMI of more than 50 kg/m2.

Use similar forms of words (for example, 'Do not offer…') for recommendations on interventions that should not be used because the GDG is confident that they will not be of sufficient benefit for most patients.

Example

  • Do not offer antibiotic prophylaxis against infective endocarditis to people at risk undergoing dental procedures.

If an intervention is strongly recommended but there are two or more options with similar cost effectiveness, and the choice will depend on the patient's values and preferences, a 'should' recommendation can be:

  • combined with a 'could' recommendation (see section 9.3.3.3), for example by using wording such as 'Offer a choice of drug A or drug B' or

  • followed by a 'could' recommendation, for example 'Offer drug treatment. Consider drug A or drug B.'

9.3.3.3 Recommendations for interventions that could be used

For recommendations on interventions that 'could' be used, the GDG is confident that the intervention will do more good than harm for most patients, and will be cost effective. However, other options may be similarly cost effective, or some patients may opt for a less effective but cheaper intervention. The choice of intervention, and whether to have the intervention at all, is therefore more likely to vary depending on a person's values and preferences, and so the healthcare professional should spend more time considering and discussing the options with the patient. It may be possible to make 'strong' recommendations for subgroups of people with different values and preferences. NICE's report Social value judgements: principles for the development of NICE guidance (2nd edition; 2008) states the following:

'Although NICE agrees that respect for autonomy and individual choice are important for the NHS and its users, this should not mean that NHS users as a whole are disadvantaged by guidance recommending interventions that are not clinically and/or cost-effective.'

Use direct instructions for recommendations of this type where possible (see section 9.3.1), rather than using the word 'could'.

Use 'consider' to indicate that the recommendation is less strong than a 'should' recommendation.

Examples

  • Consider combination chemotherapy to treat patients with advanced breast cancer for whom a greater probability of response is important and who understand and are likely to tolerate the additional toxicity.

  • Consider carbamazepine and oxcarbazepine but be aware of the risk of exacerbating myoclonic or absence seizures.

Do not use 'consider offering', because of potential confusion with the wording of strong recommendations. Also, it might be misinterpreted to mean that a healthcare professional may consider offering an intervention without discussing it with the patient.

To minimise confusion, only use 'consider' to indicate the strength of a recommendation. Avoid other possible uses of 'consider'. For example, if a particular clinical sign or symptom should make a healthcare professional think about a diagnosis, use 'be aware of the possible diagnosis…', 'explore a diagnosis of…' or similar, rather than 'consider a diagnosis of'. Use 'take other factors into account' or similar, instead of 'consider other factors'. 'Assess' and 'think about' are other possible alternatives to 'consider'.

9.3.4 Emphasise the patient's involvement

To emphasise the patient's role in decision-making and the need for them to consent to treatment, generally use verbs such as 'offer', 'consider' and 'discuss' in recommendations, rather than 'prescribe' or 'give'. As described above, 'consider' is used for recommendations on interventions that could be used, and implies that more discussion will be needed.

Use 'people' or 'patients' rather than 'individuals', 'cases' or 'subjects'. Where possible, use 'people' rather than 'patients' for people with mental health problems or chronic conditions. 'Service users' can be used for people with mental health problems if 'patients' is the only alternative. Do not use 'patients' in relation to healthy pregnant women.

9.3.4.1 Recommendations about patient-centred care

The NICE guideline includes a standard section on patient-centred care that covers informed consent and taking into account the patient's individual needs. This section also cross-refers to NICE guidance on patient experience in adult NHS services, which covers subjects such as treating the person as an individual, communication, information and shared decision-making. NICE has also produced guidance on service user experience in adult mental health, which is cross-referred to in guidelines on mental health. The patient experience and service user experience guidance can be cross-referred to in recommendations, but specific recommendations should not be made on issues covered in that guidance unless there are particular reasons to do so that relate to the guideline topic. Examples include:

  • if there are issues relating to provision of information to patients, or to patients' support needs, that are specific to the condition covered by the guideline

  • if certain drugs are prescribed 'off-label' (see section 9.3.6.3) and more detailed forms of consent than usual are required from patients.

9.3.5 Use plain English

In general, follow the principles of effective writing as described in the 'Writing for NICE' booklet, which is available on the NICE webboard for NCCs.

Avoid vague words and phrases, such as 'may' and 'can', or general statements such as 'is recommended', 'is useful/helpful', 'is needed' and 'treatment options include'. Instead, use an active verb that tells readers what they should do, and indicates the strength of the recommendation.

Examples

  • Instead of 'an intervention may be offered', say 'consider the intervention'.

  • Instead of 'an intervention is recommended', say 'offer the intervention'.

  • Instead of 'an intervention is helpful', say 'offer the intervention' or 'consider the intervention' (see section 9.3.3).

'Appropriate' is often redundant: for example 'give appropriate advice', because we would never recommend giving inappropriate advice.

9.3.6 Recommendations on drugs, including off-label use

Guideline developers should follow NICE's standard procedure when referring to drugs. This includes using standard wording when off-label use of drugs is recommended.

9.3.6.1 Use generic names

Give the recommended international non-proprietary name (rINN), as listed in the British national formulary (BNF). Usually, only the generic name is needed. Occasionally (for example, if referring to a specific preparation or device), the proprietary name may be given in parentheses at first mention. Do not give the manufacturer's name.

9.3.6.2 Do not give dosages

Readers are expected to refer to the summary of product characteristics (SPC) for details of dosages. Include dosage information only if there is evidence that a particular drug is often prescribed at the wrong dosage, or there is clear evidence about the effectiveness of different dose levels. If off-label use is being recommended and there is no relevant dosage information in the BNF, include details of the dosage regimen in the full guideline. SPCs can be found in the Electronic Medicines Compendium.

9.3.6.3 Off-label use

Make it clear if the recommended use is outside the drug's licensed indication ('off label').

Recommendations are usually about the uses of drugs (often referred to as the licensed indications) for which the drug regulatory authority has granted a marketing authorisation, either in the UK or under the European centralised authorisation procedure. However, there are clinical situations when the use of a drug off-label may be judged by the prescriber to be in the best clinical interests of the patient. Off-label use may be recommended if the clinical need cannot be met by a licensed product and there is a sufficient evidence base and/or experience of using the drug to demonstrate its safety and efficacy to support this. Off-label prescribing is particularly common in pregnant women and in children and young people (see below), as these groups have often been excluded from clinical trials during drug development. When prescribing a drug off-label, the prescriber should follow relevant professional guidance (for example, the General Medical Council's Good practice in prescribing medicines – guidance for doctors) and make a clinical judgement, taking full responsibility for the decision for the patient under his or her direct care. In addition, the patient (or those with authority to give consent on their behalf) should be made fully aware of these factors and provide informed consent, which should be documented by the prescriber.

A licensed drug is accompanied by an SPC, which describes the indications, cautions and contraindications for a drug based on an assessment of safety, quality and efficacy by the regulatory authority. The NCC and GDG should check recommended uses against the licensed indications listed in the SPC, and include a footnote if the drug does not have a UK marketing authorisation for the use being recommended. The footnote should make it clear that the drug is not licensed for the stated use.

This standard wording for the footnote captures the above points:

  • At the time of publication ([month year]), [name of drug] did not have a UK marketing authorisation for this indication. The prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Good practice in prescribing medicines – guidance for doctors for further information.

Additional information can be added as needed – for example, if off-label use is recommended and the drug is commonly used in UK clinical practice, a phrase such as 'Although this use is common in UK clinical practice' can be added. Other examples of footnote wording are shown in box 9.1. In cases where the SPC for a drug specifically mentions a caution or contraindication for its use but the GDG wishes to recommend the drug, this should be stated clearly in the recommendation or footnote. The evidence that the GDG has considered in reaching the conclusion that use in these circumstances can be justified should be clearly set out in the full guideline.

If a guideline includes recommendations for off-label use of drugs, the introduction to the NICE version should include standard wording (as in the NICE guideline template) about the responsibilities of the prescriber and the need to follow relevant professional guidance (for example, the General Medical Council's Good practice in prescribing medicines – guidance for doctors).

If there is no information on dosage regimens available in a recognised source (such as the BNF), the NCC should document dosage information in the full guideline and alert the NICE implementation team to ensure that this is disseminated to prescribers.

Prescribing drugs outside their licensed indications to children and young people

In certain circumstances drugs are prescribed to children and young people outside their licensed indications (off-label use) because the clinical need cannot be met by licensed drugs; for example, for an indication not specified in the marketing authorisation, or administration of a different dose. The Standing Committee on Medicines (a joint committee of the Royal College of Paediatrics and Child Health and the Neonatal and Paediatric Pharmacists Group) has issued a policy statement on the use of unlicensed drugs and the use of licensed drugs for unlicensed applications in children and young people. This states clearly that such use is necessary in paediatric practice and that doctors are legally allowed to prescribe drugs outside their licensed indications where there are no suitable alternatives and where use is justified by a responsible body of professional opinion (Joint Royal College of Paediatrics and Child Health/Neonatal and Paediatric Pharmacists Group Standing Committee on Medicines 2010).

Therefore, where there is no alternative treatment and only where there is a sufficient evidence base and/or experience of using the drug to demonstrate its safety and efficacy, a clinical guideline may recommend use of a drug outside its licensed indications for treating a child or young person. It is expected that prescribers will use the SPC to inform their prescribing decisions for individual patients, and they should be able to justify using a drug outside its licensed indications. Informed consent should be obtained from the child and/or their parent or guardian as appropriate and documented.

Footnotes for recommendations addressing off label-use of drugs in children and young people should follow the format described above and in box 9.1.

Box 9.1 Examples of footnotes to guideline recommendations about the off-label use of drugs

Where use is outside the licensed indication:

At the time of publication (August 2011), spironolactone did not have a UK marketing authorisation for this indication. The prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Good practice in prescribing medicines – guidance for doctors for further information.

[Adapted from: Hypertension: clinical management of primary hypertension in adults. NICE clinical guideline 127 (2011).]

Vaginal PGE2 has been used in UK practice for many years in women with ruptured membranes. However, the SPCs (July 2008) advise that in this situation, vaginal PGE2 is either not recommended or should be used with caution, depending on the preparation (gel, tablet or pessary). Healthcare professionals should refer to the individual SPCs before prescribing vaginal PGE2 for women with ruptured membranes. The prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Good practice in prescribing medicines – guidance for doctors for further information.

[Adapted from: Induction of labour. NICE clinical guideline 70 (2008).]

Where the SPC mentions a specific caution or contraindication:

Metformin is used in UK clinical practice in the management of diabetes in pregnancy and lactation. Clinical experience supports its effectiveness and safety but this is not currently reflected in the SPC. The SPC (March 2008) advises that when a patient plans to become pregnant and during pregnancy, diabetes should not be treated with metformin but insulin should be used to maintain blood glucose levels. For use of metformin in these situations, the prescriber should follow relevant professional guidance, taking full responsibility for the decision. Informed consent should be obtained and documented. See the General Medical Council's Good practice in prescribing medicines – guidance for doctors for further information.

[Adapted from: Diabetes in pregnancy: management of diabetes and its complications from pre-conception to the postnatal period. NICE clinical guideline 63 (2008).]

9.3.7 Recommendations on waiting times and ineffective interventions

Guideline developers should follow NICE's standard advice for recommendations on waiting times. It is also acceptable to make recommendations that advise stopping the use of an ineffective intervention.

9.3.7.1 Waiting times and other policies set by other bodies

Avoid giving targets for waiting and referral times: refer to relevant targets set by the Department of Health or the Welsh Government, and where possible direct readers to the relevant document rather than including the target in the recommendation. This is because policy can change, making a guideline that includes such targets out of date. If no target exists, recommendations may include a maximum time if the GDG considers this to be essential.

Sometimes a recommendation will need to specify a waiting time, referral time or time of intervention because this relates to the safety and/or effectiveness of a clinical intervention. In this case, check that the recommendation does not conflict with relevant targets set by the Department of Health or the Welsh Government and ensure that the clinical reason for specifying the time is made clear.

9.3.7.2 Ineffective interventions

Recommend stopping ineffective interventions: state explicitly if particular treatments or activities should not be carried out or should be stopped (see box 9.2).

Box 9.2 Example of a recommendation about stopping ineffective practice

Do not routinely offer pharmacological or mechanical VTE prophylaxis to patients with cancer having oncological treatment who are ambulant.

[From: Venous thromboembolism: reducing the risk. NICE clinical guideline 92 (2010).]

9.3.8 Using tables in recommendations

Do not use tables to summarise several actions in one recommendation. Such summaries make it more difficult to link the recommended actions to the evidence summaries. A recommendation may include a small table to improve clarity; for example, to present information that should be shared with patients, or if the information is most easily understood when tabulated. An example is shown in box 9.3.

Box 9.3 Example of a table within a recommendation

Use predicted 6-month mortality to categorise the risk of future adverse cardiovascular events as follows:

Predicted 6-month mortality

Risk of future adverse cardiovascular events

1.5% or below

Lowest

>1.5 to 3.0%

Low

>3.0 to 6.0%

Intermediate

>6.0 to 9.0%

High

over 9.0%

Highest

[From: Unstable angina and NSTEMI: the early management of unstable angina and non-ST-segment-elevation myocardial infarction. NICE clinical guideline 94 (2010).]

9.4 Prioritising recommendations

NICE's standard clinical guidelines can cover large clinical areas and, as a result, often contain a considerable number of recommendations relevant to the many review questions. The GDG will need to identify a subset of these recommendations as key priorities for implementation. These may be used to guide implementation activities (see chapter 13) and may be useful in the subsequent development of NICE quality standards. The number of recommendations prioritised in this way will vary depending on the guideline, but is normally between 5 and 10. There is no 'ranking' within this set of recommendations.

Key priorities for implementation are usually those that are likely to do at least one of the following:

  • have a high impact on outcomes that are important to patients

  • have a high impact on reducing variation in care and outcomes

  • set challenging but achievable expectations of health services

  • focus on key infrastructural and clinical requirements for high-quality care

  • include actions that are measurable

  • lead to more efficient use of NHS resources

  • promote patient choice

  • promote equality.

In addition, the GDG should attempt to identify recommendations that are particularly likely to benefit from support from NICE's implementation programme. Criteria overlap with those above, but include whether a recommendation:

  • relates to an intervention that is not part of routine care

  • requires changes in service delivery

  • requires retraining of staff or the development of new skills and competencies

  • highlights the need for practice to change

  • affects and needs to be implemented across a number of agencies or settings (complex interactions)

  • may be viewed as potentially contentious, or difficult to implement for other reasons.

There should be a clear record of which criteria were considered particularly important by the GDG for each prioritised recommendation. This should be reported in a short paragraph in the full guideline.

9.5 Formulating research recommendations

The GDG is likely to identify areas in which there are uncertainties or where robust evidence is lacking. NICE has published a Research recommendations process and methods guide, which details the approach to be used across NICE's guidance-producing programmes to identify key uncertainties and associated research recommendations.

For standard clinical guidelines where there may be many hundreds of uncertainties, it will not be possible to document every uncertainty in detail. Similarly, although GDGs could write research recommendations for dealing with each uncertainty, this is not likely to be feasible. Therefore the GDG should select up to five key research recommendations for inclusion in the NICE version of the guideline; more research recommendations may be listed in the full guideline. Further information about how these should be derived can be found in the research recommendation process and methods guide.

9.6 Further reading

Brown P, Brunnhuber K, Chalkidou K et al. (2006) How to formulate research recommendations. British Medical Journal 333: 804–6

Claxton K, Sculpher MJ (2006) Using value of information analysis to prioritise health research: some lessons from recent UK experience. Pharmacoeconomics 24: 1055–68

Glasziou P, Del Mar C, Salisbury J (2003) Evidence-based medicine workbook. London: British Medical Journal Books

Guideline Implementability Appraisal (GLIA) [online]

Joint Royal College of Paediatrics and Child Health/Neonatal and Paediatric Pharmacists Group Standing Committee on Medicines (2010) The use of unlicensed medicines or licensed medicines for unlicensed applications in paediatric practice [online]

Lord SJ, Irwig L, Simes RJ (2006) When is measuring sensitivity and specificity sufficient to evaluate a diagnostic test, and when do we need randomized trials? Annals of Internal Medicine 144: 850–5

Sackett DL, Straus SE, Richardson WS (2000) Evidence-based medicine: how to practice and teach EBM, 2nd edition. Edinburgh: Churchill Livingstone

Schünemann HJ, Best D, Vist G et al. for the GRADE Working Group (2003) Letters, numbers, symbols and words: how to communicate grades of evidence and recommendations. Canadian Medical Association Journal 169: 677–80

Scottish Intercollegiate Guidelines Network (2002) SIGN 50. A guideline developer's handbook. Edinburgh: Scottish Intercollegiate Guidelines Network



[14] Information throughout this manual relating to the role of the National Collaborating Centres in guideline development also applies to the Internal Clinical Guidelines Programme at NICE.