13 Research methods, statistics, critical review and evidence-based practice13.12 Guidelines and reporting
13.12.3 Guidelines: advantages, limitations and appraisal
4 min read1 figure
Written by Resident Doctor Committee · Published
Advantages and limitations of guidelines and protocols
Guidelines and protocols
Attribute | Guideline | Protocol |
|---|---|---|
What it states | Recommended care for defined circumstances | Agreed steps and who does them |
How it is used | Applied with clinical judgement | Followed consistently, with defined exceptions and escalation |
Advantages
Good guidance summarises evidence, supports safety and reduces unwarranted variation.
It gives trainees sensible defaults, supplies audit criteria and helps patients discuss options.
Protocols suit urgent or repeated tasks, such as rapid tranquillisation.
Limitations
Trials often exclude people with comorbidity or suicidality and follow participants for only 6 to 12 weeks, so applying them to routine patients needs judgement.
Industry sponsorship, selective reporting and slow updating also shape the evidence.
An average effect does not predict one person's response, and recommendations for several conditions may conflict.
Defensive adherence is following a recommendation mainly because a departure would be hard to defend.
Automation bias leads a clinician to accept a checklist's output despite contrary clinical evidence.
Departing from guidance
Document the patient-specific reasons, the alternatives considered and the plan agreed.
In negligence, guidelines are evidence, but neither compliance nor departure decides liability alone.
Changing clinician behaviour
Patient-specific reminders at the point of care, tied to the individual patient at the moment of the decision, change behaviour more than passive dissemination, endorsement or education alone.
Local development and specific educational dissemination also help.
Appraising clinical practice guidelines
AGREE II
AGREE II appraises how a guideline was developed and reported, using 23 items in six domains, each scored 1 to 7, plus two overall assessments.
Reliability improves with at least two appraisers, preferably four, and domain scores are reported separately, not as one total.
A high score does not guarantee that every recommendation is clinically correct or current.
AGREE REX adds nine items on the clinical applicability, values and preferences, and implementability of recommendations.
Scope and authorship
Read the objectives, target population, intended users and exclusions alongside the recommendations, because adult depression guidance is not automatically right for a 16-year-old.
Stakeholder involvement asks whether the relevant professions and patients' values were represented.
Editorial independence asks how funding and competing interests were handled, for example by declaration, restricted participation or exclusion from relevant discussions.
A bare list of consultancies does not establish independence, but an interest does not automatically invalidate every contribution.
Rigour and evidence
Traceability means following a recommendation back to its supporting review and rationale, and a missing link is a limitation, not proof that no evidence exists.
GRADE usually rates certainty for each outcome from high to very low, weighing bias, inconsistency, indirectness, imprecision and publication bias.
Evidence of very low certainty can still support sound advice if the uncertainty is visible.
Most recommendations combine evidence with judgement: SIGN marks good practice points and formal consensus advice, and NICE rationale sections say where judgement shaped a recommendation.
Committees should weigh benefits, harms, certainty, values, resources, equity, acceptability and feasibility, and explain the trade-offs.
Keep the recommended care separate from the implementation gap, since weekly therapy can still be the right care when the wait is nine months.
Strong versus conditional recommendations
Reader | Strong | Conditional (weak) |
|---|---|---|
Patient | Most would want it, and only a few would not | Most would want it, but many would not |
Clinician | Most patients should receive it | Different patients need different choices, with help to fit their values |
Policy maker | Can be adopted as policy in most situations | Needs substantial debate and many stakeholders |
Strength is not certainty: serious harm and an obvious precaution can justify strong advice on limited evidence, while small but certain benefits can leave preference-sensitive choices.
NICE generally uses "offer" for stronger advice and "consider" for conditional choices, and "must" usually marks a legal requirement or a particularly serious consequence. SIGN states strength outright.
NICE advises against routinely offering antidepressants first-line for less severe depression but allows them as the person's informed preference, so this is not a blanket prohibition.
Currency and applicability
Check update dates, evidence search dates, current status and safety advice, because a partial update can leave other recommendations resting on older reviews.
NICE reviews guidance by surveillance and prioritisation, not a fixed interval, so an old date does not always mean obsolete guidance.
An evidence update differs from an editorial refresh or a review that kept the existing guidance.
NICE, SIGN and the Scottish Medicines Consortium do different jobs, and Wales and Northern Ireland have their own adoption arrangements.
England and Wales use the Mental Health Act 1983 as amended, Scotland its 2003 Act, and Northern Ireland its 1986 Order plus commenced parts of its 2016 capacity legislation.
An enacted reform is not fully operational until commenced, so check commencement and current codes.
Where condition-specific recommendations conflict, make one integrated decision for the patient rather than stacking treatments.
When guidelines disagree
Guidelines differ over search dates, eligibility criteria, interpretation of evidence, values, resources, licensing or jurisdiction, and newer is not automatically better.
Licensing differs from NHS appraisal: NICE TA854 does not recommend esketamine nasal spray for treatment-resistant depression because clinical and economic uncertainties prevented a favourable cost-effectiveness judgement, not because it lacks pharmacological effect.
Document the guidance used, the alternatives, the patient-specific reasons and the person's preferences, and settle value choices by discussion with the patient.