Strong vs Weak Recommendations: How Evidence Is Graded

Your clinician mentions that a guideline gives a "weak recommendation" for a test, or a "grade C" for screening. That may sound like the advice is not very good. In fact, these labels carry useful meaning. They tell you how sure the experts are, and how much your own preferences should shape the choice. This guide explains the main grading systems in plain language.

Two separate ratings

Most modern guidelines rate two things. [1]

• Certainty of evidence: how confident the experts are that the studies show the true effect.
• Strength of recommendation: how confident the experts are that following the advice will do more good than harm for most people.

These two ratings are linked but not the same. Good evidence often leads to strong recommendations, but not always.

The GRADE system

Many groups around the world use GRADE, including many Canadian guideline groups and the World Health Organization. [1]

Certainty of evidence

GRADE rates evidence in four levels. [1]

• High: further research is very unlikely to change our confidence in the result.
• Moderate: further research may change the result.
• Low: further research is likely to change the result.
• Very low: the result is very uncertain.

Strength of recommendation

GRADE uses two levels. [1] [2]

• Strong: "We recommend". Most informed people would choose this option. Clinicians can offer it to most people.
• Weak, also called conditional: "We suggest". Many informed people would choose this option, but some would not. The choice depends more on your values and situation.

A recommendation can be for or against a test or treatment.

What goes into the strength

Panels weigh four main things. [2]

• The balance of benefits and harms.
• The certainty of the evidence.
• How much people's values and preferences vary.
• Costs and use of resources.

What doctors look at

When your clinician reads a recommendation, the grade tells them how to talk with you. [2]

• A strong recommendation: they will usually advise the option, unless you have a special reason to differ.
• A weak or conditional recommendation: they should walk you through the benefits and harms and help you choose based on what matters to you.

Canadian examples

The former Canadian Task Force on Preventive Health Care used GRADE. Two examples show how it works.

• Prostate cancer screening with PSA: strong recommendation against for men under 55 and 70 or older, and weak recommendation against for men aged 55 to 69. [4] The weaker wording for 55 to 69 signals that some men in that age group may reasonably choose screening.
• Colorectal cancer screening with stool tests or sigmoidoscopy: strong recommendation for adults aged 60 to 74, and weak recommendation for adults aged 50 to 59. [5]

The Task Force ended its mandate in March 2026, but its published guidelines remain available as reference.

USPSTF letter grades

The US Preventive Services Task Force uses letter grades. [3]

• A: recommended. High certainty of substantial net benefit.
• B: recommended. High certainty of moderate net benefit, or moderate certainty of moderate to substantial net benefit.
• C: offer or provide selectively, based on the person. At least moderate certainty that the net benefit is small.
• D: recommended against. Moderate or high certainty that there is no net benefit or that harms outweigh benefits.
• I: insufficient evidence. The balance of benefits and harms cannot be determined.

A C grade works much like a weak recommendation: the decision should be personal. An I statement is not a recommendation against. It means more research is needed. [3]

Heart guideline classes

American College of Cardiology and American Heart Association guidelines use a Class of Recommendation (1, 2a, 2b, or 3) and a Level of Evidence (A, B, or C). Class 1 is the strongest. Class 3 means no benefit or harm. Level A means evidence from several high quality trials.

Why grades differ between groups

Two guideline groups can look at the same studies and give different grades. They may weigh harms such as false positives or overdiagnosis differently. They may judge how well the studies apply to their own population. They may also update at different times. When grades differ, ask your clinician which guideline they follow and why. Neither grade is wrong just because it differs. Each reflects the judgment of a panel about the same uncertain evidence, and that judgment can reasonably vary.

How to use these labels

• A strong recommendation for an option is good reason to take it seriously.
• A weak recommendation invites a conversation. Your values count most here.
• Low certainty evidence means the advice may change as new studies appear.
• An I statement means nobody knows yet. Ask what your clinician suggests while we wait for better evidence.

Questions to ask your clinician

• How strong is the recommendation for this test or treatment?
• How certain is the evidence behind it?
• If it is a weak recommendation, what would lead someone to choose differently?
• What are the main benefits and harms for someone like me?
• Do Canadian and US guidelines grade this the same way?
• Is new evidence expected soon?

Bottom line

Guidelines rate both the certainty of evidence and the strength of each recommendation. A strong recommendation means most informed people would choose the same path. A weak or conditional one means your values should lead. Letter grades and classes follow the same idea. Use these labels to know when to ask more questions and share the decision.

Primary AI can show you the guideline and study sources behind this topic in plain language, so you can bring better questions to your appointment.