How we grade evidence
The grade describes how confident we can be that an intervention works. Not how new it is, how expensive it is, or how enthusiastically it is marketed.
The five grades
- Strong evidenceConsistent benefit across multiple independent, adequately powered, randomized controlled trials, or a systematic review of them.
- Moderate evidenceAt least one adequately powered, randomized, blinded controlled trial showing a clinically meaningful benefit.
- Limited evidenceControlled trials exist but are small, unblinded, short, inconsistent, or rely on surrogate outcomes.
- Untested in dogsNo completed randomized controlled trial in the target population. Unknown, not disproven.
- Evidence of little or no benefitAdequately powered trials or pooled analyses found no clinically meaningful effect. This is a finding, not a gap.
What we weight heavily
- Randomization and blinding. Owner-reported outcomes in canine pain trials routinely show placebo response rates above 15%. An unblinded trial cannot separate a working drug from a hopeful owner.
- Objective outcomes. Force-plate gait analysis measures how much weight a dog puts on a limb. It does not care what anyone believes. Trials using objective endpoints get more weight than those using questionnaires alone.
- Adequate power and duration. A 12-dog, 14-day study is a pilot, not an answer.
- Independent replication. One trial is a result. Two independent trials agreeing is knowledge.
What we disclose but do not penalize outright
Industry funding. Most veterinary trials are manufacturer-funded, because almost nobody else pays for them. Refusing to cite sponsor-funded work would leave this site nearly empty. Instead, every evidence entry names its funder, and sponsor-funded results without independent replication are capped at grade B.
Why grade D exists
Most evidence resources quietly omit interventions that failed. That is a mistake. An owner spending money on a supplement every month deserves to know that the pooled analysis found no effect, and needs to hear it in the same place they research everything else. Grade D means trials were done and came back negative. It is a real finding and it is arguably the most useful information on this site.
What a grade is not
A grade is not a recommendation for your dog. A grade-A intervention can be wrong for an individual animal with the wrong comorbidities, and a grade-C intervention may be exactly right when better options are contraindicated. The grade tells you how strong the general evidence is. Your veterinarian applies it to a specific animal.
Regulatory approval and evidence grade are also different things. Approval reflects what a regulator accepted at a point in time under a particular standard, and standards differ across decades and across approval pathways. Some approved products here sit at grade C for that reason.
Review cadence
Every intervention carries a last_reviewed date. Entries older than twelve months are flagged for
re-review, because a page that was accurate in 2024 may not be now. The anti-NGF safety discussion is a live example.