How good is the evidence for what we actually do?

The gap map covers each of these, worked through the same way and to the same standard. All of them are being expanded as the literature is covered, and further specialties are added as the work goes on.

If you work in one of these areas and would be willing to draft entries or check them, that speeds the work up considerably. Write to contact@evidencegap.vet.

The first mark says how far the evidence behind a question can be relied on. It follows GRADE, the system used for the same purpose in human medicine.

The second mark says which way that evidence points. The two are kept apart because a question can rest on weak evidence and still point clearly, and because nobody looked and somebody looked and found nothing are different answers that are routinely treated as the same one.

Study design sets the starting point of an assessment and not its conclusion. A trial that was funded by the manufacturer, measured a surrogate outcome and has never been replicated does not outrank a well-conducted observational study by virtue of being a trial. The specific reasons certainty was not rated higher are recorded against every question, so that each judgement can be examined rather than accepted. The full method is on the map page.

• Commentary · Canine cardiology · August 2026

  Pimobendan in preclinical stage B2 mitral valve disease

  The trial was reconstructed from its published survival curves. Reclassifying nine dogs removes the significance of the primary result, the event proportions alone are not significant, and the subsequent full approval rests on a single-arm study that was measured against a benchmark it was never required to beat.

• Commentary · Feline cardiology · August 2026

  Steroids and heart failure in cats: science, or an old wives' tale?

  The belief that corticosteroids precipitate heart failure in cats rests on two papers published in 2004 and 2006. Neither paper demonstrates the effect, the controlled studies that followed found no cardiac change, and the largest cohort reported no excess of heart failure in treated cats.

• Editorial appraisal · Feline cardiology · August 2026

  Rapamycin (sirolimus) for subclinical hypertrophic cardiomyopathy in cats

  The published report and the FDA record disagree with each other about whether the endpoint of the single supporting trial was specified before the data were examined. The regulatory summary states that it was not.

• Editorial appraisal · Canine cardiology · June 2026

  Torasemide as an alternative to furosemide in canine congestive heart failure

  Non-inferiority demonstrated on a composite endpoint is a weaker claim than this result is generally taken to support. This appraisal is in preparation.

1. Absence of direct evidence

   Some interventions in routine use have never been tested against a control in the species they are given to. Dietary sodium restriction in canine congestive heart failure is one such practice, and although it is close to universal, the trial that would justify it has not been performed.

2. Borrowed evidence

   Other recommendations are carried across from human medicine, or from one species to another, without data establishing that the finding transfers. Dose, pharmacokinetics and disease phenotype rarely translate cleanly between species, and the assumption that they do is seldom stated explicitly.

3. Consensus in place of data

   A third group consists of screening intervals, treatment thresholds and monitoring conventions that originated in expert judgement and were subsequently cited as though they had been established by study. Such conventions may well turn out to be correct, but they should be identified as consensus rather than as evidence.