'Half of what we know is wrong... but which half?'
'Follow the strongest evidence, wherever it leads.'

Advances in knowledge are made possible by questioning the accuracy of what is currently known, constantly challenging existing views and attempting to prove them wrong. Nothing can ever be finally proved to be true. It can only be shown that repeated studies fail to prove it false. As crazy as it may have sounded when Pasteur insisted that micro-organisms exist and cause infection or when Lister claimed that scrubbing for surgery might decrease surgical infections, it was an imaginative departure from what was known and accepted. Why was it so hard to update thinking decades after it was proven that there was no 'tooth worm'? Why do we believe that lemmings throw their little bodies over cliffs in mass suicides when they don't? Because misinformation gets published, then referenced and re-published.

We are in a phase of knowledge in which we are attempting to practise evidence-based veterinary dentistry (EBVD). We first form an answerable question to a clinical problem, then search for the best evidence to answer it. The information acquired is critically evaluated, then applied to the patient. Finally, the results are evaluated. This is the best use of scientific research to provide solutions and treatments for existing problems. Unfortunately, there is no good evidence available for many of our daily decisions. This problem was much worse in the past; many 'facts' that have been accepted for years were based on inaccurate information. Sometimes even the most basic and universally accepted concepts are unverified myths. Therefore we must be mindful of some fundamental tenets of science:

 Avoid dogma by critically reviewing previously interpreted data

 Apply consistent reasoning

 Question the validity of authoritative pronouncements

'It is the right and duty of dentists to constantly criticize dogma and investigate the justification for current standards of practice.
...to challenge...recommendations is a professional obligation. It is also a definite necessity.'
Dr. John Hardie

The opinions of experts (and therefore conferences and their manuscripts) are very low-level evidence. But few of us have the time to properly research and evaluate available evidence so we listen to experts. This is a necessary division of labour. So it becomes important how you choose to whom to listen. It should be based on how they acquire their knowledge and how reliable is their information. That helps guard against 'the authority fallacy' in which we trust someone simply because they have become well known. That said, each presenter has invested a significant amount of their time to research their topic and will honestly share the results with you, so their information is likely to be relatively good.

In EBVD the first step to answer our clinical question is to find the highest level of available evidence. If high level evidence does not exist, then lower levels of evidence are considered. Unfortunately, the highest level of evidence (for example a systematic review of a series of well designed randomised controlled studies) does not exist for nearly every decision that we face. To identify the strength of evidence, some of the factors that should be considered include: clinical significance; clinical relevance; publication bias; confounding factors; chance; and interpretation.

Clinical Significance

The data from a systematic review on periodontal surgery for root coverage demonstrated that connective tissue (CT) grafts were significantly better than guided tissue regeneration (GTR) for reducing recession. The analysis of the pooled data from the included studies showed a mean difference of 0.43 mm with a 95% confidence interval from 0.62 to 0.23, chi square for heterogeneity 7.8 (df = 5) P = 0.17. Impressive. So should we use free CT grafts and discontinue GTR? No! Although there is a (statistically) 'significant advantage' to the CT over GTR, it is not a clinical one. 0.5 mm is not clinically significant. More importantly, bony repair and supportive bone quality, alveolar ridge height and periodontal ligament (PDL) repair were not evaluated; it only looked at recession.

Clinical Relevance

A meta-analysis from a systematic review on GTR for infrabony defects showed that there was an additional benefit of 1.1 mm gain in attachment of GTR over only open-flap surgery. Does this mean that you can expect 1.1 mm more attachment in each of your patients when you add GTR? No! These are pooled data from many patients... some will do better and others worse.

Publication Bias

One publication bias is the fact that studies with positive results are more likely to get published than studies with negative results. From a letter to the editor of a science journal: '...although the truth is defined as much by what isn't as by what is, we tend to reward people who consistently confirm their research hypotheses... the ability to pick a winner is a significant asset... it is naïve to expect a selfless drive to tell the truth to govern the multi-billion dollar industries of science.' All the missing data that showed no benefit skew the result of a systematic review of published studies towards showing a benefit. This bias is being addressed now that many top medical journals are taking steps to ensure that medical firms publish negative results.

Confounding Factors

Confounding factors are controlled through randomisation, restriction to similar groups or matching groups. Failure to control for them should be found during the analysis or the review process. One example was when the newspaper reported on a study that showed that women taking birth control pills were less likely to have certain illnesses. When the chief statistician reanalysed the numbers he found that the association between the pill and lower incidence of disease was actually associated with age, not whether or not the participants took the pill. They had compared older women who never took the pill with younger women who did. He 'nearly spit out his cereal' when he read the newspaper report of the study that had not yet gone through peer review or been accepted for publication.

Chance

What is the likelihood that the finding could be caused purely by chance? This is where we use statistics. Since we cannot test every member of a population, small groups are tested and statistics applied to avoid sampling errors. It is a very common error in studies of restorative materials to make conclusions about a class of materials based on testing one or only a few members of the class.

Interpretation

No matter how well a study is designed, carried out and analysed, incorrect conclusions can still be drawn. For example the study that showed dental radiographs were a risk factor for low birthweight babies did not address why the dental radiographs were made. Possibly to evaluate periodontal disease? And we know that periodontal disease is a risk factor for low birthweight babies.

Some concepts that are no longer correct due to changes in materials technology include:

 'Extension for prevention in cavity preparation for caries treatment'

 'Cavity preps should never leave unsupported enamel rods'

 'Crown preparations should have a 3-5 degree taper'

 'Gingivitis is reversible, but periodontitis is irreversible'

 'Gingivectomy is an appropriate treatment to remove a periodontal pocket'

Some of the many myths that have been proven wrong or that should be investigated include:

 'The most important part of root canal treatment is obturation of the apical third'

 'The best canal flushing material is 50% sodium hypochlorite'

 'Surgical root canal (root tip fill) should be done when an endodontic procedure has failed'

 'Root canals should be filed and filled 1 mm short of the radiographic apex'

 'Nickel-titanium is the best material for root canal files'

 'The root canal is always in the centre of the root'

 'The canal should always be instrumented to a circular cross-section'

 'Clean white dentine shavings indicate that filing is complete'

 'Filing should be done using a step-back technique'

 'K-files are rotated counterclockwise with balanced force, then quarter turn clockwise to advance, then counterclockwise again'

 'Root canal pulp obliteration means there is no root canal to navigate and treat'

 'Ca(OH)₂ should be placed on very thin dentine as an indirect pulp cap'

 'After vital direct pulp capping, the presence of a dentine bridge proves success'

 'Untreated gingivitis causes periodontitis'

 'Periodontitis causes systemic diseases' (well-this may be true but it has not yet been proved)

 'Teeth should not be polished since it causes significant enamel loss'

 'Mucogingival flaps should have a wide base'

 'Mucogingival flaps should be apically repositioned'

 'Newer bonding agents are stronger'

 'Amalgam is inferior to composite resin as a restorative material'

 'Composite resins move towards the activating light when cured'

 'The faster the cure rate the better'

 'Undercuts should be made for retention in cavity preps using an inverted cone burr'

 'Resorptive lesions in cats cause severe pain, halitosis, ptyalism and anorexia'

 'Consil^TM should be placed in every extraction site'

 'We should all be using local anaesthesia injections'

 When and how to use systemic antibiotics remains widely varied

Question everything regardless of how well established it is. Evaluate the quality of the supporting evidence. Be willing to change your beliefs when the evidence shows that they were wrong.