From Evidence to Causality: How do We Determine Causality?
Conclusion

When faced with real life situations such as the above exercise, it is imperative that health care professionals are able to answer such questions accurately in order to maintain credibility with their patients. Scientific evidence is dynamic and ever changing as new research is conducted. What we know to be true right now may change and thus keeping current with the scientific literature is important.

With information currently accessible to everyone on-line, patients are constantly searching for answers to their own health questions but often find inaccurate information. It is our role to ensure that they receive the most current and accurate information about their oral health. This means that oral health professionals need to have a better understanding of the principles of causation and of how to interpret the literature and ultimately apply it towards this end. Some suggestions on how to approach this daunting task are as follows:

• Choose either clinical practice guidelines, meta-analyses, systematic reviews, or critical summaries. If these are not available then look for original research studies that are peer reviewed, rather than relying on narrative reviews as they may misinterpret the literature and are only the opinion of the authors.
• If only original research is available, identify the study design by reading the article (i.e., RCT, cohort, case-control, etc.).
• Apply the criteria for causation:
  • Did the exposure precede the appearance of the disease? (Temporality)
  • Do all studies have the same results? (Consistency)
  • Are there studies that contradict the hypothesis? (Coherence)
  • Is it plausible to explain how a relationship could exist? (Plausibility)
  • There is no other plausible explanation (Specificity)
  • Pay attention to whether confounders were measured in the study (and were they controlled?) (Experiment/type)
  • Pay attention to the limitations of the study described by the authors i.e., size of the study, selection of the participants, etc. (Experiment)
  • Is there a statistical association (RR >2; OR >4) between the exposure and the disease that is statistically significant (i.e., p<.05)? (Strength of Association)
  • Have lab or animal studies shown the same results? (Analogy)
  • Are there studies showing a dose-response relationship.... i.e., the more severe the variable in question, the greater the chance of the disease occurring (outcome). (Dose-response relationship)

Remember, that proving causality is a complex undertaking! "Unknown effects of biomedical individualism as well as the social determinants of disease cast a large 'web of causation' that are difficult to take into account."¹ Thus, if not absolutely certain, avoid using any words that imply ‘causation’ when talking to your patients.