Association and Causality PDF

 Association = identifiable relationship (statistical connection) between an exposure and a disease  Causation states that there is a true relationship between risk factor and disease. A cause leads to disease.  An association alone does not make the  relationship between r...

 Association = identifiable relationship (statistical connection) between an exposure and a disease  Causation states that there is a true relationship between risk factor and disease. A cause leads to disease.  An association alone does not make the  relationship between risk factor and disease causal.  If there is a difference in disease occurrence between groups, first question to be asked is whether this difference is statistically significant. If this difference is significant, then a statistical association is said to exist between the factor and the disease. This statistical association could be: 1. Artifactual (Spurious) association or 2. Indirect (non-causal) association. Or 3. Causal association Artifactual (Spurious) association is a false association that can result from chance or from bias in study methods. One should attempt to confirm this association by replication. If such an association doesn't hold up in this replication, then it may be considered as spurious. Thus,whenever a statistically significant association is found, it must be examined carefully to be sure that it is not attributable to some artifact or bias. Indirect(non-causal) association: It occurs when a factor and disease are associated only because both are related to some common underlying condition. Alteration in the indirectly associated factor will not produce alteration in the frequency of disease unless the change affects the common underlying condition as well. Example on indirect association: Altitude and cholera. Fetid air in low altitudes is the indirect factor. Water impurity in low altitude areas is the risk factor for cholera and not the fetid air. So low altitude is indirectly associated with cholera through the common underlying condition which is the water pollution. Causal association There are several criteria used to evaluate the likelihood that an association is causal: 1. Strength of the association as measured by the relative risk(Risk ratio) or the odds ratio. The larger this ratio, the greater the likelihood of this association to be causal.  < 2.0 very weak association 2.0 – 3.9 weak – moderate association > 4.0 moderate – strong association 2. Dose-Response relationship. With increasing levels of exposure to the factor, there is a corresponding rise in the disease. 3. Consistency of the association. An association discovered in one study persists in other studies conducted by using other methods and on different population. Association is observed repeatedly in different populations, different places and times, in different types of studies, conducted by different investigators First study and lack of consistency does not rule out causality 4. Temporality. Exposure to the factor should precede the occurrence of disease and allows for the necessary period of induction or latency.  Age-standardized death rates from bronchogenic carcinoma, by smoking history Adapted from Hammond EC, Horn D. JAMA 1958; 166:1294-1308. 5. Specificity of the association. One manifestation follows from only one cause. Ex. Angiosarcoma of liver and exposure to vinyl chloride, adenocarcinoma of vagina in female offspring resulting from DES ingestion by mothers during pregnancy. Although specificity is a strong evidence for causality, yet lack of it is of less significance. 6.Coherence with existing information(Biological Plausibility). ▪ Plausible – Higher incidence of disease in individuals who are more sexually active  The disease could be a sexually transmitted disease ▪ Not Plausible – Polio vaccine reduces fertility  Conflicts with current knowledge of polio vaccine production, biological mechanism for fertility and contraception