Statistical Measures of Epidemiology

 Analysis of epidemiology is exposure disease-relationship dependable. Measurement of epidemiology is not dependent on hypothesis testing of P tests. P refers to the probability of occurrence of cancer.

Relative risk:

The mostly used measurement analysis is relative risk. It involves occurrence of disease and no disease. Ratio of exposed agent to unexposed agent is referred as relative risk. It is important to measure the correlation between the causative agent and disease occurrence. However they are absolute measurement. If the ratio is 1:1 or relative risk is 1.0, then it shows there is no association between disease and exposure to agents. If the ratio is more than 1.0 then disease occurs with unexposed agent. If the ratio is less than 1.0 then exposure requires some prevention. Other types of relative risk measurement are odds ratio, rate ratios, SMR (Standardized mortality ratio), SIR (Standardized incidence ratio), proportional mortality ratio and hazard ratio. Confidence interval helps in easy prediction of odd ratio and relative risk. They are measured in terms of variable data functions and sample size.

Attributable risk:

This risk measurement gives estimate about actual rate of exposure. It is usually expressed in terms of number or as percentage. Basic assumption of attribute risk is causal relationship. This indicates that risk can be avoided if exposure and or causative agent are removed from the population. It helps in better understanding of specific exposure risk.

Bias of epidemiology statistical measures

Suspicion of bias is due to inaccurate measurement analysis. Main objective of any case study shall be aimed to reduce the bias effect. Following are three types of bias

• Confounding bias: Random trials are performed to reduce extraneous factor in order to predict the exact relationship between exposure agent and disease occurrence.
• Misclassification or information bias: Inaccurate analysis or incomplete study report of exposure agent, disease treatment are referred as information bias.
• Selection bias: The study report from unrepresented population and are not exact prediction of exposure disease relationship analysis is referred as sample selection bias.      

Determining causality

Exact epidemiologic studies and measurements are possible by considering the following.

• Strength of the exposure disease association
• Temporal relationship between exposure and disease onset
• Biologic coherence
• Dose response gradient
• Consistency of results within and across studies 

Questions:

• How to evaluate the exposure disease relationship?
• Describe bias and its control in epidemiologic studies.
• What does a statistical measurement of epidemiology refer to?