Concept of Error:
In epidemiology: refers to a phenomenon in which the result or finding of the study does not reflect the truth of the fact.
Types of Error:
- Random (chance) Error – associated with precision
- Systematic Error/Bias – associated with selection
Common Sources of Error:
- Selection bias
- Absence or inadequacy of controls
- Unwarranted conclusion
- Ignoring the periods of exposure to risk
- Improper interpretation of associations
- Mixing of non-comparable records
- Error of measurement
Random error/ Chance variation
- Error that generally occurs in sampling procedure.
- It is a divergence, due to chance alone, of an observation on a sample from the true population value, leading to lack of precision in the measurement of an association.
Out of a sample of 100 people, 3 consecutive sample drawn randomly may contain:
- 0% diseased people
- 10% diseased people
- 70% diseased people
This is called random error where the error is due to chance.
The only way to reduce it is to increase the size of sample.
Elimination of error is not possible
Sources of random error:
- Individual biological variation
- Sampling error
- Measurement error
Types of Random Errors
- Type I Error – alpha error
- Type II Error – beta error
How to reduce Random Error?
Increase the size of the study.
Any process or attempts in any stage of the study from designing to its execution to the application of information from the study which produces results or conclusions that differ systematically from truth.
A. Selection Bias
- A distortion in true study finding due to improper selection procedures or it is due to an effect of selection process
- Most common type of bias.
Some potential sources of selection biases:
- Self selection bias
- Selection of control group
- Selection of sampling frame
- Loss to follow up
- Improper diagnostic criteria
- More intensive interview to desired subjects etc.
B. Information Bias
It is distortion in true study finding due to improper information/lack of information or misclassification.
Potential sources of Information Bias:
- Invalid instrument
- Incorrect diagnostic criteria
- Recall laps error
- Interviewing techniques
- Losses to follow up, attrition/experimental mortality, etc.
C. Confounding Bias
- Special type of Bias
- The term “confounding” – effect of extraneous variable that entirely or partially explains the apparent association between the study exposure and the disease.
- It is a bias that results when a study factor effect is mixed, in the data, with effects of extraneous variable or the third variables.
A variable is a confounder if:
- It is an independent risk factor (cause) of disease.
- It is unevenly distributed among the exposed and the non-exposed
- It is not on the causal pathway between exposure and the disease.
Methods of Controlling Confounding in Epidemiological Study
In two stages:
In designing stage
In analysis stage
Statistical modeling (multivariate)