Workshop 4: biases

Overview

Teaching: 10 min
Exercises: 20 min

Questions

• What are the different forms of biases?

• How can we minimise bias?

Objectives

• Learn to spot different types of biases and deal with them

Biases in studies

• Selection biases
• Response Biases

Selection Bias

• Investigators have differently recruited the comparison groups
• Example: in a study on breast cancer and diet,
• cases recruited from cancer ward
• Controls recruited from healthy volunteers
• The comparison groups have to be as similar as possible
• Case control study:
• If the controls were to develop disease,
• They would be branded as cases
• How to ensure that you eliminate selection bias?
• Matching for case-control and cohort study
• Randomisation in Randomised controlled trials
• Blinding procedures in Randomised controlled trials

Response Bias

• Difference in response patterns between:
• Comparison groups
• Example: In a case control study of breast cancer:
• Cases were more accurate than controls:
• In terms of reporting their dietary patterns
• How should you eliminate response bias?
• Use objective measurements as much as you can
• If you use survey instrument, make sure
• It has been validated, and
• Establish its reliability statistics
• Train the interviewers
• Avoid self-administered instruments if possible

Biases should be addressed prior to study commencement

Challenge!

Challenge yourself

Read the following methods section of article and decide what biases may be present Mazumder, D. N. G., Haque, R., Ghosh, N., De, B. K., Santra, A., Chakraborty, D., & Smith, A. H. (1998). Arsenic levels in drinking water and the prevalence of skin lesions in West Bengal, India. International Journal of Epidemiology, 27(5), 871-877.

Read the file you downloaded File for review

Answer the questions:

Identify at least two biases in this research

Key Points

• Biases are systematic errors in study designs

• Biases should be eliminated at the stage of study designs

• Biases will drive the results of studies either towards the null or in unpredictable directions