Please enable JavaScript.

Coggle requires JavaScript to display documents.

Types of bias (Positive evidence of causality (Temporality (cause must…

- - - - Bas in case selection
        
        Admission bias
        
        Exposure associated with increased admission
        
        Survival bias
        
        Exposure associated with increased mortality
        
        Diagnostic bias
        
        Increased diagnosis rather than increased disease
      - Bias in control selection
        
        Non-response
        
        Different populations
    - - Same as case control but Diagnostic bias is not an issue
    - - Survival bias and different population not an issue
      - Non-response doesn't cause problems
      - Diagnostic bias
      - Loss to follow-up (attrition)
      - Verification bias
    - - Inadequate concealment of allocation *
      - attrition
      - withdrawals
      - loss to follow-up
      - competing risks
      - protocol violation
      - could happen 1-during randomization* 2-after randomization
- - - - recall bias
      - not an issue in prospective cohort and experimental
    - - Interviewer bias
      - all study designs affected
    - - Recall bias
      - interviewer/ascertainment/participation bias
    - - outcome verification
  - - - Dilutes effect; effect estimate closer to null
      - does not affect outcome
- - - - Analysis Phase strategies
        
        stratification
        
        not equal but on one side of the crude
        
        confonder and effect modification
        
        unequal and on either side of the crude
        
        effect modification
        
        all equal
        
        No confounder no effect modification
        
        equal to each other but not to crude
        
        confounder
        
        Adjustment
      - Design phase strategies
        
        Similiar cohorts
        
        cases equally represented
        
        Restriction
        
        only those with confounder
        
        Matching
        
        for every case with the confounder add a case without the confounder