Bias

The types of Selection Bias in Cohort and Case-Control:

1) Control-Selection Bias (occurs in case-controls) 2) Self-Selection Bias: causes differential participation 3) Differential surveillance, diagnosis, or referral (may happen in both cohort or case-control) 4) Healthy Worker Effect 5) Differential Loss to Follow Up (Cohort)

Selection Bias in Cohort Studies:

1) Healthy Worker Effect 2) Differential Loss-to-Follow up

Selection Bias in Case-Control Studies

1) Inappropriate control selection (control selection bias) (can bias towards or away from the null) 2)differential participation (can bias towards or away null) 3)differential surveillance, diagnosis, or referral (can bias towards or away from the null)

The types of information bias

1) Recall Bias 2) Interviewer Bias 3) Measurement Error

Solutions to avoiding Interviewer Bias:

1) use blinding/masking to prevent interviewers/people collecting info on outcome to study hypothesis or to knowing whether or not someone is exposed. 2) use high quality, standardized, closed-ended questionnaires 3) examine pre-existing data 4) Provide adequate and rigorous training for interviewers

Bias

A systematic error in design or conduct of study that leads to an incorrect estimate of association.

Control-Selection Bias Definition (a Selection Bias in a Case-Control Study)

Bias that can occur if controls are more (or less) likely to be selected if they are exposed (or unexposed). This bias occurs when controls fail to represent the exposure distribution in the source population from which the cases were identified, and they do not accurately represent the same source population as the cases.

Differential Surveillance, Diagnosis, or Referral Definition (a Selection Bias in a Case-Control Study)

Bias that can occur if patients who are exposed are more likely to be diagnosed with a disease. This bias occurs when the people who play a role in or are responsible for, disease ascertainment base their diagnosis on whether or not the participant has the exposure of interest. A sooner diagnosis equals an increased incidence rate, and if the cases have exposure they may be watched more closely or more often, which causes an earlier diagnosis.

Recall Bias definition

Bias that can occur if people with disease remember/report their exposure differently than people without the disease. May skew the OR in a case-control study.

Measurement (Misclassification) Error definition:

Bias that can occur if study participants are placed into the wrong exposure or disease category. ** The most common source of error

Differential Loss to Follow-Up (a Selection Bias in Cohort Studies)

Bias that can occur if study participants exit a study for reasons related to both exposure and disease. If people are lost to follow, we cannot know whether or not they've developed the disease! Can definitely skew the RR

Interviewer Bias definition

Bias that can occur if there is a systematic difference in soliciting, recording, or interpreting information.

Differential Participation/Self Selection Bias Definition (a Selection Bias in a Case-Control Study)

Bias that can occur if willingness or ability to participate is related to both exposure and disease status. Typically people who say yes to participation are more likely to have the exposure and the disease

Healthy Worker Effect (a Selection Bias in Cohort Studies)

Bias that may occur if the risk of disease is compared in person who are working (exposed) to persons in the general population (the comparison group).

The effect of differential misclassification error on the null

Biased upwards or downwards.

The effect of non-differential misclassification error on the null

Biases towards the null. If exposure is misclassified, then the extent of the exposure misclassification is the same for cases/controls / diseased/nondiseased. If disease is the thing being misclassified then the extent of the misclassification is the same for exposed and unexposed

Source of Bias

Can be caused by investigator or study participants during design or conduct of study. Can occur in experimental, cohort, and case-control studies. Not inherent characteristic of the study population. All studies have some bias.

Effects of Bias

Creates the appearance of an association when there is none, or mask an association that really exists. Selection bias and information bias cannot be fixed in the analysis; confounding bias can be fixed..to a point.

When does Selection Bias often happen?

In case-control or retrospective cohort studies. BECAUSE exposure and outcome have already occurred at time of subject selection into study.

The perils of Selection Bias (how to control for them)

Little/Nothing can be done to fix selection bias once it has occurred. Selection Bias must be avoided through careful study design and conduct. Selection bias cannot be controlled for in the analysis.

the Perils of Information Bias (and how to reduce them)

Little/nothing can be done once occurred. info bias must be avoided through careful study design and conduct info bias cannot be controlled for in the analysis.

Definition of Selection Bias

Mainly results from procedures used to select subjects into a study. In a Case-Control selection bias is when selection/participation of cases and controls is related to EXPOSURE status. In a Cohort study selection bias is when selection/participation of the exposed and unexposed subjects are related to disease status. Leads ultimately to an observed association that differs from what would have been obtained from population targeted for study.

Information Bias Definition

Occurs after subjects have entered a study. Results from systematic differences in the way information on exposure or disease is obtained from study participants. In a case-control study: different techniques are used to collect information from cases and controls. In a cohort study: different procedures are used to collect info from exposed and unexposed groups. Ultimately leads to participants being incorrectly classified as exposed or unexposed, or diseased or not diseased.

When the direction of bias is AWAY from the null

Positive association is biased away from the null. The association is overestimated. RR is bigger than it should be. Preventative association is biased away from the null value, making the true association overestimated. Overestimating the RR, making the benefit look better than it should be.

When the direction of bias is TOWARDS the null

Positive association is biased towards the null. The association is thus underestimated. This makes the RR look smaller than it actually is. A preventative association is biased towards the null, making the true association underestimated. Makes the benefit look smaller.

The Types of Bias

Selection Bias: how are they selected to be chosen in the study, and do they end up in the 2x2? Information Bias: Where do they end up in the 2x2? You can misclassify based on outcome/exposure etc.

When your denominator in your OR is smaller this means that...

You are biasing away from the null

When your denominator in your OR is big this means that...

You are biasing towards the null

Measurement Misclassification Error Solutions

improve accuracy of collected info.