Epib Midterm 2

Experimental Study

Manipulation of variables, assigning participants to exposure group

Experimental Studies

Quasi-Experimental and Randomized Control Trial

Observational Study

Investigator watches and there is no manipulation or exposure

Observational Studies

Cohort, Case Control, Cross Sectional

Steps in prospective cohort study

1. Start research

2. Data collection on risk factors and outcomes

3. Risk factors leading to outcomes

Steps in retrospective cohort study

1. Development of outcomes

2. Existing data

3. Compare groups based on risk factor exposure

RR = 1

Association between exposure and disease unlikely to exist

RR > 1

Increased risk of disease among those who have been exposed

RR < 1

Decreased risk of disease among those who have been exposed

Strengths of prospective cohort studies

• Better exposure and cofounder data

• Less vulnerable to bias

Weaknesses of prospective cohort studies

• More expensive

• Time consuming

• Not efficient for diseases with long latency

Strengths of retrospective cohort studies

• Cheaper

• Faster

• Efficient with diseases with long latency periods

Weaknesses of retrospective cohort studies

• Inadequate exposure/cofounder bias

• More vulnerable to bias

Case control study

• Estimate the magnitude of an association between an exposure and an example

• Subjects are defined based off the absence or presence of a disease or outcome of interest

Case

Have outcome/disease

Control

Do not have the outcome/disease

Risk Ratio

The chance of something happening / the chances of all things happening

Odds Ratio

The chance of something happening / the chance of it not happening

Case control strengths

• Efficient for rare diseases and diseases with long induction and latency

• Can evaluate multiple risk factors for the same disease, useful for poorly understood diseases

• Typically inexpensive because small number of subjects

Case control weaknesses

• Inefficient for rare exposures

• Vulnerable to bias because of retrospective nature

• Can only study one outcome

Ecological Fallacy

• Drawing conclusions about individuals based on data collected at a group level, not at the population level

• Cannot make conclusions about an individual based on group data

Cross-Sectional Studies

• Ecological investigation that looks at the relationship between diseases and variables of interest in a defined population at a specific moment in time

• Prevalence study

• Participants are selected based off particular variables of interest

Cross-Sectional Study Strengths

• Measure prevalence

• Single point in time so investigators do not need to follow-up with individuals over time

• Useful for establishing preliminary evidence for future planning of studies

• Population based, so helps investigators study determinants of health

Cross-Sectional Study Weaknesses

• Cannot follow-up or establish a causal relationship (time measurement)

• Prone to bias (ecological fallacy)

• Selection bias

Hawthorne Effect

“Observer Effect” - individuals may change their behavior because they know they are being watched

Recall Bias

Individuals are unable to remember information from the past, may be different between case and controls

Interviewer Bias

Systematic differences in soliciting, recording, or interpreting information to participants

Healthy Worker Bias

Those who are employed tend to have lower morality rates compared to the general population

Loss to Follow-Up Bias

Bias results from people with a shared characteristic leaving the study for some reason

Placebo Effect

Participants placed in this group improve because they are told that they will/they think that they will

Systematic Error

Continuing error or when collecting data that is consistently inaccurate so it creates an inaccurate association between an exposure of interest and a disease

Random Error

• Non-systematic error that is not consistent, occurs due to chance and cannot be changed

• Causes: poor precision, variability in measurement, sampling error

Randomization

Each participant has the same probability of being put into the treatment group

Goals of Randomization

• Eliminates bias in treatment assignment

• Provides a valid basis for statistical inference

• Permits the use of probability theory to express likelihood that differences in outcome between groups are due to change

• Treatment is the only difference between study groups

Random Sampling/Selection

Method of choosing participants to be in the overall study

Cofounding

• Associated with the exposure

• Associated with the outcome

• Not a direct consequence of an action or event on causal pathway

1. Strength of Association

Strong relationship between variables

2. Temporality

Logically necessary for a cause to precede and effect in time

3. Consistency

Multiple observations that support the association

4. Theoretical Plausability

Easier to accept an association as causal when there is a rational and theoretical basis for the conclusion

5. Coherence

Associations must be coherent with other knowledge

6. Specificity

Ideally, the effect has only one cause, showing that an outcome is best predicted by one primary factor

7. Dose-Response Relationship

Direct relationship between the risk factor and people’s status on disease variable

8. Experimental Evidence

Exposure and outcome have been associated through well-controlled randomized trials (applicability may be limited by ethical and feasible randomization)

9. Analogy

The proposed association mirror another association already established in the literature

Prospective Cohort Study

Measure association between an exposure and risk of disease in the population, moving forward in time

Retrospective Cohort Study

Measure association between an exposure and risk of disease in the population, moving backward in time

Blinding/Masking

• Being unaware of treatment group assigned

• Single-blind design: subject not aware of group assignment

• Double-blind design: neither subject nor researcher aware of group assignment

Compliance

Following the study protocol exactly as required through the course of the trial

Matching

Selecting pairs or clusters of individuals who are comparable on important variables (sex, age, medical history)

Randomized Controlled Trials

Investigator randomizes participants to a treatment or control condition to generate evidence for a causal relationship between an exposure and disease

Randomized Controlled Trial Strengths

• Considered the gold standard of study designs due to high quality of evidence

• High internal validity

• Treatment and outcome status is clear

• Ideal study design if assigning treatment is feasible and ethical

Randomized Controlled Trial Weaknesses

• Lack of generalizability/external validity

• Experimental setting does not resemble the real world

• Expensive

• Logistically challenging

Selection bias

• Healthy worker effect

  Survivors bias

  Self-selection bias

  Non-response bias

  Loss to follow-up bias

  
  

Information Bias

• Hawthorne effect / surveillance bias

  Recall bias

  Interviewer bias

  Measurement error

  Reporting bias