HS300: Data interpretation

Stages in epidemiologic study

1. Observe possible association

2. Develop a hypothesis about a cause and effect relationship

3. Test the hypothesis through a formal epidemiologic study

4. And next…………?

Issues around data interpretation

1. Is the reported study scientifically valid? This relates to study design and methodology

2. Selective nature of presentation of results? Are you getting a balanced report (lay press) Positive results more likely to be published

Study validity

• The degree to which inferences drawn from a study are warranted

• Takes into account:

1. Study methods

2. Representativeness of study sample

Internal validity

• The extent to which differences in outcome between groups in a study can be attributed to the hypothesized effects of an exposure, an intervention, or other causal factor being investigated.

• Proper selection of study group (sample)

• Lack of error in measurement: random error, systematic bias, confounding have been ruled out

• Appropriate measurement of:
  1. Exposure
  2. Outcome
  3. Association between exposure & outcome

External validity

Ability to extrapolate results from a sample population to a target
population

Random error

Leads to a false association between exposure and disease that arises from “chance”, that seems to have no assignable cause”. Unlike bias and confounding, random errors are unsystematic, because they arise from an unpredictable, non-discernable process

Sampling error

• The possibility that the sample collected is not representative of the
  target population. Is relevant to all types of epi studies

• A non-representative sample may occur even in random sampling – bad luck

• How might one reduce sampling error?

• To Avoid: increase sample size

Poor precision

• Study factor is not measured sharply

• Example: Bogalusa Heart Study – prospective cohort study on CVD. Measure blood pressure (BP) in children. Six BP measurements by two observers,

• Average BP

• By taking the average of 6 readings, improved precision and reduced random measurement error

• To Avoid: Increase number of measurements

Systematic error

• This type of error is introduced by the investigative study process.

• 3 Types:
  1. Selection Bias: more likely to include some over others
  2. Information/Observation Bias
  3. Confounding
  A more serious problem to study validity

Selection and information bias

result from an error committed by the investigator in the design or conduct of the study that leads to a false association between exposure and disease

Confounding bias

stems from the natural mixing of effects between exposure and disease and a third variable called a confounder. Distorts the relationship between exposure and disease

Selection bias

• Relation between exposure and disease is different for those who participate in the study and those, who theoretically would be eligible for the study but, who do not participate.

• Relevant to both:

• Selection & retention (how many people stay) of study subjects in the study

Bias in epidemiology

• Not prejudice or inclination by investigator

• Rather → ignorance or poor decision

• Important because: It results in an incorrect or invalid estimate of
  the measure of association derived in the study results

Where is selection bias most likely?

Case control and retrospective cohort studies
because both exposure & outcome have
occurred by the time of study subject selection