HMB342 Lecture 4 Observational Studies

What is epidemiology?

The study of the determinants and distribution of disease in populations

What is systematic error also referred to as and what can it cause if it is not properly accounted for?

also referred to as bias, can cause inaccurate results by diluting a true association, generating a false association, or distorting the direction of the association

What are some types of selection bias?

sampling bias, non response bias, survival bias, attrition bias

What are some types of information bias?

misclassification bias, recall/response bias, social desireability bias, interviewer bias

Describe case reports/series

a type of pre-study design thats goal is to document a comprehensive and detailed description of cases under observation, the goal is to generate hypothesis to be tested in further experiments

What are some of the cons of case reports

no comparison group, limited generalizability, prone to physician interpretation

Describe ecological studies

the first level of observational studies, rely on aggregate data, used to generate hypotheses

What is a con of ecological studies?

prone to biases since they rely on aggregate data

What is the ecological fallacy?

error that arises when we ascribe characteristics of a group to individuals that do not possess those characteristics; so we cannot conclude if an association between an exposure and an outcome is causal, only assess correlation

Describe cross-sectional studies

studies that are used to generate hypotheses that can involve interviews surveys/questionnaires or using health administrative data, they serve as a preliminary analysis prior to an in-depth research study and the exposures and outcomes are determined simultaneously for each participant

What are some cons of cross-sectional studies?

1) difficult to determine a time sequence for events

2) are susceptible to selection bias (sampling bias, survival bias, no response bias)

3) are susceptible to information bias (recall bias, interviewer bias)

What are some benefits of cross-sectional studies?

estimate the prevalence of risk factors, diseases and health related outcomes, can measure other outcomes like health status and health service utilization, can help determine the extent of health problems

What is the Canadian community health survey?

it is an ongoing nation wide cross-sectional study that incorporates direct self reporting through electronic questionnaire

Describe case-control studies

selected based on the outcome (cases have the outcome in the present and have had exposures in the past), exposure status can be obtained via interviews or health administrative data

When are case-control studies an efficient study design?

efficient when you are studying more than one exposure/cause/factors, studying rare diseases, or chronic diseases

What are some factors that have to be considered when conducting a case-control study?

source of the cases, incident versus prevalent cases, source of the controls, capturing the exposure in the control group

If exposure is associated with the outcome, then a/ total number of cases ____ b/total number of cases

a/ total number of cases > b/total number of cases

Describe cohort studies

type of study used to calculate incidence, cohort includes a group with a shared characteristic, effect is not present at the beginning of the study

What are the strengths of cohort studies?

provide clarity on temporality, can be used to study multiple outcomes, can be used to study rare exposures and recurrent diseases

What are some limitations of cohort studies?

expensive, prone to non-response bias and misclassification bias, long time requirement, drop outs, difficult to study rare outcomes

How do you determine the unit of analysis in analytical study designs?

groups or individuals?

How do you determine the directionality of a study?

forwards directionality = exposure -> outcome

backwards directionality = outcome -> exposure

simultaneous directionality

How do you determine temporality of an analytical study?

ask yourself if the exposure precedes the outcome

How do you determine the timing of an analytical study?

retrospective (past, present to past)

present

prospective (present to future)

What is sampling bias?

type of selection bias where some members of the population are systematically more likely to be selected/sampled than others

What is non response bias?

selection bias when participants are unable or unwilling to participate

What is survival bias?

selection bias that results when there is a focus on the surviving entities and ignore the ones that are not surviving

What is attrition bias?

selection bias that occurs when there are characteristics that differ significantly between participants lost to follow up and those who remain in the study

What is misclassification bias?

information bias that occurs when a participant is assigned to a different exposure or disease status than they should be assigned

What is recall bias?

information bias that occurs when there is a significant difference in the accuracy of recalling information (ie. in self reporting)

What is social desirability bias?

information bias that occurs when participants respond/provide information in a way that makes them look favorable to others rather than most accurately

What is interviewer bias?

information bias that occurs when the interviewer consciously or unconsciously influences the participant's response

What is misclassification bias?

information bias that occurs when a participant