Epidemiology week 1: study design

case series

cross-sectional

case control

cohort

RCT classic design

case-crossover RCT

trials w external controls

which two study designs take from past, present, and future?

case study and case series

which study design take from past and present?

cross-sectional

which study design is retrospective?

case-control

which designs allow for future application/understanding?

prospective cohort and RCT

what are the three descriptive studies?

• case study/report

• case-series

• cross-sectional

what are the three analytical studies?

• cross-sectional

• case-control (retrospective)

• cohort (prospective/longitudinal)

what are the two experimental/interventional studies?

• controlled trials

• trials w no controls or external contorls

describe the eight rankings for hierarchy of evidence best to worst

1. systematic review & meta-analyses

2. RCT

3. cohort

4. case-control

5. cross-sectional

6. case repor/series

7. expert opinion

8. anecdotal

definition:

• Characteristic of the individual (age, gender, weight)

• variable related to environment (air pollution)

• lifestyle (smoking, exercise)

• social status (poverty or income level)

• participant’s background (education level, race/ethnicity)

exposure aka risk factor

definition: result of effect of interest

outcome

T or F: there is only 1 exposure and 1 outcome per study

true

which descriptive study?

• no control used

• precursor to future studies (hypothesis generating)

• characteristics of pt recorded and published

• short duration

case study

which descriptive study:

• no control used

• precursor to future studies (hypothesis generating)

• series of pt recorded and published o observe trends or patterns for same condition or Tx

• short duration

case series

advantages:

• useful for generating hypothesis for larger studies

disadvantages:

• cannot evaluate association generalizability is an issue

case report and case series

definition: descriptive studies population at one point in time

descriptive cross-sectional

which descriptive study:

• referred to as prevalence study

• examine prevalence of disease or condition

• short or long duration

• precursor to cohort and case control studies

cross-sectional study

surveys, epidemiologic studies, incidence or mortality studies are example of

cross-sectional studies

which descriptive study is precursor to cohort and case-control studies?

cross-sectional studies

advantages:

• less time consuming

• large samples

• can be done w limited resources

disadvantages:

• Cannot evaluate causal associations as we cannot

  determine direction of association

• Participation bias

• Interview or reporting bias

• slice of time = may not be representative of the full picture

descriptive cross-sectional

definition: studies that test the research hypothesis about exposure-outcome relationship

analytical studies

studies that can measure the association and magnitude of association btwn exposure and outcome

analytical studies

definition: association btwn exposure and outcome at one point in time is evaluated, magnitude of relationship estimated but not direction

analytical cross-sectional

in descriptive or analytical cross-sectional can research hypothesis/research question can be evaluated statistically?

analytical

in descriptive or analytical cross-sectional can odds ratios for relationship btwn exposure and outcome can be estimated?

analytical

advantages:

• commonly used to describe burden of disease, disease trends, behavioral patterns

• correlation btwn variables can be determined and (blank) relationships can be analyzed

disadvantages:

• requires large sample size

• cannot establish causal association

• cannot establish time sequence or temporality

cross-sectional analytical

which analytical study:

• compares diseased indiv to non-diseased in relation to exposure status

• retrospective

• analyze factors/exposures related to outcome in cases and controls

case-control

which analytical study:

• the disease or outcome has already occurred

• cases and controls are grouped based on the outcome/diseased status

case-control

define odds ratio:

• they can be used to evaluate the association btwn exposure-outcome

• estimated based on prob of disease among the exposed indiv relative to prob of diseased among unexposed indiv

advantages:

• efficient, short, less expensive

• convenient for studying many exposures

• efficient for rare diseases and for diseases w long latency

• measures of association can be calculated to evaluate exposure-outcome relationship (odds ratio)

disadvantages:

• retrospective nature makes it difficult to establish time sequence of exposure-outcome

• retrospective nature leads to potential biases

case-control

what are the four relevant biases that can affect case-control study outcomes?

• review bias

• interview bias

• reporting bias

• misclassification bias

which bias:

• cases remember/recall & report their previous exposure experience diff from controls

• can occur in cross-sectional

• most common type of bias in case-control

recall

case-control is most prone to which bias?

recall bias

which bias:

• differences that occur in recording or interpreting info from participants

  • subject status may lead to increased questioning

• present in cross-sectional, case-control, cohort

interview

which bias:

• under-reporting of socially undesirable

  • subjects may refuse to answer

• over-reporting of exposures, etc

reporting

which bias:

• participants are incorrectly assigned/categorized into wrong exposure or outcome

• inaccurate case definitions or inclusion/exclusion criteria

• lead to diluted or exaggerated effects

• most common in case-control studies

misclassification

which bias is the most common in case-control studies

misclassification

can temporality be established in cross-sectional studies?

NO!

can temporality be established in case-control studies?

no but technically YES! if info about exact time period of exposure and date of diagnosis is available

which analytical study design?

• prospective or longitudinal studies

• identify exposure right off

• compare disease rates exposed vs non-exposed

• can calculate incidence (proportions) and incidence rates

cohort study

can time sequence of events be established for cohort studies?

yes temporality and causality can be determined

advantages:

• can be more ethical

• good for rare exposure

• can look at multiple outcomes for single exposure

• can estimate incidence rates over time as we have info

• can be used to establish temporality and causality

disadvantages:

• not efficient for rare disease

• long duration = t and money

• exposure status may change overtime

• prone to loss of follow-up bias

cohort

cohort or case control:

• participants grouped by disease status

• less t and money

• can be used for rare disease

• multiple exposures can be studied

• does not work well for rare exposures

• prone to recall bias/info

case-control

cohort or case control:

• participants grouped by exposure status

• long duration = ↑t and money

• can be used for rare exposure

• may not work for rare disease

• multiple outcomes evaluated

• prone to loss to follow-up bias due to dropouts

cohort

advantages:

• efficient way to summarize

• results can be broadly generalized

• most reliable bc summary of info

• highest quality of info

disadvantages:

• laborious

• time consuming

• may be difficult to combine info when methods and measures vary across studies

systematic reviews

which study:

• statistical combo of results from two or more studies

• quantitative formal epidemiological study design used to systematically assess previous research studies to derive conclusions about that body of research

• summarized data is used to make interpretations or causal inferences

meta analysis

advantages:

• most ideal, efficient, highest quality

• most most ideal is quantitative combo or RCT

• highest statistical power

disadvantages:

• (blank) is not possible in all systematic reviews

• for new research few studies available

• diffs in reporting, indiv bias, variability in research methods and/or clinical approach

• requires advanced stat methods and software

meta-analysis

what category of studies:

• longitudinal or prospective studies that examine the effect of an intervention on outcome of interest, investigator assigns and intervenes (aka intervention studies)

experimental studies

what are the two types of experimental studies?

• controlled vs trials w no controls

• controlled trials: random vs non-randomized

  • parallel, cross-over, split-mouth

what are the three types of controlled trials

• parallel

• cross-over

• split-mouth

what is the gold standard of experimental studies?

RCT

what are the three types of randomization?

1. simple

2. stratified (categories)

3. block (restricted randomization to ensure equal number in each group)

these are advantages of what?

• covariates are equally distributed

• minimizes bias in treatment allocation

• minimizes bias in outcome assessment

randomization in RCT

what is blinding and what are the three types?

concealing info about Tx assignment from key participants in the trial

• single: patients only

• double: patients and investigation

• triple: patients, investigation, data analysis

what do placebos contain?

sugar or non-essential ingredients

advantages:

• ability to make causal inferences

• provides strongest empirical evidence of a Tx efficacy

• tailor-made to answer a specific research

• randomization of participations to the test and control arms minimize bias and confounding of unknown variables

disadvantages:

• expensive and complicated

• dropout rates may be high if intervention unpleasant\

• interventions must be beneficial

• prior knowledge required about level of improvement

• large sample size required

experimental studies

which RCT?

• comparisons btwn two intervention groups

• participants remain in same group during the study

parallel

which RCT?

• each subject can serve as his/her own control

• same subject will receive exposure/intervention

cross over

which RCT?

• participants receive intervention in one quadrant of mouth and comparison intervention in another quadrant

split-mouth

which RCT:

• Study subjects meeting inclusion criteria are enrolled

• Treatment versus placebo randomly assigned

• Switch the groups to receive treatment versus placebo

• Groups who received treatment will switch to placebo group and vice versa (cases act as their own control so no need for separate control group)

case-crossover

definition: To remove bias from contamination of effect of one treatment on the other it is important to wait for the effects of the first treatment to disappear

washout period

what is it called when effects linger even after Tx termination?

carryover effect

what aids in minimizing carryover effect?

washout period

which RCT?

• Common in dental clinical research

• RCT trials where interventions are randomly allocated to different areas in the oral cavity

• Experimental and control interventions can be allocated by teeth, surfaces of teeth, arches an / quadrants

• Higher statistical power

• Each patient serves as his/her own control

• Difficulty in obtaining patients with symmetrical disorders

split-mouth

explain process of factorial RCT

which other types of trials:

• Cases selected from hospital or clinic setting

• Controls from a previous study or existing large cohort study

• Information readily available

• Less expensive

• Saves time and resources

trials with external controls

which other type of trial:

• Goal is usually to study primary prevention of a disease

• Example: Experiments of new vaccines

  (Salk vaccine trial 1954)

field trials

which other type of trials:

• Exposure is assigned to groups pf people in the community

• Here exposure could be comparing regions with fluoride in water to regions without fluoride in water

community intervention trials