Research Design Flashcards
If the factor (cause) is present, the effect (disease) will always occur.
Sufficient cause:
The factor (cause) must be present for the effect (disease) to occur; however, a necessary cause may be present without the disease occurring.
Necessary cause:
: If the factor is present, the probability that the effect
will occur is increased.
Risk factor
The factor exerts its effect in the absence of intermediary factors (intervening variables).
Directly causal association
The factor exerts its effect through intermediary factors.
Indirectly causal association
The relationship between two variables is statistically significant, but no causal relationship exists because the temporal relationship is incorrect (the presumed cause comes after, rather than before, the effect of interest) or because another factor is responsible for the presumed cause and the presumed effect.
Noncausal association
○ Process of answering question that can be answered by appropriate collected data
Research
○ Rules that govern the process of collecting and arranging data for analysis
Research Design
○ Creating educated ‘guesses’ about a phenomenon
Hypothesis generation
○ Making predictions from hypothesis and examining data to determine if predictions are correct.
Hypothesis testing
To describe the pattern of health problems accurately or enable a fair, unbiased comparison to be made between a group WITH and a group WITHOUT a risk factor, disease, or a preventive or therapeutic intervention
Research Design
○ If cause is present, disease will occur ○ Example: genetic abnormalities lead to some fatal diseases (HLA-B27 gene for Ankylosing Spondylitis)
Sufficient Cause
○ Cause must be present for disease to occur, although it does not always result in the disease ○ Example: MTB as prerequisite for tuberculosis (though some people can be asymptomatic carriers)
Necessary Cause
○ Exposure, behavior, or attribute that, if present and active, increases the probability of a disease to occur
Risk Factor
■ First and most basic requirement for a casual relationship to exist between outcome of interest and presumed case
Association
■ Difference must be large enough to be “unlikely” if the exposure really has no effect (event is not random)
Statistically Significant
○ Occurs when factor under consideration exerts effect without intermediary factors (e.g.blow to the head)
Direct Causal Association
○ When one factor influences or more other factors through intermediary variables (e.g.poverty)
Indirect Causal Association
○ If presumed cause occurs after the effect
Noncausal Association
○ Each of two variables may reciprocally influence the other.
Bidirectional Causation
: A differential error that produces findings consistently distorted in one direction as a result of nonrandom factors.
Bias
A nondifferential error that produces findings that are too high and too low in approximately equal frequency because of random factors.
Random error
: The confusion of two supposedly causal variables, so that part or all of the purported effect of one variable is actually caused by the other.
Confounding
: The interaction of two or more presumably causal variables, so that the total effect is greater than the sum of the individual effects.
Synergism
A phenomenon in which a third variable alters the direction or strength of association between two other variables.
Effect modification (interaction)
- first step in clinical trial involving assembling participants to be studies
Assembly bias
participants are allowed to select the study group to join
Selection bias
investigators assign participants to study groups in a non-random way
Allocation bias
- failure to detect a case or risk factor of disease (false-negative reactions)
Detection bias
wrong choice of measurement methods (false increase/decrease)
Measurement bias
- members of one group are more likely to remember than the other group
Recall bias
○ Investigator simply observes study participants ○ Assignment of treatment group vs control remains outside of investigator’s control
Observational Studies
Observational Studies can be classified to
Descriptive and Analytic
- no hypothesis specified, use existing date
Descriptive
- specified hypothesis, use of new data
Analytic
○ Investigator has control over participants or variables
Experimental Studies
○ Involves an investigation of clinical issues by using anthropologic techniques:
. Qualitative Studies
Qualitative studies: anthropologic techniques:
■ Ethnographic observation
■ Open-ended semi structured interviews
■ Focus group discussions (FGDs)
■ Key information interviews (KIIs)
Gaol: Provide rich, narrative information that tells story
Qualitative study
○ Survey of a population at a single point in time
○ Interviews at home, through telephone, mailed surveys, emails, or web-based questionnaires
Cross-Sectional Surveys
Advantages:
• Fairly quick and easy to perform
• Determines knowledge, attitudes, and health practices
Cross-Sectional Surveys
Disadvantages:
• Difficulty in determining cause and effect
Biased toward longer-lasting and mild disease cases
Cross-Sectional Surveys
CROSS-SECTIONAL SURVEY BIASES
Neynam bias (AKA late-look bias)
Health participant bias
Sever and rapidly fatal diseases are less likely to be found when doing a survey
Length bias in screening programs, which tend to find (and select for) less aggressive illnesses
Neynam bias (AKA late-look bias)
Not good in testing effectiveness of interventions like vaccination programs where people concerned about their health would less likely expose themselves to diseases, and not as direct result of intervention
Health participant bias
- collected soon after symptoms appear in the patient
Acute sera (1st sample)
- collected 10 to 28 days later when disease subsides.
Convalescent sera (2nd sample)
High IgG, No IgM =
infection occurred in distant past
High IgM, Low IgM =
current or very recent infection
High IgM, High IgG =
fairly recent infection
Relate frequency with characteristics and outcome of interest in the same geographic area
Cross-Sectional Ecological Studies
Unit of analysis: populations (not individuals)
Useful for suggesting hypothesis (associations)
Cannot be used to draw causal conclusions
Cross-Sectional Ecological Studies
■ Arriving at general conclusions based only on analyses of group data
Ecological Fallacy
○ Measures trend in disease rates over many years in a defined population
○ Use ongoing surveillance or frequent repeated cross-sectional survey data
○ Epidemiologists can determine the impact of these changes on disease rates.
Longitudinal Ecological Studies
○ Clearly identified group of people to be studied
Cohort Studies
Assembling / choosing people specifically or taking a random sample of a given population
Characteristic: groups are typically defined on the basis of exposure and are followed for outcomes.
Cohort Studies
The Present
Assemble cohorts in the present and collect data on present risk factors (present exposures)
The Future
Collect data at a time in the future on outcomes that arise
Prospective (longitudinal)
cohort study
The Past
When exposures occur defines cohorts
The Present
Assemble cohorts in the present based on past risk factors (past exposures), and collect data on present outcomes
Retrospective cohort study
The Past
When exposures occur that may be associated with outcomes
The Present
Assemble cases and controls in the present based on presentoutcomes, and collect data on past risk factors (past exposures)
Case-control study
The Present
Associations between presentexposures and present outcomes (both occurring at a single point in time)
Cross-sectional study
○ Investigator selects the case group and the control group on the basis of a defined outcome
Case-Control Studies
Compares the groups in terms of their frequency of past exposure to possible risk factors
Can estimate relative risk (odds ratio)
Useful when study needs to be performed quickly and inexpensively or when disease is rare (prevalence<1%) ; Major drawback: potential recall bias
Case-Control Studies
○ Cohort of participants is first defined
○ Baseline characteristics of participants are obtained by interview, physical examination, and pertinent laboratory or imaging studies
Nested Case-Control Studies
Participants who develop the condition = cases
Those who don’t develop the condition = control
Data from two groups are compared using appropriate analytical methods (patient-time risk)
Another variant: case-cohort study
Nested Case-Control Studies
○ Patients are enrolled in a study and randomly assigned to one of the following 2 groups:
■ intervention or treatment group
■ control group (given placebo or standard treatment)
Randomized Controlled Clinical Trials (RCCT/RCT)
Considered “gold standard” for studying interventions due to minimal bias in patient information obtained
Randomized Controlled Clinical Trials (RCCT/RCT)
– only participants are unaware
Single-blind study
– participants and investigators are unaware
Double-blind study
– participants, investigators, analysts are all unaware (most optimal)
Triple-blind study
- is an inert substance or treatment which is designed to have no therapeutic value.
Placebo
○ Intervention is usually preventive rather than therapeutic and conducted in the community
○ Participants randomly allocated to receive preventive measure or to receive placebo
○ Followed overtime to determine the rate of disease in each group
Randomized Controlled Field Trials
○ Sometimes referred to as “data fishing”
○ Danger of finding data that does not exist
Data dredging
○ Committee responsible for reviewing all proposed research and ensuring that it is ethical
Institutional Review Board (IRB)
