[06.10] Quantitative Study Designs_ Observational Study Designs V2.pdf

Hill’s Criteria for Causation

What are the basic epidemiological principles addressed first, before discussing confounding variables?

Temporal relation, plausibility or biologic plausibility, consistency, strength of association, dose-response relationship, reversibility, study design, judging the evidence

What are the eight criteria for causation listed in the sources?

Descriptive studies

What type of epidemiological studies focus on Populations?

Ecological, case reports, case series, cross sectional

What four study types are listed under Descriptive studies focused on Individuals or Populations?

Analytical studies

What type of epidemiological studies include Observational and Experimental designs?

Cross sectional, case control, cohort

What three specific study designs are listed under Analytical/Observational studies?

Randomized control trials

What specific study design is listed under Analytical/Experimental studies?

Temporal relation

Which of Hill's criteria states that the hypothesized cause must precede the occurrence of the disease?

Lack of temporal ambiguity

What is the criterion related to ensuring the cause precedes the effect?

The stronger the observed association, the less likely it is due to bias

What is the relationship between the strength of association and bias?

Dose-response relationship

What criterion is supported if the disease frequency increases with the dose or level of exposure?

The absence of a dose-response effect may not rule out causation from alternative explanations (i.e., Threshold effect)

What caveat is given regarding the absence of a dose-response effect?

Consistency

If all studies dealing with a given relationship produce similar results, what supports a causal interpretation?

Biological plausibility

If the hypothesized effect makes sense in the context of current biological knowledge, what supports a causal interpretation?

Current state of biological knowledge may be inadequate

Under what condition might biological plausibility not determine plausibility, requiring consideration of other Hill's criteria?

Coherence of evidence

What criterion is supported if the findings do not seriously conflict with our understanding of the natural history of the disease or other accepted facts about its occurrence?

Knowledge of pathological diseases and physiology

What areas of knowledge come into play when considering coherence of evidence?

Specificity of the association

What criterion is supported if the study factor is found to be associated with only one disease, or the disease is found to be associated with only one factor?

Cannot rule out causal hypothesis, since many factors have multiple effects and most diseases have multiple causes

What is the limitation of the specificity criterion?

Experimentation

What criterion is supported by the use of experimental evidence such as clinical trials in humans, animal models, and in vitro laboratory experiments?

Analogy

What criterion is used when similar relationships have been shown with other exposure-disease relationships?

Confounding

What is a distortion in the estimated exposure effect that results from differences in risk between the exposed and the unexposed that are not due to the actual exposure?

Review previous studies

What is the best way to know if a variable is a possible confounder?

It must be a risk factor for the disease, It must be associated with the exposure variable in the source population, It must not be affected by the disease (i.e., not an intermediate variable/step in the causal path)

What three conditions must be met for a variable to be screened as a confounder?

Alcohol consumption

In Case #8 regarding smoking and lung cancer, what was found to be a confounder because people who smoked were also alcohol drinkers?

Formation of the plaque

In the hypertension example (Case #9), what is the intermediate step in the causal pathway between high-fat diet and hypertension, meaning it is not considered a confounder?

Restriction, stratification, matching, multivariable regression analysis

What four statistical techniques are listed for controlling a confounder during data collection or analysis?

Limit the participation of respondents to those who are similar with respect to the confounding variables

How does restriction control a confounder?

Study of the association between exposure and outcome within different strata of the confounding variables

How does stratification control a confounder?

Forming matched sets of two groups of subjects who share a similar value of the propensity score

How does matching control a confounder?

Builds a multivariable regression model for the outcome and exposure as well as other confounding variables, allowing the variable of interest to be examined while others are held constant statistically

How does multivariable regression analysis control a confounder?

Can it be due to selection or measurement bias, Can it be due to confounding, Could it be a result of chance, Could it be causal

What four sequential questions should be asked to determine an association in an observed association?

Causal inference

What is the term for the process of determining whether observed associations are likely to be causal?

Independent variables

What type of variables can be called Factor, Exposure, Cause, Intervention, or Treatment?

Dependent variables

What type of variables can be called Outcome (of a disease or a risk factor), Effect, or Response?

Case Reports, Case Series, Cross Sectional Surveys, Case Control Studies, Cohort (Follow up) Studies, Randomized Controlled Trials

What is the hierarchy of epidemiological design strategies from lowest to highest?

Cohort (Follow up) Studies and Randomized Controlled Trials

Which two study designs at the bottom of the hierarchy can assert themselves with causality?

Populations

What are Descriptive studies concerned with, according to Figure 1?

Individuals

What are Ecological studies concerned with, according to Figure 1?

Individuals

What are Case reports, Case series, and Cross-sectional studies concerned with, according to Figure 1?

Analytic studies

What category of studies is the cross-sectional design considered part of when it is observational?

Limited to a certain point in time, with no follow up period

What characterizes the observation period of a cross-sectional study?

Outcome and exposure variables are measured at the same time

What key characteristic makes it relatively difficult to establish causal relationships in a cross-sectional study?

Prevalence of health outcomes

What specific measure of frequency can a cross-sectional study find out?

Incidence (measure of new cases)

What measure is a cross-sectional study unable to measure?

Prevalence of one or multiple health outcomes in a specified population

What is the main goal of a Descriptive Cross-sectional Study?

Comparing outcome differences between exposed and unexposed subjects

What is the purpose of an Analytical Cross-sectional Study?

Relatively quick and inexpensive to conduct, No (or rare) ethical difficulties, Multiple outcomes and exposures can be studied, Easy to generate hypotheses

What are four strengths or advantages of cross-sectional studies?

Data on all variables are only collected at one time point, Difficult to make a causal inference, Unable to investigate the temporal relation between outcomes and risk factors, Not good for studying rare diseases, Susceptible to biases such as nonresponse bias and recall bias

What are five weaknesses or disadvantages of cross-sectional studies?

Cross-sectional study

What study design is not good for studying rare diseases and has no follow-up period?

Longitudinal period

What type of period defines cohort studies due to the follow-up of participants?

Prospective or retrospective

What are the two types of cohort studies?

Population that does not have a disease

What is the starting point for participants in a cohort study design?

Prospective cohort studies

What type of cohort study requires researchers to look forward/ahead?

Retrospective cohort studies

What type of cohort study requires researchers to look back, often involving a review of records?

Fixed cohort and Dynamic cohort

What are the two types of prospective cohort studies?

Predetermined participants that cannot be added to

What defines a fixed cohort?

Subjects are enrolled and followed up separately, with different time zeroes

What defines a dynamic cohort?

Identifying the temporal relationship between cause and effect (the exposure and outcome of the disease)

What is the instrumental function of prospective cohort studies in generating valid scientific evidence?

Respondents already have the outcome of the disease

In a retrospective cohort study, what is true about the respondents at the start?

Find out if the outcome of interest is due to the exposure (cause and effect)

What is the goal of a retrospective study?

Can investigate multiple outcomes associated with multiple exposures, Able to study change in exposure/outcome over time, Good for examining rare exposures, Can measure the incidence of outcome, May be able to infer causality

What are five advantages of cohort studies?

Control design, sampling, data collection, and follow-up methods, Can measure all variables of interest

What are two specific advantages of prospective cohort studies?

Expensive to conduct and time-consuming

What is the main disadvantage of prospective cohort studies?

Time-efficient and inexpensive, Easy to obtain large sample size

What are two advantages of retrospective cohort studies?

Less control over study variables, Susceptible to information bias and recall bias, Subject to how study records were kept

What are three disadvantages of retrospective cohort studies?

Relative Risk (RR)

What is the measure of association used in cohort studies?

Incidence among exposed / Incidence among unexposed

What is the formula for Relative Risk?

Risk Ratio (RR) calculated from cumulative incidence OR Rate Ratio (RR) calculated from incidence rate

What are the two ways RR can be calculated in cohort studies?

0 to ∞ (infinity)

What is the range of possible values for Relative Risk?

1

What is the null value for Relative Risk?

No association between exposure and disease, Incidence rates are identical between groups

What does an RR = 1 indicate?

Positive association, Exposure is considered a risk factor for the disease

What does an RR > 1 indicate?

Negative association, Exposure is a protective factor for the disease

What does an RR < 1 indicate?

Relative risk, Risk ratio, Rate ratio, Risk reduction

What are the four meanings RR could stand for?

Odds ratio and Hazard ratio

What two other terms are sometimes called relative risk?

A smoker’s risk factor for getting lung cancer is 2.5 times higher than the risk of a non-smoker

What is the interpretation of RR = 2.5 in a study of smoking and lung cancer?

A physically active person is half as likely to experience cardiovascular death than a physically inactive person

What is the interpretation of RR = 0.5 in a study of physical activity and cardiovascular death?

Long-term, Large sample size, Attrition (dropout rate), Expensive, Need a good information management system (IMS)

What are five challenges listed in cohort studies?

20%

What is the standard buffer rate for attrition when computing sample size for cohort studies?

Disease outbreak investigations, Rare diseases, Rare outcomes

What three scenarios are case-control studies best known to be used for?

Case group and control group

What two groups, defined by the presence or absence of the outcome of interest, are identified in a case-control study?

Investigate associations between a disease and one (or more) hypothesised risk factors

What is the aim of case-control studies?

Representative of patients who develop the disease

What characteristic must cases have in a case-control study?

Defining the study question, Selecting the cases, Selecting appropriate controls, Measuring exposure, Analyzing the data, Interpreting the results, and assessing potential sources of error

What are the six main steps in a case-control study?

Controls will need to be drawn from the same population

When defining the population from which cases are drawn, what requirement applies to the controls?

Hospital-based or population-based

What are the two types of study settings to determine when defining the case population?

Incident or Prevalent

What characteristic of cases must be examined?

Representativeness of cases, Cases are independent of exposure status

What two things should be considered when making a case definition?

Controls must be representative of the population that produced the cases

What is the key point regarding the selection of controls?

Not more than 4 controls per case

What is the rough rule for the number of controls per case?

Matching

What technique is used to select control subjects in a manner that matches some characteristics (potential confounders) to those of the individuals in the case group?

Age, race, gender, sex

What are four examples of common matching variables?

Odds Ratio (OR)

What measure is used in case-control studies to compare the odds of exposure between cases and controls?

Odds among cases / Odds among controls

What is the formula for the Odds Ratio?

The odds of being exposed to smoking are 2 times higher for lung cancer cases compared to controls

What is the crude interpretation of OR = 2 in a smoking and lung cancer study?

Substitute for a cohort study in assessing the risk

How is a case-control study generally regarded?

The odds ratio is interpreted as an estimate for the relative risk

Under what condition can the OR be interpreted as an estimate for the RR?

The outcome (disease) is rare OR The RR is close to 1

What two conditions allow the Odds Ratio to provide a good estimate of the Risk Ratio?

0 to ∞

What is the range of possible values for the Odds Ratio?