Epidemiology: Exposure & Outcome

• Epidemiology applies scientific methods to study the relationships between exposure and outcome

Overview of the Scientific Method

• A three-stage process to investigate a research question.
• Step one is selecting subjects for study.
• Step two is finding associations between data and the sample.
• Step three is making conclusions about a scientific theory and causation.

Epidemiologic Study Designs

• Study designs encompass case reports, case series, analysis of secular trends, cross sectional, case–control studies, cohort studies, randomized clinical trials, non-randomized studies, systematic review and meta-analysis.

Case Reports

• Documents events observed in single patients.
• Describes a single patient's exposure and particular outcome.
• Useful for raising hypotheses about exposure effects to be tested with rigorous study designs.
• Cannot usually determine if the outcome was due to the exposure, or would have happened anyway.
• Doesn't always deal with serious threats, and may enliven the medical literature.
• Example: A published case report about a young woman on oral contraceptives who suffered a pulmonary embolism.

Case Series

• Aggregates individual cases into a single report
• The appearance of similar cases in a short period may indicate an epidemic.
• Features unique to the exposure cannot be known due to the lack of a control group.
• Not useful in determining causation, but provide clinical descriptions of a disease, or patients receiving an exposure.

Characteristics of Case Reports and Case Series

• Generally unplanned, and lack a research hypothesis and control groups.
• Advantages:
• Describes rare outcomes.
• Identifies emerging health conditions.
• Best to describe outcomes, not causality.
• Disadvantages:
• Are purely descriptive and lack a definite conclusion.
• May be misleading and suggest non-existent relationships.

Analysis of Secular Trends

• Examines exposure trends as presumed causes, and disease trends as presumed effects, testing if trends coincide.
• Can be examined over time, or across geographic boundaries.
• Data from a single region analyzed to see how trends change over time.
• Data from a single time period analyzed to compare how the data differs between regions or countries.
• Useful for rapidly providing evidence for, or against, a hypothesis.
• Lacks individual data, utilizing only aggregated group data.
• Unable to differentiate which factor is likely to be the true cause.
• Example: Sales data for oral contraceptives compared to venous thromboembolism, using recorded vital statistics.
• Example: Lung cancer mortality rates in the US have been increasing in women since the 1960s.

Cross-Sectional Studies

• Observational studies that examine only one point in time, so have no time sense.
• Exposure and outcome are assessed simultaneously.
• Also known as prevalence surveys.
• Provides a snapshot of the population.
• Each patient seen once.
• Researcher is not intervening in subjects.
• Example: Estimating diabetes prevalence in Jeddah, Saudi Arabia, by assessing diabetes status among subjects.
• Advantages:
• Are cheap fast, and simple.
• Determines prevalence.
• Disadvantages:
• Cannot determine causality.
• Not optimal for rare or short-term diseases.

Case-Control Studies

• Compares cases with a disease to controls without the disease, looking for differences in antecedent exposures.
• Example: Comparing young women with venous thromboembolism to controls without, looking for differences in oral contraceptive use.
• Useful when:
• Studying multiple possible causes of a single disease.
• Studying a relatively rare disease.
• Obtains information on exposures retrospectively.
• Information on past exposure to potential risk factors is generally obtained by:
• Abstracting medical records.
• Administering questionnaires.
• Conducting interviews.
• Limitations on the validity of retrospectively collected exposure information.
• Advantages:
• Inexpensive and require short time.
• Appropriate for rare diseases.
• Can study multiple exposures for a disease.
• Disadvantages:
• Study only one outcome.
• Prone to selection and information recall bias.
• Subject to confounding.

Cohort Studies

• Identifies subsets of a defined population; then follows them over time.
• Seeks to identify differences in outcome.
• Comparative, observational study where subjects are grouped by their exposure status and followed longitudinally.
• Example: Comparing women of reproductive age who use oral contraceptives to users of other methods, looking for differences in the frequency of venous thromboembolism.
• Can be performed:
• Prospectively: Simultaneously with the events under study.
• Retrospectively: After the outcomes under study had already occurred using medical records, questionnaires, or interviews.
• Advantages:
• Studies multiple outcomes for single exposure.
• Appropriate for rare exposure.
• Exposure precedes the disease.
• Has less selection bias.
• Disadvantages:
• Costly and time consuming.
• Possibility of loss to follow-up.

Types of Cohort Studies

• Prospective Cohort: Results will likely happen in future.
• Retrospective Cohort: Looking back, or dealing with past events.
• Ambidirectional Cohort: Looking back and continue to happen.

Randomized Clinical Trials

• Experimental study.
• Investigator controls the therapy each participant receives.
• Investigators randomly allocate patients between study groups - "treatment" and "comparison" groups.
• The non-exposed group can be untreated (placebo), or given a standard treatment.
• Example: Randomly allocating women to use, or not use, oral contraceptives and examining whether they differ in their incidence of subsequent venous thromboembolism.
• Can be impossible to perform ethically, and logistically.
• Associations are more likely to be causal associations.
• Considered the “gold standard” by which other designs must be judged.
• Advantages:
• Most accurate in establishing causality.
• Controls confounders, known and unknown, by randomization.
• Disadvantages:
• Possibility of loss of follow-up
• Ethical concerns
• Expensive, in terms of time and money

Randomized Clinical Trials vs Cohort Studies

• Participants selected based on their exposure status in both study types.
• The main difference between a cohort study, and a clinical trial is:
• Cohort studies are observational in nature; the researcher does not intervene in assigning the exposure.
• Clinical trial is experimental; the researcher intervenes and assigns the exposure to subjects.

Experimental Study Examples

• Randomized clinical trial to determine if giving magnesium sulfate to pregnant women in preterm labor decreases the risk of their babies developing cerebral palsy.
• Randomized community trial to determine if fluoridation of the public water supply decreases dental cavities.

Nonrandomized Interventional Study Designs

• Also known as Quasi-Experimental Designs.
• Used when it is not logistically feasible or ethical to conduct a randomized controlled trial.
• An experimental study where people are allocated to different interventions, using methods that are not random.
• Example:
• Violent games effects on children's behavior, when parents wants to choose.
• Effects of gender on aggressiveness, when it cannot assign.

Systematic Review and Meta-analysis

• Systematically collects and publishes all papers on a topic in a summary.
• A systematic review is a detailed, systematic, and transparent means of gathering, appraising, and synthesizing evidence to answer a well-defined question.
• A meta-analysis is a statistical procedure for combining numerical data from multiple separate studies.
• A meta-analysis should only ever be conducted in the context of a systematic review.
• Systematic reviews may not include a meta-analysis when:
• The review question is better answered with qualitative data.
• Studies are too different to combine.
• Examples:
• “How acceptable are psychosocial interventions for maltreated children?”
• “How do self-help group programs impact on women’s empowerment?”