Systematic Review and Meta-Analysis Techniques

Systematic Review

• Larger sample size than individual studies allows for subgroup analysis.
• Results of systematic reviews are more accurate than individual studies due to larger sample size.
• Meta-analysis is a specific type of systematic review that uses quantitative assessment of outcome parameters.

Meta-Analysis

• Not all systematic reviews include a meta-analysis.
• Requires studies to include (approximately) the same intervention, patient population, and outcome measures.
• Studies must meet acceptable methodological quality standards.
• Three statistical methods for meta-analysis: Inverse Variance-weighted method, Mantel-Haenszel, and Peto.

Inverse Variance-weighted Method (IV)

• Suitable for all outcome measures (OR, RR, RD, MD, etc.).
• Applicable for both binary and continuous data.

Mantel-Haenszel (MH)

• Used when data is sparse (low event rates and small trials).
• Only suitable for binary outcomes.

Peto

• Suitable for odds ratios and when the event of interest is rare.
• Not suitable for large treatment effects

Fixed Effects Model

• Assumes a fixed treatment effect (same value in each study).
• Differences in results are attributed to chance.
• Assumes studies estimate the same effect size.
• Requires testing the fixed effects assumption with a test of homogeneity.

Random Effects Model

• Assumes effect sizes may differ.
• Differences in results are not purely due to chance.
• Incorporates more uncertainty.
• DerSimonian and Laird method is used to incorporate heterogeneity.
• Confidence intervals in random effects models are wider than fixed effects models.

Relative Risk (RR)

• Two types: risk ratio and rate ratio. = 1: No increased risk, no association.

• 1: Increased risk, positive association.

• < 1: Decreased risk, negative association.
• Reported along with a p-value.
• Cannot be performed with case-control studies.

Odds Ratio (OR)

• Used in both cohort and case-control studies. = 1: Exposure not associated with disease.

• 1: Exposure positively associated with disease.

• <1: Exposure negatively associated with disease.
• Rate Ratios and Risk Ratios are similar for rare diseases.

Bias

• Systematic error that leads to incorrect conclusions about the true effect of an exposure on an outcome.
• Undermines the internal validity of research.
• Selection bias: Results from differences in study groups in important aspects (sampling, attrition, volunteer).
• Information Bias: Misclassification or inaccurate reporting of information (reporting, recall, observer).

Confounding Variable

• Real but misleading associations.
• Related to both the cause and effect of the study.
• Stratification can be used to reduce the impact of confounding variables.

Internal and External Validity

• Internal validity: Whether the conclusions about the cause and effect relationship in a study are valid.
• External Validity: The extent to which study results can be generalized to another context.

Sensitivity and Specificity

• Measures of the accuracy of a diagnostic test.
• Sensitivity: Ability of a test to correctly identify individuals with the condition.
• Specificity: Ability of a test to correctly identify individuals without the condition.
• Positive Predictive Value (PPV): Likelihood that someone with a positive test result has the condition.

Positive Predictive Value (PPV)

• Probability someone with a positive test result actually has condition.
• High sensitivity screening tests (low false negatives).
• High specificity confirmatory tests (low false positives).

Health Economic Evaluations

• Comparative analysis of two or more alternative investment possibilities (interventions, strategies, policies).
• Goal is to identify, measure, value, and compare costs and consequences of alternatives.

Types of Healthcare Decisions

• Market Approval Decisions (EU Regulation)
• Market Access and Pricing Decisions (Companies)
• Reimbursement of New Pharmaceuticals and Medical Devices (Ministers of Health, National Insurance)
• Physician Decisions about Medical Treatments (Physician/Patient)

Hierarchy of Evidence Development

• IV: Expert Opinions
• III: Non-experimental Studies
• IIb: Quasi-experimental Studies
• IIa: Controlled Studies without Randomization
• Ib: Well-Performed RCTs
• Ia: Systematic Reviews of Well-Performed RCTs

Market Authorization

• Process of evaluating the safety and efficacy of drugs, etc.

Development of a Value Dossier

• Collecting and synthesizing evidence for the cost-effectiveness of a treatment.

Requesting Reimbursement

• Considerations include: budget, affordability, pricing, access, and return on investment.

Markov Model

• Mathematical model that simulates the progression of individuals or systems through various states over time.
• Key Components:
  • Mutually Exclusive: Only one state at a time.
  • Cycle: Individuals move between states at each time period.
  • Transition Probabilities: Likelihood of moving from one state to another.
  • Memoryless Property: Transition probabilities depend only on the current state.
• Example: A simple chronic disease can be modeled with health states (Healthy, Sick, Dead), transition probabilities (moving between states), cycle length, and costs/utilities for each state.

Deterministic Decision Analysis (DDA)

• Uses mean values for each input parameter, assuming exact values are known.

This quiz covers the essential concepts of systematic reviews and meta-analyses, focusing on different methods used in the analysis. You will learn about the importance of larger sample sizes, subgroup analysis, and the specific statistical methods for conducting a meta-analysis. Understand the distinctions and applications of the Inverse Variance-weighted method, Mantel-Haenszel, and Peto methods.