Literature Review Types

• Literature review evaluates scholarly writings on a topic based on research objectives and a hypothesis.
• It provides historical perspective, context for the research question, and identifies gaps in knowledge.
• It is not just a summary of individual studies.

Background Section

• A background section briefly introduces the problem, its clinical significance, and justification for the trial.
• It defines key terms and includes relevant information and analysis/critique of current literature.
• It ends with the research question of the proposal.

Narrative Review

• A narrative review gives a broad summary of a topic rather than answering a research question.
• It is broad, potentially biased, with no clear search or selection process and is qualitative.
• It can sometimes be evidence-based.

Systematic Review

• A systematic review collects, critiques, and synthesizes studies relevant to answering a research question or estimating a treatment effect.
• It has a clear research topic and question and is comprehensive with a specific search strategy and rigorous critical appraisal.
• It is quaNtitative and provides evidence-based conclusions.
• It differs from meta-analysis by not always including a quantitative summary.

Meta-Analysis

• Meta-analysis is the statistical analysis of data from multiple studies that address the same research question.
• The goal is to systematically combine results across trials for more statistical power and better precision to provide an overall interpretation of findings.
• Combining data from multiple trials may provide definitive results that individual studies cannot.
• Meta-analysis cannot fix the flaws of the studies it pulls from, potentially leading to flawed results.
• It can be misleading when there are differences in patient populations, treatments, or study designs. In these cases, a systematic review is more appropriate.
• Negative endpoints might not be reported, and positive studies are more likely to be published.

Forest Plot

• A forest plot provides a graphical summary of individual and pooled estimates in meta-analysis.
• The box represents point estimates, with the size indicating sample size.
• The line represents the 95% confidence interval.
• The diamond represents combined estimates, with the horizontal points indicating the 95% CI.
• If the diamond crosses the vertical line of no effect, it indicates no significant difference.

Funnel Plot

• A funnel plot provides a graphical assessment of publication bias.
• It's a scatter plot of effect estimates from individual studies against the study's size or precision (measured by standard error).
• With no bias, smaller studies scatter at the bottom, and larger studies scatter on top.
• Severe asymmetry in the plot indicates potential publication bias.

Statistical Methods

• The fixed effects method assumes all studies estimate the same true treatment effect, with differences across studies only due to chance, accounting for homogeneity.
• The random effects method assumes that studies are different and their treatment effects vary around an overall average treatment effect, accounting for heterogeneity. It produces wider confidence intervals than fixed effects.

Heterogeneity

• Heterogeneity is the variability in the treatment effect across studies.
• It can be caused by study population, length of follow-up, study quality, data collection methods, random variability, or differences in interventions.
• Different statistical tests can be used for meta-analysis/testing heterogeneity.
• The Breslow-Day test assesses homogeneity by testing if the odds ratio is consistent across studies or if it varies by subgroup.
• The Q-stat is a formal test of homogeneity.
• The I-squared index indicates the percentage of variation in effects estimates due to between-study variability rather than chance and has higher power than chi-square when the sample size is small.
• Meta-regression looks at factors influencing study differences.
• Subgroup/sensitivity/influence analysis looks at how specific factors affect results.