Systematic Reviews: Evidence Based Eating

Question 1

RCTs: methods to reduce confounders

Answer 1

• randomisation

- blinding

Question 2

Problem with RCTs in nutrition epidemiology

Answer 2

• intervention is difficult to blind, unethical, or difficult to impose
• expensive: often select patient samples already at risk for disease to shorten study

Question 3

How to conduct a systematic review

Answer 3

1. Define research question
2. Define search terms
3. Define inclusion/exclusion criteria
4. Select studies and document reasons for inclusion/exclusion
5. Assess any risk of bias
6. Extract data
7. Conclude

Question 4

How to conduct meta-analysis

Answer 4

Like SLR but quantify results

Question 5

Issues with SLR/meta analysis

Answer 5

• only as good as single studies, need to show evidence of sensitivity analysis to ensure not too dependent on a single study
• publication bias: journals unlikely to publish null findings

Question 6

How to form causal viewpoints in nutritional epidemiology

Answer 6

1. Evidence of consistent association i.e in SLR
2. Strength of association, usually expect RR/OR of ~20%
3. Dose response: ie does more fibre decrease risk of CBD further
4. Biological plausibility
5. Temporality: this could be affected by trial design