Lecture 30: Systematic Reviews

Question 1

What is evidence based practice? And what is the “recipe” for it?

Answer 1

Goal of treatments:
-successful interventions with MAX benefits
-Successful interventions with absence of or MIN risks

Recipe:
1. High quality research
2. Practitioner Expertise
3. Patient/ community values (ex of indigenous values different from western medicine so won’t be as accepting)

Question 2

How to determine the quality of medicine?

Answer 2

Quality and bias are important

-Same triangle of types of studies, systemic reviews at top and Randomized control trials (RCT) 2nd therefore highest quality with lowest bias
-Expert opinion and case series, case reports at bottom

Question 3

What is a systemic review?

Answer 3

-A form of a knowledge synthesis
-A systemic approach to synthesize scientific evidence for a specific topic
-Employs a clear, rigorous, and transparent approach to learn about a topic
-Goals is so transparent that if you were to show your study to someone they should be able to replicate and get same

Question 4

What is a meta-analysis?

Answer 4

-Statistical method used to combine the results of various studies included in the systemic review
-You can do a systemic review without the meta-analysis BUT you can’t do a meta-analysis without a systemic review

Question 5

Who needs to be included in a systemic review?

Answer 5

-A team leader (assign a delegate)
-Experts in the topic of interest (at least 1)
-A librarian or library scientist
-A member with statistical knowledge (for meta-anaylsis specifically)

Question 6

What are the steps to conduct a systemic review?

Answer 6

1. Define the specific research question (very clear research question CAN be ambiguous bc then have too many results)
   -certain intervention/predictor or outcome of interest
2. Prepare and execute a search strategy
   -Done with library scientist, typically design and register a protocol
   -Specific keywords, year, language restrictions, databases
   -Full text articles (published and unpublished but have to do extra stuff for unpublished)
3. Study selection
   -Systematic way to select or reject papers (a second person should be used for quality assurance)
   -Screening of abstract and titles based on inclusions criteria
   -So now go from potentially 1000 to 75 papers
4. Quality assessment
   -Various standardized checklists available to check quality
   -Ensure there are no biases, and that study is of “quality”
   -Done by 2 reviewers (inter-rater reliability)
5. Data synthesis
   -collecting necessary info: author, date, purpose, methods, results (stats), conclusions etc
   -Collecting effect sizes to obtain overall effect (odds ratio and condense intervals)
6. Data interpretation
   -Present a data selection decision tree (PRISMA)
   -Present narrative info in tables
   -Present effect sizes in forest plot

Question 7

What are the benefits of systemic reviews?

Answer 7

1. Transparency: must abide to a protocol allowing for replicability (less room for biases)
2. Many studies considered simultaneously: less time consuming for practitioner, more confidence in decision
3. Meta-analysis: large samples, yielding higher statistical power, generalizability of studies can be tested (instead of having 3 groups of 500 participants would now have 1500 instead)

Question 8

What are limitations of systemic reviews?

Answer 8

-if Garbage goes in garbage goes out
-Selection bias (why have multiple ppl omitting/selecting specific studies on purpose)
-Heterogeneity (from different countries)
-Publication bias and inclusion of “grey literature” debate

Question 9

What are the benefits of systemic reviews in observational studies?

Answer 9

-Studies of risk factors generally can’t be randomized because they relate to inherent human characteristics or practices
-Exposing subjects to harmful risk facts is unethical

Question 10

What are the limitations of systemic reviews?

Answer 10

-Inability to control
-Various study designs