Fundamentals of Research Flashcards
How do we know anything?
- scientific method drives knowledge
- starts with observation/question, then hypothesis, tested, result, and conclusion
- update collective thinking
observational
independent variable not under the control of the observer, just observing the dependent variables
experimental
independent variable directly manipulated by experimenters and then observe dependent variables
What are examples of observational studies?
- case control
- cross-sectional
- cohort
What are examples of experimental studies?
- non-randomized controlled trial
- randomized controlled trial (parallel/crossover group)
What is the gold standard of research design?
one placebo-controlled, double-blind, randomized trial of a sufficient and generalizable sample; can provide the highest quality of evidence
What are some pros and cons of background information, expert opinions, and non-EBM guidelines?
Pros:
- easy to access
- can be agreed upon knowledge
- trust in the expert can be high
Cons:
- can be wrong information
- people can say whatever they want (no quality control)
- not necessarily based on evidence
- not necessarily based on scientific method
- biases rampant
What are some pros and cons from individual case reports?
Pros:
- real data from patient/case
- new observations possible
- generate hypotheses
- useful for rare disorders
Cons:
- small amount of data
- poor for cause/effect understanding
- no ability to generalize
What are some pros/cons from case series or studies?
Pros:
- more useful than case reports as more data
- very useful for rare disorders/conditions
- observational, can record data
- easy to do
Cons:
- no control group
- difficult to generalize
- selection bias (biases of who is available to come to your clinic)
- often a single time point
What are some pros and cons of a cohort study?
Pros:
- can determine temporal sequence (for cause/effect)
- determine incidence (how often X happens)
- can study rare exposures
- can evaluate multiple exposures and outcomes
- avoids selection bias
Cons:
- very expensive and time consuming (ex. follow people for years and years)
- not useful for rare diseases/outcomes
- not useful for diseases with long prodromal phase
- loss of follow-up can lead to bias
What are the pros/cons of a non-randomized controlled trial?
Pros:
- generally cheaper and quicker than RCTs
- can show evidence of benefit/harm of an intervention
- randomization might be unethical (ex. smoking during pregnancy/ 9/11 individuals who had health outcomes)
Cons:
- poor ability to generalize
- may be reasons for group membership
- blinding may not be possible
What are the pros/cons of randomized controlled trials?
Pros:
- randomization limits bias and confounders (includes variation)
- group membership not influenced by choice
- comparative
- only study type that can definitively show cause & effect (how people were at beginning vs. end)
Cons:
- expensive
- major problems with dropout/poor adherence = affects credibility of experiment (strive for 80-90%)
- may need huge study to actually test the question (smaller difference you want, more people you need)
What are the pros/cons of evidence-based practice guidelines?
Pros:
- generally from trusted sources (ex. APA)
- generally accurate information
- can integrate information across many different studies and give one conclusion
Cons:
- still requires transparency/trust to be useful (still may have people talk about things for a reason)
- often slow to adapt to new information (sometimes only update guidelines every 10 years)
- no quantitative data (no numbers; just yes/no)
What are the pros/cons of systematic reviews?
Pros:
- typically comprehensive
- can identify gaps in evidence
- show methodological problems/concerns
Cons:
- no quantitative assessment
- can be biased
- may miss information
What are the pros/cons of meta-analyses?
Pros:
- provide a single effect size estimate (how much?)
- can also include moderator/modifier estimates (does X matter?)
- provide systematic evidence
- can compare between treatments
- the most useful!
- synthesizes all studies conducted into one study
Cons:
- may not be asking the right question
- reliant on quality of studies that go into it
