3. empirically supported tx: research methods + evidence based Flashcards
RCT steps to empirically evaluate psychotherapy, past and present definitions of ESTs
what are the 4 methods of evaluating psychotherapies
- case studies
- naturalistic studies
- quasi-experiments
- RCTs
in what context would it be best to use a case study?
best when studying completely new therapies
what are naturalistic studies? pros? cons?
- over long period of time
- usually with same patient
- con: usually the same people
what is quasi-experiment, and what is it usually used for?
- not a complete experiment
- comparing 2 experiments or tx
what does RCT stand for, what are their origins, what’s it used for?
- randomized controlled trials
- adopted from medicine
- used to determine cause and effect
name the 7 steps of RCT testing
- develop protocol
- choose comparison to tx of interest
- select participants of interest
- random assign participants to conditions
- administer tx and assess fidelity
- evaluate outcomes at end of tx
- evaluate outcomes at follow up time points
developing the protocol…
what are the 2 questions you need to ask yourself when developing the protocol?
- what is the tx? (what’s the model? what’s the tx technique?)
- how will it be administered? (is it standardized?)
what are the 3 types of comparison txs?
- waitlist
- supportive psychotherapy
- CBT (gold standard)
what is the difference between CBT and supportive psychotherapy?
support implies building a relationship w the patient as a baseline. usually includes just conversational therapy. CBT is the tested therapy in which there’s a little sprinkle sprinkle
what are the 2 things you need to consider when selecting participants?
- balancing concerns with internal and external validity
- finding participants that are representative of the population to which it’ll be generalized
what is internal validity?
quality of experimental design and control for extraneous factors
what is external validity?
will the results be generalizable?
what are the 2 factors that usually prevent researchers to have representative samples when selecting participants?
- demographic factors (usually causasion undergrads)
- comorbid dx (usually ask for non-comorbid)
why do we need random assignment?
minimizes pre-existing differences between groups that could affect the outcome
difference between single blind and double blind
double blind: both part and exp don’t know who gets what
single blind: part doesn’t know if they’re getting tx, exp knows
explain how to administer double blind experiment in the context of psychotherapy
… u can’t. lol.
what is a fidelity check? what is it used for?
fidelity checking is to make sure that the tx is administered adequately. researchers will use recordings and rate how well it’s being administered
what’s the biggest influence that could alter how a tx is being administered in the context of psychotherapy?
individual therapist factors
what kind of validity is being ruptured if the therapists aren’t administering the tx as outlined?
internal validity
how do you evaluate the outcome of a tx? on a px
there are 4 ways
- no longer meeting the DSM dx criteria
- decrease in target sx (not necessarily a full recovery)
- decrease in comorbid sx
- increase in functioning
what are 4 ways to evaluate the success of your tx?
- statistical significance
- effect size
- therapist and site effects
- people who drop out
what 2 things influence statistical significance?
- magnitude of effect
- sample size
what is an effect size?
magnitude of difference, independent of sample size
what’s the formula for Cohen’s d?
d = (x1-x2)/Sp
d = cohen’s d
x1 = mean exp
x2 = mean contr
Sp = SD of control group
Cohen’s d: what is a small, medium, large effect?
small: 0.2
medium: 0.5
big: 0.8
it can go over 1
how much change is needed for clinical significance?
relative to what we’re studying
ex: if it’s suicide attempts? anything. lol.
what are site effects?
results can be different based on location in which tx is administered
what do we do if people drop out? what are some assumptions we can have? what do we do, statistically?
perform intention-to-treat analyses
did they get better before the end? did it not help?
two things that can happen once tx stops?
- relapse
- sleeper effect
what’s the sleeper effect?
during tx, might not see any improvements. after tx, continue to do better (concept of consolidation during sleep)
according to the DeRubeis et al study, what was the conclusion drawn about psychological tx and medication?
psychological tx was more enduring than meds
why might some people relapse after tx?
during tx, you’re continuously held accountable, learning new skills, practicing said skills. after, all the accountability stops.
what’s one element that needs to be in the report of an RCT?
i usually skip these when skimming research articles lol
CONSORT flow diagram
what’s in the flow diagram?
4 questions
- who’s in the study
- what happened to them
- how well was it designed
- is something wrong with the design
the last two are the same tho so cut urself some slack
what did the first ever meta-analysis conclude?
Smith + Glass 1977
typical therapy client was better than untreated clients
what does EST stand for?
empirically supported tx
what were the criteria that the chambless & hollon article imposed for well established tx?
minimum 2 “good” between group designs that show tx better than meds, psychotherapy, or other tx
OR
equivalent to established tx
OR
bunch of single case designs with good exp design and comparison to other tx AND
conducted with tx manuals or other clear descriptions
characteristics of sample must be clearly defined
effects must be demonstrated by min 2 different investigators or teams
according to chambless & hollon
what’s a probably efficacious tx?
- 2 exp show tx is better to waitlist control
OR - 1 or more exp meet criteria but have not been replicated by independent investigators
OR - small series of single case design experiements have been conducted
what were tolin’s critiques of chambless & hollon?
- eval based on 2 studies is too low of a bar
- focus is on sx reduction
- no guidance on which EST to choose
which is better:
meta-analysis or single case study?
meta-analysis
which is better:
RCT or single case study?
RCT
what’s the difference between a strong and a modest tx?
strong: well established
modest: probably efficacious
