Measuring Flashcards
What outcomes would you measure?
- mental health (of disease under study), mainly symptoms
- mental health (comorbid symptoms)
- mechanisms of therapy change (e.g. cognitions in CBT)
- functioning in daily life
- quality of life
- satisfaction with treatment
mental health (of disease under study)- how would you measure this?
Diagnosis vs. symptoms:
Diagnosis: DSM interview –> dichotomous, well understood and accepted, high burden
Symptoms: questionnaire (by therapist or patient) –> continuous (more sensitive to change), lower burden, more difficult to interpret (no gold standard to interpret severity)
how would you measure other outcomes?
- questionnaire
- hospital records (with permission)
- behavioral tests or other tasks (e.g. cognition, is sb still afraid of spiders?)
How do you decide that your treatment is effective?
you decide based on your primary outcome
primary outcome
- the outcome on which you decide whether your treatment is effective or not
- also used to calculate how many people you need to include
secondary outcome
- all other outcomes which you think are relevant
How many questionnaires/ or tests can you use?
- no gold standard
- danger too many questions: drop-out because of burden, unethical when you don’t use data
- better to use few but well selected measures
- always test how long it will take patients to fill it out before RCT
How to choose the best instrument/ questionnaire
- literature on your topic/ target group
- find a number of (frequently used) questionnaires
- search for additional articles on the questionnaires itself
- does it measure the concept you are interested in?
- as short as possible
- broadly used
- psychometrically sound (valid, reliable, sensitive to change)
What if there is no questionnaire (in your language)?
-translate: backwards and forwards
- establish psychometric properties again
only in very rare cases:
-develop your own questionnaire
- establish psychometric properties
how to administer self-report questionnaires
- paper-and-pencil
- interview
- online
when to measure
- baseline always necessary
- during treatment
- after treatment (post-test) also necessary
- follow-ups (e.g., 3 months, 6 months, 1 year)
Why would you (not) measure during treatment?
- interested in overall treatment-effect: not measuring during treatment
- when treatment is very long: maybe measure during treatment
- interested in mediating variables: measure during treatment (e.g. in insomnia treatment, what changes first sleep or depression?)
Why would you do follow-ups and how often / for how long?
- you usually want your treatment to work for a long(er) period
- measure as long as possible
- be aware: not possible in RCTs with waitlist control group
- measuring too frequently: people will drop out
blinding
being unaware of treatment group
- originally: single vs. double blind trial (patients blinded only vs. patient and experimentor blinded)
What can you blind?
- therapist (often not possible)
- patient (often not possible)
- outcome assessor
- statistician (person doing analysis)
always blind:
outcomes that need interpretation:
- diagnostic interview
- therapist rated questionnaire
- test
what other things do you need to measure in a RCT?
- safety (suicidal ideation)
- in- and exclusion criteria
- descriptives of the sample
- adverse events
- treatment uptake and integrity
in- and exclusion criteria
- you need to define it and give a clear cut-off (e.g., what is suicidal?)
why do you need descriptives of the sample?
- to see if it is a representative sample (of your target group)
- important for you interpretation
- important for the readers: do they have similar patients?
gender, age, SES, ethnicity, marital status, history of disease (how long, therapy, pills), comorbid diseases
what are adverse events?
- increase of health problems
- might or might not be sure to your treatment
e. g, accidents, mental health problems, suicide
why measure adverse events?
- should be reported to ethical committee
- important if they are treatment related
treatment uptake
= how much of the treatment has been done
e.g., treatment protocol prescribes 10 face-to-face sessions but patient only has done 2
but also: how much was the patient involved in the treatment? e.g., how much homework did they do?
online: how many times logged in, stayed online etc.
treatment integrity
= was the therapy delivered as planned
video taped, checklists…
