This sounds very interesting. For interviews, even semi-structured, you often design the questions yourself. Maybe base those on individual aspects of the therapy and explore these?

I assume you would want a 'before' and 'after' measurement of anxiety to show efficacy.  This can be a self-reported Likert scale questionnaire.  

Here is a good article you can use as a benchmark for designing your study: http://hkpsychotherapy.org/the-effectiveness-of-psychotherapy

Camille, you could include Likert scale as tool to include in your semi-structured interviews with older adults as they are more responsive to

To my experience, semi-structured interviews rely greatly to the skills and expertise of the interviewer. As for ols age people, thre is one more confounding variable, that of cognitive performance. So , personally, I do not think that  kind of interview hrlps much  to estimate the outcome ogf this kind of therapy to older adults. of course, there ar eresearchers that advocate the use of  cognitive psychotherapy in older adults with depression, but the results are more evidence in the mood part than the cognitive part ( Simon SS et al Int J Geriatr Psychiatry. 2014 Dec 17. doi: 10.1002/gps.4239. [Epub ahead of print] 

the dyad as Violeta suggested,

Furthermore, efficacy, validity and reliabilty can be ascertained as you are conducting the interview face to face, if not by proxy.

Regards. Mariam

I think a semi-structured interview will give you a great deal of qualitative information about your participants and what elements of therapy worked best for them.  I agree with Georgios that doing semi-structured interviews requires experienced interviewers - inexperienced interviewers are more prone to introduce confounding variables and less likely to elicit information.  However, as Mariam mentions, it will be difficult to measure efficacy with a semi-structured interview that you will have to design, see Clarissa's answer above.  So, a mixed methods design is likely to work best.  Maybe you can give us more information, such as: age range of participants, structure of therapy sessions, and, as Georgios mentions, any cognitive impairments.  You can also use a different scale, in conjunction with, or instead of a Likert Scale, depending on what exactly you want to measure: Outcome Questionnaire-45 (OQ-45), Lambert et al., 2004 designed specifically to measure efficacy of therapy.

Hi Clarissa,

Thank you for your time. Your answer was helpful, and I think that basing these questions on individual aspects of CBT, like education and relaxation, is a start.

Kind regards, 

Hi Violeta,

Thank you for your answer. It wasn't clear in my question, but my future research is not about testing the efficiency of the CBT for older adults (many studies have done that and Likert-scales "before and after" are a good way of doing it). As you rightly pointed out, I'm looking for qualitative information that would lead for better-adapted CBT for older adults. Kind regards,

Hi Mariam,

I didn't know that the efficacy, validity and reliabilty can be ascertained as you are conducting the interview face to face, so thanks a lot! Kind regards,

When doing face to face interview work it is very hard to be scientific and get quantitative measures. The project at issue here -- what may be important is what the results is to be used for .Is it for publication and if so how scientific is the publication or is it for qualitative reporting? In my experience projective techniques such as TAT and Rorschach in pre and post test formats with some projective questions could work well. Let me know how it goes.

Hi,

Indeed, it is for qualitative reporting. Thanks for your answer! 

Before conducting these experiments, I suggest that all participants should have a cognitive screening first [GPCog, Trail Making Test B, etc.], as persons with dementia should not be exposed to CBT.  I assume you have already taken this precaution.  J.V. White, PhD, gerontologist