Test-retest reliability and responsiveness of health status questionnaires

Hi, I would like to validate an end-of-life questionnaire, and have some questions to ask regarding measurement properties (1) test-retest reliability and (2) responsiveness.

I understand that (1) assesses the degree to which the participant’s performance is repeatable; i.e., how consistent their scores are across time, I have chosen the time points; week 4 and week 8 for this. And (2) is defined as the ability of a questionnaire to detect clinically important changes over time, I have chosen the time points; baseline and week 4, where patients had a palliative intervention.

In Long et al, they have a global change question, and only patients that indicated “no change” are included for test-retest reliability. In my study, we collected ECOG scores at all time points. Should I only include patients whose ECOG score remained the same across the 2 time points? Or is it okay to include all patients?
image

Similarly for responsiveness, should I only include patients whose ECOG score remained the same or improved?

Thanks, would appreciate any advice. :slight_smile:

EDIT: included the ECOG scale for reference

Please specify which change = 0 you are conditioning on. Is it week 4 minus baseline? And wouldn’t you need two baseline measurements to condition on? I’m wary of any conditioning on non-baselines.

  1. Patient Health Questionnaire (PHQ)

The PHQ is a self-administered questionnaire designed to screen for depression in adults over the age of 14 years old. The PHQ consists of 9 questions that assess symptoms of depression including feelings of sadness, hopelessness, worthlessness, guilt, sleep problems, appetite changes, concentration difficulties, suicidal thoughts, and psychomotor agitation or retardation. A score of 10 or greater indicates a probable diagnosis of major depressive disorder.

Please tell us how that relates to the ECOG discussion. And note that the cutoff of 10 was never validated.