I'm currently validating a nonverbal mood measure that shows strong correlations with the HADS-D, and which might be of use for you if language is an issue. It does not require language to use. Description and preliminary validation findings below. The prototype can be downloaded free of charge.

Conference Paper Developing Mood Scales Suitable for Use in Stroke Patients w...

I would  not see a particular problem with that scale as long as you administer the instrument in the same way with all patients.  Singular administrator practiced.  Or were you thinking multiple administrators?

Thank you Paul!!!

Have you ever thought about validating for ICU patients? In the postoperative period of cardiac surgery, after intubation, the patientes spend hours, or even one or two days, with difficulty speaking. For me, it could be very important. I need evaluate pre and pos-cardiac surgery this values!

Randall,

are two administrators, me and my student, trained by me!

That would work.  A slew of like data is collected from children and adolescents with cross validation showing no bias introduced.  Unless someone knows of a specific study of which I am unaware, I do not see any issues as long as the two of you are as consistent as possible.  Of course there are antithetical arguments but I see them having little merit.

Hi Eduardo,

For now, the focus of the validation of D-VAMS is on stroke patients with aphasia, where screening for depression is frequently very difficult or impossible. Once the initial validation study is complete and the findings published, I would expect its use in other populations to be explored also. Since the HADS was designed for hospital patients, the D-VAMS – which was also designed with comorbid physical symptoms in mind – would seem likely to be a good substitute for the HADS-D where communication is very difficult. However it's probably going to be a good while before a specific validation study looking at patients in postoperative care emerges. 

One other alternative does come to mind, though: You could read out the HADS question and their options in Yes/No, format, so that the patient can nod or shake their head, or use some other binary, non-verbal response to select the option they want. Though slower, this would enable people who can understand what you are saying to respond to the HADS questions. 

Certainly, if they cannot complete it alone for any reason, there is absolutely no problem with going through it with them supervised, item by item. Just bear in mind that people sometimes find it difficult to share their feelings, and may report things being better than they are ('faking good'). 

Hi Eduardo,

The only precaution is that you must read the statements word for word and try to keep your face from showing any emotion or signs that would prompt an answer.

Good, William!

We try expressing any emotion the least possible! Good! 

Good, William!

We try expressing any emotion the least possible! Good! 

Good, William!

We try expressing any emotion the least possible! Good! 

• Thank you! Some researches in Brasil say not. But the most do as structure interview!

Thank you! Some researches in Brasil say not. But the most do as structure interview!

While I would not have a problem with you reading the questions to the patients in an interview format, I would have a problem with you using the HADS in the first place. We have shown that it in fact measures a single dimension of distress (mostly), so if you want to really discriminate between anxiety and depressive symptoms I would recommend other measures - such as the PHQ-9 and the GAD-7, which are both freely available.

Frank

Frank,

Are the PHQ-9 and the GAD-7 validated to inpantients or, specifically, in preoperative period?

if patients can not read your questions , fewer will understand , therefore your score would not be reliable or valid . I suggest making a version of the HADS with drawings of facial expressions to answers questions sketched and validated in your country version . Submit to the evaluation by experts in psychometrics or who validated the scale and get the validity of judges . Then you could confiabilizar with psychometric properties.

I've thought about using this strategy, adapting to a cup of empty to full. But I think  the items of the scale are very, not being a simple analysis of a Likert-type scale only with numbers.

The problem is that if you use the vacuum cup filled would be biased question . Technically though the scale is ordinal level , the reactive response is reactive rather nominal nature, because it can only choose one answer for reagent. The sum would ordinal nature itself.