The HADS instrument is also fairly reliable. It is designed for medical patients and avoids questions about physical symptoms of disease. What it does not assess however is whether anxiety or depression were pre-morbid or whether they are a normal reaction to the psychosocial overlays common to all diseases. For example as someone else said it is depressing to be sick, to be stigmatized, to lose the ability to live your previous life etc and this is true of any disease. However, the BPS theory assumes there is no divide between emotional and physical so they assess answers filtered through their ideology. It's called confirmation bias. A better method is in-depth interviews where the patient can explain that the question doesn't have a reply that fits their circumstances, but these are much more expensive to do because they are so time intensive and variable. So instruments such as surveys and questionnaires serve as general proxies and do have some reliability in some situations. The devil is always in the details.