Challenges in the Assessment of Negative Symptoms and Cognitive Impairment in Schizophrenia

Please answer the question below before watching the video. 

This video describes common challenges faced during the assessment of negative symptoms and cognitive impairment in people living with schizophrenia.

Please take the quiz again to asses your understanding after watching the video.

Video Transcript

One of the challenges in assessment, though, is that there are some things that people with schizophrenia just aren't very good at self-assessing. One of the things that they're the worst at self-assessing is, actually, their cognitive abilities. In this study here, Rich Keefe and I developed a twenty-item rating scale to rate the functional impact of certain different cognitively demanding tasks that are challenging to perform.

We asked people with schizophrenia to answer the twenty questions. We asked their mothers to answer the twenty questions. The interviewer who talked to the patient and their mother—she generated ratings, too.

Then we had the patient perform a test of functional skills, and what you see with those correlations with neurocognitive performance is that the correlation between self-reported cognitive functioning and objectively measured cognitive functioning was zero. But if you asked the patient's mother—and you asked the right questions and came to the right answer—the correlation was as high as if you asked them to directly perform a functional capacity assessment, meaning that the report from an informed informant who's observing the patient regularly is just about as good as a neuro-psych assessment, and takes 5 minutes instead of  3 hours.

So, that's an important point. But again, often, there is no mother around. And if they're in a group home or so, people are not as closely observing the patient.

So, we need some tests that are not as large as some of the cognitive battery tests that we use in studies.

And you got to ask the right informant. The title of this paper was not all informants seem equally informative.

So, what we did was we asked—this is only one of the slides we asked high-contact clinicians and caregiver relatives to rate patient's functioning. But then what we did was we simulated what commonly happens in drug company clinical trials where we had the patient deliver their own informant, who was a friend of theirs or a roommate or something. As you see there, the correlation was zero between the informal informant and the patient's actual functioning measure with objective tests.

So, the correlation was zero between the informal informant and the patient report. The correlation was zero between the informal informant and the patient's objective functioning. So, this is not an informative informant. Patients for whom they had a case manager or a caregiver, the correlations were extraordinarily high and not zero like this.

We have another issue, I think, and that is the gap between the measurements and clinical trials and the measurements we can use clinically.

And so we are, as you know, the PANSS is the name of the game for these studies that we're doing in schizophrenia. That is thirty items: seven positive, seven negative, and sixteen general items. And then you put all of the numbers together, and then you have a score, and you do an analysis on it.

We actually subjected this to some scale testing and saw that, well, the thirty items are not scalable. And I'll tell you what that is in a second.

Even the fourteen items that have been floating around or eight items, the Andreas and the remission criteria, they're not scalable. What does that mean? Very frequent symptoms should, in the population, be less severe, and more severe symptoms should be less frequent because otherwise people couldn't walk around. Only if that's given and gives you a nice overall image of what the patient can do and what the symptoms are, you would be able to call this a scalable scale that you can numerically add together.

And so just by looking at this distribution, and we did that in acutely ill patients, in treatment-resistant patients, in KD, also in the outpatients, and we're now doing it in adolescents. Other groups have repeated this very conveniently. The PANSS-6 emerges at six items, and conveniently, it is three items we should always, on the positive side, ask about delusions, hallucinations, and conceptual disorganization. And we have three negative items that are also very common: affective blunting, alogia, and asociality.

Now, these are six items, but you only need to ask three questions because three are observational: affective, blunting, alogia. You talk to the patient, you look at them—and also the conceptual disorganization—you hear their language. Do they string the words together? So, you only need to ask three questions to get your PANSS-6 that is correlated with the PANSS-30 very tightly. We don't need the PANSS-30, at least in order to capture these two positive and negative symptom domains. So you would have to ask: do you hear voices? Do you feel somebody's out to get you or is there anything unusually happening that others people don't agree with? And then, also, have you been socially active with anyone? And these three negative items actually very nicely dovetail with the PANSS items that are now in the NSA-4.

The NSA-16 has sixteen negative symptoms, but there's a short form. You have both the NSA-4 and the PANSS-6, and these are questions we should always ask during a clinical interview anyway. And that could then give us some basis for measurement-based care that we see whether our treatments actually are helpful and to what degree.