HONOLULU— The vast majority of patients with schizophrenia have substantial functional disability for the entire course of their adult lives. While the role of neurocognitive impairment in functional outcome in schizophrenia is now widely accepted, surprisingly little data address the prediction of return to competitive work or school. However, “there is a small percentage—somewhere around 20% to 30% of patients with schizophrenia—who do well in the long term if you look at vocational function. There are no epidemiologic samples in the literature, but there are patients with schizophrenia who work,” said James Gold, PhD. He and Keith Nuechterlein, PhD, presented the results of two studies of vocational outcome in patients with schizophrenia at the 31st Annual Meeting of the International Neuropsychological Society.

NEUROCOGNITIVE PREDICTORS

Dr. Nuechterlein, Professor and Chief of the Cognition, Psychophysiology, and Neuropsychology Lab at the University of California, Los Angeles, and colleagues, examined neurocognitive predictors of vocational outcome in schizophrenia. “Questions about the role of neuropsychological deficits in predicting real-life outcome in schizophrenia have become increasingly prominent,” he said. “I want to talk about a study that was done with young, first-episode or recent-onset patients that is longitudinal in design.”

The overall model examined the various factors that might aid in understanding work functioning in schizophrenia, Dr. Nuechterlein explained. The study cohort of 47 had a mean age of just under 25 (range, 18 to 40). Subjects had an average of about a year of college or trade school and came from ethnic and racial backgrounds of a fairly wide range. Seventy-two percent of the patients were male.

The researchers defined “recent onset” temporally. All patients were within two years of the onset of their first psychotic episode. For the majority of patients, the diagnosis was schizophrenia, with another subgroup having schizophreniform disorder diagnoses that usually converted to schizophrenia within a year and a few patients having schizoaffective disorders, Dr. Nuechterlein said.

THE THREE FACTORS OF VARIANCE

A neurocognitive test battery was administered three to four times, typically having been initiated at a clinical stabilization point about two to three months after hospitalization. “A key aspect of this study is that all patients were seen in a common treatment environment so that we could provide them all with this package: risperidone, group social skills training—which, I should point out, was not cognitive retraining but focused on social and coping skills—general case management, and family education,” Dr. Nuechterlein said.

The researchers selected nine months after baseline measurements as “a reasonable time for these individuals to try to get back to work or school. Obviously they can’t do that instantly, so any benefits from early changes in these neurocognitive deficits would not translate into benefits in work-functioning gains. But nine months later seems like a reasonable period.” A larger variable set was reduced to three cognitive factors, which Dr. Nuechterlein termed focused attention and effortful memory, rapid perceptual encoding and recognition, and divided attention and short-term maintenance. “With those three independent factors, we asked in our overall model, Can we predict from these three uncorrelated factors the later return to work functioning?”

The researchers found that the three independent cognitive factors, when considered as a group, accounted for 44% of the variance in predicting which subjects resumed paid employment or returned to regular school within nine months. Through logistic regression, they could correctly classify 19 of the 23 patients who did return to paid employment or school, as well as 10 of the 12 who did not. “Our sensitivity was 83% and specificity was also 83% for classification by those three cognitive factors identified nine months earlier,” Dr. Nuechterlein added.

SIGNIFICANT CONTRIBUTIONS

When asked whether all three factors contribute equally to this prediction, Dr. Nuechterlein revealed that it was the first two factors—focused attention and effortful memory, and rapid encoding and recognition—that predicted significantly. Divided attention, “at least as we measured it, does not in itself predict significantly,” he said. Further, seven of the 11 test scores that the researchers chose to examine also predicted significantly.

Dr. Nuechterlein remarked that “44% is really a substantial amount of variance, especially when you consider that there are other factors and practical constraints: This doesn’t even get into whether some of the patients have continuous problems with disorganized or negative symptoms, whether some have very supportive families, while some don’t.... All of those things presumably have something to do with work functioning. So I think this really suggests that our field is going to have great impact in schizophrenia in answering who is going to do well in their everyday functioning,” he concluded.

GOOD VOCATIONAL OUTCOME

“I’ve decided to study patients with good vocational outcomes as a model because I don’t think right now we have a good notion of what real rehabilitation or drug targets are that would be very reliably related to good functional outcome in schizophrenia—and it would seem like the best way to figure that out is to look at patients who have done well and determine what’s different about them,” said Dr. Gold. Another consideration “is that maybe good-outcome schizophrenia patients have the minimal necessary and sufficient biological abnormality but [that abnormality] is uncontaminated by illness severity and disability. In the literature, people look at family members to predict liability, and then they study patients to look at illness. What I’m suggesting is that there’s actually a third dimension: which is that you have liability factors—patients who have good outcome are ill, but then you also have factors that are related to disability, and those things are conceptually distinct, and this is a way to try to take that puzzle apart. In short, maybe the neurobiology of illness is not completely identical with the neurobiology of disability in people with schizophrenia,” he added.

Dr. Gold, Associate Professor, Department of Psychiatry, Maryland Psychiatric Research Center, University of Maryland in Baltimore, and his colleagues examined 38 good-outcome patients (ie, they had been employed at competitive integrated jobs 20 or more hours per week for 18 of the previous 24 months); 38 poor-outcome patients (unemployed for the preceding 24 months other than occasional odd jobs or street hustling); and 43 healthy controls. “Excluded from the study were people who were in school or homemakers. Patients also had to be ill at least five years,” Dr. Gold noted.

The mean age across the three groups was 40, with a mixture of whites and African-Americans forming the cohort. Good-outcome patients had completed a higher level of education and had experienced the onset of schizophrenia three years later than poor-outcome patients did. Good-outcome patients performed a variety of jobs, Dr. Gold said. “Overall, not highly skilled jobs, but they’re real jobs: These people are getting a paycheck, and they have schizophrenia.”

The groups underwent a battery of cognitive tests, with the most robust associate of outcome status being measures of processing speed. “If you adjust for performance on the reading subtest of the Wide Range Achievement Test, you can attenuate the processing speed difference, but you certainly can’t knock it down,” said Dr. Gold. “Is it just speed? Yes and no. It’s not just a motor speed, it’s a cognitive speed, whatever that construct is. So the speed differences come when processing symbols or when selecting responses, not just when making a response.”

Additionally, “there are large IQ differences on every factor from the WAIS [Wechsler Adult Intelligence Scale] and on the factors from the WMS [Wechsler Memory Scale],” he reported. “It is very curious that we don’t find much of the difference on verbal memory. In fact, our good-outcome patients are very impaired on their verbal learning and memory. They do significantly better on logical memory, but verbal list learning differences are not significant.”

ILLNESS OR DISABILITY?

Comparing the odds of patients’ having good-outcome schizophrenia as opposed to their being healthy controls, Dr. Gold and colleagues attempted to determine markers of illness. “Here the memory scores get a little more peppy. You have odds ratios above 3. Processing speed is still important, but if you contrast your odds ratios—schizophrenia odds versus disability odds—you can see that memory is much more a marker of illness in this sample, whereas processing speed is much more a marker of disability. If you look at poor-outcome schizophrenia patients relative to controls, the odds ratio for being a poor-outcome patient with schizophrenia is 40 for processing speed.”

Also, poor performance on the Identical Pair Continuous Performance Test “in this context is either marking illness or a form of liability that is common to the sort of families that these folks come from,” said Dr. Gold. “Good- and poor-outcome patients completely track on top of each other. They differ from controls on each level; they do not differ from each other.”

On the span of apprehension test, the poor-outcome patients performed worse than the good-outcome patients, who performed worse than the healthy controls. “Those differences were all significant,” Dr. Gold noted, “and then there was a provocative result on smooth pursuit eye movements, where it is only poor outcome patients with schizophrenia who have impaired smooth pursuit. So here in this one sample you have these three putative genetic liability markers that are tracking in very different ways.”

TREATMENT, REHABILITATION, AND OTHER IMPLICATIONS

As for theoretical implications, “Abstractly, you might be able to argue that there are distinct and also partially independent dimensions of cognitive impairment in schizophrenia, and that processing speed is the most robust marker of disability among patients with the illness,” Dr. Gold proposed. The real challenge is to use these findings to foster rehabilitation. “Treatments of any type that provide only subtle differences, I think, are of very little relevance. For those of us who have any involvement with drug trials and drug companies, subtle advantages will have very limited relevance for work outcome.”

Processing speed would appear to be the most promising target in terms of pharmacologic trials, he added. “If you want an outcome measure, these tests are easy, they’re quick, they’re highly reliable, and they appear to matter. The other treatment-related implication is that some of the impairment in our patients is likely developmental in origin. We have big verbal differences between groups.…I don’t think those are amenable to treatment—but you might be able to do something about processing speed.”

Return to table of contents