"Notably, half the persons in remission at time 1 were not in remission at time 2. These findings suggest that sustained remission is much less than suggested by cross-sectional studies, and optimism regarding outcome may need to be tempered.
"Secondly, [the results] suggest that symptoms in later life are not stagnant and that there is considerable flux in symptoms," he commented.
In bivariate analysis, 7 of 12 of the baseline predictor variables were significant at T1: baseline remission, total number of intimates, community integration, residential status, self-esteem score, number of psychiatric medications, and number of entitlements. Significant loss in the past 5 years at T2 also proved to be a significant predictor.
However, on logistic regression analysis, only 3 of the baseline predictors were found to predict readmission at follow-up: a higher community integration score (8.8 vs. 7.3; odds ratio, 1.52), higher number of entitlements (3.9 vs. 3.3; OR, 1.57), and a lower number of psychiatric medications (1.6 vs. 2.3; OR, 0.63). After network size at T1 was controlled for, remission at T1 was significantly correlated with network size at T2 (10.3 vs. 6.5 contacts for persons in remission at T1 vs. nonremission at T2). Baseline remission did not predict any other clinical or social variables at T2, after their baseline levels were controlled for.
"The fact that community integration can impact on subsequent clinical remission suggests that social interventions can be very important in older adults with respect to both clinical and social well-being. The significance of entitlements in this population underscores the need for older adults to secure and maintain various safety net supports as they grow older and more physically frail," Dr. Cohen said.
The finding that more psychotropic medications at baseline were associated with lower remission rates at follow-up is probably attributable to greater symptoms among the nonremission group, rather than the medications making patients worse. However, "it does suggest that the medications may not be as powerful in the older age group," he noted.
Post hoc analysis revealed that compared with the 36 patients who remained in nonremission, the 16 who improved showed higher Instrumental Activities of Daily Living scores at baseline (24.3 vs. 21.9), had a greater number of physical disorders at baseline (2.3 vs. 1.0), and had more persons in their network who could be counted on (7.3 vs. 4.6). When the investigators compared the 25 patients who remained in remission with the 25 who went from remission at baseline to nonremission at follow-up, they found that those who worsened showed lower rates of community integration scores (7.8 vs. 8.9), he said.
There were no significant changes in positive symptom remission (65% at baseline vs. 72% at follow-up), with 51% in remission at both assessments, 15% not in remission at either assessment, 14% going from remission to nonremission, and 20% from nonremission to remission. There were also no significant changes in negative symptom remission (64% at baseline vs. 68% at follow-up), with 54% in remission at both assessments, 23% not in remission at either assessment, 10% going from remission to nonremission, and 14% from nonremission to remission.
"Little is known about the factors that predict the movement from remission to nonremission status and vice versa. Our preliminary data – with a fairly small subsample – suggest that social and functional factors are important," Dr. Cohen concluded. "More studies are needed to identify those factors that move persons toward remission."
Dr. Cohen’s research was funded in part by a grant from the National Institute of General Medical Sciences. He reported no other disclosures.