Studies of social support networks have almost exclusively measured only their positive aspects. In this research, we investigated both the helpful or positive and the upsetting or negative aspects of social networks in a longitudinal study of spouses caring for a husband or wife with Alzheimer's disease, a progressive senile dementia. Measures of helpful and upsetting aspects of the care givers' networks, derived from interviews and daily interaction ratings, were studied for their relations with overall network satisfaction and depression at an initial interview period (n = 68) and at a follow-up period about 10 months later (n = 38). Results from hierarchical multiple regression analyses, in which care givers' age and sex and a measure of the spouses' health status were controlled, showed that the care givers' degree of upset with their networks was strongly associated with lower network satisfaction and increased depression at both time periods. Helpful aspects bore little or no direct relation to either depression or network satisfaction. Helpful aspects of the network did, however, interact with network upset in predicting network satisfaction, and depression (combined probabilities test, p less than .05). Longitudinal predictions of follow-up depression, after age, sex, care givers' health status, and initial depression levels were controlled, showed that changes in upsetting aspects of one's network were predictive of changes in depression over time. We interpreted these results within an attributional framework that emphasizes the salience of upsetting events within a social network.

Twenty-seven patients with medically refractory paroxysmal disorders underwent EEG-video-audio (EVA) monitoring in an inpatient neurology-neurosurgery unit over 1-15 (mean 8.9) days. Fast visual review of all EEG records (5,784 h) and subsequent analysis of synchronized EVA patterns demonstrated a total of 208 partial epileptic seizures (ES) in 12 individuals and 87 psychogenic episodes (PE) in 15 subjects. Clinical ES lasted 83.3 s on the average and were most frequent from day 7 to 9 of monitoring (42.3%) and during sleep (56.4%). PE were longer in duration (mean 724.5 s), most numerous during the first 2 days of monitoring (41.4%), and occurred exclusively during wakefulness. Subjects with PE signaled (by pressing on a push button) more events (35.6%) than did the individuals with ES (27.9%). Multiple observers raised the proportion of alarms to 69.0% of PE compared to 39.9% of ES. Following the alarm, nurses reached the patients' bedside within a brief time (mean 22.2 s). To differentiate partial ES from PE or to establish the association of these disorders, EVA monitoring is best performed around the clock over a period of 1-2 weeks. The limited number of paroxysmal events, especially ES, signaled by the patients should be considered when designing studies of the effectiveness of pharmacologic, surgical, and other treatments.

No abstract

No abstract

Decisions about the appropriate balance between centralized and decentralized staffing and responsibilities in multisite evaluations should be based on scientific, administrative, and political considerations.