Adherence to treatment for bipolar disorder may be enhanced by interventions that address issues of appropriately taking medications to manage illness. For optimum outcomes, promotion of adherence must be integrated into the medication management of bipolar illness.

The authors examine the characteristics of threatening and otherwise inappropriate communications sent to members of the U.S. Congress by a sample of 86 subjects, 20 of whom threatened assassination. We quote excerpts from these letters and provide quantitative data on such variables as the volume, duration, form, and appearance of such communications; the enclosures; the subjects' perceived relationships to the recipients; the thematic content of the communications; and the messages and threats communicated. Comparisons between 43 subjects who pursued encounters with members of Congress and 43 who did not revealed 17 factors associated with such pursuit. In this population, threateners were significantly less likely to pursue an encounter than inappropriate letter writers who did not threaten, regardless of the type of threat or the harm threatened. Inappropriate letters to members of Congress are compared with those directed to Hollywood celebrities. Mentally disordered persons writing to public figures often mention and sometimes threaten public figures other than those to whom the letters are addressed, which raises important issues regarding notification of endangered third parties and the sharing of information among protective agencies.

Patients with schizophrenia who commit violent acts have insight deficits, including lack of awareness of the legal implications of their behavior. Targeted interventions to improve insight and treatment compliance in this population are warranted.

The associative network theory of memory [2] is outlined along with the concepts of mood congruity and state dependent learning. Two experiments are reported which investigate the occurrence of these effects where memory for pain is concerned. In experiment 1 the performance of 25 chronic pain patients was compared with that of 25 non-patient controls on a test involving both immediate and delayed recall of a mixed list of stimulus words of 3 types: pain-related, negative or neutral. No significant group differences were found in overall rates of immediate recall. As predicted, however, pain patients recalled more pain-related words than non-patient controls (P less than 0.001). On delayed recall the same significant group x word-type interaction was obtained (P less than 0.02), but in addition the non-patient controls recalled significantly more words overall (P less than 0.02). These results provide some evidence for the occurrence of a mood congruity effect. Experiment 2 investigated state dependent learning and mood congruity effects in experimentally induced pain. Twenty volunteer subjects were allocated to 1 of 4 conditions in which a wordlist (as in experiment 1) was presented following either a painful stimulus (cold pressor test) or a non-painful one (warm water) and was then recalled immediately following further exposure to stimulus conditions which were congruent with the original stimulus (warm/warm and cold/cold conditions) or non-congruent (warm/cold and cold/warm conditions). A 3-way split plot ANOVA yielded no significant main effects for group or word-type, but a significant interaction emerged between state at encoding and at recall (P less than 0.04).(ABSTRACT TRUNCATED AT 250 WORDS)