Most efforts to improve adherence assume that the patient is the problem and focus on patient education or on the use of incentives. Specific behaviors of the nurse that could enhance adherence and are sensitive to culture and gender have rarely been identified. This proposed study will examine the effects of: 1) behavioral training for the patients alone vs 2) behavioral training for the nurses alone vs 3) the combined effect of behavioral training for both patients and nurses. Based on the behavioral principles of cues, reinforcers, and modeling, these adherence strategies are devised to provide modeling cues and reinforcers for specific patient behavior which are sensitive to ethnic and racial culture, gender and class. We propose to use a randomized, four-group design, involving 50 subjects per group. Women will represent 50% of the sample and African Americans will represent 20% of the sample. The main outcome measures are monthly appointment keeping, medication intake monitored daily and unobtrusively with the MEMS electronic monitor, and direct counting of patient and nurse verbal behaviors observed by videotaping nurse-patient interactions. Ancillary components of the study include: a) pre- and post-test measurements of patient and nurse self-efficacy and knowledge, and b) development of a coding scheme for body language. The complimentary skill building of both the patient and the nurse is designed to yield maximum effectiveness. Such randomized adherence studies that focus on direct measurement of nurse and patient behaviors, with an emphasis on sensitivity to ethnic and racial culture, gender and class have not been undertaken. By capitalizing on the combination of unobtrusive monitoring of medications, clinic attendance, and specific nurse and patient behaviors, the proposed study will produce a unique data set for the study of patient adherence in TB that may also be relevant to other diseases where there are problems with patient adherence.
