Congress Abstracts 2003

56

NURSING BEHAVIORAL INTERVENTION TRIAL RESULTS IN FEWER HOSPITALIZATIONS AMONG CANCER PATIENTS UNDERGOING CHEMOTHERAPY. Barbara Given, PhD, FAAN, C.W. Given, PhD, Mohammad Rahbar, PhD, and Cathy Bradley, PhD, Michigan State University, East Lansing, MI; Sharon Kozachik, MSN, RN, Johns Hopkins School of Nursing, Baltimore, MD; Sanjchoon Jeon, MS, Michigan State University, East Lansing, MI.

Purpose: This report describes how a cognitive behavioral intervention delivered by nurses and directed toward lowering the severity of symptoms reduced numbers of hospitalizations reported by patients during this 20-week trial.

Conceptual Framework: A cognitive behavioral model using a problem-solving format guided intervention strategies that focused on the severity of 14 symptoms. Strategies included; information, self-care management, and counseling and support individualized for each patient. Nurses assessed effects of interventions at subsequent visits and modified interventions as necessary.

Design: A randomized design assigned 118 patients to the experimental and 119 to the control group. Sample: Patients with solid tumors who signed consents were accrued. Mean age was 59 with 75% being female, and 66% had stage 3 or higher disease.

Methods: Age, gender, comorbidity, severity of 15 symptoms each rated on a ten point scale (0-150) were measured at intake, 10, and 20 weeks. Hospitalizations were based on patient report over 20 weeks.

Data Analysis: A Zero-inflated Poisson regression model was used to compare between groups the distribution of numbers of hospitalizations during the 20-week intervention. Absence or presence of hospitalizations was considered a binary random variable. For the number of hospitalizations a Poisson random variable with the probability of zero inflated using group, age, lung compared with other sites, as covariates. The binary component included only symptom severity as a covariate.

Findings: All measures and treatment protocols between the experimental and control group were equivalent. Twenty control group patients accounted for 37 admissions; 9 experimental patients accounted for 14 admissions. Zero-inflated Poisson regression modeled these events and group was significant after adjusting for age, lung vs. other sites, and symptom severity.

Conclusions: These findings indicate that nursing interventions that reduce symptoms can have a positive impact on patient outcomes including, use of services and subsequent costs.

Implications: The “value added” of these interventions must be considered as a part of cancer treatment.