Problem solving to improve adherence and asthma outcomes in urban adults with moderate or severe asthma: a randomized controlled trial.

摘要

BACKGROUND: Improving inhaled corticosteroid (ICS) adherence should improve asthma outcomes. OBJECTIVE: In a randomized controlled trial we tested whether an individualized problem-solving (PS) intervention improves ICS adherence and asthma outcomes. METHODS: Adults with moderate or severe asthma from clinics serving urban neighborhoods were randomized to PS (ie, defining specific barriers to adherence, proposing/weighing solutions, trying the best, assessing, and revising) or standard asthma education (AE) for 3 months and then observed for 3 months. Adherence was monitored electronically. Outcomes included the following: asthma control, FEV(1), asthma-related quality of life, emergency department (ED) visits, and hospitalizations. In an intention-to-treat-analysis longitudinal models using random effects and regression were used. RESULTS: Three hundred thirty-three adults were randomized: 49 +/- 14 years of age, 72% female, 68% African American, 7% Latino, mean FEV(1) of 66% +/- 19%, and 103 (31%) with hospitalizations and 172 (52%) with ED visits for asthma in the prior year. There was no difference between groups in overall change in any outcome (P > .20). Mean adherence (61% +/- 27%) decreased significantly (P = .0004) over time by 14% and 10% in the AE and PS groups, respectively. Asthma control improved overall by 15% (P = .002). In both groups FEV(1) and quality of life improved by 6% (P = .01) and 18% (P < .0001), respectively. However, the improvement in FEV(1) only occurred during monitoring but not subsequently after randomization. Rates of ED visits and hospitalizations did not significantly decrease over the study period. CONCLUSION: PS was not better than AE in improving adherence or asthma outcomes. However, monitoring ICS use with provision of medications and attention, which was imposed on both groups, was associated with improvement in FEV(1) and asthma control.