Predictors of protocol adherence in a pediatric asthma clinical trial

Background: Declining protocol adherence can threaten the validity of a clinical trial. Objective: We sought to explore patient and family factors important for protocol adherence in the 133 patients followed at one of the 8 Childhood Asthma Management Program (CAMP) clinical centers. Difficulties with timely return of diary cards (diary card problem), with keeping or frequently rescheduling appointments (appointment problem), and with commitment to all aspects of the trial (commitment problem) were tracked prospectively during the treatment phase of CAMP, which ranged from 20 to 40 months at the time of the analysis. Methods: We performed a cross-sectional analysis. Results: During the course of this investigation, no St Louis CAMP patients dropped out of the study, although signs of eroding participation were observed in 44% of patients. For this cross-sectional analysis, the percentage of patients exhibiting protocol-adherence problems was greater the longer patients had been in the trial: 33.3% at 20 to 25 months, 39.5% at 26 to 30 months, 51.4% at 31 to 35 months, and 69.2% at 36 to 40 months (P < .01). The diary card problem was present in 22.2% of the patients enrolled in the trial for 20 to 25 months compared with 66.7% for patients enrolled for 36 to 40 months (P < .005). Appointment and commitment problems were present in smaller percentages of patients and did not change by time in the trial (P = .41 and .22, respectively). A logistic regression analysis of demographic characteristics indicated that age at randomization and time in the trial were significant factors: for every 2-year increase in age, a child was twice as likely to have a commitment problem (odds ratio [OR], 1.96; 95% CI, 1.50-2.57), and for each additional 5 months of participation in the study, a child was twice as likely to have a diary card problem (OR, 1.91; 95% CI, 1.76-2.07). There was no influence of family income, patient race, or patient sex on the occurrence of any of the 3 protocol-adherence problems. A similar analysis of psychologic characteristics of the child and family indicated (1) a 2-fold increase in the risk of a diary card problem with a 10% increase in the percentage of total commissions on the attention scale of the Gordon Diagnostic Study (OR, 2.18; 95% CI, 2.02-2.35), (2) a 2-fold decrease in the risk of an appointment problem with a 10-unit increase in the Child Manifest Anxiety Scale (OR, 0.46; 95% CI, 0.44-0.49), (3) a 2-fold decrease in risk of an ap