Background: While a scientifically sound study design is the basis for all clinical trials, activities such as patient recruitment present ongoing challenges. The availability of electronic data greatly enhances recruitment efforts, but also may bring new problems. The Journey for Control of Diabetes: the IDEA Study, a randomized clinical study conducted in two distant healthcare settings, utilizes an advanced electronic data system to identify and recruit eligible participants. We believe that communication of difficulties encountered could help researchers who face similar recruitment challenges.
Methods: The inclusion criteria included: Type 2 diabetes (DM), an A1C = 7 within 6 months of enrollment, and diabetes education-naïve within the last year. Preliminary estimates of patient availability were calculated with the following assumptions: 1) two ICD-9 codes for Type 2 DM (10,000 patients); 2) 52% with A1C = 7 (5,200); 3) 60% with no DM education in the last 1-2 years (3,120); 4) 15% excluded with Type 1 or gestational DM (2,652); and 5) 15% willing to participate (397). Recruitment of 311 subjects was estimated to take five months.
Results: Recruitment has been much slower than predicted (50% of subjects recruited at seven months). Recruitment results to date are: 1) 3,706 individuals identified as eligible using electronic data; 2) 156 enrolled; 3) 1,921 (50.4%) individuals not interested, 4) recent DM education (8.3%), and 5) Type 1 DM (2.1%). The electronic data on the pool of potential subjects is often out of date by the time the subject is screened and randomized. Refreshing the electronic data creates a moving target of potential candidates. A1C data and date of last diabetes education change at a rate faster than the recruitment mailing, screening and randomization process. Therefore, more time, programming, and staffing resources have been needed than originally anticipated to recruit study participants.
Conclusions: This analysis of recruitment data demonstrates that initial recruitment capability estimates did not account for the inefficiencies created by outdated data and refreshed electronic encounter data. Daily tracking of all criteria for database accuracy has been a difficult and time-intensive task.