Background:
Patient recruitment for randomized clinical trials is often challenging, especially when the alternatives being studied may require significant changes or have large consequences for the patient. We describe our experience recruiting patients for a clinical trial comparing the Comprehensive Care Program (CCP), a care delivery model in which physicians focus their practice on caring for patients at risk of hospitalization to allow them to provide their patients with both inpatient and outpatient care to standard care by different doctors in the inpatient and outpatient settings. Eligible patients must have traditional Medicare and been hospitalized in the past year. Patients must also be open to the possibility of leaving their current primary care physician (PCP) to receive care from a CCP physician. To assess efficacy of recruitment strategies, we looked at the number of enrolled patients based on how they were identified and evaluated cost-effectiveness.
Methods:
Recruitment began in November 2012 with a target enrollment of 2000 patients in two years, an average of 2.74 patients per day. The primary source of enrolled patients in the first quarter was through physician referrals and screening for patients on the general medicine service at the University of Chicago Medical Center (UCMC). Strategies were refined based on simple predictive models of risk of hospitalization.
Results:
Among 1665 patients enrolled as of November, 2015, inpatient recruitment is the longest running strategy accounting for 476 patients or 0.43 patients per day. At this pace, it would have taken 12.74 years to reach our target enrollment. During the summer of 2013, a new strategy of approaching patients in the emergency department (ED) was implemented. This led to the highest yield of enrolled patients with 787 patients or 0.96 patients per day. ED recruitment reduced anticipated recruitment time to 3.94 years. Other strategies included physician referrals (227), recruiting patients of leaving house staff and faculty at UCMC (69), and community outreach efforts (39). These other strategies averaged 0.31 patients per day; this further reduced anticipated recruitment time to 3.22 years. The success of ED recruitment can be attributed to the high volume of eligible patients and patients who use the ED as a main source of care because they have no previous PCPs. Patients admitted from the ED then enrolled are counted as inpatient recruits. The low number of enrolled patients and the high labor requirements of community-based efforts made that approach less cost-effective than other strategies.
Conclusions:
Many strategies were employed over time to recruit patients for CCP, engaging diverse individuals from section chiefs to aldermen. Recruitment strategies were most successful and cost-effective within UCMC, making community-based recruitment lower priority. Despite that, our other strategies did add non-trivial numbers of patients to our internal recruitment efforts.