This study evaluated the nature and efficacy of compliance with emergency department (ED) patient referral recommendations. This was a prospective, nonrandomized, descriptive analysis of all ED patients referred mandatorily to an established urban hospital follow-up network. Compliance was measured by analysis of hospital records determined as appointment completion. Patient demographics, urgency of complaint, hospital relationship (new versus established), diagnosis (International Classification of Diseases-9CM), specialty, and method of payment, defined as clinic or private referral, were determined. Comparisons between groups used Fisher's exact test and chi 2 analysis (alpha = 0.05). There were 2,185 patients encountered with 1,443 (66%) discharged for referral, and an overall compliance rate of 27.8% (401 patients). Patients had a mean age of 36.9 years; 50.6% were male, 94.4% were established patients, 51.1% were clinic cases, and 96.7% had nonurgent complaints. Patients encountered had higher rates of compliance if female (33.9%), greater than 40 years of age (43.4%), with urgent complaints (46.8%), and if referred to private physicians (37.0%) (P less than .001). Compliance also correlated with the diagnosis of fracture (63.3%) or laceration (45.6%); and specialty referral to obstetrics-gynecology (28.4%) and general surgery (22.4%) consultants (P less than .01). Most patients demonstrate low compliance (28%) with follow-up recommendations, even with a directed ED referral system.