No clinical tools currently exist to stratify patients' risks of patient-directed discharge (PDD). This study aims to identify trends and factors associated with PDD, representation, and readmission. This was an IRB-approved, single-centered, retrospective study. Patients aged > 18, admitted to medicine service, were included from January 1st through December 31st, 2019. Patients admitted to ICU or surgical services were excluded. Demographics, insurance information, medical history, social history, rates of events occurrences, and discharge disposition were obtained. Of the 16,889 encounters, there were 776 (4.6%) PDDs, 4312 (25.5%) representations, and 2924 (17.3%) readmissions. Of those who completed PDDs, 42.1% represented and 26.4% were readmitted. Male sex, age ≤ 45, insurance type, homelessness, and substance use disorders had higher rates of PDD (OR = 2.0; 4.2; 4.5; 6.2; 5.2; p < 0.0001, respectively). Patients with homelessness, substance use disorders, mental health disorders, or prior history of PDD were more likely to represent (OR = 3.6; 2.0; 2.0; 1.5; p < 0.0001, respectively) and be readmitted (OR = 2.2; 1.6; 1.9; 1.5; p < 0.0001, respectively). Patients aged 30-35 had the highest PDD rate at 16%, but this was not associated with representations or readmissions. Between July and September, the PDD rate peaked at 5.5% and similarly representation and readmission rates followed. The rates of subsequent readmissions after PDDs were nearly two-fold compared to non-PDD patients in later half of the year. 51% of all subsequent readmissions occur within 7 days of PDD, compared to 34% in the non-PDD group (OR = 2.0; p < 0.0001). Patients with primary diagnosis of abscess had 16% PDDs. Factors associated with PDD include male, younger age, insurance type, substance use, homelessness, and primary diagnosis of abscess. Factors associated with representation and readmission are homelessness, substance use disorders, mental health disorders, and prior history of PDD. Further research is needed to develop a risk stratification tool to identify at-risk patients.