Over 100 million Americans suffer from chronic pain resulting in annual cost of roughly $635 billion. Limited treatments are available for this widespread disease. The data supporting these treatments lack generalizability to patients with more serious medical and psychological comorbidities who are often excluded from explanatory efficacy trials. Our work aims to integrate randomized comparative effectiveness research with patient care. The investigators will randomize the patients and collect data using an open-source learning healthcare system already in use in our department to monitor patients' progress: Collaborative Health Outcomes Information Registry (CHOIR). Collaborative Health Outcomes Information Registry uses the National Institute of Health Patient Reported Outcomes Measurement Information System item banks for comparative metrics through computer adaptive testing. The investigators will leverage the advantage of this novel system to compare effectiveness of duloxetine and desipramine in decreasing pain in patients with chronic pain. The investigators will also compare adherence of patients to these two commonly used medications over a period of six months. This will evaluate feasibility of integrating randomized comparative effectiveness research with patient care in subspecialty clinics. Collaborative Health Outcomes Information Registry can then be applied for numerous future trials to advance our knowledge in perioperative and pain medicine.