Objectives: To quantify patients’ preferences for attributes of pharmaceutical treatments for osteoarthritis (OA) pain and chronic low back pain (CLBP) in the United States, exploring preference heterogeneity. Methods: A discrete-choice experiment was administered online to respondents with a self-reported physician diagnosis of OA and/or CLBP with moderate-to-severe pain. Respondents were presented with a series of choices between two hypothetical treatments defined by six attributes: symptom control; treatment-related risks of severe joint problems, heart attack, and physical dependency; mode and frequency of administration; and cost. Sample preferences and subgroup analysis (SA) for 10 prespecified subgroup pairs defined by demographic/clinical characteristics and survey comprehension were estimated using a random-parameters logit model. These results were compared with those of a latent class analysis (LCA). Results: The survey was completed by 602 respondents: 201 with OA, 202 with CLBP, and 199 with both. Mean (standard deviation) age was 64 (10.8), and 59% were female. Symptom control, risk of physical dependency, and cost were most important and were statistically significantly more important than risk of heart attack, risk of severe joint problems, and mode and frequency of administration. SA found systematically different (P <0.05) preferences for only one subgroup (defined by the number of correct answers to the comprehension questions). LCA identified 3 classes: class 1 focused on avoiding risk of physical dependency (39.1%), class 2 on improvements in symptom control (35.0%), and class 3 on cost (25.9%). Respondents with opioid experience and respondents who incorrectly answered 3 or more comprehension questions were less likely to be in class 1 and class 2, respectively. Conclusions: SA and LCA suggest subgroups of patients have different preferences. LCA highlighted 1 patient characteristic explaining preference heterogeneity (opioid experience) not identified in SA. Discussion of risks and benefits of treatments should take differences in individual preferences into account.