OBJECTIVES: Hospital-at-home programs for COPD exacerbations aim to provide care efficiently by shortening or avoiding hospital admissions.  The objective was to quantify Dutch patient and informal caregiver preferences for different aspects of hospital-at-home. METHODS: In a discrete-choice experiment, respondents made 14 choices between regular hospital admission (7 days) and two programs in which 3 hospital days were followed by a 4-day treatment at home. The home treatment was described by a set of attributes (see results). Hospital treatment was constant across choice sets.  Respondents were patients and their informal care givers who participated in an RCT on the cost-effectiveness of hospital-at-home versus regular hospital care. The data were analyzed in latent-class conditional logit models, which allowed for heterogeneity across groups. RESULTS: A total of 202 questionnaires were returned. 25% of patients and caregivers always opted for hospital treatment, 46% always chose hospital-at-home. For both groups, the best fit was provided by a model with four latent classes, depending on preference for  hospital and caregiver burden. All attributes had the expected sign and a significant effect on choices, except for number of home visits. Attribute levels with the strongest impact were hospital preference (for patients, coefficients (depending on class): -5.62 to +3.3) ,  a  5h/day caregiver burden (-3.5 to -0.11) and co-payment of €100 (1.11).  Also influential were specialized training for the homecare nurse (0.52), visits by many different nurses (-0.43), high readmission risk (-0.41), GP instead of hospital as contact for emergencies (-0.63),  €50 co-payment  (-0.48), 3h/day caregiver burden (-0.32), medium readmission risk (-0.24). Results were similar for informal care givers. CONCLUSIONS: A considerable proportion of patients and caregivers have a fixed preference for either admission or hospital-at-home, regardless of the specifics of the program.  When choosing between hospital-at home programs, co-payments and the burden on informal caregivers are the principal attributes.