OBJECTIVES: Multiple barriers to effective pain management are present in the nursing home setting. The purpose of this analysis was to determine the extent to which residents in pain declined to request pain medication from the staff, and the reasons provided by the residents to explain this behavior. DESIGN: Every 3 months, a 20% sample of residents in 12 nursing homes was administered a short pain interview, then observed for pain indicators. Medical records were reviewed at the same time for documentation about pain and its treatment. All residents were asked if they had pain (or a similar word) now or in the past 24 hours. They were also asked if they had pain but did not request pain medication. If affirmative, the resident was asked to provide up to three reasons for not requesting medication. SETTING: The study was conducted in 12 Colorado nursing homes, located in both urban and rural settings. PARTICIPANTS: A total of 2033 nursing home residents completed pain interviews and/or were observed for pain indicators by trained research assistants. These interviews took place before, during, and after implementation of an intervention to improve pain practices. MEASUREMENTS: A cognitive organizing structure was used to categorize resident responses into a coherent classification. Individual responses were assigned by team members to the appropriate category using a consensus process. The final classification scheme consisted of 10 categories of reasons why residents do not request pain medication. RESULTS: More than one-half of residents (59.5%) reporting pain in the past 24 hours did not request medication for that pain. Subjects in pain were most likely to state medication concerns or stoicism as the reasons for not requesting pain medication. Concerns about staff reactions to a request or perceptions that the staff was too busy were also mentioned frequently by the residents. Subgroup analyses suggested that residents in pain but not requesting pain medication were significantly more likely to be in rural rather than urban nursing homes (67.9% vs. 52.9%, P < or = .01), and white as compared to nonwhite ethnicity (60.6% vs. 52.1%, P < or = .05). They also tended to be older on average (80.4 +/- 12.1 years vs. 77.9 +/- 12.7 years, P < or = .01) than residents who did request pain medication. Finally, residents in pain but not requesting pain medication were significantly more likely to report having both continuous (c) and intermittent (i) pain (71.8% [c + i] vs. 61% [c] or 56.5% [i], P < or = 0.01). CONCLUSION: Interventions to reduce pain in nursing home residents need to be responsive to the concerns of the residents. It must be acknowledged that resident preferences and beliefs may lead to declined pain interventions regardless of the staff's motivation to make the resident more comfortable. Staff nurses also need to make a more concerted effort to systematically assess pain and offer pain medication to residents rather than rely on resident requests.