Using the minimum data set to select nursing home residents for interview about pain

OBJECTIVES: To determine how many nursing home residents can provide stable responses to a simple pain interview and whether a Minimum Data Set (MDS) cognitive performance measure can be used to identify these residents. DESIGN: Cross-sectional descriptive study. SETTING: Thirty-three community-based nursing homes. PARTICIPANTS: Eight hundred ninety-five nursing home residents. MEASUREMENTS: Resident completion rate, stability, and interrater reliability of a four-item interview derived from the Geriatric Pain Measure were calculated. Demographic data and MDS items concerning pain and memory were obtained from medical records. RESULTS: Overall, 835 residents were able to answer all four interview questions. At the lowest MDS recall score of 0, 52.7% of residents were able to complete all questions. All residents able to respond to the interview achieved high stability (kappa=0.633, P<.001). Overall, 62.8% of residents with daily pain or activity-limiting pain on interview did not have daily or moderate to severe pain recorded on the MDS. Residents who had lower MDS recall scores were significantly less likely (P=.004) to be appropriately identified on the MDS. CONCLUSION: Residents with a low MDS recall score were significantly less likely to be noted on the MDS as having serious pain despite being able to complete a simple yes/no interview about pain in a stable fashion. Nursing staff should attempt to ask all residents direct questions about pain. Surveyors may restrict direct questioning to those residents with an MDS recall score of 1 or higher if time is an important consideration. Adjustment for MDS-derived prevalence of pain based on residents' cognitive status is questionable.