Correspondence of Verbal Descriptor and Numeric Rating Scales for Pain Intensity
An Item Response Theory Calibration
Published In: The Journals of Gerontology, v. 65A, no. 7, July 2010, p. 778-785
Posted on RAND.org on December 31, 2009
BACKGROUND: Assessing pain intensity in older adults is critical and challenging. There is debate about the most effective way to ask older adults to describe their pain severity, and clinicians vary in their preferred approaches, making comparison of pain intensity scores across settings difficult. METHODS: A total of 3,676 residents from 71 community nursing homes across eight states were asked about pain presence. The 1,960 residents who reported pain within the past 5 days (53% of total, 70% female; age: M = 77.9, SD = 12.4) were included in analyses. Those who reported pain were also asked to provide a rating of pain intensity using either a verbal descriptor scale (VDS; mild, moderate, severe, and very severe and horrible), a numeric rating scale (NRS; 0 = no pain to 10 = worst pain imaginable), or both. We used item response theory (IRT) methods to identify the correspondence between the VDS and the NRS response options by estimating item parameters for these and five additional pain items. RESULTS: The sample reported moderate amounts of pain on average. Examination of the IRT location parameters for the pain intensity items indicated the following approximate correspondence: VDS mild NRS 1-4, VDS moderate NRS 5-7, VDS severe NRS 8-9, and VDS very severe, horrible NRS 10. CONCLUSION: This IRT calibration provides a crosswalk between the two response scales so that either can be used in practice depending on the preference of the clinician and respondent.