Pain assessment for cognitively impaired older adults: do items of available observer tools reflect pain-specific responses?

Background A number of observational tools are available to assess pain in cognitively impaired older adults, however, none of them can yet be regarded as a "gold standard". An international research initiative has created a meta-tool compiling the facial, vocalization and body movement items of the majority of available tools. Objective of this study was to investigate the pain specificity and the validity of these items. Method N = 34 older adults with or without cognitive impairment were videotaped in three different conditions (one reference, two painful conditions) in their nursing homes. They were further asked to self-report their pain in each condition. The occurrence of non-verbal behaviours was coded as present or absent using the items of the meta-tool. Results The majority of non-verbal behaviours was not pain sensitive as they occurred less than three times across participants and conditions. Of the remaining items, two facial items ("pained expression" and "raising upper lip"), one vocalization item ("using pain-related words") and one body movement item ("guarding") were found to be pain specific and valid. One additional item, the vocalization item "gasping", was pain specific, but not associated with pain self-report, and three additional items, the facial items "frowning" and "narrowing eyes" and the vocalization item "mumbling" were correlated with pain self-report but did not help to separate pain from non-pain conditions. Conclusions Systematic evaluation of items of existing observational pain assessment tools under naturalistic conditions seems a promising approach in the process of further investigating and improving tools. Significance Only few items stemming from observational pain assessment tools were found to be pain sensitive and specific as well as valid in this study. The investigation of existing tools not only on tool but additionally on item-level can provide helpful insights and thereby can help to improve the original tools and establish a gold standard for nonverbal pain assessment in older adults with cognitive impairments.