Psychometric evaluation of observational tools in the assessment of pain in older adults with severe dementia

Abstract

Specialized observational tools that assess pain in specific patient populations (e.g., older adults living with dementia, people with developmental disabilities) or that assess specific types of pain in populations with limited ability to communicate (e.g., orofacial pain) are available and routinely used in clinical practice and research settings. This investigation involved the comparison of various observational pain assessment tools in the assessment of pain in people with dementia. Despite available psychometric evaluations, it has not been demonstrated that observational pain assessment tools, specifically developed for patients with moderate to severe dementia, are better equipped to assess various types of pain in this population than observational tools developed for other populations (e.g., adults with intellectual disabilities), or tools developed to assess a specific type of pain (e.g., orofacial pain). In the first part of this investigation, tools developed to assess pain in people with dementia were compared to a tool developed for other populations with limited ability to communicate. The ability of an orofacial painspecific tool to detect general types of pain was also assessed. Finally, the validity and reliability of a newly developed ‘meta-tool’, the Pain Assessment in Impaired Cognition scale (PAIC15), was also compared to that of well-established tools for pain in dementia. The second part of this investigation focused on whether university students who are not trained in the health professions could use these tools (developed specifically for use by healthcare staff) validly and reliably. Results demonstrated that all observational pain assessment tools could discriminate painful from non-painful situations when applied to a dementia sample irrespective of whether they were developed specifically for people with dementia. The general pain tools and the orofacial pain tool performed comparably. Convergent and discriminant validity analyses indicated that all pain tools correlated moderately to highly with one another and showed low correlations with a measure of agitation. The psychometric properties of the relatively less researched PAIC15 were also comparable to that of the well-established pain measures but were not found to be superior. Untrained observers were also able to use the tools under investigation in a valid fashion suggesting that, with professional guidance and under professional supervision, lay people can gather useful standardized information about pain in a loved one with dementia and share this information with health care providers. The results of this investigation do not support the need for specialized pain assessment tools for different adult populations with cognitive impairments.