For cognitively intact older adults, pain assessment relies on the reliability and validity of self-report1
and behavioural observation.56
Pain expression in those with dementia, however, may pose threats to the validity of traditional approaches to pain assessment. For example, patients with cognitive impairment generally report less pain than cognitively intact older adults even though there is no evidence that cognitive impairment reduces the ability to feel painful stimuli. 84 112
On the other hand, older adults with dementia may display behavioural indicators of pain (e.g. bracing) in the absence of self-reported pain.97
When evaluating pain in older adults with dementia, experts recommend incorporating several methods: self-report, proxy report, and behavioural scales.1
Pain measurement in older adults with dementia should take into account the severity of cognitive loss.93
Evidence for the reliability of current pain self-report in those with mild to moderate dementia is strong. 15 105 106
The reliability of historical pain reporting has not been evaluated in those with dementia. As cognitive decline progresses to more advanced stages that are associated with further deterioration of memory and verbal abilities, the utility of self-report scales becomes more limited and proxy pain assessments are increasingly relied upon.1
Professional caregivers (nurses or nursing assistants) tend to underestimate the presence of pain,30
but not at higher levels of pain intensity.98
In contrast, family caregivers tend to report more pain than that reported by the cognitively impaired individual. The accuracy of proxy pain assessment also may be impacted by dementia-associated behavioural changes. For example, high levels of agitation are associated with increased likelihood of pain rating disagreement between patients and caregivers.98
Formal pain behaviour observation instruments have been developed to assist with pain assessment in those unable to report pain.43
Domains frequently included are changes in facial expression, verbalizations or vocalizations, body movements (guarding, pacing, rocking, and rigid tense body posture), changes in interpersonal interactions (withdrawn, disruptive, aggressive), changes in activity patterns or routines (changes in appetite, sleep, or routines), and mental status changes (confusion and crying).1
No one instrument has sufficiently developed psychometric properties to be recommended for routine clinical use. 43 45
Importantly, the specificity of the behaviours identified using these scales for identifying pain has not been evaluated. For example, psychological symptoms (e.g. depression, anxiety, and fear) and unmet physical needs (e.g. hunger, social isolation, and soiled diaper) could cause behavioural expressions that might be misconstrued as pain.
Although behavioural changes such as altered movement patterns (e.g. bracing, guarding) are often relied upon as indicators of pain, these changes may occur as part of dementia itself. For example, progressive AD may be accompanied by parkinsonian rigidity, spasticity, and spontaneous non-startle myoclonic jerks.117
Rigidity is also a core feature of Lewy body dementia.72
Vascular dementia may be accompanied by a variety of cognitive, behavioural, and neuromuscular changes depending on the location and degree of neuronal injury or loss.9
It is not difficult to imagine, therefore, that these changes might impact the specificity of behavioural pain ratings.
We recently evaluated the validity of traditional pain behaviours (guarding, bracing, rubbing, grimacing, and sighing) by examining if pain status and/or cognitive status were independently associated with the frequency of observed behaviours. The number of pain behaviours was recorded as participants completed a structured protocol that simulated activities of daily living.111
The two pain groups were CLBP and pain-free and the two cognitive status groups were mild to moderate dementia and cognitively intact. Participants with CLBP, independent of cognitive status, displayed significantly more grimacing and guarding behaviours than the pain-free group. Participants with dementia, independent of their pain status, displayed significantly more guarding, bracing, and rubbing behaviours than cognitively intact participants. It is noteworthy that rubbing is considered a stereotypical movement in persons with dementia.80
Although this behaviour is more common in persons with frontotemporal dementia, it does occur in other more common neurodegenerative diseases such as AD.74
Thus persons with dementia may exhibit behaviours that could just as easily be related to pain as to the underlying neurodegenerative disease itself.