The findings have led experts from The Netherlands to call for nurses to be given more education about how to assess and treat chronic pain. They would also like to see relatives being given more information about pain and for both parties to use other methods, like encouraging greater mobility and providing soothing massages, to alleviate pain.
Researchers led by the Pain Expertise Centre at the Erasmus Medical Center studied 174 nursing home residents - 124 who had cognitive impairments and 50 did not. They also spoke to 171 nurses and 122 relatives.
Six nursing homes took part in the five-year study, with the researchers speaking to the patient, someone responsible for care and, wherever possible, a relative who had regular contact with the patient. In some cases nurses reported back on more than one patient.
Patients were included if they had a pain rating of more than four out of ten, assessed by themselves - if they were not cognitively impaired - or a staff nurse if they were. The majority of the 110 women and 64 men, who had an average age of 82, experienced pain as a result of musculoskeletal and circulatory problems.
Previous studies have shown that some people with mild or moderate cognitive impairment are still able to use simple zero to ten scales, where zero is no pain at all and ten is the worse pain imaginable.
In this study, all of the 50 patients in the non-impaired group were able to report pain levels in the previous week, together with 19 in the impaired group, making a total of 69. Fifteen patients in the impaired group were able to report pain levels at rest, making a total of 65.
"When the team interviewed the residents without cognitive impairments they found that all of them reported pain in the last week, but that only 89 per cent of the caregivers and 67 per cent of the relatives were aware of that pain" says Dr Rhodee van Herk. "However, if they were aware that the patient had experienced pain, the nurses and relatives gave it a median score of six out of ten, with the same score reported by the patients."
Nurses and relatives were less unaware of pain levels when the patient was at rest. They gave their pain levels a median score of zero, compared with the patients, who gave it a median score of four out of ten. However relatives were more aware of pain issues than nurses, with their median scores ranging from zero to five, compared with nurses, who reported a median score of zero to two.
In general, there was more agreement between residents and relatives on pain levels than between relatives and nurses. The only exception was pain at rest in the impaired group, when relatives and nurses were in greater agreement.
Eight-three per cent of nurses and 65 per cent of relatives were certain about the level of pain the patient had experienced in the preceding week and 83 per cent of caregivers and 58 per cent of relatives were certain about pain at rest.
"Our study shows that nurses and relatives find it hard to accurately assess pain in nursing home residents, especially if the resident has a cognitive impairment, such as dementia, or is unable to speak" concludes Dr van Herk.
"Pain seemed to differ, not only on an individual basis but also in different daily situations. It is clear that pain at rest is a particular issue that needs addressing as residents rated this much higher than caregivers and relatives.
"Using a simple pain intensity scale, like the zero to ten scale employed in our study, is clearly not enough. We would like to see nurses use a combination of the existing pain intensity scale, together with multi-dimensional pain observations scales to judge how much discomfort a patient is experiencing.
"We are also keen to ensure that relatives receive more information on chronic pain and that both relatives and nurses explore pain relief methods that don't always rely on drugs.
"These can include massage, applying warmth, encouraging residents to be more mobile and distractions such as music and storytelling."