This is the first study that has explored GPs’ assessment and management of pain in people with dementia. Our findings suggest that GPs are confident in many aspects of assessing pain in people with dementia such as the value of observing behavioural and physiological indicators and the importance of surrogate reports but that they are challenged by many aspects of assessing and managing pain in dementia. The majority of GPs surveyed believed that a person with dementia cannot self-report pain and the vast majority of GPs were unfamiliar with dementia-specific pain assessment tools. In the absence of either a self-report or a standardized observational tool to assess pain, the responding GPs appeared to rely significantly on surrogate reports from family members and nursing home staff when assessing pain in dementia. Although the majority of responding GPs welcomed the idea of guidance in the area of pain assessment and management in dementia, in the free-text comments many questioned the value of a standardized, observational pain tool. Furthermore, when managing pain in people with dementia the GPs were particularly uncertain about the role of opioid medication and the consequences of the use of opioids in people with dementia. While the majority of GPs agreed that in order to achieve optimum pain relief analgesia should be prescribed regularly, over a quarter of GPs surveyed did not agree with this statement. The inference being that these GPs are favouring ‘as required’ analgesic medication, a sub-optimal method of pain control. Although we did find that GPs with a larger nursing home commitment were more likely to be familiar with dementia-specific pain assessment tools, in general, the experience level of the GP was not associated with increased levels of knowledge or a more positive attitude towards pain assessment and management in dementia.
Comparison with existing literature
Our findings indicate that GPs’ value good communication with family members and nursing home staff when managing pain in people with dementia. This finding is similar to previous studies with community pharmacists [
10] and nurses [
35]. In previous research on GPs’ educational needs in dementia [
29] both GPs and family caregivers emphasized the importance of good channels of communication in dementia care. In our study, the GP respondents also emphasized the importance of a report from nursing staff or family carer when assessing pain in the qualitative free-text responses. In our study nursing staff were seen by the GPs to facilitate pain assessment. This is in contrast to findings from a previous study that examined nurses knowledge and attitudes to pain management in dementia where nurses identified a lack of GP support as a barrier to successful pain management [
35]. Previous research with GPs found that consistency of care was an important factor in improving relationships between GPs and nursing staff [
34]. In that study structured visits by the GP to the nursing home were seen to facilitate the provision of continuity of care and led to good communication channels between the GP and nurses [
34]. Effective communication between nursing staff and GPs is an essential component of any optimisation of pain assessment and management in nursing home settings.
The majority of respondents in our study believed that people with dementia could not accurately provide a self-report of pain. However, self-reporting of pain is considered the gold standard method of pain assessment [
39] and can be a reliable way of assessing pain in people with dementia [
39‐
41]. Best practice recommendations advise that where possible attempts should always be made to elicit self-reports of pain from the person with dementia [
39]. Although in the very advanced stages of dementia many individuals may be unable to self-report pain [
14], people with mild-moderate [
24,
42] and in some cases severe dementia [
43] have been found to provide valid self-reports of pain. Our finding echoes previous research with nursing home managers which found that only 8.3% of respondents felt that people with dementia could self-report pain [
44]. The large majority of GP respondents agreed with the statement that
‘a person with dementia is not able to provide a self-report of pain’. This finding could mean that an attempt is not being made to elicit a self-report from a person with dementia. However, this needs to be explored further, ideally with qualitative research, to establish the impact this attitude has on how GPs assess pain in people with dementia.
Nearly all GP respondents agreed that patient observation was a critical part of pain assessment in dementia. Despite this belief the vast majority of respondents were not using any validated, standardised, observational approach. Observational methods are central to clinical assessment, especially when a person lacks the ability to self-report, however, there is a risk of observer bias if there is no standardized approach to the observation [
40]. A large array of pain assessment tools exist for use in people with dementia - a recent systematic review of pain assessment tools included 28 such tools [
45]. However, the vast majority of GP respondents were unaware of these tools. Similar to previous findings from a study with community pharmacists [
10], the more experience the GP had with dementia the more likely he or she was to be aware of these tools. The lack of awareness of dementia-specific pain assessment tools is a particularly noteworthy finding since the majority of respondents thought that having such a pain assessment tool would be helpful. This highlights an incongruity between research in this area and real-life clinical practice. These tools appear to be rarely used in general practice.
The responding GPs lack of familiarity with pain assessment tools may be surprising to researchers in the area but may be unsurprising to front-line GPs. GPs do not readily embrace assessment tools [
46,
47]. They are inductively trained to rely on their clinical skills. This is in contrast to nursing staff who are specifically trained to use and rely on assessment tools. In the free-text responses some GPs feared the additional workload such tools could bring and questioned what clinical value they would add. Similarly, recent qualitative research exploring GPs’, hospital physicians’ and nurses’ perspectives on the use of observational pain tools in people with dementia identified a number of barriers to using observational pain tools, one being the perceived lack of value in using them [
48]. Furthermore, participants in that qualitative study described using the pain tools to comply with local recommendations but not actually using the results to inform treatment decisions [
48]. This echoes some of the concerns raised by participants in our study that a pain tool would become another source of paper-work rather than a tool to aid clinical decision making. GPs are not usually the healthcare professional tasked with completing a pain assessment tool; in a nursing home setting that role would typically fall to the nursing staff. GPs do still need to be aware of these tools in order to interpret the findings and generate appropriate management plans in discussions with nursing staff. However, clinicians and nurses can find the results of pain assessment tools difficult to interpret [
49]. Implementing these observational pain assessment tools in isolation, without adequate guidance on how to interpret the results, will not lead to improved treatment of pain [
14]. The tools in themselves will not result in improved care unless they are combined with guidance that will help clinicians to translate a pain score into an appropriate treatment plan.
Respondents appeared unsure about the safety of opioid analgesics in people with dementia. This is similar to previous research conducted with nurses [
35], nurse managers [
44] and community pharmacists [
10] all of whom had similar concerns regarding the safety of prescribing opioids to people with dementia. Guidelines on pain management in the older adult do recommend considering opioid analgesics for patients with moderate to severe pain, particularly if the pain is causing functional impairment or reducing quality of life [
3]. The uncertainty the GP respondents felt regarding the safety of opioid medication in dementia is perhaps understandable; prescribing opioids to an older adult is not without its complexities. Age is a significant predictor of opioid related harm [
50,
51]. Adverse effects of opioids can include respiratory depression, sedation, constipation, nausea and dizziness [
52]. Many of these adverse effects increase with age [
53] and frailty [
51] both of which are associated with dementia. Furthermore, these adverse effects can be particularly problematic to identify in a person with advanced dementia because of their reduced ability to communicate. However, when managing pain in people with dementia the adverse effects of opioids needs to be weighed up against the harmful effects of undertreating pain. Although a narrow majority agreed with the statement that opioids were ‘
safe in people with dementia’, a larger majority of respondents agreed with the statement that there is
‘a greater risk of side effects from opioid analgesics when used in people with dementia’. Many respondents appeared to feel that, despite the increased risks, the use of opioids in people with dementia was still “safe”. How this belief influences the GP’s prescribing is unclear. Previous research suggested that people with dementia may receive less opioids [
54]. A qualitative study which examined GPs’ perspectives on prescribing opioids for chronic pain found that fear of causing harm was a barrier to prescribing opioids to the older adult [
55]. Despite this a recent systematic review identified that, internationally, prescribing of opioids to nursing home residents has increased over time [
56]. Similarly, a large Danish study found that buprenorphine and fentanyl patches were more commonly prescribed to people with dementia [
57]. It is possible that these opioid patches are perceived as having less side effects than oral opioids or as being more tolerated by people with dementia. However, these transdermal opioid patches typically contain a stronger opioid dose than oral opioids, and since the adverse effects of opioids are dose related these patches are likely to result in more, not less, side effects. Prescribers’ perspectives of the role of different opioid medications in the management of pain in dementia is an important area that could be explored in future research.
Although optimal treatment of pain is achieved when analgesics are given on a regular basis [
58], over a quarter of respondents in our study did not agree with this statement. A similar finding was reported in previous research with nurses [
44] where 20% of respondents either agreed, or neither agreed nor disagreed, with a statement that ‘optimal pain treatment of pain relief is achieved when analgesics are given in a PRN (or as required) way’. If the value of regular administration of analgesia is not recognized this could result in the prescribing of analgesics in the less-effective ‘when required’ way. Research conducted in nursing homes in Northern Ireland in 2015 found that in the majority of residents with dementia, analgesic medication was prescribed ‘when required’ and not regularly [
24]. Likewise, research from the U.S. found that cognitively impaired nursing home residents were less likely than their cognitively intact peers to be prescribed regular, scheduled analgesia and were more likely to be prescribed analgesic medication in an ‘as required’ way [
18]. To receive this ‘as required’ medication a person would either need to ask the nurse for pain relief or the nurse would need to identify that the person was in pain and give them pain relief [
18]. In the case of a person with advanced dementia neither of these situations are very likely to occur. In view of their cognitive and communication difficulties a person with advanced dementia is unlikely to self-request analgesia and we know from existing evidence that there are a number of barriers to nurses identifying and initiating pain relief in a person with dementia [
35,
44].
Areas for future research & implications for policy & practice
This study highlights several areas of GPs assessment and management of pain in dementia that could be explored with future research. There were some findings that warrant further exploration in future research such as the use of ‘as required’ medication and the belief that people with dementia cannot self-report pain. Future qualitative research with GPs in this area would help gain a deeper understanding of the context and nuances that are involved in this complex area. It would be important to ascertain how GPs’ knowledge of and attitudes towards pain in people with dementia impacts on their actual prescribing of analgesia to people with dementia. A quantitative study of current prescribing of analgesia to people with dementia both in nursing homes and in the community would be an important complementary study. These studies would lead to a more extensive understanding of the problem and would help to inform the development of effective interventions to improve the management of pain in dementia in primary care settings.
There is a role for educational interventions for GPs that focus on pain assessment and management in dementia. The results of this current study will inform the ongoing, national roll-out of educational interventions for GPs in dementia care that have being developed by our research team [
59,
60] as part of the implementation of the Irish National Dementia strategy [
61]. Previous educational interventions for GPs in the area of pain management in dementia have used teleconferencing technology based on the Project ECHO© model and have been found to improve healthcare professionals’ knowledge and self-efficacy of pain assessment and management in advanced dementia [
62]. Such educational interventions would be particularly acceptable to general practitioners as they eliminate the need for travel and don’t require any prolonged periods away from practice. Another important finding was in relation to educational initiatives was that GPs see nurses as facilitators to optimum pain management, whereas in previous studies [
35,
44] nurses identified GPs as barriers to optimum pain management. This highlights a communication gap between these two professional groups, both of whom play a pivotal and complementary role in the assessment and management of pain in dementia. This finding underlines the importance of inter-professional educational initiatives in this area. Palliative care is another professional group that have a significant role to play in the management of pain in dementia, particularly in complex cases, yet we know that many people with dementia are not routinely assessed to determine their palliative care needs [
63]. An inter-professional educational approach including all the relevant professional groups could help improve communication and bridge professional divides [
64], which would play an important role in improving the care provided to people with dementia who are living with chronic pain.
Although relevant and important we know that educational interventions alone have limited effect in changing GPs’ behaviour in dementia care [
65,
66]. To effectively improve GPs’ performance in dementia care, education needs to be combined with adequate reimbursement and organisational incentives [
67]. A previous systematic review highlighted how time constraints and inadequate remuneration act as barriers to optimum diagnosis and management of dementia in primary care [
68]. In Ireland general practice receives only 4.5% of the overall health budget, significantly lower than other European countries [
69]. Additionally, recent government-led austerity cuts in Ireland have dramatically cut funding to general practice, these cuts have particularly affected general practitioners providing care to nursing home residents. In our study the dissatisfaction of GPs with the current under-resourcing of dementia care was evident in the free-text responses. Many GPs discussed the challenges of providing optimum care to people with dementia in the context of current resource limitations, reporting that they can no longer provide care to nursing home residents. Like many other European countries [
70‐
72], Irish general practice is currently facing a recruitment crisis [
73]. This recruitment crisis is particularly affecting rural general practice, therefore, it was notable that rural based GPs were significantly more likely to provide care to nursing home residents than non-rural GPs. In the face of this recruitment crisis and in the context of current inadequate reimbursement the future of GP led nursing home care is uncertain. Future policy needs to focus on adequate resourcing of dementia care in both community and nursing home settings.
The study results have several immediate practical implications for GPs caring for people with dementia. The implications for GPs assessing pain in people with dementia include; being more aware of the increased risk of pain in people with dementia, considering pain as a potentially reversible trigger for BPSD, providing people with dementia an opportunity to self-report pain and familiarising themselves with the pain assessment tools that may already be in place in the nursing homes they attend in order to facilitate more effective communication with the nursing home staff. From a pain management perspective one practical implication for GPs is that if pain is identified, or suspected, then it is important that the person’s pain is not under-treated. Inappropriate prescribing in people with dementia does not just mean over-prescribing. It also pertains to under-prescribing of appropriate medications that can improve comfort and overall quality of life.
Strengths and limitations
Although a 49% response rate is modest, it is typical of postal surveys with this professional group [
74]. Our response rate is similar to the response rate in a previous national study on GPs attitudes to diagnosis in dementia [
75] and significantly higher than a recent online survey of Irish GPs referral patterns in dementia care [
76]. To adequately power this study a sample size of 175 respondents was required. Therefore, with 157 respondents the study was marginally underpowered. Furthermore, this study was a census study of a specific geographical area in the southern region of the Republic of Ireland. This may affect the generalisability of the study, however, the respondents’ demographic characteristics, in terms of years of experience and practice location, are representative of Irish GPs nationally [
77]. A large proportion of respondents had a nursing home commitment. There is currently no Irish data available on the number of GPs who attend nursing homes, therefore, it is difficult to ascertain whether this represents a respondent bias. It is possible that GPs with a nursing home commitment were more likely to respond to the survey as they may have been more interested in the research topic. Finally, this was a cross-sectional survey of self-reported knowledge and there is evidence to suggest that when self-reporting physicians may underestimate their knowledge in an area [
78]. The nuances of clinically complex areas such as this are difficulty to fully address with a single study that rely primarily on self-reporting measures. However, since there is very little research exploring GPs experiences of managing pain in people with dementia our chosen methodological approach is a necessary and reasonable place to start.