The impact of a pain assessment intervention on pain score and analgesic use in older nursing home residents with severe dementia: A cluster randomised controlled trial

International Journal of Nursing Studies 84 (2018) 52–60

Contents lists available at ScienceDirect

International Journal of Nursing Studies journal homepage: www.elsevier.com/locate/ijns

The impact of a pain assessment intervention on pain score and analgesic use in older nursing home residents with severe dementia: A cluster randomised controlled trial

T

⁎

Hanne Marie Rostada, , Inger Utnea, Ellen Karine Grova, Milada Cvancarova Småstuena, Martine Putsb, Liv Halvorsruda a b

OsloMet – Oslo Metropolitan University, Oslo, Norway Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, Toronto, Canada

A R T I C LE I N FO

A B S T R A C T

Keywords: Aged Aged 80 and over Dementia Frail elderly Intervention Long-term care Nursing home Pain Pain assessment Pain management

Background: Pain is highly prevalent in older adults, especially those in institutional settings such as nursing homes. The presence of dementia may increase the risk of underdiagnosed and undertreated pain. Pain assessment tools are not regularly used in clinical practice, however, there are indications that the regular use of pain assessments tools may influence the recognition of pain by nursing staff and thereby affect pain management. Objectives: To assess whether regular pain assessment using a pain assessment tool is associated with changes in i) pain scores and ii) analgesic use in nursing home residents with severe dementia. Design: Cluster-randomised controlled trial. Setting: The study was conducted in 16 nursing homes in four counties in Norway. Participants: A total of 112 nursing home residents aged 65 years and older with dementia who lacked the capacity for self-reporting pain or were non-verbal. Methods: The experimental group were regularly assessed pain with a standardised pain scale (the Doloplus-2) twice a week for a 12-week intervention period. The control group received usual care. The primary outcome was pain score measured with the Doloplus-2, and the secondary outcome was analgesic use (oral morphine equivalents and milligram/day paracetamol). Data on the outcomes were collected at baseline and at the end of week 12. The nursing staff in both the experimental and the control groups received training to collect the data. Linear mixed models were used to assess possible between-group difference over time. Results: No overall effect of regular pain assessment was found on pain score or analgesic use. The mean score of Doloplus-2 and analgesic use remained unchanged and above the established cut-off in both groups. Conclusion: The current intervention did not change analgesic use or pain score compared with the control condition. However, there is not sufficient evidence to conclude that regular pain assessment using a pain assessment tool is not clinically relevant. Furthermore, our results indicated that pain continued to be inadequately treated in nursing home residents with severe dementia. Therefore, further research on how standardised pain assessment can be used to support effective pain management in this population is needed.

What is already known about the topic?

• Pain is a frequently reported symptom in older adults with de• • ⁎

mentia, and research has shown that it is often ineffectively managed. Pain assessment tools have the potential to influence the recognition of pain by nursing staff and thereby affect pain management. In general, pain assessment tools are not regularly used in clinical

practice. What this paper adds

• Regular pain assessment using a pain assessment tool did not affect •

the pain score or analgesic use of nursing home residents with severe dementia. The mean pain score and analgesic use remained unchanged in both

Corresponding author at: OsloMet – Oslo Metropolitan University, P.O. Box 4, St. Olavs Plass, N-0130 Oslo, Norway. E-mail address: [email protected] (H.M. Rostad).

https://doi.org/10.1016/j.ijnurstu.2018.04.017 Received 9 January 2018; Received in revised form 8 April 2018; Accepted 26 April 2018 0020-7489/ © 2018 Elsevier Ltd. All rights reserved.

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

knowledge of and focus on how to recognise pain, which in turn support clinical decision-making affecting analgesic administration and subsequently lowering pain scores. Thus, this study aimed to assess the effectiveness of regular pain assessment on analgesic use and pain score in NHRs with severe dementia. Specifically, the following research questions were formulated:

groups.

• The pain scores were above the established cut-off in both groups; •

this result indicates insufficient pain management in nursing home residents with severe dementia. Further research is needed on how pain assessment tools can be implemented to support pain management.

1. Introduction

1) Are there differences in pain score between the experimental and the control group? 2) Are there differences in analgesic use between the experimental and the control group?

Many studies have reported alarmingly high pain prevalence in older nursing home residents (NHR) with dementia, estimating that ≥50% experience pain (Zwakhalen et al., 2009; Torvik et al., 2009; Björk et al., 2016; Rostad et al., 2017a). The estimates vary considerably because of patients’ characteristics, methodological differences and assessment tools used. Evidence suggests that pain remains under-assessed, under-diagnosed and inadequately managed in this vulnerable population (Gibson and Lussier, 2012; Hadjistavropoulos et al., 2014; Achterberg et al., 2013). The presence of dementia may increase the risk of under-diagnosis and under-treatment of pain, as the reduced ability to verbally communicate may cause healthcare professionals to wrongly assume that pain is not present (Hadjistavropoulos et al., 2014). Pain assessment is the first step to adequate pain management (Hadjistavropoulos et al., 2014; Achterberg et al., 2013; Torvik et al., 2015; McAuliffe et al., 2012), but assessing pain in this population remains challenging because of patients’ impaired memory, changes in cognitive processing and a reduced ability to communicate verbally (Achterberg et al., 2013; Lichtner et al., 2014). The other barriers in establishing good pain management practice for patients with dementia are inaccurate beliefs about pain management, poor knowledge and training of staff (Achterberg et al., 2013) and difficulties with, and lack of, pain assessment in this population (Achterberg et al., 2013; Martin et al., 2005; McAuliffe et al., 2009; Kaasalainen et al., 2007). Standardised assessment tools have the potential to structure documentation, guide clinical decision making and facilitate effective clinical communication of patient needs. Over the past decade, a number of tools have been developed to assess pain in non-verbal older adults (Lichtner et al., 2014; Herr et al., 2006). In general, pain assessment tools are not regularly used in clinical practice (Hadjistavropoulos et al., 2014; Torvik et al., 2015; Lillekroken and Slettebø, 2013). Previous research found that pain assessment in older adults with dementia usually depends on the subjective impression of healthcare professionals (McAuliffe et al., 2009; Zwakhalen et al., 2007a). A few studies have investigated the impact of regular pain assessment on pain score and pain management strategies (Fuchs-Lacelle et al., 2008; Zwakhalen et al., 2012; Monacelli et al., 2013; Ando et al., 2016) with differing results. The regular use of a pain assessment tools resulted in increased usage of pro re nata (PRN) analgesics (FuchsLacelle et al., 2008), analgesic therapy in general (Monacelli et al., 2013; Ando et al., 2016) and a reduced pain score (Fuchs-Lacelle et al., 2008; Monacelli et al., 2013; Ando et al., 2016). Other studies demonstrated that the regular use of pain assessment tools did not result in the increased use of regularly scheduled analgesic (Fuchs-Lacelle et al., 2008) or in the frequent use of pain management interventions (Zwakhalen et al., 2012). Therefore, if and to what extent regular pain assessment results in changes in pain management strategies applied and in a subsequent reduction in pain scores in NHR with dementia remains uncertain. Furthermore, many previous studies were of relatively short intervention duration (Zwakhalen et al., 2012; Ando et al., 2016) and used small sample sizes (Zwakhalen et al., 2012; Monacelli et al., 2013; Ando et al., 2016). Only one study was a randomised controlled trial (Fuchs-Lacelle et al., 2008). Therefore, more studies are warranted. We hypothesised that nursing home staff who regularly assess residents for pain using a pain assessment tool would have increased

2. Materials and method 2.1. Trial design A single-blinded, parallel cluster randomised controlled trial was used to evaluate the effectiveness of regular pain assessment on pain prevalence and analgesic use in Norwegian NHRs with severe dementia. A cluster was defined as a single nursing home (NH). The reason for using a cluster design was the risk of contamination if individual randomisation was used. The intervention was carried out by the NH staff, and the outcomes were measured on the individual NHR. This clinical trial is registered at ClinicalTrials.gov (ref. no: NCT02945865). The Regional Ethics Committee approved the procedure of this study (ref. no. 2014/1431, REC South East) prior to the recruitment of the clusters. Written consent was obtained from the residents’ next of kin before the residents were included in the study. If the registered nurses (RN) considered study participation to be a potential burden for the resident, the resident would be withdrawn from the study. Reporting adhered to the CONSORT extension for cluster trials (Campbell et al., 2012) and the TIDieR checklist (Hoffmann et al., 2014). 2.2. Randomisation, allocation and blinding An independent statistician randomised (computer generated allocation) every NH in four counties in the southeast part of Norway (N = 161 in November 2014). Random allocation to one of two study arms was performed before the invitation to participate was e-mailed to the NHs. The NHs were consecutively invited according to the computer-generated allocation sequence from November 2014 to January 2016. We stopped inviting NHs after reaching the required number according to the sample size calculation. Ninety-seven NHs were invited and 16 participated (Fig. 1). The NHs, the individual residents and their next-of-kin were blinded to group allocation. Owing to ethical issues and the risk of contamination, the participating parties were told that the study was to be conducted on two groups receiving different study procedures rather than to test the effectiveness on an intervention. Information was given on the overarching aim of the study: To obtain more knowledge of the importance of pain and pain assessment in people with severe dementia in NHs. Written consent was obtained from the individual participants’ next-of-kin after randomisation and before each resident was enrolled in the study. 2.3. Sample Data were collected in 16 NHs. The only eligibility criterion for the clusters was that they did not routinely use a pain assessment tool. The eligibility criteria for the NH staff delivering the intervention were as follows: i) RN or nursing assistant with a course in drug administration (i.e. he or she could obtain the delegated responsibility form RNs use to administer medication) ii) Worked at the NH for a minimum of six months 53

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

Fig. 1. Participant flow through the study.

iii) Employed ≥50% full-time equivalent iv) Fluent in written and oral Norwegian

process (Fig. 1). In step 1, a RN at the NHs included residents according to the following eligibility criteria:

In Norway, both RNs and nursing assistants are authorised healthcare professionals. Nursing assistants have completed a vocational course in upper secondary education, and RNs hold a bachelor’s degree (three years) from a university or university college. In Norway, nursing assistants and RNs are responsible for every aspect of care for a certain number of patients rather that specific tasks. They have similar sets of responsibilities in a NH (i.e. measurement of vital signs, monitoring patients’ condition, dressing, feeding, etc.). However, nursing assistants provide more basic patient care, as RNs are also responsible for administering medication and performing procedures (i.e. IV therapy, wound care and catheterisation). Individual residents within the clusters underwent a two-step

i) Aged 65 years or older ii) Dementia diagnosis in medical records iii) Inability to self-report pain or non-verbal. Specifically, the residents’ primary nurse was asked to use chart information, clinical judgment and knowledge of the residents to assess eligibility. In step 2, the investigator screened the residents included in step 1 for clinically significant a) pain and/or b) agitation and/or c) neuropsychiatric symptoms at baseline. Clinically significant pain was defined as a score ≥5 on the Doloplus-2 pain scale (Lefebvre-Chapiro, 2001). A score ≥4 on a single item (frequency x severity) on the Neuropsychiatric Inventory, Nursing Home version (Selbæk et al.,

54

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

2.6.3. Training The nursing assistants (in the experimental group only) and the RNs (in both groups) participated in a three-hour face-to-face course, in which a member of our research team (the first author) gave a presentation on pain (including prevalence, consequences and challenges) and pain assessment (what, why and how) in older adults with dementia and taught them how to use the Doloplus-2 according to the instructions accompanying the scale. The RNs and nursing assistants also received the presentation and the scale’s instructions in writing. The scale’s instructions can be accessed through the Doloplus-2 webpage (Doloplus, 2018). The presentations on pain and pain assessment are available upon request to the corresponding author.

2007) indicated clinically significant neuropsychiatric symptoms. Clinically significant agitation was defined as a score ≥39 on the Cohen–Mansfield Agitation Inventory (CMAI) (Husebø et al., 2014a,b). Residents who had a short-term stay (< 4 weeks) were excluded. 2.4. Sample size calculation We were unable to calculate the sample size using the ICC because we did not have this estimate, which is usually acquired from a pilot study (Killip et al., 2004). A simple approach was used: the sample size was calculated assuming individual randomisation and inflated by an add-on of 20% to account for randomisation by cluster. The primary outcome was change in pain score from baseline to week 12 measured with the Doloplus-2 pain scale. We found no previous reports of the smallest clinically relevant difference for the Doloplus-2. Thus, we chose a change of one point as clinically relevant in this study. We would need 33 individuals per arm to keep 80% power to detect a difference of one point as statistically significant at the 5% level. To account for 20% add-on due to cluster randomization, 40 individuals per group were need. Finally, we added 20% to account for drop-out, resulting in the need for 48 in each group.

2.7. Data collection Data were collected from March 2015 to June 2016. The primary and secondary outcomes were assessed at baseline and at the end of the intervention period (week 12). The RNs provided anonymised copies of the residents’ medication lists, and the NHs were asked to have the same RN complete both the baseline and end-of-study assessments with the Doloplus-2. At baseline, the RNs used a standardised form to collect data on demographic variables and residents’ medical diagnoses. On the basis of the medical diagnoses, we composed a variable called ‘pain-related diagnoses’ based on previous research on diagnosis commonly/potentially related to pain; cancer, (sequela from) fractures, stroke, osteoporosis, arthritis, migraine, angina pectoris, herpes zoster, nerve pain (e.g. post-polio syndrome, post-herpetic neuralgia and sciatica) pressure ulcers, low-back pain, amputation, prostheses (knee and hip), peptic ulcers and genito-urinary infections (Nygaard and Jarland, 2005; Rustøen et al., 2005). ‘Chronic pain’ was included even if it was listed without a cause. Additional data on physical health were collected by using the General Medical Health Rating (GMHR), a valid and reliable tool for rating medical comorbidity in people with dementia (Lyketsos et al., 1999). The GMHR was used to review the residents’ current and past medical history unrelated to the dementia syndrome and rated by the RNs on a scale of 1 to 4 (1 = Excellent, 2 = Good, 3 = Moderate and 4 = Poor). Data on the residents ability to perform basic activities of daily living (toileting, feeding, dressing, grooming, physical ambulation, and bathing) was collected by RNs using The physical selfmaintenance scale (PSMS) (Lawton and Brody, 1970). Together with the participating NHs, we discussed and planned how to integrate the intervention (regular pain assessment with the Doloplus-2) into everyday practice to prevent disrupting or negatively affecting normal patient care. The resident’s primary RN or nursing assistants caring for the resident during the dayshift on Tuesdays and Thursdays completed the Doloplus-2 registration at the end of their shift, usually between 2 p.m. and 3 p.m., along with other nursing documentation. A member of our research team (HMR) was on-site and available for questions during the baseline assessment. The participating NH staff received follow-up e-mails in which they were asked if they had questions and/or concerns and encouraged to make contact if needed. A document with an overview of the assessment points was provided along with each participating residents’ assessment form to maintain or improve fidelity.

2.5. Participant involvement Seven RNs working in dementia care (none working or affiliated with any of the participating NHs) were consulted during the development of the study protocol. The RNs were asked about what they considered to be the appropriate intervention dose and duration. Furthermore, they were asked whether the intervention should consist of both standardised assessment and a standardised pain management plan as done in previous studies (Liu and Lai, 2017; Kovach et al., 1999; van Kooten et al., 2017). The RNs were sceptical about a standardised pain management plan for pain management because, firstly, it could be perceived as undermining the NH staff’s knowledge about pain management and the individual residents. Secondly, they argued that treatment needed to be individualised. Thirdly, the standardised pain scale used in this study (the Doloplus-2) has a defined cut-off score that can support the initiation of treatment. Therefore, based on the RNs feedback, our intervention was the regular use of an observational pain behaviour tool only. 2.6. Intervention 2.6.1. Experimental condition The NHRs allocated to the experimental group were assessed for pain twice a week (Tuesday and Thursday) for 12 weeks. The Doloplus2 pain scale was administered by the residents’ primary RN or nursing assistants (e.g. the nurse who had the best knowledge of the residents’ behaviour) at the end of their shift, usually between 2 p.m. and 3 p.m., along with other nursing documentation. Choice of intervention dose was based on findings from previous research (Zwakhalen et al., 2012) and on the discussions we had with RNs in the field (see ‘Participant involvement’) as available clinical practice guidelines and recommendation simply states it should be done on ‘regular basis’ or ‘ongoing’ (Hadjistavropoulos et al., 2007, 2010).

2.8. Outcomes 2.6.2. Control condition The control group received usual care. Usual care practice for pain assessment in Norwegian NHs is non-standardised, and likely to encompass a wide variety of practice styles. In NHs in the control group, no nursing assistants participated in the study, only RNs. The RNs attended a course identical in content and delivery to that given to the intervention group in order to be able to collect data on the primary outcome measure.

2.8.1. Primary outcome The primary outcome was the total pain score measured with the Doloplus-2 pain scale, which is based on observation of behaviour in older adults with cognitive impairment (Lefebvre-Chapiro, 2001). The 10-item scale is grouped into three dimensions, namely, somatic, psychomotor and psychosocial reactions to pain. Each item was scored from 0 to 3, where 0 indicated ‘absent’ or ‘as usual’ and 3 indicated the 55

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

2.8.2. Secondary outcome The secondary outcome was analgesic use. Analgesics were defined as Anatomical Therapeutic Chemical (ATC) codes N02A (opioids), N02B (other analgesics and antipyretics) and M01A (anti-inflammatory and anti-rheumatic products). For opioids, doses were converted into oral morphine equivalents (OMEQ) and summed up to obtain an index of mean intake of oral morphine in milligram/day. The Norwegian Medicines Agency’s table for ‘Equianalgesic doses of opioid analgesics and duration of action’ (Norwegian Medicines Agency, 2016) was used as a guide to convert to the daily dose of OMEQ. For other analgesics and antipyretics, the daily dose of paracetamol (acetaminophen) was calculated in milligram/day. Only frequency was presented for antiinflammatory and anti-rheumatic products because few participants used anti-inflammatory and anti-rheumatic products. The calculation of analgesic use only included analgesics provided as scheduled, not analgesics prescribed as PRN or non-pharmacological initiatives. In Norway, 0.55 h is the national average physician hours per week per NH resident (Helsedirektoratet, 2017), and a key task of these physicians is medication review. Therefore, if the nursing staff considered a need for a change in the analgesic regimen, they would have the opportunity to bring this matter to the physician’s attention relatively quickly to modify the prescription of the scheduled administered analgesics. Thus, we would have been able to capture these changes during our outcome data collection.

Table 1 Demographic and clinical data characteristics at baseline (n = 112). Characteristic

Total sample (n = 112)

Control (n = 62)

Experimental (n = 50)

Age Mean (SD) Sex N (%) Female Male

83.6 (7.1)

83.1 (6.4)

84.1 (7.8)

78 (69.6) 34 (30.4)

41 (66.1) 21 (33.9)

37 (74.0) 13 (26.0)

44 (39.3) 16 (14.3) 10 (8.9) 1 (0.9) 41 (36.6)

24 (38.7) 12 (19.4) 5 (8.1) – 21 (33.9)

20 (40.0) 4 (8.0) 5 (10.0) 1 (2.0) 20 (40.0)

21 (19.3) 45 (41.3) 40 (36.7) 3 (2.8) 51.6 (18.5)

10 (16.7) 28 (46.7) 19 (31.7) 3 (5.0) 48.2 (14.8)

11 (22.4) 17 (34.7) 21 (42.9) – 55.8 (21.7)

29.2 (18.5)

26.8 (18.2)

32.2 (18.7)

39 (34.8)

22 (35.5)

17 (34.0)

73 (65.2)

40 (64.5)

33 (66.0)

Type of dementia N (%) Degenerative Vascular Mixed Secondary Unspecified GMHR a) N (%) Excellent Good Moderate Poor Agitation (CMAI score) Mean (SD) Neuropsychiatric symptoms (NPI-NH score) Mean (SD) ADL N (%) Independent to low dependency (PSMS score ≤17) Moderate-to-high dependency (PSMS score ≥18) Pain related diagnoses Yes No Pain score Mean (SD)

b)

N (%) 59 (54.6) 49 (45.4) 8.0 (5.4)

32 (54.2) 27 (45.8) 8.0 (4.9)

27 (55.1) 22 (44.9) 8.1 (6.1)

76 (67.9) 36 (32.1)

45 (72.6) 17 (27.4)

31 (62.0) 19 (38.0)

N (%) 72 (65.5) 38 (34.5) 53 (48.2) 2 (2.8)

36 (58.1) 18 (29.0) 30 (48.4) 1 (1.6)

36 (75.0) 20 (41.7) 23 (47.9) 1 (2.1)

38 (34.5)

26 (41.9)

12 (25.0)

10.8 (26.1) 1277.3 (1436.1)

10.2 (21.9) 1282.3 (1464.4)

11.7 (30.8) 1270.8 (1414.1)

32 (51.6) 30 (48.4)

21 (43.8) 27 (56.3)

25 (40.3) 37 (59.7)

13 (27.1) 35 (72.9)

Presence of pain Present (Doloplus-2 ≥5) Not present (Doloplus2 < 5) Regular scheduled analgesics Yes Opioids Paracetamol Anti-inflammatory and anti-rheumatic products No Analgesic use c), Mean (SD) OMEQ Paracetamol Scheduled psycholeptics Yes No

c)

c)

Data were analysed by an independent statistician using SPSS version 24 (Armonk, NY: IBM Corp.). STATA (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC) was used to run logistic regression for repeated measures. All the tests were two sided, and all associations were considered statistically significant when p < 0.05. Data on the baseline characteristics are presented as means and standard deviations (SD) for continuous variables and as proportions with percentages for categorical data. The differences between dropouts and those who completed the study were analysed using the chi-square test for pairs of categorical variables. The Mann–Whitney–Wilcoxon test was used for continuous data. Analyses of primary and secondary outcomes were conducted using the intention-to-treat principal. The primary and secondary outcome variables were measured at two time points (baseline and 12 weeks) to assess changes over time. Linear mixed models were used to account for dependencies in the data as each participant could be measured twice. All available data were used because fitting linear mixed models does not require imputation of missing data. The model assessed ‘pain score’ (primary outcome) as a dependent variable, ‘time’ and ‘group’ (experimental or control) as fixed factors and ‘gender’ and ‘age’ as covariates to estimate the differences in pain score between the groups. We performed a sensitivity analysis, in which clusters were included in the model as a fixed effect to account for the cluster effect within each NH. The same linear mixed model was fitted to estimate the differences in analgesic use (secondary outcome), that is, one model for OMEQ and one for paracetamol. Additionally, logistic regression for repeated measures was used to model pain prevalence (pain yes/no) as the dependent variable.

c)

N (%) 53 (48.2) 57 (51.8)

Scheduled anti-depressants Yes No

2.9. Statistical methods

N (%) 72 (65.5) 38 (34.5)

Missing data on a) 3; b) 4; c) 2 participants. ADL: Activities of daily living; CMAI: Cohen–Mansfield Agitation Inventory; GMHR: General Medical Health Rating; OMEQ: NPI-NH: the Neuropsychiatric Inventory, Nursing Home version; Oral morphine equivalents; PSMS: the physical self-maintenance scale.

3. Results

highest score for the behaviour in question. The total score ranged 0–30. A cut-off ≥5 indicated pain (Doloplus, 2018). The reported time needed to complete the Doloplus-2 assessment from previous studies was 5–10 min on average (Monacelli et al., 2013; Pickering et al., 2010). The Doloplus-2 has been translated to Norwegian (Hølen et al., 2005) and psychometrically tested in a non-verbal NH population (Torvik et al., 2010).

In total, 121 NHRs were eligible for inclusion; one died and eight withdrew before data collection started, leaving 112 participants (Fig. 1). Among the 112 participants, 15 NHRs (13%) did not complete the study period: 5 died in the control group and 7 in the experimental group, one was hospitalised and two moved to a different NH. No statistically significant differences were revealed between those who were lost to follow-up and those who completed the study in any of the demographic and clinical variables. 56

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

4. Discussion

For the total sample, the mean age was 83.6 years, and the majority was females (69.9%) (Table 1). Over one-third of the sample (39.3%) had a degenerative type of dementia (Alzheimer’s, frontotemporal or Lewy body), and 36.6% was categorised as ‘unspecified’. About 40% of the NHRs was assessed as having poor and moderate physical health (GMHR). Two-thirds of the sample were moderately to highly dependent in activities of daily living (ADL). More than half (54.6%) had one or more pain-related diagnoses. The mean pain score was 8 (≥5 indicating pain), and 67.9% had a pain score ≥5. Two-thirds of the total sample had regularly scheduled analgesics. The only substance prescribed in the ‘other analgesics and antipyretics’ group was paracetamol, which 48.2% of the sample used. One-third of the NHRs were prescribed opioids. Only 3% were prescribed anti-inflammatory and anti-rheumatic products. Of those NHRs who scored above the cut-off score for pain (67.9%) scoring ≥5, 50.0% used paracetamol, and 36.5% used opioids; 51.6% used a combination with paracetamol. The mean dose of OMEQ was 10.8 and that of paracetamol was 1277.3 mg/ d. Almost half of the total sample (48.2%) had scheduled psycholeptic drugs (ATC code N05A, N05B and N05C), and one-third (34.5%) was prescribed antidepressants (ATC code N06a). Adherence to the intervention protocol (pain assessment with Doloplus-2 twice a week, total possible assessment per participant = 24) was examined by the number of assessments completed for the participants in the interventions group who completed the study (n = 40). The nursing home staff completed 84% (800 out of 960) of the scheduled assessments, with a median 19 assessments per participant (range 12–24). The linear mixed model fitted with ‘time’ (two assessment points) and ‘group’ as fixed factors and ‘gender’ and ‘age’ as covariates did not demonstrate an overall effect of regular pain assessment on pain score (p = 0.86) or analgesic use (p = 0.47 for OMEQ and p = 0.85 for paracetamol). The mean score of the Doloplus-2 (primary outcome) remained unchanged in both groups. Table 2 outlines the mean pain score by group at the end of week 12. The mean pain score was over the established cut-off score (≥5) for the primary outcome measure: 7.9 and 8.0 in the control and in the experimental group, respectively, on the last day of the study period. For analgesic use (secondary outcome) fitted with ‘time’ (two assessment points) and ‘group’ as fixed factors and ‘gender’ and ‘age’ as covariates, no between-group difference was found in the linear mixed models analyses (Table 2). Furthermore, no overall effect was found on pain prevalence (pain yes/no) tested with logistic regression for repeated measures (p = 0.42). The sensitivity analysis showed that including clusters as a fixed effect did not affect our results to a noteworthy extent; e.g. estimates for the difference in pain score (experimental minus control) changed from −0.13 to −0.17. Therefore, the results presented in this paper are based on models without the cluster effect included.

This cluster-randomised controlled trial showed no significant effect of regular pain assessment on analgesic use or pain score in NHRs with severe dementia. Our findings corresponded to those of previous research reporting no effect on scheduled analgesics (Fuchs-Lacelle et al., 2008) or pharmacological treatment (Zwakhalen et al., 2012). However, other studies reported a reduced pain score over the course of the study and an increase in analgesic therapy in general (Monacelli et al., 2013; Ando et al., 2016). A plausible reason for the difference in results is the characteristics of the sample. For example, Ando et al. (2016) investigated patients admitted to a geriatric hospital because of orthopaedic diseases that typically involve pain. Thus, the focus and strategies for pain management might have been more comprehensive in such a specialised environment. Several reasons may account for why our intervention did not result in a change in analgesic use or pain score. In the following, we discuss three elements that may have affected our results and have implications for clinical practice and future research: rater factors, patient factors and pain scale used. Firstly, despite the reasonably high compliance rates for assessments, an issue in our study could be the lack of translating assessment findings to practical action. Previous research demonstrated that interpreting pain assessment results could be difficult (Liu et al., 2011). Furthermore, providing nursing staff in long-term care with pain assessment tools alone may not be sufficient to change the pain management practices (Zwakhalen et al., 2012). Assessment tools can support human observation but may not in themselves lead to decisions or action. Therefore, the following (in)action is dependent on the competencies of the healthcare professionals (Koppitz et al., 2015). Another explanation is that behaviour change could have been misinterpreted by the NH staff as caused by dementia and reduced cognitive function, not as an indication of pain. Moreover, as demonstrated in previous studies, staff could have considered the patient’s pain score as not sufficiently high as to administer or request for analgesic, or they could have been unsure about suitable pain management strategies (Zwakhalen et al., 2007a; Corbett et al., 2012; Barry et al., 2012). Limited knowledge and a misconception among nursing staff, mainly about pharmacological treatment, have been reported (Zwakhalen et al., 2007b). Moreover, previous research found that healthcare professionals experience uncertainty about the accuracy of pain assessment in older adults with cognitive impairment, thus making them reluctant to administer analgesics as a result of this uncertainty (Kaasalainen et al., 2007). Furthermore, we used both RNs and nursing assistants to deliver our intervention. Even though the nursing assistants could administer analgesics through the delegation of an RN, previous research showed that nursing assistants felt excluded from taking pain management decisions and management initiatives. Their assessment and suggestions for interventions often went unrecognised by RNs (Liu, 2014). Whether assessment results are affected by raters’ professional qualification is also issue. Nursing assistants could potentially observe more pain cues than RNs because residents tended to manifest pain behaviours in activities usually aided by nursing assistants, such as mobilisation, shower and dressing (Chen et al., 2010a). Furthermore, the educational background for nursing assistants is significantly different from RNs. This difference may have affect their judgment about the residents' behaviour. Even though both the RNs and nursing assistants received training, their skills in, and knowledge, of pain assessment may differ. Thus, the raters’ professional qualification could have affected the assessment results and the number and type of pain management interventions applied. Secondly, close to two-thirds of the NHRs included in our study were moderately to highly dependent in ADL, and all of them had severe cognitive impairment due to dementia. The Doloplus-2 was developed for older adults presenting with verbal communication

Table 2 Effect of regular pain assessment compared with usual care on outcome variables. Control Mean (SE)

Experimental Mean (SE)

Difference exp. minus contr. (95% CI)

p-value

Pain

7.87 (0.49)

8.00 (0.57)

0.86

OMEQ

11.13 (2.56)

13.93 (2.95)

Paracetamol

1281.39 (129.83)

1246.82 (150.28)

−0.13 (−1.35, 1.62) −2.80 (−10.43, 4.83) 34.57 (−355.94, 427.08)

0.47 0.85

OMEQ: oral morphine equivalents; exp: experimental group; contr: control group; SE: standard error; CI: confidence interval. 57

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

disorders, in which ADL dependency and cognitive impairment are highly prevalent. However, these issues may prevent these individuals from expressing their pain through behaviour or body language because of issues such as physical functioning, immobility and contractures (Kaasalainen, 2007). Furthermore, many of our participants received psycholeptics and antidepressants, which could influence and mask their pain behaviour. These factors could have affected and made it especially challenging for the nursing home staff to observe any pain behaviours, consequently underestimating the residents’ pain and need for analgesics. Thirdly, the Doloplus-2 was used to measure our primary outcome. According to the Doloplus-2 guidelines, items on the scale should not be scored if they do not apply to the patient (Doloplus, 2018). Therefore, maximum scores differ among patients and may affect the results when making comparisons (Hølen et al., 2005). Questions have been raised about whether the results of the Doloplus-2 assessment, including the established cut-off, can inform clinical decisions and actions. Moreover, evidence is limited regarding the responsiveness of the Doloplus-2. An unresponsive instrument may fail to detect true improvement. Research on the content validity of the Doloplus-2 is rare, and thus we do not know whether clinicians and experts in the field find that the items of the scale adequately reflect the construct of pain in older adults with severe dementia (Rostad et al., 2017b). Previous studies also reported that some items of the Doloplus-2, especially the psychosocial items, were difficult to understand and score (Chen et al., 2010b; Zwakhalen et al., 2006). However, on the basis of the literature (Lichtner et al., 2014; Torvik et al., 2010; Zwakhalen et al., 2006), we considered Doloplus-2 to be the most appropriate of the tools available in Norway for our study. The mean pain score remained higher than the established cut-off in both groups at the end of the intervention period; this result may indicate a persistent under-treatment of painful conditions in our sample. Among the two-thirds assessed in pain, only half received paracetamol and one-third received opioids. Under-treatment of pain in older adults with dementia is a frequent observation and concern in previous research (Thakur et al., 2016; van der Steen, 2010; Brecher and West, 2016). Achieving optimal and safe pain control in older people is a complex and multi-factorial issue because of comorbidities, polypharmacy and significant pharmacokinetic and pharmacodynamic changes that occur with advanced age (McLachlan et al., 2011). The causes of pain are usually complex, and obtaining an accurate diagnosis can be difficult, especially when the person is not able to provide pivotal information, such as the quality of pain (stabbing, burning, shooting, etc.), its location, severity and other important aspects. Accordingly, the cause and type of pain can be misdiagnosed, and inappropriate drug use may occur. Moreover, discussing how the current treatment guidelines advocating a ‘start low and go slow’ approach to analgesic dosing in older patients may partly contribute to inadequate pain management may be necessary. The dosing strategy is often misinterpreted as ‘start low, stay low’ (Hanlon et al., 2009) and may cause the reluctance to prescribe or administer analgesics in sufficiently high doses (McLachlan et al., 2011). Furthermore, insufficient follow up on the treatment effect can cause unnecessary suffering and burden to the patients. Analgesics can cause potential burden or harm, however, unrelieved pain in older adults with dementia can greatly impact on the individual and society. Pain has a detrimental impact on physical, psychological and social functioning and quality of life (Rostad et al., 2017a; Zanocchi et al., 2008; Lin et al., 2011), which in turn can affect the management of health care and costs.

in everyday NH practice, and thus the study occurred in a setting in which the findings could be applied. Moreover, this work was a multisite study with a low attrition rate and with a sample that we considered to represent the key characteristics (i.e. distribution of age and sex) of the population of interest. This study also has limitations that needs to be addressed. Firstly, pain is a subjective experience. Acknowledging that an observationbased assessment may insufficiently and even erroneously reflect patients’ pain experience is important. Another limitation is that data on analgesic use were solely based on what was scheduled, not what was actually administered. Scheduled medications were probably administered as planned, but exceptions might have occurred. Another issue is possible confounding of PRN analgesics or non-pharmacological strategies. Furthermore, as we do not know whether a change in PRN analgesics or non-pharmacological strategies occurred, it's difficult to conclude that there was no improvement in the residents’ pain therapy. In addition, increase of OMEQ was considered a positive outcome as it likely indicates that a higher analgesic dose was administered. However, a higher dose does not necessarily mean more efficient pain relief. Moreover, QMEQ would not reflect a switch of medication which can be more beneficial to the residents compared to an increase in dosage of an already prescribed analgesic. Adjuvant pain medications, such as antiepileptic medications, are increasingly used for neuropathic pain (Baftiu et al., 2016). However, we did not include these drugs when evaluating analgesic use because we had few cases in our material (7%) who used such medications, and because we did not know the rational for prescribing these drugs as this information was missing in the individual residents’ medication lists. We do not have the exact number of RNs and nursing assistants who completed the assessments or their demographic information (e.g. age, sex and professional qualification) The interpretations done of the assessment results in this study have been a product of both the rater’s observations (the resident’s behaviour), but also the rater’s personal knowledge, attitudes and experiences affected by their age, sex, cultural background and more, and we were unfortunately unable to examine how that may have impacted our results. We also lacked information on the number and the characteristics of the NHRs who refused to participate because the recruitment was conducted by RNs at the NHs. Thus, we are unable to draw any conclusion regarding nonresponse bias; the difference between those who participated and those who declined participation (Polit and Beck, 2018), which threatens the generalizability of our findings. Furthermore, we have no information on ‘usual care’ conditions. Usual care in terms of pain assessment among older nursing home residents with severe dementia is variable and not based on a consensus. These varying practice makes usual care difficult to describe and we do not really know how pain assessment in this population is performed. Finally, this study may have contamination issues. The control condition was introduced to our intervention, as the nursing home staff were taught how to use the Doloplus-2 for assessment at baseline and at the end of week 12 to collect data on the primary outcome. We had no guarantee that this scale was not used regularly in the control conditions. However, no indication of regular use was found in any of the NHs assigned as the control condition based on follow-up throughout and after the study and the number of provided and completed assessment forms.

4.1. Strengths and limitations

The underassessment and under-treatment of pain is a pressing ethical and professional concern, but both are modifiable factors. A standardised pain assessment tool could help in deciding on, initiating and evaluating treatment, which are an important part of nurses’ responsibilities. However, introducing standardised plans or tools to support or cover nursing practice is neither straightforward nor

4.2. Clinical implications and future research

This study has several strengths. The clusters were randomly included in the study, and the raters were blinded to group allocation. Furthermore, the study was conducted in real-life conditions. Data were collected by the people who would be assessing and intervening on pain 58

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

uncontroversial (Pedersen et al., 2010). Nonetheless, using a systematic approach to assess whether pain is likely present or not is crucial when patients are unable to self-report pain. Standardised tools are not meant to replace the clinical judgment of the nursing staff but rather to complement it, and their use is dependent on the nursing staff’s knowledge of the patients’ past and present life situation and health issues. Further research is needed to explore how, when and for whom pain assessment tools can be used to support pain management in older NHRs with severe dementia. Future investigators can also examine a multi-method approach to pain assessment. Moreover, applying a design to explore changes in individuals’ pain score and analgesic use following a pain assessment intervention over a longer observational period is an interesting field for further research.

neuropsychiatric symptoms and ADL dependency among persons living in nursing homes; a cross-sectional study. BMC Geriatr. 16 (1), 154. Brecher, D.B., West, T.L., 2016. Underrecognition and undertreatment of pain and behavioral symptoms in end-stage dementia. Am. J. Hosp. Palliat. Care 33 (3), 276–280. Campbell, M.K., Piaggio, G., Elbourne, D.R., Altman, D.G., 2012. Consort 2010 statement: extension to cluster randomised trials. BMJ 345, 5661. Chen, Y.H., Lin, L.C., Watson, R., 2010a. Validating nurses’ and nursing assistants’ report of assessing pain in older people with dementia. J. Clin. Nurs. 19 (1–2), 42–52. Chen, Y.-H., Lin, L.-C., Watson, R., 2010b. Evaluation of the psychometric properties and the clinical feasibility of a Chinese version of the Doloplus-2 scale among cognitively impaired older people with communication difficulty. Int. J. Nurs. Stud. 47 (1), 78–88. Corbett, A., Husebo, B., Malcangio, M., Staniland, A., Cohen-Mansfield, J., Aarsland, D., et al., 2012. Assessment and treatment of pain in people with dementia. Nat. Rev. Neurol. 8 (5), 264–274. Doloplus. The Doloplus web-site. Available from: http://www.doloplus.fr/index.php. Fuchs-Lacelle, S., Hadjistavropoulos, T., Lix, L., 2008. Pain assessment as intervention: a study of older adults with severe dementia. Clin. J. Pain 24 (8), 697–707. Gibson, S.J., Lussier, D., 2012. Prevalence and relevance of pain in older persons. Pain Med. 13 (Suppl. 2), S23–S26. Hølen, J.C., Saltvedt, I., Fayers, P.M., Bjørnnes, M., Stenseth, G., Hval, B., et al., 2005. The Norwegian Doloplus-2, a tool for behavioural pain assessment: translation and pilotvalidation in nursing home patients with cognitive impairment. Palliat. Med. 19 (5), 411–417. Hadjistavropoulos, T., Herr, K., Turk, D.C., Fine, P.G., Dworkin, R.H., Helme, R., et al., 2007. An interdisciplinary expert consensus statement on assessment of pain in older persons. Clin. J. Pain 23, 1–43. Hadjistavropoulos, T., Fitzgerald, T.D., Marchildon, G.P., 2010. Practice guidelines for assessing pain in older persons with dementia residing in long-term care facilities. Physiother. Can. 62 (2), 104–113. Hadjistavropoulos, T., Herr, K., Prkachin, K.M., Craig, K.D., Gibson, S.J., Lukas, A., et al., 2014. Pain assessment in elderly adults with dementia. Lancet Neurol. 13 (12), 1216–1227. Hanlon, J.T., Backonja, M., Weiner, D., Argoff, C., 2009. Evolving Pharmacological Management of Persistent Pain in Older Persons. Blackwell Publishing Inc, Malden, USA. Helsedirektoratet. Legetimer per beboer i sykehjem: helsenorge.no; updated 29. June 2017. Available from: https://helsenorge.no/Kvalitetsindikatorer/kvalitetsindikatorpleie-og-omsorg/legetimer-for-beboer-i-sykehjem. Herr, K., Bjoro, K., Decker, S., 2006. Tools for assessment of pain in nonverbal older adults with dementia: a state-of-the-science review. J. Pain Symptom Manage. 31 (2), 170–192. Hoffmann, T.C., Glasziou, P.P., Boutron, I., Milne, R., Perera, R., Moher, D., et al., 2014. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ 348, 1687. Husebø, B.S., Ballard, C., Cohen-Mansfield, J., Seifert, R., Aarsland, D., 2014a. The response of agitated behavior to pain management in persons with dementia. Am. J. Geriatr. Psychiatry 22 (7), 708–717. Husebø, B.S., Ostelo, R., Strand, L.I., 2014b. The MOBID-2 pain scale: reliability and responsiveness to pain in patients with dementia. Eur. J. Pain 18 (10), 1419–1430. Kaasalainen, S., Coker, E., Dolovich, L., Papaioannou, A., Hadjistavropoulos, T., Emili, A., et al., 2007. Pain management decision making among long-term care physicians and nurses. West. J. Nurs. Res. 5, 561–580. Kaasalainen, S., 2007. Pain assessment in older adults with dementia: using behavioral observation methods in clinical practice. J. Gerontol. Nurs. 33 (6), 6. Killip, S., Mahfoud, Z., Pearce, K., 2004. What is an intracluster correlation coefficient? Crucial concepts for primary care researchers. Ann. Fam. Med. 2 (3), 204–208. Koppitz, A., Waldboth, V., Dreizler, J., Imhof, L., 2015. Dying with dementia: most frequent symptoms. A review of the literature. Int. J. Health Prof. 2 (1), 49–63. Kovach, C.R., Weissman, D.E., Griffie, J., Matson, S., Muchka, S., 1999. Assessment and treatment of discomfort for people with late-stage dementia. J. Pain Symptom Manage. 18 (6), 412–419. Lawton, M., Brody, E., 1970. Assessment of older people: self-maintaining and instrumental activities of daily living. Nurs. Res. 19 (3), 278. Lefebvre-Chapiro, S., 2001. The DOLOPLUS-2 scale – evaluating pain in the elderly. Eur. J. Palliat. Care 8 (5), 191–194. Lichtner, V., Dowding, D., Esterhuizen, P., Closs, S.J., Long, A.F., Corbett, A., et al., 2014. Pain assessment for people with dementia: a systematic review of systematic reviews of pain assessment tools. BMC Geriatr. 14, 1. Lillekroken, D., Slettebø, Å., 2013. Smertekartlegging og smertelindring hos pasienter med demens: utfordringer og dilemmaer. Nordic J. Nurs. Res. 33 (3), 29–33. Lin, P.C., Lin, L.C., Shyu, Y.I., Hua, M.S., 2011. Predictors of pain in nursing home residents with dementia: a cross-sectional study. J. Clin. Nurs. 20 (13–14), 1849–1857. Liu, J.Y., Lai, C.K., 2017. Implementation of observational pain management protocol for residents with dementia: a cluster-RCT. J. Am. Geriatr. Soc. 65, 3. Liu, J.Y., Briggs, M., Closs, J., 2011. Acceptability of pain behaviour observational methods (PBOMs) for use by nursing home staff. J. Clin. Nurs. 20 (13–14), 2071–2073. Liu, J.Y., 2014. Exploring nursing assistants' roles in the process of pain management for cognitively impaired nursing home residents: a qualitative study. J. Adv. Nurs. 70 (5), 1065–1077. Lyketsos, C.G., Galik, E., Steele, C., Steinberg, M., Rosenblatt, A., Warren, A., et al., 1999. The General Medical Health Rating: a bedside global rating of medical comorbidity in patients with dementia. J. Am. Geriatr. Soc. 47 (4), 487–491. Martin, R., Williams, J., Hadjistavropoulos, T., Hadjistavropoulos, H.D., MacLean, M., 2005. A qualitative investigation of seniors' and caregivers' views on pain assessment

5. Conclusion Our study failed to document effectiveness of regular pain assessment on pain score or analgesic use in our sample. However, there is not sufficient evidence to conclude that regular pain assessment using a pain assessment tool is not clinically relevant. The lack of statistically significant differences might be due to limitations concerning how our intervention was designed and conducted considering the tool used and the patient and rater factors. Moreover, the mean pain score remained higher than the established cut-off in both groups, thus indicating that pain management could have been insufficient. Further research is needed on effective pain assessment and management strategies in older NHRs with dementia. Funding This study was funded by the OsloMet – Oslo Metropolitan University. One of the authors is supported by a Canadian Institutes of Health Research New Investigator Award. Conflicts of interest None. Acknowledgements The authors like to thank all the residents, their next-of-kin and the staff of the participating nursing homes for their willingness and enthusiasm that made this study possible. We also thank Kristian Svendsen (Post Doc, Pharmacoepidemiology, The Arctic University of Norway), Julie Wendelbo Aanensen (Pharmacist and Project Manager, The Centre for Development of Institutional and Home Care Services, Oslo) and Arton Bafitu (PhD student, Department of Life Sciences and Health, Pharmacy, Oslo Metropolitan University) for sharing their expertise and for their intellectual contribution in preparation of the manuscript. References Achterberg, W.P., Pieper, M., van Dalen-Kok, A.H., de Waal, M., Husebo, B.S., Lautenbacher, S., et al., 2013. Pain management in patients with dementia. Clin. Interv. Aging 8, 1471–1482. Ando, C., Ito, Y., Amemiya, S., Tamura, K., Kako, K., Tsuzura, S., et al., 2016. Effectiveness of the Japanese DOLOPLUS-2: a pain assessment scale for patients with moderate-to-severe dementia. Psychogeriatrics 16, 315–322. Baftiu, A., Landmark, C.J., Rusten, I.R., Feet, S.A., Johannessen, S.I., Larsson, P.G., 2016. Changes in utilisation of antiepileptic drugs in epilepsy and non-epilepsy disorders—a pharmacoepidemiological study and clinical implications. Eur. J. Clin. Pharmacol. 72 (10), 1245–1254. Barry, H.E., Parsons, C., Peter Passmore, A., Hughes, C.M., 2012. An exploration of nursing home managers' knowledge of and attitudes towards the management of pain in residents with dementia. Int. J. Geriatr. Psychiatry 27 (12), 1258–1266. Björk, S., Juthberg, C., Lindkvist, M., Wimo, A., Sandman, P.-O., Winblad, B., et al., 2016. Exploring the prevalence and variance of cognitive impairment, pain,

59

International Journal of Nursing Studies 84 (2018) 52–60

H.M. Rostad et al.

homes. Int. J. Geriatr. Psychiatry 22 (9), 843–849. Thakur, E.R., Amspoker, A.B., Sansgiry, S., Snow, A.L., Stanley, M., Wilson, N., et al., 2016. Pain among community-dwelling older adults with dementia: factors associated with undertreatment. Pain Med. 225. Torvik, K., Kaasa, S., Kirkevold, Ø., Rustøen, T., 2009. Pain in patients living in Norwegian nursing homes. Palliat. Med. 23 (1), 8–16. Torvik, K., Kaasa, S., Kirkevold, Ø., Saltvedt, I., Hølen, J.C., Fayers, P., et al., 2010. Validation of Doloplus-2 among nonverbal nursing home patients-an evaluation of Doloplus-2 in a clinical setting. BMC Geriatr. 10 (9). Torvik, K., Nordtug, B., Brenne, I.K., Rognstad, M.-K., 2015. Pain assessment strategies in home care and nursing homes in mid-Norway: a cross-sectional survey. Pain Manage. Nurs. 16 (4), 602–608. van Kooten, J., Smalbrugge, M., van der Wouden, J.C., Stek, M.L., Hertogh, C.M., 2017. Evaluation of a pain assessment procedure in long-term care residents with pain and dementia. J. Pain Symptom Manage. 54 (5), 727–731. van der Steen, J.T., 2010. Dying with dementia: what we know after more than a decade of research. J. Alzheimers Dis. 22 (1), 37–55. Zanocchi, M., Maero, B., Nicola, E., Martinelli, E., Luppino, A., Gonella, M., et al., 2008. Chronic pain in a sample of nursing home residents: prevalence, characteristics, influence on quality of life (QoL). Arch. Gerontol. Geriatr. 47 (1), 121–128. Zwakhalen, S.M., Hamers, J.P., Abu-Saad, H.H., Berger, M.P., 2006. Pain in elderly people with severe dementia: a systematic review of behavioural pain assessment tools. BMC Geriatr. 6, 1. Zwakhalen, S.M., Hamers, J.P., Peijnenburg, R.H., Berger, M.P., 2007a. Nursing staff knowledge and beliefs about pain in elderly nursing home residents with dementia. Pain Res. Manage. 12 (3), 177–184. Zwakhalen, S.M., Hamers, J.P., Berger, M.P., 2007b. Improving the clinical usefulness of a behavioural pain scale for older people with dementia. J. Adv. Nurs. 58 (5), 493–502. Zwakhalen, S.M., Koopmans, R.T., Geels, P.J., Berger, M.P., Hamers, J.P., 2009. The prevalence of pain in nursing home residents with dementia measured using an observational pain scale. Eur. J. Pain 13 (1), 89–93. Zwakhalen, S.M., Hof, C.E., Hamers, J.P., 2012. Systematic pain assessment using an observational scale in nursing home residents with dementia: exploring feasibility and applied interventions. J. Clin. Nurs. 21 (21–22), 3009–3017.

and management. Can. J. Nurs. Res. 37 (2), 142–165. McAuliffe, L., Nay, R., O’Donnell, M., Fetherstonhaugh, D., 2009. Pain assessment in older people with dementia: literature review. J. Adv. Nurs. 65 (1), 2–10. McAuliffe, L., Brown, D., Fetherstonhaugh, D., 2012. Pain and dementia: an overview of the literature. Int. J. Older People Nurs. 7 (3), 219–226. McLachlan, A.J., Bath, S., Naganathan, V., Hilmer, S.N., Le Couteur, D.G., Gibson, S.J., et al., 2011. Clinical pharmacology of analgesic medicines in older people: impact of frailty and cognitive impairment. Br. J. Clin. Pharmacol. 71 (3), 351–364. Monacelli, F., Vasile Nurse, A., Odetti, P., Traverso, N., 2013. Doloplus-2 pain assessment: an effective tool in patients over 85 years with advanced dementia and persistent pain. La Clinica Terapeutica 164 (1), 23–25. The Norwegian Medicines Manual for Health Personnel, L20 Smertestillende Legemidler [L20 Analgesic drugs], 2016. The Norwegian Medicines Manual for Health Personnel [updated 31.10.2016; cited 26.11.2017] Available from: http:// legemiddelhandboka.no/legemidler/80005. Nygaard, H.A., Jarland, M., 2005. Are nursing home patients with dementia diagnosis at increased risk for inadequate pain treatment? Int. J. Geriatr. Psychiatry 20 (8), 730–737. Pedersen, R., Ellingsen, G., Monteiro, E., 2010. Standardized nursing work: works in practice but not in theory. Stud. Health Technol. Inf. 157, 91–96. Pickering, G., Gibson, S., Serbouti, S., Odetti, P., Gonçalves, J.F., Gambassi, G., et al., 2010. Reliability study in five languages of the translation of the pain behavioural scale Doloplus®. Eur. J. Pain 14 (5), 545. Polit, D.F., Beck, C.T., 2018. Essentials of Nursing Research: Appraising Evidence for Nursing Practice, 9th ed. Wolter Kluwer, Philadelphia, US. Rostad, H.M., Puts, M.T., Småstuen, M.C., Grov, E.K., Utne, I., Halvorsrud, L., 2017a. Associations between pain and quality of life in severe dementia: a Norwegian crosssectional study. Dement. Geriatr. Cogn. Dis. Extra 7 (1), 109–121. Rostad, H.M., Utne, I., Grov, E.K., Puts, M., Halvorsrud, L., 2017b. Measurement properties, feasibility and clinical utility of the Doloplus-2 pain scale in older adults with cognitive impairment: a systematic review. BMC Geriatr. 17 (1), 257. Rustøen, T., Wahl, A.K., Hanestad, B.R., Lerdal, A., Paul, S., Miaskowski, C., 2005. Age and the experience of chronic pain. Clin. J. Pain 21 (6), 513–523. Selbæk, G., Kirkevold, Ø., Engedal, K., 2007. The prevalence of psychiatric symptoms and behavioural disturbances and the use of psychotropic drugs in Norwegian nursing

60