Received: 5 November 2020 | Revised: 29 January 2021 | Accepted: 21 March 2021 DOI: 10.1002/nop2.880 RESEARCH ARTICLE Pain management nursing practice assessment in older adults with dementia Alícia Minaya-­Freire1,2 | Mireia Subirana-­Casacuberta2,3 | Emma Puigoriol-­Juvanteny1,4 | Anna Ramon-­Aribau2 1 University Hospital Of Vic-­V ic Hospital Consortium (HUV-­CHV). C. Francesc Plà “El Abstract Vigatà”, Vic, Barcelona, Spain Aim: To assess pain management nursing practice in older adults with dementia 2 Research Group on Methodology, Methods, through electronic health records (EHR). Models and Health and social Outcomes (M3O), Faculty of Health Science and Design: Retrospective study. Welfare, Centre for Health and Social Care Methods: Data were collected from EHR related to pain management in older adults Research (CESS), University of Vic-­Central University of Catalonia (UVIC-­UCC). with dementia treated at the Acute Geriatrics Unit (AGU) of a university hospital in C. Sagrada Família, Vic, Barcelona, Spain early 2018. 3 Nursing Department, Parc Taulí Hospital Universitari, Institut d’Investigació i Results: EHR related to the pain of 111 patients were reviewed. Pain intensity was Innovació Parc Taulí I3PT, Universitat assessed at admission in 88% of patients and a median of 1.9 times per day of stay. Autònoma de Barcelona, Sabadell, Spain 4 A disproportionate number of the assessments (39%) occurred during the late shift. Tissue Repair and Regeneration Laboratory (TR2Lab), Faculty of Sciences and A median of 1 drug per day was administered. Pain was recorded in 28% of patients’ Technology, University of Vic -­Central care plans, and non-­pharmacological interventions were recorded in 12%. University of Catalonia, Faculty of Medicine, University of Vic -­Central University In conclusion, exist variability in pain management nursing practice in older adults with of Catalonia. C. Sagrada Família, Vic, dementia. Admission diagnosis correlated with the analgesic administration schedule, Barcelona, Spain the number of drugs administered and the number of pain nursing annotations. Correspondence Mireia Subirana-­C asacuberta, Nursing KEYWORDS Department. Parc Taulí Hospital Universitari. Institut d’Investigació i Innovació Parc Taulí acute care, comfort, dementia, nurses, nursing care, pain management I3PT. Universitat Autònoma de Barcelona. Sabadell, Spain. Email: msubiranac@tauli.cat Funding information This project has been funded with the support of the “Osona Foundation for Health Research and Education (FORES) –­ Consultori Bayés.” 1 | I NTRO D U C TI O N 25%–­76% in people living in the community and as high as 83%–­ 93% in people living in nursing homes (Sáez-­L ópez et al., 2015), Pain and dementia are two health problems linked to ageing and due in part to the difficulty of evaluating and treating pain in two of the main causes of the reduction in quality of life in older the older population. This prevalence is similar to that of older people. Pain in older people tends to be chronic and mainly os- people with dementia elsewhere in Europe (Achterberg et al., teoarticular, although of mixed origin (Sirsch et al., 2015). The 2013, 2020; Corbett et al., 2014; Kruschinski et al., 2016), in prevalence of chronic pain in older people in Spain is estimated at Canada (Lynch, 2011) and in the United States (Patel et al., 2013). This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2021 The Authors. Nursing Open published by John Wiley & Sons Ltd. Nursing Open. 2021;8:3349–3357.  wileyonlinelibrary.com/journal/nop2 | 3349 3350 | MINAYA-­FREIRE et al. Managing pain in older people with dementia is complex, because Montoro-­Lorite & Canalias-­Reverter, 2015, 2018; Schofield, 2018; their ability to recognize and express pain may be affected (Álvaro Tsai et al., 2018). In a recent publication (Herr et al., 2019), the first González, 2010) and because pain sometimes manifests through step in the hierarchy is the assessment of the causes of potential behaviours that can be misinterpreted or overlooked as signs of pain, such as musculoskeletal or neurological problems. Once pain is pain (Karlsson et al., 2015). Pain can unleash or increase neuropsy- detected, it is necessary to establish treatment using individualized, chiatric symptoms such as agitation, restlessness and depression patient-­ centred pharmacological and non-­ pharmacological inter- (Regier & Gitlin, 2018). Frequently nurses associate these symp- ventions (Registered Nurses’ Association of Ontario (RNAO), 2013). toms more with dementia than with pain, and consequently, pa- Pharmacological therapy in older patients should aim to relieve pain tients are treated with antipsychotics instead of with analgesics and promote autonomy and functionality. The older population with (Corbett et al., 2014; Gilmore-­B ykovskyi & Bowers, 2013; Regier & dementia presents comorbidity and polypharmacy, which makes Gitlin, 2018). The poor pain management in people with cognitive pharmacological treatment difficult due to the risk of drug inter- decline is associated with sleep disorders, functional disabilities, ventions and greater risk of adverse effects (American Geriatrics falls, depression, cognitive impairment, anxiety, restlessness, de- Society, 2015 Beers Criteria Update Expert Panel, 2015; Eiche & creased physical activity, isolation, polypharmacy and malnutrition Schache, 2016; Sáez-­López et al., 2015). Non-­pharmacological ther- (Eiche & Schache, 2016; Kruschinski et al., 2016; Montoro-­L orite & apies are often used to enhance and/or supplement drug treatment Canalias-­Reverter, 2015; Sáez-­L ópez et al., 2015; Schofield, 2018). (Registered Nurses’ Association of Ontario (RNAO), 2013), and they Therefore, effective pain monitoring is needed in older people need to be given by trained and experienced professionals, such as with dementia to try to lessen pain's impact on their health and nurses, who are oriented to providing comfort (Kolcaba, 2003). In well-­b eing. people with dementia, non-­pharmacological therapies require more follow-­up (Achterberg et al., 2020; Schuler, 2019; Tick et al., 2018). After treatment, pain must be reassessed and recorded in pa- 2 | BAC KG RO U N D tient's health records. Entire pain management process must be clearly recorded so that pain can be monitored and patient's re- For pain to be monitored effectively, it should be considered as the sponse to interventions can be evaluated (Montoro-­Lorite & fifth vital sign and should be assessed systematically in all patients Canalias-­Reverter, 2018; Registered Nurses’ Association of Ontario using validated scales adapted to patients’ state and characteristics. (RNAO), 2013). Documentation should be systematic, standard- In 1996, The American Pain Society (APS) declared pain as “the 5th ized and accessible to all professionals (Montoro-­Lorite & Canalias-­ vital sign” with the objective to improve pain treatment and therefore Reverter, 2018). Proper documentation makes it possible to it must be evaluated alongside the other four vital signs (blood pres- formalize the process of pain assessment. It is essential for providing sure, heart rate, temperature and respiratory rate). Since then, many individualized care and ensuring the continuity of care, from a legal pain management protocols have been based on this assertion and and a professional perspective (Herr & Garand, 2001). Moreover, it have included pain assessment in routine monitoring of vital signs, favours patient-­centred care. using different approaches to pain assessment (Scher et al., 2018). In recent years, progress has been made in managing pain for Despite this statement, which it has not had the expected impact patients with dementia. Numerous studies have evaluated PAINAD on clinical practice, besides frequently, some professionals are not scale in different care settings (Fry et al., 2017; Goebel et al., 2019), trained to assess and treat pain properly (Morone & Weiner, 2013). reviewed pharmacological and non-­ pharmacological treatments According to the hierarchy of pain assessment techniques of for pain (Schuler, 2019), and updated protocols and guidelines for Pasero and McCaffery (2011), pain assessment in patients with de- pain management in older people with dementia (Herr et al., 2019; mentia should begin with a direct question about the presence and Schofield, 2018; Whitlock et al., 2017). intensity of pain, because in mild to moderate states of dementia, However, there are few studies focussed on nursing records person is able to describe and quantify his or her pain using sim- about pain and pain management, and most of these deal with ple scales such as the Numeric Rating Scale (NRS), the visual analog oncological pain (Song et al., 2015) or postsurgical pain (Erden scale or the faces rating scale (Achterberg et al., 2013, 2020; Herr et al., 2017; Heikkilä et al., 2016). A recent study reviewed records et al., 2011; Schofield, 2018). When these scales are used in patients related to episodes of pain of nursing home residents with dementia with dementia, questions must be adapted to the level of cognitive in Australia (Andrews et al., 2019), but as far as we know, no study decline, and sufficient time must be provided for the patient to has focussed on records related to older patients’ pain treated in process the question and formulate a response. It is also import- an Acute Geriatrics Unit (AGU) from the perspective of pain as ant to have the help of the caregivers (Herr et al., 2011; Sáez-­López the fifth vital sign. Reviewing nursing records related to pain man- et al., 2015). In non-­communicative patients, behavioural observa- agement enables us to view objectively what interventions nurses tion is recommended using scales such as Pain Assessment Checklist carried out, with what frequency and in connection to what other for Seniors with Limited Ability to Communicate (PACSLAC), Pain activities. Assessment in Advanced Dementia (PAINAD) or DOLOPLUS 2 (Coca This study is part of a broader mixed-­methods study that aims & Abeldaño Zuñiga, 2020; Fry & Elliott, 2018; Herr et al., 2011; to improve nursing practice in the pain management of patients MINAYA-­FREIRE et al. | 3351 with dementia treated in an AGU. This article presents the pre-­ 3.2 | Data Collection intervention phase results. The aim of this study is to assess pain management nursing practice in older adults with dementia through For data collection, a tool consisting of two sections was designed. electronic health records (EHR). The first section corresponded to the sociodemographic and clinical data of the patients: age, gender, length of stay (days), functional status, reason for hospitalization, surgical procedure, death, type 2.1 | Research question of dementia, stage of dementia, presence of delirium during admis- sion, analgesic administration schedule (fixed, on demand or none) The research questions were as follows: a) How and with what fre- and type of analgesic prescribed. We analysed the assessments for quency do AGU nurses assess pain in patients with dementia? b) the presence of delirium (Confusion Assessment Method, CAM) and How do they treat patients’ pain? c) Do nurses reassess pain in pa- the degree of functionality (Barthel Index, BI). These data were ex- tients with dementia? d) What factors influence pain management tracted from the medical history (discharge report), the drug pre- nursing practice? scription registry and the nurse evaluation records. A second section, where all pain-­ related variables were de- scribed in Table 1, was collected. For the pain variables collection, 3 | M E TH O DS the different phases of pain management were taken into account: evaluation, treatment, re-­evaluation and record keeping. 3.1 | Design, settings and participants To guarantee the validity and reliability of the data, the research team established a protocol for data collection that included the Retrospective descriptive study based on clinical practice nursing definition of the variables and the way in which they would be col- records (EHR) in the pain management of patients with demen- lected. We studied clinical variables for pain intensity according to tia admitted to the AGU of a university hospital in the province of the NRS or PAINAD (the two scales used in the hospital to assess Barcelona (Spain). We reviewed EHR of patients age ≥65 with a de- pain in adults) and the timing and frequency with which nurses as- mentia diagnosis (noted in the patient's discharge report), released sessed pain. NRS is a unidimensional tool. Asks persons to rate their from the AGU from January–­March 2018. AGU is a specialized geri- pain from 0–­10 and it is scored 0–­10 with the anchors of 0 being “no atric patient care unit that includes all patients aged ≥85 years and/ pain” and 10 being “pain as bad as it can be.” PAINAD is an obser- or with cognitive impairment and/or with advanced chronic condi- vational behavioural tool used in elderly with cognitive impairment. tions and/or who are institutionalized. It consists of 24 beds and a It consists of 5 items: breathing, facial expression, body language, staff of 18 Registered Nurses divided into three shifts and a ratio of negative vocalizations and consolability. Each item is rated on a scale one nurse for 12 patients. of 0–­2 for a total score from 0 (no pain)–­10 (severe pain); score 1 or TA B L E 1 Description of pain variables Baseline pain assessment Pain value recorded during the first 24 hr of stay Baseline pain intensity Numeric value from 0–­10 (11 means “not ratable” and was counted as “not assessed”) Maximum pain during stay Numeric value from 0–­10 (the highest value recorded during the patient's stay). Value 11 was considered invalid Number of pain assessments per day The total number of pain assessments recorded per day Pain intensity scale used during stay Numeric Rating Score (NRS) or Pain Assessment In Advanced Dementia (PAINAD) or both Number of drugs given per day Total number of doses of analgesics administered per day Care plan diagnosis If pain appeared in the care plan as a diagnosis or potential complication Non-­pharmacological interventions Record of non-­pharmacological interventions in the care plan Number of pain assessments>3 Extent to which pain intensity >3 was reassessed during the same shift Percentage of nursing annotations We reviewed all nursing annotations for the patient's that mentioned pain stay and counted how many specifically mentioned the presence or absence of pain. We excluded those that spoke of the presence or absence of unspecified discomfort 3352 | MINAYA-­FREIRE et al. 2 indicates some pain. In order to obtain information on pain vari- had a diagnosis of dementia in their medical report (inclusion crite- ables, the registry of clinical variables registry, drug administration rion), and therefore, these patients were included in the study. register, nursing clinical annotations and the patients' care plan were The 111 patients reviewed (Table 2) had a mean age of 87 (SD 5.1) consulted. years and 56% were women. They were hospitalized for a median of We reviewed all patient care documentation about each shift 6.0 days. over the each patient's entire length of stay. Sociodemographic data and pain variables recorded in the clinical data were analysed TA B L E 2 Description of the sociodemographic and clinical in an automated mode. The variables related to the prescription profile of patients (n = 111) and administration of drugs and the nurses’ clinical annotations Variables n (%) were reviewed manually by the first author. There is no automated Gender (Female) 62 (55.9%) system for classifying clinical annotations. Nurses are not provided Age, mean (SD) 87.0 (5.1) with a registration system with pre-­existing fields, they evaluate Length of stay (days), median (IQR) 6.0 (4.0–­8.0) patient pain and records pain intensity in clinical variables section. Then, they plan the care based on the result of pain evaluation and Barthel index, median (IQR) 50 (38.8–­75.0) adjusted according to the needs detected in the nursing evalua- Reason for admission tion. In our AGU, according to the multidisciplinary pain protocol, Infection (respiratory, urinary or abdominal) 67 (60.4%) pain intensity should be assessed at least once per day (late shift) Fracture/Osteoarticular Problem 12 (10.8%) and assessment frequency should be increased if pain is detected. Cardiovascular Problem 22 (19.8%) Respiratory Problem 8 (7.2%) Other 2 (1.8%) 3.3 | Data Analysis Surgical Procedure (yes) 6 (5.4%) Death (yes) 8 (7.2%) We conducted a descriptive analysis of the sample. Categorical var- Type of dementia iables were expressed in frequencies and percentages. Quantitative Cognitive decline of unknown type 52 (46.8%) variables were presented in means and standard deviation (SD) for Alzheimer 31 (27.9%) those with a normal distribution and median and inter-­ quartile range (IQR) for those with a non-­normal distribution. For the bivari- Mixed 10 (9.0%) ate analysis, we used a chi-­square test (Fisher's exact test) to ana- Vascular Dementia 10 (9.0%) lyse the relationship between two categorical variables. To analyse Parkinson's 4 (3.6%) the relationship between a categorical variable and a quantitative Degenerative 2 (1.8%) variable, we used the non-­parametric Mann–­Whitney U test (when Alzheimer's or Lewy Body Dementia 2 (1.8%) the categorical variable had two categories) and Kruskal–­Wallis Dementia stage test (when the categorical variable had more than two categories). Mild 25 (22.5%) Finally, to analyse the correlation between two quantitative vari- Mild-­Moderate 6 (5.2%) ables, we used Spearman's correlation coefficient. For the entire Moderate 31 (27.9%) statistical analysis, a confidence level of 95% was assumed (p <.05). Moderate-­Severe 11 (9.9%) Severe 25 (22.5%) Not given 13 (11.7%) 3.4 | Ethics Delirium (yes) 84 (75.7%) The study was approved by the Ethics Committee for Clinical Analgesic administration schedule Research of the Osona Foundation for Health Research and Fixed 32 (30.2%) Education. The researchers were employed at the AGU and used As needed 74 (69.8%) their passwords codes to access EHR. Once the data were extracted, No prescription 5 (4.5%) we unlinked it from patient information so that patients could not Prescribed Drugs (on demand or fixed) be identified. Only the research team was allowed to use the data Paracetamol 106 (95.5%) entirely for scientific ends. Dipyrone 31 (27.9%) Morphine chloride 24 (21.6%) Dexketoprofen 6 (5.4%) 4 | R E S U LT S Tramadol 7 (6.3%) Oxycodone 2 (1.8%) During the first trimester of 2018, 284 patients were discharged from Other (including comfort guidelines) 16 (14.4%) the AGU. Clinical records showed that 111 (39%) of these patients MINAYA-­FREIRE et al. | 3353 Table 3 shows results related to pain management. The inten- A median of 1 drug per day were administered. Pain figured as a sity of baseline pain was evaluated in 88% of patients. Pain in- potential complication or as diagnosis of acute pain in 28% of nursing tensity was assessed a median of 1.9 times per day of stay, and care plans and 12% included a record of non-­pharmacological inter- proportionally more of these assessments were made during the ventions (cold packs, changing positions and/or end-­of-­life comfort late shift (39%). care provided at the end of life). We reviewed 2,346 annotations written in free text in patients’ nursing records, of which in 592 (25.2%) nurses referred to pain. The mean percentage of nursing an- TA B L E 3 Pain Management Results (n = 111) notations that mentioned pain was 23.8 (SD 19.7). Variables n(%) Table 4 shows relationships between pain variables. There are Baseline pain assessment 98 (88.3%) several significant relationships between the number of drugs ad- Baseline pain intensity (n = 98) ministered per day, gender, reason for admission, surgical procedure and the type of analgesic prescription. Patients received more drugs No Pain (NRS/PAINAD=0) 92 (93.9%) if they were female, if they were admitted with a diagnosis of frac- Mild (NRS/PAINAD=1–­3) 6 (6.1%) ture or osteoarticular problem, if they underwent a surgical proce- Moderate (NRS/PAINAD=4–­6) 0 (0.0%) dure, if they died and if they had a fixed analgesic administration Intense (NRS/PAINAD >6) 0 (0.0%) schedule. Women made up 66.7% of patients undergoing surgical Pain intensity scale used during stay (n = 110) procedure, 50% of deaths and 75% of those admitted for a facture NRS only 26 (23.6%) or an osteoarticular problem. Of those that had a fixed analgesic pre- PAINAD only 4 (3.7%) scription, 78% were women (data not shown in table). Both scales 80 (72.7%) According to Spearman's correlation coefficient, there is an asso- Maximum pain during stay, NRS (n = 104) a ciation between the maximum pain intensity according to the NRS, No Pain (NRS=0) 73 (70.2%) the number of drugs administered (ρ = .399; p <.001) and the per- Mild (NRS=1–­3) 27 (26.0%) centage of nursing annotations about pain (ρ = .509; p <.001). The Moderate (NRS=4–­6) 2 (1.9%) greater the pain intensity according to the NRS, the greater the num- ber of drugs administered and the greater number of annotations Intense (NRS>6) 2 (1.9%) about pain in the patient's nursing records (data not shown in the Maximum pain during the stay, PAINAD (n = 84) table). We did not find statistically significant relationship between No Pain (PAINAD=0) 45 (53.5%) maximum pain intensity according to PAINAD and the number of Mild (PAINAD=1–­3) 34 (40.5%) drugs administered per day or the percentage of nursing annotations Moderate (PAINAD=4–­6) 4 (4.8%) about a patient mentioning pain. Severe (PAINAD >6) 1 (1.2%) Percentage of nursing annotations about pain was higher when Number of pain assessments per day, median 1.9 (1.5 –­2.3) the admission diagnosis was fracture or osteoarticular problem or (IQR) when the patient had undergone surgery during hospitalization. % of assessments per shift, mean (SD) Early 26.4(13.5) Late 39.9 (16.5) 5 | D I S CU S S I O N Night 33.8 (14.7) Number of administrations per day, median (IQR) 0.8 (0.14–­2.11) Patients treated at the AGU during the study period were patients Care plan with a high mean age who needed help to perform basic activities of Potential complication 32 (28.8%) daily life and most were admitted because of an infection. Therefore, Diagnosis of acute pain 1 (0.9%) they represented a profile of patient likely to suffer from chronic pain (Achterberg et al., 2013, 2020; Corbett et al., 2014; Kruschinski Not given 78 (70.3%) et al., 2016; Lynch, 2011; Patel et al., 2013; Sáez-­López et al., 2015; Non-­pharmacological interventions recorded Sirsch et al., 2015). Moreover, patients were likely to be suffering Cold application 9 (8.1%) from acute pain linked to the health problem that motived admission Position change 1 (0.9%) to the AGU. End-­of-­life care 3 (2.7%) According to study results, the nurses of the AGU assessed Not given 98 (88.3%) the baseline pain of 88% of patients when they were admitted % of nursing annotations about pain, median 18.5 to the unit and reassessed a median of 1.9 times per day during (IQR) (10.8–­3 0.0) their hospital stay, with a greater frequency during the late Number of pain assessments>3 11 (0.8%) shift. These results contrast with those obtained by Andrews Number of pain reassessments>3 5 (45.5%) et al., (2019), which revealed that nurses asked about pain in only a The value 11 has been excluded. 2% of patients with dementia in a nursing home. Perhaps this is 3354 | MINAYA-­FREIRE et al. TA B L E 4 Relationship between sociodemographic and clinical variables in pain management Analgesic prescription Pain assessments Number of drugs Percent of nursing if necessary per day, median given per day, annotations that mention n = 111 n (%) (IQR) median (IQR) pain, median (IQR) Gender Male 49 39 (84.8) 1.9 (1.6–­2.4) 0.4 (0–­1) 16.7 (8.6–­27) Female 62 35 (58.3) 1.8 (1.4–­2.2) 1.4 (0.4–­2.7) 20.5 (13.1–­35.9) P-­value .003* .263 <.001* .087 Reason for admission Infection 67 53 (82.8) 1.8 (1.4–­2.1) 0.8 (0.1–­1.7) 17.1 (10.8–­25) Fracture/Osteoarticular 12 0 (0.0) 2.2 (1.9–­2.5) 4.8 (3.3–­5.5) 66.8 (50.8–­8 0.5) problem Cardiovascular problem 22 14 (66.7) 2.0 (1.7–­2.4) 0.7 (0–­2.4) 21.8 (9.8–­31.8) Respiratory problem 8 6 (85.7) 1.8 (1.5–­2.1) 0.3 (0–­0.6) 16.7 (9.6–­25.6) Other 2 1 (50) 1.6 (1.5-­.) 2l. 1 (0-­.) 21.7 (0-­.) P-­value <.001* .145 <.001* .001* Surgical procedure No 105 73 (73.0) 1.9 (1.5–­2.2) 0.7 (0.1–­1.9) 18.5 (10.7–­28.6) Yes 6 1 (16.7) 2.2 (1.7–­2.7) 4.8 (3.2–­5.4) 64.1 (34.9–­8 4.7) P-­value .009* .123 .003* .005* Death No 103 69 (70.4) 1.9 (1.5–­2.3) 0.7 (0.1–­2.0) 18.8 (11.1–­3 0) Yes 8 5 (62.5) 2.0 (1.8–­2.3) 1.6 (0.7–­2.9) 11.8 (1.9–­3 0.4) P-­value .695 .500 .099 .302 Analgesic administration schedule Fixed 32 -­-­-­ 2.0 (1.8–­2.4) 2.8 (1.8–­4.4) 19.5 (10.2–­60.2) As needed 74 -­-­-­ 1.8 (1.4–­2.1) 0.3 (0–­1) 18.5 (11–­28) P-­value -­-­-­ .053 <.001* .266 *P-­value<0.05 due to the differences between care organization in acute setting Nurses used the validated NRS (self-­report) and/or PAINAD (ob- and nursing homes (e.g. patient/nurse ratio or constant monitor- servational) scales for pain intensity assessment (using both scales in ing). However, our results are similar to those reported by Song 72% of patients). According to best practice guidelines (Registered et al., (2015) for cancer patients in hospital setting. These differ- Nurses’ Association of Ontario (RNAO), 2013), it is important to use ences may be due to the fact that in an acute setting, vital signs the same tool to be able to assess the presence and intensity of pain assessment –­including measuring pain as the fifth vital sign –­is precisely and to evaluate the patient's response to interventions. standard nursing procedure. Even so, the AGU’s person-­centred The use of both scales might be attributable to the presence of de- care model treats pain as a geriatric syndrome and defines patient lirium during admission (76% according to the CAM scale) and the comfort as the aim of nursing care. In this context, the fact that capacity of the patient with neuropsychiatric symptoms to report nurses did not assess pain with the same frequency during the pain feeling. These facts could explain why nurses adapted the scale various shifts may indicate that pain management in the unit is less to the patient. Goebel et al., (2019) endorse the use of the PAINAD than optimal and that probably not all nurses consider the need scale in patients with delirium. of pain assessment in patient with dementia. The pain protocol Only for 12% of patients were non-­pharmacological interven- should be revised to base nursing care on the available evidence. tions listed in care plan, although that does not necessarily mean Several authors report that nurses lack knowledge of how to man- that such interventions did not occur. In older patients with demen- age pain in patients with dementia and that they need continuing tia –­a patient profile in which the ability to treat with drugs is limited education (Achterberg et al., 2013; De Witt Jansen et al., 2017a, –­the tailored use of non-­pharmacological therapies can relieve or 2017b; Lichtner et al., 2016; Schofield, 2018). This suggests that prevent pain and decrease drug use. It would therefore be desirable the variability in practice discovered in our study may be due to to improve the recording of such interventions in order to assess lack of training. their effectiveness and ensure continuity of care between shifts. MINAYA-­FREIRE et al. | 3355 Strikingly, patients that had no pain or little pain when they were themselves, pain treatment can be delayed. This could explain our find- admitted to the AGU, but some patients had moderate or intense ings: nurses administered more drugs and pay more attention to pain pain at some time during their stay. This finding pointed out the need when patients were able to report pain feeling and request painkillers. for further analysis about factors that affect optimal pain manage- Pain has a very negative impact on health and well-­being of the ment in our unit. In addition, only 45% of pain values greater than 3 person with dementia, so it is necessary to systematize pain man- led to the patient being reassessed during the same shift, meaning agement, based on the best available evidence to avoid variability in that nurses did not follow up on interventions they performed for nurse practice and to be able to provide safe, high-­quality care. For pain management and they did not develop an individualized care that, it is essential that pain be considered the fifth constant. plan that would take into account successful interventions or possi- ble adverse effects. In the EHR we reviewed, there was no evidence of a compre- 5.1 | Limitations hensive pain assessment that identified the location, type and characteristics of pain. Only in some cases was the location of pain Data for some variable were collected manually from several infor- documented in the nursing annotations that were written in free mation systems and by the first author. This task was very labori- text. These data are comparable to claims by Dowding et al., (2016) ous, despite the apparently small sample size (n = 111), given that and Lichtner et al., (2016) that pain-­related documentation is frag- it covered a wide range of types of information, including clinical mented, inaccessible and poorly organized and does not provide variables, annotations about the prescription and administration of enough information to enable professionals to carry out effective drugs, clinical nursing annotations and patient care plans. When re- pain management. Recording all the interventions performed by the viewing clinical annotations, it would have been very useful to have nurse in patient's nursing record is an arduous, time-­consuming task. an automated system capable of classifying them by headers (Moen When annotations are written in free text without a specific struc- et al., 2018, 2020), making it possible to extract the information in ture, they are difficult to review. Currently, there are automated a more systematic way. Even so, we tried to include all records and systems that make possible to classify annotations in clinical nurs- annotations relevant to pain and were consistent in our techniques ing records by topic. Using such system could reduce nurses’ time for collecting data. need for recording patients’ clinical course. Such system could also On the other hand, this study report results of a single AGU and be useful for extracting data for research (Moen et al., 2018, 2020). centre, and this limits generalizability. According to our analysis, the reason for hospital admission in- fluenced the type of prescription, the number of drugs per day that were administered and the number of nursing annotations that 6 | CO N C LU S I O N mentioned the absence or presence of pain. According to Lichtner et al., (2016), the reason for admission influences professionals’ ex- The main conclusion is the variability in nursing practice in pain pectations about pain and how it should be treated. A surgical patient management of older people with dementia, which sure impact on is expected to have more pain and therefore pain in such patients nursing care quality and safety. Probably pain it is not considered by tends to be treated routinely and almost exclusively with drugs ac- nurses the fifth constant in this population. The reason for admission cording to a fixed schedule, contrasting with patients who are ad- and gender influenced the prescription of analgesics and the number mitted for other reasons, such as infection (Sampson et al., 2015). of drugs administered. Additionally, more drugs were administered Additionally, when the patient was at the end of life, nurses admin- and nurses wrote more annotations about pain when patients self-­ istered fixed pain treatment without questioning the presence of reported higher intensity pain (NRS), but not when nurses observed pain, which is consistent with a report by Gilmore-­Bykovskyi and higher intensity pain (PAINAD). Bowers (2013). At the same time, the percentage of female patients Given these outcomes, it is important to study in depth how pain who were prescribed analgesics was higher, and they also received scales are used to measure pain, and their frequency (taking into ac- more drugs than men. This finding might be explained by the fact count that it should be considered fifth constant and measured with that women have traditionally been considered more fragile and less the same frequency as the rest of vital signs) and their impact on pain-­tolerant than men or by the fact that most of the nurses in our subsequent decision-­making. It is also important to analyse barriers unit are women. The issue of gender should be investigated further and facilitators that influence pain management in patients with de- due to their probably impact on pain interventions. mentia from the perspective of nurses and try to find practical, tai- When patients reported more pain (NRS), more drugs were given lored solutions for the acute geriatric setting in order to guaranteed and more nursing annotations mentioned pain were recorded. This did patients’ well-­being. not occur when pain was identified through observation (PAINAD), even though PAINAD assessments showed that 47% of patients had AC K N OW L E D G E M E N T S pain, in contrast to only 30% of patients who reported pain on the This project has been funded with the support of the “Osona NRS. According to Gilmore-­Bykovskyi and Bowers (2013), in patients Foundation for Health Research and Education (FORES) –­Consultori with dementia and difficulties reporting their pain and speaking up for Bayés.” 3356 | MINAYA-­FREIRE et al. C O N FL I C T O F I N T E R E S T with dementia. British Medical Bulletin, 111(1), 139–­148. https://doi. org/10.1093/bmb/ldu023 The authors have no conflicts of interest to disclose. De Witt Jansen, B., Brazil, K., Passmore, P., Buchanan, H., Maxwell, D., McIlfactrick, S. J., Morgan, S. M., Watson, M., & Parsons, C. (2017). AU T H O R S C O N T R I B U T I O N Nurses’ experiences of pain management for people with advanced Conception and design: Alícia Minaya-­Freire, Mireia Subirana-­ dementia approaching the end of life: A qualitative study. Journal of Clinical Nursing, 26(9–­10), 1234–­1244. https://doi.org/10.1111/ Casacuberta, Emma Puigoriol-­Juvanteny, Anna Ramon-­Aribau, jocn.13442 Data Collection: Alícia Minaya-­Freire, Emma Puigoriol-­Juvanteny, Dowding, D., Lichtner, V., Allcock, N., Briggs, M., James, K., Keady, J., Analysis and interpretation of data: Alícia Minaya-­Freire, Mireia Lasrado, R., Sampson, E. L., Swarbrick, C., & José Closs, S. (2016). Subirana-­C asacuberta, Emma Puigoriol-­Juvanteny, Anna Ramon-­ Using sense-­making theory to aid understanding of the recognition, Aribau, Drafting the Manuscript: Alícia Minaya-­Freire, Mireia assessment and management of pain in patients with dementia in acute hospital settings. International Journal of Nursing Studies, 53, Subirana-­C asacuberta, Emma Puigoriol-­Juvanteny, Anna Ramon-­ 152–­162. https://doi.org/10.1016/j.ijnur​stu.2015.08.009 Aribau, Final approval of the version to be published: Alícia Minaya-­ Eiche, J., & Schache, F. (2016). Pain Management in geriatric patients. Freire, Mireia Subirana-­C asacuberta, Emma Puigoriol-­Juvanteny, Deutsche Medizinische Wochenschrift, 141(9), 635–­6 41. https://doi. Anna Ramon-­Aribau. org/10.1055/s-­0 041-­106307 Erden, S., Arslan, S., Deniz, S., Kaya, P., & Gezer, D. (2017). A review of postoperative pain assessment records of nurses. Applied Nursing DATA AVA I L A B I L I T Y S TAT E M E N T Research, 38, 1–­4. https://doi.org/10.1016/j.apnr.2017.08.003 The data that support the findings of this study are available on Fry, M., Arendts, G., & Chenoweth, L. (2017). Emergency nurses’ evalu- request from the corresponding author, [MSC]. The data are not ation of observational pain assessment tools for older people with cognitive impairment. Journal of Clinical Nursing, 26(9–­10), 1281–­ publicly available due to their containing information that could 1290. https://doi.org/10.1111/jocn.13591 compromise the privacy of research participants. Fry, M., & Elliott, R. (2018). Pragmatic evaluation of an observational pain assessment scale in the emergency department: The Pain Assessment ORCID in Advanced Dementia (PAINAD) scale. Australasian Emergency Care, 21(4), 131–­136. https://doi.org/10.1016/j.auec.2018.09.001 Alícia Minaya-­Freire https://orcid.org/0000-0002-4103-363X Gilmore-­Bykovskyi, A. L., & Bowers, B. J. (2013). Understanding nurses’ Mireia Subirana-­Casacuberta https://orcid. decisions to treat pain in nursing home residents with demen- org/0000-0003-1577-4159 tia. Research in Gerontological Nursing, 6(2), 127–­ 138. https://doi. Emma Puigoriol-­Juvanteny https://orcid. org/10.3928/19404​921-­20130​110- ­02 Goebel, J. R., Ferolito, M., & Gorman, N. (2019). Pain screening in the org/0000-0002-5510-2480 older adult with delirium. Pain Management Nursing, 20(6), 519–­525. Anna Ramon-­Aribau https://orcid.org/0000-0003-3303-9932 https://doi.org/10.1016/j.pmn.2019.07.003 Heikkilä, K., Peltonen, L.-­M., & Salanterä, S. (2016). Postoperative REFERENCES pain documentation in a hospital setting: A topical review. Achterberg, W., Lautenbacher, S., Husebo, B., Erdal, A., & Herr, K. (2020). Scandinavian Journal of Pain, 11(1), 77–­89. https://doi.org/10.1016/j. Pain in Dementia. PAIN Reports, 5, e803. https://doi.org/10.1097/ sjpain.2015.12.010 PR9.00000​0 0000​0 00803 Herr, K., Coyne, P. J., Ely, E., Gélinas, C., & Manworren, R. C. B. (2019). Achterberg, W., Pieper, M. J. C., van Dalen-­Kok, A. H., de Waal, M. W. Pain assessment in the patient unable to self-­report: clinical practice M., Husebo, B. S., Lautenbacher, S., Kunz, M., Scherder, E. J. A., & recommendations in support of the aspmn 2019 position statement. Corbett, A. (2013). Pain management in patients with dementia. Pain Management Nursing, 20(5), 404–­417. https://doi.org/10.1016/j. Clinical Interventions in Aging, 8, 1471–­1482. https://doi.org/10.2147/ pmn.2019.07.005 CIA.S36739 Herr, K., Coyne, P., McCaffery, M., Manworren, R., & Merkel, S. Álvaro González, L. C. (2010). El neurólogo frente al dolor en la de- (2011). Pain assessment in the patient unable to self-­report: po- mencia. Neurologia, 30(9), 574–­ 585. https://doi.org/10.1016/j. sition statement with clinical practice recommendations. Pain nrl.2012.02.001 Management Nursing, 12(4), 230–­ 250. https://doi.org/10.1016/j. American Geriatrics Society 2015 Beers Criteria Update Expert Panel pmn.2011.10.002 (2015). American Geriatrics Society 2015 Updated Beers Criteria for Herr, K., & Garand, L. (2001). Assessment and measurement of pain in Potentially Inappropriate Medication Use in Older Adults. Journal older adults. Clinics in Geriatric Medicine, 17(3), 457–­478. https://doi. of the American Geriatrics Society, 63(11), 2227–­2246. https://doi. org/10.1016/S0749​- ­0690(05)70080​-­X org/10.1111/jgs.13702 Jansen, B. D. W., Brazil, K., Passmore, P., Buchanan, H., Maxwell, D., Andrews, S. M., Dipnall, J. F., Tichawangana, R., Hayes, K. J., Fitzgerald, McIlfatrick, S. J., Morgan, S. M., Watson, M., & Parsons, C. (2017). J. A., Siddall, P., Poulos, C., & Cunningham, C. (2019). An Exploration Exploring healthcare assistants’ role and experience in pain assess- of Pain Documentation for People Living with Dementia in Aged ment and management for people with advanced dementia towards Care Services. Pain Management Nursing, 20(5), 475–­481. https://doi. the end of life: A qualitative study. BMC Palliative Care, 16(1), 6. org/10.1016/j.pmn.2019.01.004 https://doi.org/10.1186/s1290​4-­017-­0184-­1 Coca, S. M., & Abeldaño Zuñiga, R. A. (2020). Instruments for pain as- Karlsson, C. E., Ernsth Bravell, M., Ek, K., & Bergh, I. (2015). Home health- sessment in patients with advanced dementia: A systematic review care teams’ assessments of pain in care recipients living with demen- of the evidence for Latin America. Palliative and Supportive Care, tia: A Swedish exploratory study. International Journal of Older People 18(6), 741–­747. https://doi.org/10.1017/S1478​95151​900107X Nursing, 10(3), 190–­200. https://doi.org/10.1111/opn.12072 Corbett, A., Husebo, B. S., Achterberg, W. P., Aarsland, D., Erdal, A., & Kolcaba, K. (2003). Comfort theory and practice: A vision for holistic health Flo, E. (2014). The importance of pain management in older people care and research. Springer Publishing Company. MINAYA-­FREIRE et al. | 3357 Kruschinski, C., Wiese, B., Dierks, M.-­L ., Hummers-­Pradier, E., Schneider, el anciano. Revista De La Sociedad Española Del Dolor, 22(6), 271–­274. N., & Junius-­Walker, U. (2016). A geriatric assessment in general https://doi.org/10.4321/S1134​-­8 0462​01500​0600008 practice: Prevalence, location, impact and doctor-­patient percep- Sampson, E. L., White, N., Lord, K., Leurent, B., Vickerstaff, V., Scott, tions of pain. BMC Family Practice, 17, 8. https://doi.org/10.1186/ S., & Jones, L. (2015). Pain, agitation, and behavioural problems in s1287​5- ­016- ­0 409-­z people with dementia admitted to general hospital wards: A longitu- Lichtner, V., Dowding, D., Allcock, N., Keady, J., Sampson, E. L., Briggs, dinal cohort study. Pain, 156(4), 675–­683. https://doi.org/10.1097/j. M., Corbett, A., James, K., Lasrado, R., Swarbrick, C., & Closs, S. J. pain.00000​0 0000​0 00095 (2016). The assessment and management of pain in patients with Scher, C., Meador, L., Van Cleave, J. H., & Reid, M. C. (2018). Moving be- dementia in hospital settings: A multi-­c ase exploratory study from yond pain as the fifth vital sign and patient satisfaction scores to im- a decision making perspective. BMC Health Services Research, 16(1), prove pain care in the 21st century. Pain Management Nursing, 19(2), 427. https://doi.org/10.1186/s1291​3-­016-­1690-­1 125–­129. https://doi.org/10.1016/j.pmn.2017.10.010 Lynch, M. E. (2011). The need for a canadian pain strategy. Pain Research and Schofield, P. (2018). The assessment of pain in older people: UK national Management, 16(2), 77–­8 0. https://doi.org/10.1155/2011/654651 guidelines. Age and Ageing, 47, i1–­i22. https://doi.org/10.1093/agein​ Moen, H., Hakala, K., Peltonen, L. M., Suhonen, H., Ginter, F., Salakoski, g/afx192 T., & Salanterä, S. (2020). Supporting the use of standardized nurs- Schuler, M. (2019). Treatment of pain in people with dementia. ing terminologies with automatic subject heading prediction: A Zeitschrift Fur Gerontologie Und Geriatrie, 52(6), 607–­622. https://doi. comparison of sentence-­level text classification methods. Journal of org/10.1007/s0039​1-­019-­01578​- ­4 the American Medical Informatics Association : JAMIA, 27(1), 81–­88. Sirsch, E., Gnass, I., & Fischer, T. (2015). Diagnostics of pain in old age. https://doi.org/10.1093/jamia/​ocz150 Perspectives on a Multidimensional Phenomenon. Schmerz, 29(4), 339–­ Moen, H., Hakala, K., Peltonen, L.-­M., Suhonen, H., Loukasmäki, P., 348. https://doi.org/10.1007/s0048​2-­015-­0 026-­0 Salakoski, T., & Salanterä, S. (2018). Evaluation of a prototype sys- Song, W., Eaton, L. H., Gordon, D. B., Hoyle, C., & Doorenbos, A. Z. tem that automatically assigns subject headings to nursing narratives (2015). Evaluation of evidence-­ based nursing pain management using recurrent neural network. In Proceedings of the 9th interna- practice. Pain Management Nursing, 16(4), 456–­ 463. https://doi. tional workshop on health text mining and information analysis (pp. 94–­ org/10.1016/j.pmn.2014.09.001 100). Brussels, Belgium: Association for Computational Linguistics. Tick, H., Nielsen, A., Pelletier, K. R., Bonakdar, R., Simmons, S., Glick, R., https://doi.org/10.18653/​v1/W18-­5611 Ratner, E., Lemmon, R. L., Wayne, P., & Zador, V. (2018). Evidence-­ Montoro-­Lorite, M., & Canalias-­Reverter, M. (2015). Dolor y demencia based nonpharmacologic strategies for comprehensive pain care: avanzada. Revisión Bibliográfica. Gerokomos, 26(4), 142–­147. the consortium pain task force white paper. Explore, 14(3), 177–­211. Montoro-­Lorite, M., & Canalias-­Reverter, M. (2018). Protocolos de https://doi.org/10.1016/j.explo​re.2018.02.001 gestión del dolor en demencia avanzada. Enfermería Clínica, 28(3), Tsai, I.-­P.-­P., Jeong, S.-­Y.-­S.-­S., & Hunter, S. (2018). Pain assessment and 194–­204. https://doi.org/10.1016/J.ENFCLI.2017.06.003 management for older patients with dementia in hospitals: an in- Morone, N. E., & Weiner, D. K. (2013). Pain as the fifth vital sign: Exposing tegrative literature review. Pain Management Nursing, 19(1), 54–­71. the vital need for pain education. Clinical Therapeutics, 35(11), 1728–­ https://doi.org/10.1016/j.pmn.2017.10.001 1732. https://doi.org/10.1016/j.clint​hera.2013.10.001 Whitlock, E. L., Diaz-­Ramirez, L. G., Glymour, M. M., Boscardin, W. J., Pasero, C., & McCaffery, M. (2011). Pain assessment and pharmacological Covinsky, K. E., & Smith, A. K. (2017). Association between persistent management. Mosby. pain and memory decline and dementia in a longitudinal cohort of el- Patel, K. V., Guralnik, J. M., Dansie, E. J., & Turk, D. C. (2013). Prevalence ders. JAMA Internal Medicine, 177(8), 1146. https://doi.org/10.1001/ and impact of pain among older adults in the united states: findings jamai​ntern​med.2017.1622 from the 2011 national health and aging trends study. Pain, 154(12), 2649–­2657. https://doi.org/10.1016/j.pain.2013.07.029 Regier, N. G., & Gitlin, L. N. (2018). Dementia-­related restlessness: How to cite this article: Minaya-­Freire A, Subirana-­ Relationship to characteristics of persons with dementia and family caregivers. International Journal of Geriatric Psychiatry, 33(1), 185–­ Casacuberta M, Puigoriol-­Juvanteny E, Ramon-­Aribau A. Pain 192. https://doi.org/10.1002/gps.4705 management nursing practice assessment in older adults with Registered Nurses’ Association of Ontario (RNAO). (2013). Assessment dementia. Nurs Open. 2021;8:3349–­3357. https://doi. and Management of Pain. Third Edition. Retrieved February 28, org/10.1002/nop2.880 2016, from www.rnao.ca/bestp​ractices Sáez-­López, M. P., Sánchez-­Hernández, N., Jiménez-­Mola, S., Alonso-­ García, N., & Valverde-­García, J. A. (2015). Valoración del dolor en