Pain Cues in People With Dementia: Scoping Review

JMIR Preprints Smrke et al Decoding Pain: A Scoping Review of Pain Cues in People with Dementia Urška Smrke, Ana Milošič, Izidor Mlakar, Matic Kadiš, Satja Mulej Bratec Submitted to: JMIR Mental Health on: April 14, 2025 Disclaimer: © The authors. All rights reserved. This is a privileged document currently under peer-review/community review. Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review purposes only. While the final peer-reviewed paper may be licensed under a CC BY license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Table of Contents Original Manuscript ....................................................................................................................................................................... 5 Supplementary Files ..................................................................................................................................................................... 41 ................................................................................................................................................................................................... 41 Figures ......................................................................................................................................................................................... 42 Figure 1 ...................................................................................................................................................................................... 43 Multimedia Appendixes ................................................................................................................................................................. 44 Multimedia Appendix 0 .................................................................................................................................................................. 45 CONSORT (or other) checklists ...................................................................................................................................................... 46 CONSORT (or other) checklist 0 ...................................................................................................................................................... 46 https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Decoding Pain: A Scoping Review of Pain Cues in People with Dementia Urška Smrke1* PhD; Ana Miloši?1* MA; Izidor Mlakar1 PhD; Matic Kadiš2* MA; Satja Mulej Bratec2* PhD 1 Faculty of Electrical Engineering and Computer Science University of Maribor Maribor SI 2Department of Psychology Faculty of Arts University of Maribor Maribor SI *these authors contributed equally Corresponding Author: Urška Smrke PhD Faculty of Electrical Engineering and Computer Science University of Maribor Koroška cesta 46 Maribor SI Abstract Background: Individuals with dementia, especially those in later stages, have difficulties with verbally reporting their experience of pain. This results in both underassessment and undertreatment of pain, signalling the need for better pain recognition in persons with dementia. A promising form of pain assessment is digital monitoring, which can concurrently and more objectively detect and utilize numerous relevant pain cues. Objective: The present review aimed to identify observable cues of pain, which could be used for digital pain monitoring. We set out to examine which digital cues offered a valid insight into pain in people with dementia (RQ1) and identify how these cues were originally measured (RQ2). Methods: A standard methodological approach for scoping reviews was utilized. Relevant research papers were chosen based on SCOPUS and Web of Science databases, and relevant data on pain cues were extracted from all papers that satisfied the inclusion criteria. The gathered data were analyzed using a thematic analysis, which involved categorizing the observable cues into higherorder categories. Results: Of the 3705 publications identified in the search, 34 satisfied the inclusion criteria and were closely examined. Addressing RQ1, we could identify seven categories of behavioral and physiological cues associated with pain. Several subcategories for each main category of pain cues were identified, each involving between one and 28 relevant specific pain cues. Addressing RQ2, most of the reported pain cues were assessed via human observation. A few studies additionally included facial recognition software or combined human observation with computer vision. Conclusions: The review provides a comprehensive list of the most relevant cues that signify pain in persons with dementia and offers a foundation for the use of artificial intelligence and digital monitoring for the screening of pain in dementia. Clinical Trial: Na (JMIR Preprints 14/04/2025:75671) DOI: https://doi.org/10.2196/preprints.75671 Preprint Settings 1) Would you like to publish your submitted manuscript as preprint? Please make my preprint PDF available to anyone at any time (recommended). Please make my preprint PDF available only to logged-in users; I understand that my title and abstract will remain visible to all users. Only make the preprint title and abstract visible. No, I do not wish to publish my submitted manuscript as a preprint. 2) If accepted for publication in a JMIR journal, would you like the PDF to be visible to the public? Yes, please make my accepted manuscript PDF available to anyone at any time (Recommended). Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand that the title and abstract will remain visible to all users (see Important note, above). I also understand that if I later pay to participate in <a href="https://jmir.zendesk.com/hc/en-us/articles/360008899632-What-is-the-PubMed-Now-ahead-of-print-option-when-I-pay-the-APF-" target="_blank">JMIR’s PubMed Now! service</a> service, my accepted manuscript PDF will automatically be made openly available. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Yes, but only make the title and abstract visible (see Important note, above). I understand that if I later pay to participate in <a href="https://jmir.zendesk.com/hc/en-us/articles/360008899632-What-is-the-PubMed-Now-ahead-of-print-option-when-I-pay-the-APF-" target="_blank">JMIR’s PubMed Now! service</a> service, my accepted manuscript PDF will automatically be made openly available. No. Please do not make my accepted manuscript PDF available to anyone. I understand that if I later pay to participate in <a href="https://jmir.zendesk.com/hc/en-us/articles/360008899632-What-is-the-PubMed-Now-ahead-of-print-option-when-I-pay-the-APF-" target="_blank">JMIR’s PubMed Now! service</a>, my accepted manuscript PDF will automatically be made openly available. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Original Manuscript https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Title Decoding Pain: A Scoping Review of Pain Cues in People with Dementia Authors Urška Smrke1*, Ana Milošič1*, Izidor Mlakar1, Matic Kadiš2#, Satja Mulej Bratec2# *Authors contributed equally. #Authors contributed equally. Affiliations 1Faculty of Electrical Engineering and Computer Science, University of Maribor, Maribor, Slovenia 2Department of Psychology, Faculty of Arts, University of Maribor, Maribor, Slovenia Corresponding Author Urška Smrke Faculty of Electrical Engineering and Computer Science University of Maribor Koroška cesta 46 Maribor, 2000 Slovenia Email: [email protected] https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Abstract Background: Individuals with dementia, especially those in later stages, have difficulties with verbally reporting their experience of pain. This results in both underassessment and undertreatment of pain, signaling the need for better pain recognition in persons with dementia. A promising form of pain assessment is digital monitoring, which can concurrently and more objectively detect and utilize numerous relevant pain cues. Objective: The present review aimed to identify observable cues of pain, which could be used for digital pain monitoring. We set out to examine which digital cues offered a valid insight into pain in people with dementia (RQ1) and identify how these cues were originally measured (RQ2). Methods: A standard methodological approach for scoping reviews was utilized. Relevant research papers were chosen based on SCOPUS and Web of Science databases, and relevant data on pain cues were extracted from all papers that satisfied the inclusion criteria. The gathered data were analyzed using a thematic analysis, which involved categorizing the observable cues into higher-order categories. Results: Of the 3705 publications identified in the search, 34 satisfied the inclusion criteria and were closely examined. Addressing RQ1, we identified seven categories of behavioral and physiological cues associated with pain, most frequently facial expressions (20 of 24, 59%) and body movements/expressions (15 of 34, 44%). Several sub-categories for each main category of pain cues were also identified, each involving between one and 28 relevant specific pain cues. Addressing RQ2, 29 of 24 studies (85%) of studies assessed pain cues via human observation, while 5 of 34 (15%) applied digital methods such as facial recognition software or combined observation with computer vision. Conclusions: The review provides a comprehensive list of the most relevant cues that signify pain in persons with dementia and offers a foundation for the use of artificial intelligence and digital monitoring for the screening of pain in dementia. Keywords: artificial intelligence, digital monitoring, pain detection, pain, dementia, symptoms, review, screening https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Introduction Dementia presents a global health concern, with some estimates suggesting that it affects over 50 million people worldwide, and this figure is predicted to triple by 2050 [1,2]. The course of dementia can span from five to 12 years, during which the patient's ability to function independently gradually decreases with the progression of the disease [3]. With this progression, there is also an increase in communication problems [4]. In a clinical setting, this can lead to the underassessment of many cooccurring health conditions or states, especially for patients with moderate to severe dementia [5,6]. One of these under-recognized states is pain [7]. Pain is frequently experienced by patients with moderate to severe dementia but is often undetected and consequently untreated because of their inability to self-report [8]. Since reliable recognition and assessment of pain are essential for effective treatment, observational pain tools that emphasize observable cues have become increasingly important for the accurate assessment of pain in people with dementia [9]. While self-reporting is generally considered the most reliable method for assessing pain [9], a different approach is needed when assessing patients with dementia. There are numerous observational scales available, such as The Abbey Pain Scale [10], Doloplus-2 [11], The Pain Assessment Checklist for Seniors with Severe Dementia (PACSLAC) [12], and The Pain Assessment in Advanced Dementia Scale (PAINAD) [13]. The problem with this type of assessment, however, is that the scales have poor or unproven reliability, insufficient evidence for validity, and untested sensitivity to change. Additionally, their implementation in practice is often poor [14,15]. It is usually nurses or caregivers who assess and report on the patient’s pain, and they do not always rely on observational scales. Even when they do, their ratings of pain are not always related to any specific pain behaviors, which results in inadequate pain assessment [16–18]. As an alternative, digital monitoring of pain could be very effective in assessing pain in patients with dementia, as it promises to provide objective evidence of the presence and intensity of pain [5,8,19]. The method could use a combination of different technologies, such as automated facial recognition and analysis, smart computing, affective computing, and cloud computing (i.e., Internet of Things) for identifying the presence of pain in patients with advanced dementia [20]. As an example, Atee and colleagues [5] recently developed an electronic Pain Assessment Tool (ePAT), an application that uses a facial recognition technology to detect facial micro-expressions indicating pain and to record pain-related behaviors. A similar system is also used in healthcare – Internet of Things-enabled surveillance cameras capture real-time video data and can enhance patient care with features like sentiment analysis and emotion detection [21]. Digital monitoring has the potential to change pain assessment in individuals who are unable to verbalize their inner states, as it can be used by clinicians and caregivers in everyday clinical practice [5,8,20]. A critical foundation for effective digital monitoring is a thorough overview and categorization of cues associated with pain that can be measured using digital technologies (i.e., digital cues). These are essential for identifying pain in individuals unable to directly report it themselves. However, despite previous efforts in identifying and categorizing pain cues (e.g. by the American Geriatrics Society (AGS[22]), which identified broad categories of facial expressions, verbalizations and vocalizations, body movements, changes in interpersonal interactions, changes in activity patterns or routines, and mental status changes), the categorizations of pain cues remain inconsistent and segmented. If we want to develop better, technology-supported ways of monitoring pain, a systematic and thorough set of digital pain cues is needed. Therefore, the main goal of this paper was to identify specific cues that could help identify pain in patients with dementia by way of digital monitoring. Despite many studies that identify common pain behaviors (e.g. [22,23]), there is a lack of specific information about how certain cues are measured, or how identified behaviors are linked with pain. The current scoping review aimed to https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al address this gap by identifying pain cues that could potentially be intercepted using artificial intelligence (AI). Among other potential applications, results will be directly utilized to develop an AI-based pain identification system within the project Artificial intelligence-based health, optimism, purpose, and endurance in palliative care for dementia (AI4HOPE [24]). Two research questions (RQs) were formed for the purpose of this study. The first aimed to investigate which digital cues offer a valid insight into pain in patients with dementia, and the second focused on determining how these cues were measured. To address the RQs, we conducted a scoping review. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al The median sample size in the reviewed studies was 116 (SD = 8929,5). The target sample (excluding control groups) across the studies included between 12 and 45,568 participants (Med. = 59, M = 1537,25, SD = 7581,9), predominantly older adults over 60 years old. Most participants (N = 36525, 66%) resided in long-term care facilities or nursing homes. The majority of studies (N = 28, 82,4%) focused on individuals with moderate to severe dementia, including those unable to express pain. About half of the studies (N = 18, 52,9%) involved participants with a history of pain-related conditions. Additionally, a subset of studies (N = 5, 14.7%) involved caregivers, healthcare providers, or family members as study participants. About a third (N = 11, 34.4%) of the included studies utilized comparison groups in their research design. Most of those (N = 6, 54.5%) included control groups composed of cognitively intact individuals, with or without pain symptoms. Among the studies that mentioned the type of dementia (N = 28, 82.4%), unspecified dementia (such as “dementia”, “advanced dementia”, “moderate dementia”) was the most frequently mentioned (N = 20, 58.8%). A total of 211 distinct pain-related cues (485 overall instances) were extracted and categorized into seven broad categories, each consisting of one or more sub-categories. The main categories and their corresponding number of cues were: Behaviors (N = 49, 23.2%), Body movement/expression (N = 55, 26.1%), Facial expressions (N = 47, 22.3%), Medical status – somatic (N = 5, 2.4%), Mental state (N = 14, 6.6.%), Physiology (N = 10, 4.7%), and Speech, language and sounds (N = 31, 14.7%). The following sections describe each category in detail. Behaviors We identified five subcategories within the Behaviors category: Active behavior, Behavioral change, Inappropriate behavior, Mood related behavior and Social behavior (see Table 2 for a summary, and Table S1 for detailed information). Cues without any statistical information (i.e. general behavior change) are listed in Table S8. Table 2. Summary of relevant cues of pain in the Behaviors category. Pain cue Relevance of the cue Categor y 1 Category 2 Category 3 Direction of the association to pain Relevance of the pain cue Study Behavior Active behavior [General] + Strong [19] Falls + Weak/medium [39] Impulsive behavior + Strong [12] Normal behavior - Strong [37] Wandering + Strong [5,12,54] Washing &/or dressing + Weak/medium [52] Washing &/or dressing + Not specified [58] Behavioral change [General] + Not specified [51] + Strong [57] Changes in appetite + Strong [12,54] + Weak/medium [39] Changes in communication + Weak/medium [52] + Not specified [58] Changes in routines + Strong [5] Changes in sleep + Strong [5,12] + Weak/medium [52] + Not specified [58] Decrease in activity + Strong [6,12] + Weak/medium [39,52] + Not specified [58] Lethargy + Weak/medium [39] https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al + Strong [5] Stopping an activity + Weak/medium [41] Inappropriate behavior [General] + Strong [5,36] + Weak/medium [52] + Not specified [58] Combativeness + Weak/medium [39] Handling things inappropriate + Weak/medium [42] Refusing medications + Strong [12] Resisting care + Strong [5,12,48] Throwing things + Strong [12] Trying to leave/get to a different place + Strong [12] Mood-related behavior Aggressive behavior + Strong [5,12,54] Whiny + Weak/medium [37] Social behavior Argumentativeness + Strong [7] Consolability - Strong [13] Disruptive behavior + Strong [54] Interpersonal changes + Strong [6] Not allowing people near + Strong [12,38] Not wanting to be touched + Strong [12,38,54 ] Requesting attention + Weak/medium [42] Social life + Weak/medium [52] + Not specified [58] Striking out + Strong [54] Unsocial behavior + Strong [5] Withdrawn + Strong [54] Notes. Direction of the association to pain: + = positive, - = negative. Relevance of the pain cue was determined as follows: Strong = large effect size, Weak/medium = small to medium effect size [61], Not specified = study reported only on the significance and direction of the association, without statistical coefficients. In the subcategory of Active behavior, general active behavior, impulsive behavior, normal behavior and wandering were associated with pain. Falls and washing and/or dressing had mixed associations, while scratching was not significantly related to pain. Active behavior, measured using ePAT, was positively associated with APS [19]. Falls were rated as an important indicator of pain but did not show a direct relationship with pain [39,42]. Impulsive behavior, measured by PACSLAC, was associated with global pain intensity ratings [12]. Wandering, measured using Pain Assessment Tool in Cognitively Impaired Elders (PATCIE) or PACSLAC, was linked to the Checklist of Nonverbal Pain Indicators (CNPI), APS, and global pain intensity ratings [5,12,54]. Washing/dressing behavior, measured using Doloplus-2, was associated with the Visual Analogue Scale (VAS) rating [52,58]. In the Behavioral change subcategory, change in communication, change in sleep, and lethargy were associated with pain, while behavioral change in general, changes in appetite or routine, decreased activity, and stopping an activity showed mixed associations. In the general behavioral change category, APS distinguished pain intensity across MMSI categories [57] and a significant difference in behavioral pain observation was found during aversive vs. pleasant activities [51]. Decreased appetite was rated as important [39], and appetite changes in general, measured using PACSLAC or PATCIE, were associated with CNPI and global pain intensity ratings [12,54]. Communication changes, measured using Doloplus-2, were associated with VAS [52,58], while routine changes and sleep changes, measured using ePAT, were associated with APS [5]. Sleep changes, measured using Doloplus-2 or PACSLAC scores, were also associated with VAS or global pain ratings during events [12,52,58]. Decreased activity was rated somewhat important by nursing staff [39] and was associated with global pain intensity, as well as with Assessment of Discomfort in Dementia (ADD), CNPI, Non-communicative Patient’s Pain Assessment Instrument (NOPPAIN), PACSLAC, Pain Assessment for the Dementing Elderly (PADE) and PAINAD, when measured using PACSLAC-II or https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al PACSLAC [12,38,52,58]. Activity changes showed large effect sizes for both influenza vaccination and movement-exacerbated pain [6]. Lethargy was considered important by nursing staff [39] and prolonged resting, measured using ePAT, was associated with APS [5]. Pain intensity was predictive of stopping an activity pre-activity, but not during an activity [41]. In the Inappropriate behavior subcategory, cues such as inappropriate behavior in general, combativeness, handling inappropriate things, refusing medication and resisting care were associated with pain, while throwing things and trying to leave showed mixed associations. There was no association between pain and biting, eating inappropriate substances, grabbing, hiding, hitting, hoarding, hurting self or others, kicking, pushing, disrobing, spitting or tearing things. Inappropriate behavior in general, measured using ePAT or Doloplus-2, showed a positive association with APS and VAS and predicted pain in a regression model [5,36,52,58]. Combativeness was rated as important by nursing staff [39]. Handling inappropriate things, refusing medication, resisting care, throwing things and trying to leave were all related to pain [5,12,43,48]. In the Mood-related behaviors subcategory, aggressive and whiny behavior were associated with pain, while pleasant behavior showed no association. Aggressive behavior, measured using ePAT, PATCIE and PACSLAC, was associated with APS and CNPI [5,12,54]. Whiny behavior indicated the presence of pain [37]. In the Social behavior subcategory, argumentativeness, disruptive behavior, interpersonal changes, not allowing people near, not wanting to be touched, requesting attention, social life, striking out, unsocial behavior and being withdrawn were associated with pain. Consolability showed mixed associations, while sexual advances showed no association with pain. Argumentativeness was associated with reported pain frequency [7]. Consolability, disruptive behavior, not allowing people near, avoiding being touched, social life, striking out, unsocial behavior, and being withdrawn were associated with pain measures, such as CNPI, VAS and APS, when measured using PAINAD, ePAT, PATCIE, PACSLAC-II, PACSLAC or Doloplus-2 [5,12,13,38,52,54,58]. Interpersonal changes, measured using PACSLAC, showed higher effect sizes for both influenza vaccination and movement-exacerbated pain [6], while a decrease in constant need for attention was linked with decreased pain [42]. Body movements/expressions We identified five subcategories within the Body movements/expressions category: Body language, Body movement, Body parts and cues, Body positions/postures, and Physical cues (see Table 3 for a summary, and Table S2 for more detailed information). Cues without any statistical information (i.e. general body language, decreased movement, reluctance to move, rubbing, tense body and guarding) are listed in Table S8. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Table 3. Summary of relevant cues of pain in the Body movement/expressions category. Pain cue Relevance of the cue Category 1 Category 2 Category 3 Direction of the association to pain Relevance of the pain cue Study Body movement/ expressions Body language [General] + Weak/medium [60] + Strong [13] + Not specified [51] Change in body language + Weak/medium [57,60] Body movement [General] + Strong [6,19] Bracing + Weak/medium [39–41] + Strong [41] Difficulty chewing + Strong [39] Fidgeting + Strong [12] Flinching and/or pulling away + Strong [12,38] Freezing + Weak/medium [48] + Strong [5] Gait changes + Strong [54] Handwringing + Weak/medium [37] Leg/arm movement + Strong [5] Limping + Strong [12,38,3 9] + Weak/medium [39] Pacing + Weak/medium [42] + Strong [12] Pulling/moving away + Strong [5,12] Reluctance to move + Strong [12,38,5 4] + Weak/medium [39] Repetitive movements + Weak/medium [39] + Strong [36,39] Restlessness + Strong [5,12,42, 48,54,59 ] + Weak/medium [7,39,59 ] Rigidity + Weak/medium [37,39] + Strong [12,38,4 1] Rocking motion + Strong [54] Rocking motion - head + Weak/medium [39] Rubbing + Mixed results [40,56] + Strong [38,41,4 8,54] + Weak/medium [41,59] Shaking/Trembling + Strong [12,38] Shifting + Strong [41] + Weak/medium [41] Slow movement + Strong [12,38] Tense body + Weak/medium [37] + Strong [54] Thrashing + Strong [12,38] Tossing and/or turning + Weak/medium [39] + Mixed results [43] https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Touching a body part/area + Strong [12,39,5 4] Trembling + Weak/medium [39] Wincing + Weak/medium [39] + Strong [39] Body parts and cues [General] + Weak/medium [36] + Strong [19] Abdomen + Mixed results [43,55] Arms + Mixed results [43] + Strong [55] Hands + Mixed results [43] + Weak/medium [55] Head, mouth, neck + Mixed results [43] + Weak/medium [55] Heart, lung, chest wall + Weak/medium [55] + Mixed results [43] Legs + Mixed results [43,55] Pelvis, genital organs + Mixed results [43] + Strong [55] Skin + Mixed results [43,55] Turn over + Weak/medium [55] Body positions/posture s Abnormal/awkward sitting/standing/walking + Strong [5,39] + Weak/medium [39] Clenched fist + Strong [12,38] Fetal position + Strong [12,38] Guarding + Weak/medium [41,52,5 6] + Strong [5,12,38, 41,48,54 ] + Not specified [58] Poor posture + Weak/medium [39] Protective posture at rest + Weak/medium [52] + Not specified [58] Sitting + Mixed results [43] + Strong [55] Physical cues Abnormal skin color + Strong [39] Blood stains + Weak/medium [39] Heat from specific body part + Strong [39] Swollen joints + Strong [39] Tight belly + Strong [39] Notes. Direction of the association to pain: + = positive, - = negative. Relevance of the pain cue was determined as follows: Strong = large effect size, Weak/medium = small to medium effect size [61], Mixed results = study reported inconsistent findings, Not specified = study reported only on the significance and direction of the association, without statistical coefficients. In the Body language subcategory, general body language had mixed associations, while changes in body language showed a moderate association with pain. General body language was associated with pain via behavioral pain observations, high inter-rater agreement and positive association with VAS, when measured using PAINAD [13,51,60]. Changes in body language (e.g. fidgeting, rocking, etc.) also demonstrated high inter-rater agreement, as well as a connection with APS in the group exhibiting pain [57,60]. In the Body movement subcategory, general body movement, and cues such as difficulty chewing, fidgeting, flinching and/or pulling away, freezing, gait changes, handwringing, leg or arm movement, limping, pulling or moving away, reluctance to move, restlessness, rocking motion, shaking or trembling, shifting, slow movement, trashing, touching a body part, and wincing were associated https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al with pain. Bracing, pacing, repetitive movements, rigidity, rubbing, tense body, tossing, turning, and trembling had mixed associations, while ease of movement was not associated with pain. General movement, measured using ePAT, CNPI, NOPPAIN, PAINAD, PADE or PACSLAC, was associated with APS or pain across conditions [6,19]. Bracing, measured using CNPI, was associated with the Verbal Descriptor Scale (VDS), linked to overall pain intensity preand during activity, and rated as important by nursing staff [39–41]. Difficulty chewing was rated as important by nursing staff [39], while fidgeting and flinching/pulling away, measured using PACSLAC and PACSLAC-II, showed associations with global pain ratings, CNPI, PADE, PAINAD and NOPPAIN [12,38]. Freezing, measured using ePAT, showed an association with APS, and was supported by inter-rater reliability [5,48]. Gait changes, assessed with PATCIE, were associated with CNPI [54]. Handwringing was associated with the presence of pain [37] and leg or arm movement, measured using ePAT, was associated with APS [5]. Limping was rated as important by nursing staff and associated with global pain ratings, NOPPAIN, CNPI, PADE and PAINAD, when measured using PACSLAC or PACSLAC-II [12,38,39]. Pacing was associated with global pain ratings, when measured using PACSLAC, and was reduced following intervention [12,42]. Pulling or moving away, measured using ePAT or PACSLAC, was associated with global pain ratings and APS [5,12]. Reluctance to move, measured using PACSLAC, PACSLAC-II or PATCIE, was associated with global pain intensity, CNPI, NOPPAIN, PADE and PAINAD, and was rated as important by nursing staff [12,38,39,54]. Repetitive movements, measured using the Minimum Data Set Resident Assessment Instrument 2.0 (MDS-RAI 2.0), were linked to pain, and rated as important by nursing staff [36,39]. Restlessness, measured using the caregiver report, the Cohen-Mansfield Agitation Inventory (CMAI), ePAT, PACSLAC, PATCIE, and the Observational Pain Scale - Non-Verbal Indicators (OPS-NVI), was associated with APS, CNPI, NIH-REACH, global pain ratings and self-reported pain. It was also present in situations where pain was more likely and rated as important by nursing staff [5,7,12,39,43,48,54,59]. Rigidity and muscle tensing were associated with global pain intensity, NOPPAIN, CNPI, PADE, and PAINAD, when measured using PACSLAC, PACSLAC-II, or caregiver observation [12,37–39,41]. Rocking motions, assessed using PATCIE, were associated with CNPI, and rated as important by nursing staff [39,54]. Rubbing, observed via caregiver observation, CNPI, OPS-NVI or PATCIE, was associated with VDS, CNPI or self-reported pain, and was supported by inter-rater agreement during painful situations [40,41,48,54,59]. Shaking or trembling, trashing and slow movement, measured using PACSLAC or PACLAC-II, were associated with global pain ratings during pain events, as well as with NOPPAIN, CNPI, PADE and PAINAD [12,38]. Shifting predicted pre-activity pain [41]. Tense body was rated as important by nursing staff and was associated with CNPI when measured using PATCIE [37,39,54]. Tossing and/or turning was also rated as important and was associated with CMAI and Neuropsychiatric Inventory – Nursing Home version (NPI-NH), when measured using Mobilization-Observation-Behavior-Intensity-Dementia-2 (MOBID-2) [39,43]. Touching a body part, measured using PACSLAC or PATCIE, was associated with global pain intensity and CNPI. It was frequently reported by nursing staff [12,39,54]. Trembling and wincing were rated as important by nursing staff [39]. The Body parts and cues subcategory in general, measured using ePAT or MOBID-2, showed associations with APS, CMAI and NPI-NH, and predicted pain in a regression model [5,36,43]. Pain in the abdomen, arms, hands, head, mouth, neck, heart, lung, chest wall, legs, pelvis, genital organs or skin was associated with CMAI and NPI-NH, when measured using MOBID-2. Total MOBID-2 scores improved after a pain intervention [43,55]. In the Body positions and postures subcategory, abnormal or awkward sitting, standing or walking, clenched fists, fetal position, guarding, poor posture, and protective posture at rest, were associated with pain, while sitting showed mixed associations. Abnormal movements or posture, measured using ePAT, were associated with APS and rated as important by nursing staff [5,39]. Clenching https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al fists, fetal position and guarding, measured using PACSLAC or PACSLAC-II, all had a positive association with global pain ratings, NOPPAIN, CNPI, PADE and PAINAD [12,38]. Guarding, assessed via caregiver observation, ePAT, PATCIE or Doloplus-2, was associated with APS, VAS, self-reported pain or CNPI. It was supported by inter-rater reliability for predicting pain [5,41,48,52,54,56,58]. Poor posture was rated as important by nursing staff [39] while sitting, measured using MOBID-2, was associated with CMAI and NPI-NH, [43]. Lastly, Cohen-Mansfield & Creedon [39] reported that Physical cues ratings by nursing staff, abnormal skin, blood stains, heat from a specific body part, swollen joints and tight belly, were all rated as important indicators of pain by nursing staff. Facial expressions We categorized Facial expressions into eight distinct subcategories: Brows, Cheeks, Eyes, Forehead, Jaw, Lips and mouth, Nose and Whole face (see Table 4 for a summary, and Table S3 for more detailed information). Cues without statistical information (i.e. brow lowering, cheek raising, blinking, closing eyes, specific eye movement (up, down, left or right), tightening of eyelids, jaw drop, parting lips, specific facial expression, grimacing and sudden jerk) are listed in Table S8. Table 4. Summary of relevant cues of pain in the Facial expressions category. Pain cue Relevance of the cue Category 1 Categor y 2 Category 3 Direction of the association to pain Relevance of the pain cue Study Facial expressions Brows Brow lowering + Strong [5,8,9] Frowning + Weak/medium [48] + Strong [12,38,54,59] Cheeks Cheek raising + Strong [5] Eyes Changes in eyes + Strong [12] Closing eyes + Strong [5,8,9,38] Dirty look + Strong [12] Increased eye movement + Strong [38] Narrowing and/or closing eyes + Weak/medium [48,59] + Strong [59] Teary eyes + Strong [12] Tightening of eyelids + Strong [5,8] Forehead Creasing forehead + Strong [12,38] Jaw Restricting jaw movement while chewing + Strong [59] Lips, mouth Clenching teeth + Strong [12,54] Drooling + Strong [59] Horizontal mouth stretch + Strong [5,8] + Weak/medium [9] Opening mouth + Strong [12,38] + Weak/medium [48,59] Parting lips + Weak/medium [8] + Strong [5,9] Pulling at corner lip + Strong [5] + Weak/medium [8] Raising of upper lip + Strong [5,8,9,48] Nose Screwing up + Strong [12] https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al nose Wrinkling nose + Weak/medium [8] + Strong [5,9] Whole face [General] + Strong [8,19] Change in color + Strong [39] Facial expression + Not specified [51,58] + Strong [6,13,57,60] Facial expression - specific + Strong [38,60] Fearful expression + Weak/medium [37] Flushed, red face + Strong [12] Gloomy facial expression + Strong [54] Grim face + Strong [12] Grimacing + Weak/medium [39] + Mixed results [40,56] + Strong [12,38,41,54] Looking tense + Weak/medium [48] Pain expression + Weak/medium [52] + Strong [12,38] Pale face + Strong [12] Pale or flushed/red face + Strong [5] Prkachin & Solomon Pain Index + Weak/medium [53] Sad expression/look + Strong [36] + Mixed results [12] Scared expression + Weak/medium [37] Tighter face + Strong [12,38] Wincing + Strong [12,38] Notes. Direction of the association to pain: + = positive, - = negative. Relevance of the pain cue was determined as follows: Strong = large effect size, Weak/medium = small to medium effect size [61], Mixed results = study reported inconsistent findings, Not specified = study reported only on the significance and direction of the association, without statistical coefficients. The Brows subcategory included brow lowering and frowning, with frowning strongly associated with pain, and brow lowering showing mixed associations. Brow lowering, measured using ePAT, was associated with APS and linked to higher pain scores in the PainChek App [5,8,9]. Frowning, measured with PACSLAC, PACSLAC-II, PATCIE or OPS-NVI, was associated with global pain intensity, NOPPAIN, CNPI, PADE, PAINAD, and self-reported pain [12,38,48,54,59]. Cheekraising, measured with ePAT, was associated with APS [5]. In the Eyes subcategory, eye changes, closing eyes, dirty look, increased eye movement, narrowing and/or closing of eyes, and teary eyes were associated with pain, while specific eye movements (e.g., looking to the left), blinking, and tightening of eyelids had mixed associations. Eye changes, measured using PACSLAC, were associated with global pain intensity [12]. Closing eyes, measured using ePAT or PainChek app, was associated with APS and observed pain, and predicted pain when it was present [5,8,9]. Eye closure and increased eye movement, measured using PACSLAC-II, were associated with NOPPAIN, CNPI, PADE and PAINAD [38]. Dirty look and teary eyes, measured using PACSLAC, were associated with global pain intensity ratings [12]. Narrowing eyes, measured using OPS-NVI and supported by inter-rater agreement, was associated with observational or selfreported pain [48,59]. Tightening of eyelids, measured with ePAT, was associated with APS and https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al predicted higher pain [5,8]. In the Forehead subcategory, creasing of the forehead, measured using PACSLAC or PACSLAC-II, was associated with global pain intensity, NOPPAIN, CNPI, PADE, or PAINAD [12,38]. In the Jaws subcategory, restricting jaw movement was associated with self-reported pain when measured using with OPS-NVI [59]. In the Lips and mouth subcategory, clenching teeth, drooling, horizontal mouth stretch, opening mouth, pulling at corner of the lip and parting lips were associated with pain, while raising of upper lip had mixed associations. Clenching teeth, measured using PACSLAC or PATCIE score, was associated with global pain intensity and CNPI [12,54]. Drooling, measured using OPS-NVI, was associated with self-reported pain [59]. Horizontal mouth stretching and parting of lips, measured using ePAT or the PainChek app, were associated with APS or observational pain scores and were more likely to predict pain when present [5,8,9]. Opening mouth, measured using inter-rater agreement, PACSLAC, PACSLAC-II or OPS-NVI, was associated with self-reported pain, global pain intensity, NOPPAIN, CNPI, PADE and PAINAD [12,38,48,59]. Pulling at corner of lips and raising of upper lip, measured using ePAT, were associated with APS and more likely to predict higher pain when present [5,8]. Raising upper lip, measured using the PainChek app, OPS-NVI or inter-rater agreement also had an association with observational pain scores and self-reported pain [9,48,59]. In the Nose subcategory, screwing the nose and wrinkling the nose were linked with pain. Screwing the nose, measured using PACSLAC, was associated with global pain ratings [12], while wrinkling the nose, measured using ePAT or the PainChek app, was associated with APS or observational pain scores and was more likely to predict higher pain when present [5,8,9]. In the Whole face subcategory, the general face domain, facial expression, fearful expression, flushed face, gloomy expression, grim face, looking tense, pain expression, pale face, Prkachin and Solomon Pain Intensity, scared expression, tight face and wincing were associated with pain, while change in color, grimacing, sad expression and sudden jerks had mixed associations. Relaxed expression had no association with pain. The face domain, measured using ePAT, was associated with APS. Wholeface domain scores were significantly associated with pain, with upper face AUs (e.g., brows, eyelids, eyes) noted more frequently than lower face AUs (e.g., nose, lips) during moderate and severe pain [8,19]. Change in face color was rated as important by nursing staff [39]. Facial expressions measured using APS, Checklist of Agitation Symptoms (CAS), ADD, CNPI, Mahoney Pain Scale (MPS), PACSLAC, PADE, PAINAD and NOPPAIN, differentiated pain conditions and levels of dementia. When measured using Doloplus-2, PAINAD, or PACSLAC-II, facial expressions were associated with caregiver-reported pain, VAS, NOPPAIN, CNPI, PADE, and PAINAD. Facial expressions were also rated as reliable indicators of pain by inter-rater agreement [6,13,38,51,57,58,60]. Specifically, brow lowering, lid tightening, cheek raising, and jaw clenching, looking tense, frowning, grimacing, or appearing frightened were identified as pain indicators based on caregiver report or inter-rater agreement [60]. Fearful expression had a significant association with the presence of pain [37]. Flushed, red face and grim face, measured using PACSLAC were associated with global pain ratings [12], while gloomy facial expression, measured using PATCIE, was associated with CNPI [54]. Grimacing, measured using PACSLAC, PACSLAC-II, PATCIE CNPI, or observation, was associated with global pain rating, self-rated pain, VDS, NOPPAIN, CNPI, PADE or PAINAD. It was rated as important by nursing staff and was more frequent in patients with chronic low back pain than in pain-free participants [12,38–41,54,56]. Looking tense was present in situations where pain was more likely, based on inter-rater agreement or intraclass correlation coefficient [48]. Pain expression, measured using PACSLAC, PACSLAC-II or Doloplus2, was associated with global pain intensity, NOPPAIN, CNPI, PADE and PAINAD [12,38,52]. Pale face, measured using PACSLAC, was associated with global pain intensity [12], while pale and/or flushed (red) face, measured using ePAT, was associated with APS [5]. Prkachin and Solomon pain estimation model outperformed a baseline model in pain estimation [53]. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Scared expression, measured using Computerized Pain Assessment Tool (CPAT), was associated with the caregiver’s report of pain [37], while sad expression predicted pain in a regression model, and was associated with pain intensity when measured using PACSLAC [12,36]. Tighter face and wincing, measured using PACSLAC or PACSLAC-II, were associated with NOPPAIN, CNPI, PADE, PAINAD, and global pain intensity [12,38]. Medical status – somatic The Medical status – somatic category included two subcategories, Injuries and Medical conditions (see Table 5 for a summary, and Table S4 for more details). Table 5. Summary of relevant cues of pain in the Medical status - somatic category. Pain cue Relevance of the cue Category 1 Category 2 Category 3 Direction of the association to pain Relevance of the pain cue Stud y Medical status - somatic Injuries Dislocated limbs + Strong [39] Injuries + Strong [5] Physical changes + Weak/medium [60] Medical conditions One leg shorter + Weak/medium [39] Painful medical conditions + Strong [5] + Mixed results [51] Notes. Direction of the association to pain: + = positive, - = negative. Relevance of the pain cue was determined as follows: Strong = large effect size, Weak/medium = small to medium effect size [61], Mixed results = study reported inconsistent findings. In this category, all cues were associated with pain, except for painful medical conditions, which had mixed associations. Dislocated limbs and having one leg shorter were rated as important by nursing staff [39]. Injuries in general, and painful medical conditions, when measured using ePAT, were associated with APS [5], but the latter showed no connection to pain, when measured using MPS [51]. Physical changes such as skin tears, pressure areas, and arthritis, were reported as good indicators of pain based on high inter-rater agreement [60]. Mental state / mood The Mental state / mood category had two subcategories: Mental states and Mood indicators (see Table 6 for a summary, and Table S5 for more details). Cues that were reported without any statistical information (i.e. depression) can be found in Supplementary Table S8. Table 6. Summary of relevant cues of pain in the Mental state / mood category. Pain cue Relevance of the cue Category 1 Category 2 Category 3 Direction of the association to pain Relevance of the pain cue Study Mental state / mood Mental state Changes in mental status + Strong [6,38, 39] Confusion + Strong [5,12, 54] Delusions + Weak/medium [7] Distressed + Strong [5] Mood indicators Agitation + Weak/medium [7,39] + Strong [12] Anger + Strong [12] Anxiousness, nervousness + Weak/medium [7,12] + Strong Depression + Weak/medium [39] Fear + Strong [5] Frustrated + Strong [12] Irritable + Strong [12,39 https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al To the best of our knowledge, this review serves as a first overview of cues for the purpose of digitally monitoring cues related to pain. To utilize these cues in pain assessment, it is necessary to explore ways to design efficient algorithms and to investigate which cues are better or worse predictors of pain. The findings of this paper show that potential cues of pain exist in many different categories (e.g. behavior, body movements, facial expressions, language). This is particularly important for the development of algorithms for the purpose of identifying or classifying the identified cues of pain, as it has been found that algorithms which consider multiple modalities tend to be more successful than those focusing on a single modality in identifying other conditions, for example depression [68,69]. The identified pain cues could also be used to develop digital solutions that are on the rise for supporting people with dementia [70], e.g., based on AI algorithms, to help identify pain in individuals with dementia. AI-based systems could offer a more efficient and costeffective method for clinicians to assess pain. These solutions would also allow for continuous, realtime pain monitoring, which is not feasible with traditional observational or self-reported measures. For individuals with dementia, AI would provide a discreet and accessible way to monitor pain. It would benefit clinicians by not requiring much conscious effort and allowing for continuous assessments, as well as help patients by reducing the reliance on self-reporting, which can be especially challenging for those unable to communicate their pain. Additionally, such solutions could be incorporated into broader platforms collecting various disease-related cues over longer periods of time, benefiting researchers and clinicians in gaining insight into disease progression, in order to improve prevention strategies, interventions and personalized care for patients with dementia (e.g., [71]). Limitations The current scoping review provides a valuable synthesis of research on the observable cues of pain in individuals with dementia and their measurement. However, certain limitations must be acknowledged. Because we focused exclusively on English-language publications, predominantly involving Anglophone participants from industrialized nations, our findings may not be fully generalizable across different cultural contexts. Another limitation lies in the fact that the largest proportion of the included papers consisted of evaluation and validation studies, which may have led to a narrow focus, as these studies concentrate on the effectiveness or reliability of specific interventions or tools, but do not address all relevant aspects of the phenomenon in question. This limitation of the study distribution may have resulted in a less comprehensive understanding of the topic, particularly in terms of causality and breath of applications. However, it is important to note that there is a general lack of studies that evaluate observable cues indicative of pain, which is why such a large proportion of evaluation and validation studies was used in this review. Lastly, as a scoping review using established methodology [25] was conducted to provide a comprehensive overview of the research related to our RQs, the quality of included articles was not assessed. Future research could additionally perform a risk of bias assessment. Conclusions The current review examined the relevant literature for the purpose of identifying observable cues that could be used to detect pain in persons with dementia with the use of digital monitoring. We focused on cues that can be measured without the use of specialized equipment unavailable to the general public. The review resulted in a comprehensive set of observable cues that could be used to help identify pain in persons with dementia. We also identified inconsistencies regarding the relevance of some identified cues, as well as a lack of studies that evaluate observable cues indicative of pain, beyond the scope of evaluation and validation studies of specific pain https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al questionnaires. The conducted review could help future advancements aiming to objectively and efficiently identify and manage pain in persons with dementia, potentially via the use of AI and digital monitoring. Our findings could also inform the design of accessible devices (e.g. smartwatches and smartphones), to incorporate features that monitor and assess pain. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Acknowledgments The authors would like to thank Rigon Sallauka for his help in screening the studies. Conflict of interest The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Abbreviations ABID = Agitation Behavior Identification AD = Alzheimer's Dementia ADD = Agitation and Distress Detection AGS = American Geriatrics Society AI = Artificial Intelligence APS = Abbey Pain Scale APS-J = Abbey Pain Scale - Japanese version CAS = Comfort Assessment Scale CI = Cognitive Impairment CMAI = Cohen-Mansfield Agitation Inventory CNPI = Checklist of Nonverbal Pain Indicators CPAT = Checklist of Nonverbal Pain Indicators EEG = Electroencephalogram ePAT = Electronic Pain Assessment Tool FACS = Facial Action Coding System FPS-R = Faces Pain Scale - Revised FTD = Frontotemporal Dementia LBD = Lewy Body Dementia MDS-RAI = Minimum Data Set - Resident Assessment Instrument MOBID-2 = Mobilization-Observation-Behavior-Intensity-Dementia-2 MRI = Magnetic Resonance Imaging N = Numerus NOPPAIN = Non-Communicative Patient's Pain Assessment Instrument NRS = Numeric Rating Scale OPS-NVI = Observational Pain Scale - Non-Verbal Individuals PACI = Pain Assessment Checklist for Seniors with Limited Ability to Communicate PADE = Pain Assessment in Dementing Elderly PAINAD = Pain Assessment in Advanced Dementia PASLAC = Pain Assessment for Seniors with Limited Ability to Communicate PATCIE = Pain Assessment Tool for Cognitively Impaired Elders PD = Parkinson's Disease Dementia PRISMA-ScR = Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews RQ = Research Question SD = Standard Deviation VAS = Visual Analog Scale https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al VD = Vascular Dementia VDS = Verbal Descriptor Scale Funding This research was partially funded by the project ‘AI4HOPE: Artificial intelligence based health, optimism, purpose, and endurance in palliative care for dementia’ that has that has received funding from the European Union's Horizon Europe Research and Innovation Program (GA No. 101136769), and the Slovenian Research and Innovation Agency, ‘Advanced methods of interaction in telecommunication research programme’ (grant number P2-0069). The content of this works does not reflect the official position of the European Union or any other institution. The information and views expressed are the sole responsibility of the authors. Author Contributions Urška Smrke: conceptualization, data curation, formal analysis, investigation, methodology, supervision, validation, writing – original draft preparation, writing – review & editing. Ana Milošič: data curation, investigation, writing – original draft preparation, writing – review & editing. Izidor Mlakar: conceptualization, data curation, funding acquisition, methodology, supervision, writing – review & editing. Matic Kadiš: data curation, investigation, writing – original draft preparation, writing – review & editing. Satja Mulej Bratec: conceptualization, data curation, formal analysis, investigation, methodology, supervision, validation, writing – original draft preparation, writing – review & editing. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Not applicable. https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al References 1. Alzheimer’s Disease International. World Alzheimer Report 2019: Attitudes to dementia. London; 2019. Available from: https://www.alzint.org/u/WorldAlzheimerReport2019.pdf [accessed Apr 8, 2025] 2. Stephan BCM, Birdi R, Tang EYH, Cosco TD, Donini LM, Licher S, Ikram MA, Siervo M, Robinson L. Secular trends in dementia prevalence and incidence worldwide: A systematic review. Journal of Alzheimer’s Disease 2018 Oct 30;66(2):653–680. doi: 10.3233/JAD-180375 3. Li F, Qin W, Zhu M, Jia J. Model-based projection of dementia Prevalence in China and worldwide: 2020–2050. Journal of Alzheimer’s Disease 2021 Aug 17;82(4):1823–1831. doi: 10.3233/JAD-210493 4. Khan Z, Da Silva MV, Nunez K, Kalafatis C, Nowicki S, Walker Z, Testad I, Francis P, Ballard C. Investigating the effects of impairment in non verbal communication on neuropsychiatric‐ symptoms and quality of life of people living with dementia. Alzheimer’s & Dementia: Translational Research & Clinical Interventions 2021 Jan;7(1). doi: 10.1002/trc2.12172 5. Atee M, Hoti K, Parsons R, Hughes JD. Pain assessment in dementia: Evaluation of a point-ofcare technological solution. Journal of Alzheimer’s Disease 2017 Aug 29;60(1):137–150. doi: 10.3233/JAD-170375 6. Lints-Martindale AC, Hadjistavropoulos T, Lix LM, Thorpe L. A comparative investigation of observational pain assessment tools for older adults with dementia. Clin J Pain 2012 Mar;28(3):226–237. doi: 10.1097/AJP.0b013e3182290d90 7. Hodgson N, Gitlin LN, Winter L, Hauck WW. Caregiver’s perceptions of the relationship of pain to behavioral and psychiatric symptoms in older community-residing adults with dementia. Clin J Pain 2014 May;30(5):421–427. doi: 10.1097/AJP.0000000000000018 8. Atee M, Hoti K, Chivers P, Hughes JD. Faces of pain in dementia: Learnings from a real-world study using a technology-enabled pain assessment tool. Frontiers in Pain Research 2022 Feb 22;3. doi: 10.3389/fpain.2022.827551 9. Pu L, Coppieters MW, Smalbrugge M, Jones C, Byrnes J, Todorovic M, Moyle W. Associations between facial expressions and observational pain in residents with dementia and chronic pain. J Adv Nurs 2024 Sep 9;80(9):3846–3855. doi: 10.1111/jan.16063 10. Abbey J, Piller N, de Bellis A, Esterman A, Parker D, Giles L, Lowcay B. The Abbey pain scale: A 1-minute numerical indicator for people with end-stage dementia. Int J Palliat Nurs 2004 Jan;10(1):6–13. doi: 10.12968/ijpn.2004.10.1.12013 11. Hølen JC, Saltvedt I, Fayers PM, Hjermstad MJ, Loge JH, Kaasa S. Doloplus-2, a valid tool for behavioural pain assessment? BMC Geriatr 2007 Dec 19;7(1):29. doi: 10.1186/1471-2318-7-29 12. Fuchs-Lacelle S, Hadjistavropoulos T. Development and preliminary validation of the Pain Assessment Checklist for Seniors with Limited Ability to Communicate (PACSLAC). Pain Management Nursing 2004 Mar;5(1):37–49. PMID:14999652 13. Warden V, Hurley AC, Volicer L. Development and psychometric evaluation of the Pain Assessment in Advanced Dementia (PAINAD) scale. J Am Med Dir Assoc 2003 Jan;4(1):9–15. doi: 10.1097/01.JAM.0000043422.31640.F7 14. Achterberg W, Lautenbacher S, Husebo B, Erdal A, Herr K. Pain in dementia. Pain Rep 2020 Jan;5(1). doi: 10.1097/PR9.0000000000000803 15. Fry M, Chenoweth L, Arendts G. Assessment and management of acute pain in the older person with cognitive impairment: A qualitative study. Int Emerg Nurs 2016 Jan;24:54–60. doi: 10.1016/j.ienj.2015.06.003 16. Eritz H, Hadjistavropoulos T. Do informal caregivers consider nonverbal behavior when they assess pain in people with severe dementia? J Pain 2011 Mar;12(3):331–339. doi: https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al 10.1016/j.jpain.2010.07.012 17. Herr K, Zwakhalen S, Swafford K. Observation of pain in dementia. Curr Alzheimer Res 2017 Apr 3;14(5):486–500. doi: 10.2174/1567205013666160602234526 18. Tsai YI, Browne G, Inder KJ. Documented nursing practices of pain assessment and management when communicating about pain in dementia care. J Adv Nurs 2022 Oct 18;78(10):3174–3186. doi: 10.1111/jan.15251 19. Atee M, Hoti K, Hughes JD. Psychometric evaluation of the electronic pain assessment tool: An innovative instrument for individuals with moderate-to-severe dementia. Dement Geriatr Cogn Disord 2018;44(5–6):256–267. doi: 10.1159/000485377 20. Atee M, Hoti K, Hughes JD. A technical note on the PainChekTM system: A web portal and mobile medical device for assessing pain in people with dementia. Front Aging Neurosci 2018 Jun 12;10. doi: 10.3389/fnagi.2018.00117 21. Sardar A, Umer S, Rout RKr, Wang S-H, Tanveer M. A secure face recognition for IoT-enabled healthcare system. ACM Trans Sens Netw 2023 Aug 31;19(3):1–23. doi: 10.1145/3534122 22. AGS Panel on Persistent Pain in Older Persons. The management of persistent pain in older persons. J Am Geriatr Soc 2002 Jun 18;50(S6):205–224. doi: 10.1046/j.1532-5415.50.6s.1.x 23. Park H, Lim S. Awareness of pain assessment for elderly patients with dementia among geriatric care hospital nurses. Asia-pacific Journal of Convergent Research Interchange 2022 Jan 31;8(1):77–86. doi: 10.47116/apjcri.2022.01.06 24. Project AI4HOPE. Available from: https://www.ai4hope.eu/ [accessed Apr 8, 2025] 25. Arksey H, O’Malley L. Scoping studies: Towards a methodological framework. Int J Soc Res Methodol 2005 Feb;8(1):19–32. doi: 10.1080/1364557032000119616 26. Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, Moher D, Peters MDJ, Horsley T, Weeks L, Hempel S, Akl EA, Chang C, McGowan J, Stewart L, Hartling L, Aldcroft A, Wilson MG, Garritty C, Lewin S, Godfrey CM, Macdonald MT, Langlois E V., SoaresWeiser K, Moriarty J, Clifford T, Tunçalp Ö, Straus SE. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med 2018 Oct 2;169(7):467– 473. doi: 10.7326/M18-0850 27. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, Chou R, Glanville J, Grimshaw JM, Hróbjartsson A, Lalu MM, Li T, Loder EW, Mayo-Wilson E, McDonald S, McGuinness LA, Stewart LA, Thomas J, Tricco AC, Welch VA, Whiting P, Moher D. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 2021 Mar 29;372:n71. doi: 10.1136/bmj.n71 28. López-Illescas C, de Moya-Anegón F, Moed HF. Coverage and citation impact of oncological journals in the Web of Science and Scopus. J Informetr Elsevier; 2008 Oct 1;2(4):304–316. doi: 10.1016/J.JOI.2008.08.001 29. Zhu J, Liu W. A tale of two databases: The use of Web of Science and Scopus in academic papers. Scientometrics 2020 Apr 22;123(1):321–335. doi: 10.1007/s11192-020-03387-8 30. Burnham JF. Scopus database: A review. Biomed Digit Libr 2006 Mar 8;3(1):1. doi: 10.1186/1742-5581-3-1 31. Gavel Y, Iselid L. Web of Science and Scopus: A journal title overlap study. Online Information Review 2008 Feb 22;32(1):8–21. doi: 10.1108/14684520810865958 32. Kazi M, Chowdhury N, Chowdhury M, Turin T. Conducting comprehensive scoping reviews to systematically capture the landscape of a subject matter. Popul Med 2021 Dec 23;3(December):1–9. doi: 10.18332/popmed/143831 33. Dobrescu A, Nussbaumer-Streit B, Klerings I, Wagner G, Persad E, Sommer I, Herkner H, Gartlehner G. Restricting evidence syntheses of interventions to English-language publications is a viable methodological shortcut for most medical topics: a systematic review. J Clin Epidemiol 2021 Sep;137:209–217. doi: 10.1016/j.jclinepi.2021.04.012 34. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol 2006 Jan;3(2):77– https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al 101. doi: 10.1191/1478088706qp063oa 35. Hoti K, Atee M, Chivers P, Vahia I, Hughes J. Technology-guided assessment of vocalisations and their diagnostic value as pain indicators for people living with dementia. Age Ageing 2023 Jun 1;52(6). doi: 10.1093/ageing/afad088 36. Burfield A, Wan, Sole, Cooper. Behavioral cues to expand a pain model of the cognitively impaired elderly in long-term care. Clin Interv Aging 2012;7:207–213. doi: 10.2147/CIA.S29656 37. Cervo FA, Raggi RP, Bright-Long LE, Wright WK, Rows G, Torres AE, Levy RB, Komaroff E. Use of the Certified Nursing Assistant Pain Assessment Tool (CPAT) in nursing home residents with dementia. Am J Alzheimers Dis Other Demen 2007 Apr 1;22(2):112–119. doi: 10.1177/1533317506298907 38. Chan S, Hadjistavropoulos T, Williams J, Lints-Martindale A. Evidence-based development and initial validation of the Pain Assessment Checklist for Seniors with Limited Ability to Communicate-II (PACSLAC-II). Clin J Pain 2014 Sep;30(9):816–824. doi: 10.1097/AJP.0000000000000039 39. Cohen-Mansfield J, Creedon M. Nursing staff members’ perceptions of pain indicators in persons with severe dementia. Clin J Pain 2002 Jan;18(1):64–73. doi: 10.1097/00002508200201000-00010 40. Feldt KS. The checklist of nonverbal pain indicators (CNPI). Pain Management Nursing 2000 Mar;1(1):13–21. doi: 10.1053/jpmn.2000.5831 41. Horgas AL, Elliott AF, Marsiske M. Pain assessment in persons with dementia: Relationship between self report and behavioral observation. J Am Geriatr Soc 2009 Jan 31;57(1):126–132.‐ doi: 10.1111/j.1532-5415.2008.02071.x 42. Husebo BS, Ballard C, Cohen-Mansfield J, Seifert R, Aarsland D. The response of agitated behavior to pain management in persons with dementia. The American Journal of Geriatric Psychiatry 2014 Jul;22(7):708–717. doi: 10.1016/j.jagp.2012.12.006 43. Husebo BS, Ostelo R, Strand LI. The MOBID pain scale: Reliability and responsiveness to pain in patients with dementia. European Journal of Pain 2014 Nov 5;18(10):1419–1430. doi: 10.1002/ejp.507 44. De Witt Jansen B, Brazil K, Passmore P, Buchanan H, Maxwell D, McIlfactrick SJ, Morgan SM, Watson M, Parsons C. Nurses’ experiences of pain management for people with advanced dementia approaching the end of life: A qualitative study. J Clin Nurs 2017 May 7;26(9– 10):1234–1244. doi: 10.1111/jocn.13442 45. Kaasalainen S, Stewart N, Middleton J, Knezacek S, Hartley T, Ife C, Robinson L. Development and evaluation of the Pain Assessment in the Communicatively Impaired (PACI) tool: Part I. Int J Palliat Nurs 2011 Aug;17(8):387–391. doi: 10.12968/ijpn.2011.17.8.387 46. Kaasalainen S, Stewart N, Middleton J, Knezacek S, Hartley T, Ife C, Robinson L. Development and evaluation of the Pain Assessment in the Communicatively Impaired (PACI) tool: Part II. Int J Palliat Nurs 2011 Sep;17(9):431–438. doi: 10.12968/ijpn.2011.17.9.431 47. Karlsson CE, Ernsth Bravell M, Ek K, Bergh I. Home healthcare teams’ assessments of pain in care recipients living with dementia: A Swedish exploratory study. Int J Older People Nurs 2015 Sep 14;10(3):190–200. doi: 10.1111/opn.12072 48. Kunz M, de Waal MWM, Achterberg WP, Gimenez Llort L, Lobbezoo F, Sampson EL, van‐ Dalen Kok AH, Defrin R, Invitto S, Konstantinovic L, Oosterman J, Petrini L, van der Steen JT,‐ Strand L, de Tommaso M, Zwakhalen S, Husebo BS, Lautenbacher S. The Pain Assessment in Impaired Cognition scale (PAIC15): A multidisciplinary and international approach to develop and test a meta tool for pain assessment in impaired cognition, especially dementia. European‐ Journal of Pain 2019 Jan 13;24(1):192–208. doi: 10.1002/ejp.1477 49. Lints-Martindale AC, Hadjistavropoulos T, Barber B, Gibson SJ. A psychophysical investigation of the facial action coding system as an index of pain variability among older adults with and https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al without Alzheimer’s disease. Pain Medicine 2007 Nov 1;8(8):678–689. doi: 10.1111/j.15264637.2007.00358.x 50. Lundin E, Godskesen TE. End-of-life care for people with advanced dementia and pain: A qualitative study in Swedish nursing homes. BMC Nurs 2021 Dec 20;20(1):48. doi: 10.1186/s12912-021-00566-7 51. Mahoney AEJ, Peters L. The Mahoney pain scale: Examining pain and agitation in advanced dementia. Am J Alzheimers Dis Other Demen 2008 Jun 7;23(3):250–261. doi: 10.1177/1533317508317524 52. Pautex S, Herrmann FR, Michon A, Giannakopoulos P, Gold G. Psychometric properties of the Doloplus-2 observational pain assessment scale and comparison to self-assessment in hospitalized elderly. Clin J Pain 2007 Nov;23(9):774–779. doi: 10.1097/AJP.0b013e318154b6e3 53. Rezaei S, Moturu A, Zhao S, Prkachin KM, Hadjistavropoulos T, Taati B. Unobtrusive pain monitoring in older adults with dementia using pairwise and contrastive training. IEEE J Biomed Health Inform 2021 May;25(5):1450–1462. doi: 10.1109/JBHI.2020.3045743 54. Richey SA, Capezuti E, Cron SG, Reed D, Torres-Vigil I, de Oliveira Otto MC. Development and testing of the Pain Assessment Tool in Cognitively Impaired Elders (PATCIE): Nonverbal pain behaviors in African American and Caucasian nursing home residents with dementia. Pain Management Nursing 2020 Apr;21(2):187–193. doi: 10.1016/j.pmn.2019.08.004 55. Sandvik RK, Selbaek G, Seifert R, Aarsland D, Ballard C, Corbett A, Husebo BS. Impact of a stepwise protocol for treating pain on pain intensity in nursing home patients with dementia: A cluster randomized trial. European Journal of Pain 2014 Nov 13;18(10):1490–1500. doi: 10.1002/ejp.523 56. Shega JW, Rudy T, Keefe FJ, Perri LC, Mengin OT, Weiner DK. Validity of pain behaviors in persons with mild to moderate cognitive impairment. J Am Geriatr Soc 2008 Sep 3;56(9):1631– 1637. doi: 10.1111/j.1532-5415.2008.01831.x 57. Takai Y, Yamamoto-Mitani N, Ko A, Heilemann M V. Differences in pain measures by minimental state examination scores of residents in aged care facilities: Examining the usability of the Abbey pain scale-Japanese version. Pain Management Nursing 2014 Mar;15(1):236–245. doi: 10.1016/j.pmn.2012.09.004 58. Torvik K, Kaasa S, Kirkevold Ø, Saltvedt I, Hølen JC, Fayers P, Rustøen T. Validation of Doloplus-2 among nonverbal nursing home patients - an evaluation of Doloplus-2 in a clinical setting. BMC Geriatr 2010 Dec 20;10(1):9. doi: 10.1186/1471-2318-10-9 59. van de Rijt LJM, Weijenberg RAF, Feast AR, Vickerstaff V, Lobbezoo F, Sampson EL. Oral health and orofacial pain in people with dementia admitted to acute hospital wards: Observational cohort study. BMC Geriatr 2018 Dec 23;18(1):121. doi: 10.1186/s12877-0180810-7 60. van Iersel T, Timmerman D, Mullie A. Introduction of a pain scale for palliative care patients with cognitive impairment. Int J Palliat Nurs 2006 Feb;12(2):54–59. doi: 10.12968/ijpn.2006.12.2.20531 61. Ellis PD. Interpreting effects. The Essential Guide to Effect Sizes Cambridge University Press; 2010. p. 31–44. doi: 10.1017/CBO9780511761676.003 62. AGS Panel on Persistent Pain in Older Persons. The management of persistent pain in older persons. J Am Geriatr Soc 2002 Jun 18;50(S6):205–224. doi: 10.1046/j.1532-5415.50.6s.1.x 63. Hadjistavropoulos T, Herr K, Prkachin KM, Craig KD, Gibson SJ, Lukas A, Smith JH. Pain assessment in elderly adults with dementia. Lancet Neurol 2014 Dec;13(12):1216–1227. doi: 10.1016/S1474-4422(14)70103-6 64. Prkachin KM, Solomon PE. The structure, reliability and validity of pain expression: Evidence from patients with shoulder pain. Pain 2008 Oct 15;139(2):267–274. doi: 10.1016/j.pain.2008.04.010 65. Kunz M, Mylius V, Scharmann S, Schepelman K, Lautenbacher S. Influence of dementia on https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al multiple components of pain. European Journal of Pain 2009 Mar 9;13(3):317–325. doi: 10.1016/j.ejpain.2008.05.001 66. Herr K, Coyne PJ, McCaffery M, Manworren R, Merkel S. Pain assessment in the patient unable to self-report: position statement with clinical practice recommendations. Pain Management Nursing 2011 Dec;12(4):230–250. doi: 10.1016/j.pmn.2011.10.002 67. Husebo BS, Ballard C, Sandvik R, Nilsen OB, Aarsland D. Efficacy of treating pain to reduce behavioural disturbances in residents of nursing homes with dementia: cluster randomised clinical trial. BMJ 2011 Jul 15;343(jul15 1):d4065–d4065. doi: 10.1136/bmj.d4065 68. Guohou S, Lina Z, Dongsong Z. What reveals about depression level? The role of multimodal features at the level of interview questions. Information & Management 2020 Nov;57(7):103349. doi: 10.1016/j.im.2020.103349 69. Smrke U, Mlakar I, Lin S, Musil B, Plohl N. Language, Speech, and Facial Expression Features for Artificial Intelligence–Based Detection of Cancer Survivors’ Depression: Scoping MetaReview. JMIR Ment Health 2021 Dec 6;8(12):e30439. doi: 10.2196/30439 70. Abdulazeem H, Borges do Nascimento IJ, Weerasekara I, Sharifan A, Grandi Bianco V, Cunningham C, Kularathne I, Deeken G, de Barros J, Sathian B, Østengaard L, LamontagneGodwin F, van Hoof J, Lazeri L, Redlich C, Marston HR, Dos Santos RA, Azzopardi-Muscat N, Yon Y, Novillo-Ortiz D. Use of digital health technologies for dementia care: Bibliometric analysis and report. JMIR Ment Health 2025 Feb 10;12. doi: 10.2196/64445 71. Rashid Z, Folarin AA, Zhang Y, Ranjan Y, Conde P, Sankesara H, Sun S, Stewart C, Laiou P, Dobson RJB. Digital phenotyping of mental and physical conditions: Remote monitoring of patients through RADAR-base platform. JMIR Ment Health 2024 Oct 23;11. doi: 10.2196/51259 https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint] JMIR Preprints Smrke et al Supplementary Files Untitled. URL: http://asset.jmir.pub/assets/6d9eeeb331cfd82df1c4be9bfc40b897.doc https://preprints.jmir.org/preprint/75671 [unpublished, peer-reviewed preprint]