Comments on Pu et al. (2024) ‘Associations between facial expressions and observational pain in residents with dementia and chronic pain’

With regard to the authors' comment on the use of presence or absence of AUs as opposed to their intensities for assessing pain, there is a failure to acknowledge that other validated pain assessment tools in people with dementia such as the Pain Assessment Checklist for Seniors with Limited Ability to Communicate (PACSLAC) and the Checklist of Nonverbal Pain Indicators (CNPI) also use this binary scoring approach. This approach helps mitigate against user subjectivity and adds to the ease of use. Furthermore, pain tools (e.g., PACSLAC) with binary scoring are preferred by nurses over those with ordinal scoring format as the latter are much more complex to administer (Zwakhalen et al., 2006). The authors recommended the Prkachin-Solomon Pain Intensity (PSPI) Index (which uses ordinal AU scoring except for AU43) for grading the intensities of core facial AUs of pain (listed under point 3). Nevertheless, binary-scored pain tools such as PACSLAC II was found to have a stronger correlation with self-reported pain (gold standard of pain assessment) than the PSPI index (Hadjistavropoulos et al., 2018). In the ‘Implications for Practice’ section the multidimensionality of the PainChek® App goes unmentioned, despite it supporting best practice in pain assessment as cited by the authors. To account for variability in pain expressions and manifestations (e.g., ‘stoic’ face), the multi-domain App covers a wide range of evidence-based facial, vocal, somatic, kinetic, behavioural and functional items. These aspects are supportive of both the multidimensional nature of pain and the biopsychosocial model of pain (Atee et al., 2018). Thus, given the considerable evidence of work regards its validity (Atee et al., 2018; Babicova et al., 2021), the fact that PainChek® is regulatory cleared as a medical device in Australia, United Kingdom, Europe, Canada and Singapore, and its adoption into clinical practice in Australia and internationally (with over 4 million PainChek assessments completed to date), we believe that their comments under both ‘The limitations of the PainChek® App’ and ‘Implications for Practice’ sections are unfounded and misleading. The above points lead us to seek the authors' justification for their final two concluding remarks ‘These facial expressions were independent of age, gender, cognitive impairment and cultural background, which indicates the potential of automated real-time facial analysis as part of the pain assessment in people with dementia. However, more research is still required to develop new and valid AI-based algorithms that can be applied to support healthcare’. The PainChek® App completes the automated facial analysis in real time (3 s) and its algorithms are trained to detect nine AUs which are indicative of pain. Similar to previous studies evaluating its psychometric properties, the PainChek® App's algorithms have been demonstrated in the Pu et al. study to detect those nine AUs. (Atee et al., 2018; Babicova et al., 2021). Furthermore, while we do not dispute the ongoing need for innovation, the authors have questioned the validity of PainChek's AI-based algorithm without providing any evidence from their research to back up their conclusions. This is also while providing a contradictory statement suggesting ‘the potential of automated real-time facial assessment’ (which in fact reflects the PainChek® App, the tool that was used in the Pu et al. study). How was a higher observational pain score defined and what was the justification for that? Apart from a brief mention in the Methods under Data Analysis that the adjusted observational pain scores were the ‘33 items of pain behaviours’, the authors fail to explain the rationale for this or what constituted a higher adjusted observational pain score? These should have been clarified in detail in the Methods. Furthermore, were the adjusted scores only included for people with pain or the entire sample (i.e., those with and without pain)? Including scores of people with no pain may have skewed the results It was unfortunate that this information was not included in the Pu et al. paper as each represents a potential limitation to the study. We feel it is important that we bring these points to the attention of your readers and to allow the authors to address them. That will then permit your readers to examine the presented evidence with a fair lens and then decide whether the Pu et al. results demonstrate anything other than the PainChek® App does what it is designed to do, acknowledging though that all technologies and pain assessment tools have their own limitations. Yours sincerely JH, MA, PC and KH all contributed equally to the ideas, writing and approval of this Letter to the Editor. This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. The authors have nothing to report. JH, MA and KH are co-inventors of the original PainChek instrument (branded ePAT at the time), which was acquired and subsequently commercialized by PainChek Ltd. They are shareholders of PainChek Ltd. JH currently holds the position of Chief Scientific Officer at PainChek Ltd, while serving as an Emeritus Professor at Curtin Medical School. MA previously held the position of a Senior Research Scientist (October 2018–May 2020) at PainChek Ltd, and currently serving in the position of Research and Practice Lead at The Dementia Centre, HammondCare. KH is employed as a consultant by PainChek Ltd, while also serving as a Professor at the University of Prishtina, Kosovo. The co-inventors had authored a patent titled ‘A pain assessment method and system; PCT/AU2015/000501’ which was assigned to PainChek Ltd and who have, to date, received granted patents in the jurisdictions of China, Japan and the United States. PC is the Principal Consultant of DATaR Consulting providing independent biostatistical services, while also serving as Associate Professor at the University of Notre Dame Australia and Adjunct Research Fellow at Edith Cowan University. PC has previously been paid as an independent consultant to complete the data analysis for PainChek Ltd sponsored projects. Not applicable.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.