Sociocultural & biobehavioral influences on pain expression and assessment

点击翻译按钮获取中文摘要

This protocol measures pain-related facial responses in a diverse population to measure whether nonverbal responses to pain vary as a function of biological and sociocultural factors. We will then measure whether individuals (both healthy volunteers and medical providers) pay attention to different features of pain or assess pain differently in in-group relative to out-group individuals, and whether we can develop interventions to reduce any biases in attention or pain assessment. During FY22, we completed data collection for the protocols first two sub-studies. During FY23, we focused on data analysis for sub-study 1, and began data collection for the third sub-study. Sub-study 1 measures the association between noxious stimuli, pain, and facial responses. We collected our 100th participant in July of 2022, which completed our anticipated data collection based on the protocol, following numerous setbacks due to the COVID-19 pandemic (see FY22 report for details). Although our final enrollment did not achieve equal numbers across our key racial and gender demographics (i.e. 20 participants per group for Black male, Black female, White male, White female), difficulties in recruiting individuals from minoritized groups are well documented, and we will be explicit about our recruitment efforts and restricted eligibility and enrollment when we submit the manuscript. We will use specialized software to measure facial responses via video and to avoid implicit biases that could affect results if we used human coders. We will measure whether sex differences are observed in facial responses that are similar to sex differences in pain, as well as whether we see differences in facial responses or sex differences as a function of race, ethnicity, or identity centrality of race or sex. The PhD student who led the study graduated and started a postdoctoral fellowship during FY23, and thus we have paused analyses. We have met with scientists in the NIMH Machine Learning Section to consult on facial analyses, and anticipate that this project will be led by new postdoctoral fellows in the Section, in collaboration with the former PhD student. Following completion, sub-study 1 participants were asked whether they want their images to be included in a database that will be shown to other participants. Images of a subset of participants who opted into this database were used as stimuli for a subsequent sub-study 2, which measured how individuals view and judge pain in others that they perceive to be similar or different from them. 60 participants (perceivers) viewed 96 videos of people (targets) undergoing painful stimulation. There were 12 targets from 4 sociodemographic groups: Black male, Black female, White male, White female (3 targets per group, 2 innocuous and 2 painful videos per participant). Perceivers assessed pain using the same scales as the targets, which allowed us to measure whether sociodemographic factors influenced pain assessment accuracy. We found that perceivers attributed less pain to Black individuals relative to White individuals, and were less accurate in rating pain in Black individuals. This finding mimics the well documented health disparities in pain that are observed in clinical settings and previous research. Participants also over-estimated pain in female individuals relative to males, in contrast to disparities observed in the clinic, and we observed interactions between race and gender, consistent with intersectionality. We also found that racial differences in pain assessment were stronger for individuals who reported higher values in the modern racism scale, and that group membership (i.e. whether an individual self-reported the same racial or gender identity as the person in the video) did not impact assessment. We recently submitted these findings for publication and are awaiting reviews (Dildine et al., Submitted). We also began data-collection on a third sub-study during the FY23. This study builds on sub-study 2, such that healthy volunteers (perceivers) view videos of sub-study 1 participants (targets) experiencing pain and evaluate pain using the same scales as the targets. However, half the participants receive feedback about the targets actual pain on every trial, whereas half the participants receive no feedback. We will measure whether feedback improves pain assessment accuracy, and use computational models to evaluate how individuals learn about other individuals pain. We are also collecting eye-tracking and autonomic data while individuals perform the task. Exploratory analyses will examine relationships between nonverbal pain assessment, gaze position, and arousal. 16 participants have enrolled in the study, including 6 pilots. Data collection will terminate when 46 participants complete the task. Finally, the sub-study 3 task will form the foundation for future studies that will examine neural mechanisms of nonverbal pain assessment, test whether performance-based feedback reduces sociocultural biases in pain assessment, and test whether pain assessment and social learning differ across groups (e.g. healthcare providers versus laypeople, patients versus healthy volunteers).