Neural Mechanism of Pain Assessment Bias among Monolingual and Multilingual Medical Trainees
Poster No:
1018
Submission Type:
Abstract Submission
Authors:
Morgan Gianola1, Elizabeth Losin2, Theoni Varoudaki2, Nikta Khalilkhani2
Institutions:
1University of Miami, Coral Gables, FL, 2Penn State University, State College, PA
First Author:
Co-Author(s):
Introduction:
Medical providers show bias in their prescribing of opioid analgesics for pain, overprescribing among white patients and undertreating pain among marginalized demographic groups (Santoro & Santoro, 2018; Keister et al., 2021). Moreover, low English proficiency patients tend to receive fewer pain assessments and lower pain evaluations than English proficient patients (Schwartz et al., 2022), and language concordant care has been associated with improvements across several health outcomes (Diamond et al., 2019). While differences in pain assessment and opioid prescribing across patients' language preferences have been noted (Schwartz et al., 2022), limited research has considered clinician language ability as a potentially relevant factor in pain assessment and opioid prescribing biases.
Methods:
During fMRI scanning, 66 medical trainees (32 male) participated as clinicians in a series of simulated patient interactions and pain assessments. They viewed short videos of mock patients experiencing genuine evoked pain before estimating each patient's pain intensity and reporting their likelihood to prescribe an analgesic. All participants saw 36 patients (from a pool of 72) evenly divided across six demographic groups: male and female non-Hispanic black, Hispanic, and non-Hispanic white. Trainee participants' evaluations of patients' pain were compared against the mock patients' own ratings of their pain. Outcomes of pain evaluation accuracy and likelihood of analgesia prescribing were compared across monolingual (N=23) and multilingual (N=43, knowing two or more languages) participants with mixed effects multilevel models with random effects for participants. Fixed effect variables for patient demographics and their interaction with participant language ability were included in the models. Neuroimaging data were pre-processed and univariate GLM whole brain activity was compared across monolingual and multilingual participants using FSL version 6.0.6.5
Results:
Monolingual trainees tended to underestimate black patients' pain as multilingual participants slightly overestimated pain in this group (B=0.78, t(109)=2.26, p=.026). Alternatively both mono- and multi-lingual trainees were fairly accurate in assessing non-Hispanic white patients' pain, resulting in a significant patient race by clinician language interaction (B= -0.62, t(2302)= -2.14, p=.032; Figure 1). These effects remain after controlling for the strength of patient's pain facial expressions. While pain overestimation positively predicted prescribing likelihood overall (B= 0.35, t(2337)= 22.96, p<.0001), both mono- and multi-lingual participants were less likely to prescribe analgesics to white (compared to black) patients (B= -0.41, t(2300)= -2.08, p=.037). When making treatment decisions for white patients, multilingual participants showed significantly greater activation in a cluster encompassing portions of left superior parietal and supramarginal gyrus (z=2.3, cluster corrected p<.05; Figure 2) compared to monolingual participants.
·Figure 1: Pain evaluation accuracy of monolingual and multilingual medical trainees across different patient racial/ethnic groups.
·Figure 2: Cluster showing significantly higher activity among multilingual compared to monolingual medical trainees when determining likelihood to prescribe an analgesic medication to a white patient.
Conclusions:
This research demonstrates that pain assessment accuracy and prescribing biases across patient demographics may differ between mono- and multi-lingual clinicians. However, assessment accuracy differences didn't necessarily alter prescribing behavior. This could suggest education around opioid abuse stereotypes among more recent trainee cohorts help alleviate previously observed prescribing biases (Morden et al., 2021). This racially and ethnically diverse sample of multilinguals showed greater activity in regions involved in valuation and social decision making (Ho et al., 2021) when deciding to treat white patients, potentially reflecting a wider range of factors being weighed, compared to monolinguals, reducing prescribing likelihood. These findings could relate to greater exposure to people from different demographic backgrounds among multilingual clinicians or wider bodies of information considered when making prescribing decisions.
Higher Cognitive Functions:
Decision Making
Language:
Language Other 1
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
Perception, Attention and Motor Behavior:
Perception: Pain and Visceral
Perception and Attention Other
Keywords:
Cognition
Data analysis
Language
MRI
NORMAL HUMAN
Pain
Perception
Social Interactions
Other - Multilingual
1|2Indicates the priority used for review
Provide references using author date format
2. Keister, L. A., Stecher, C., Aronson, B., McConnell, W., Hustedt, J., & Moody, J. W. (2021). Provider bias in prescribing opioid analgesics: a study of electronic medical records at a hospital emergency department. BMC public health, 21(1), 1-9.
3. Diamond, L., Izquierdo, K., Canfield, D., Matsoukas, K., & Gany, F. (2019). A systematic review of the impact of patient–physician non-English language concordance on quality of care and outcomes. Journal of General Internal Medicine, 34, 1591-1606.
4. Schwartz, H., Menza, R., Lindquist, K., Mackersie, R., Fernández, A., Stein, D., & Bongiovanni, T. (2022). Limited English Proficiency Associated With Suboptimal Pain Assessment in Hospitalized Trauma Patients. Journal of Surgical Research, 278, 169-178.
5. Morden, N. E., Chyn, D., Wood, A., & Meara, E. (2021). Racial inequality in prescription opioid receipt—role of individual health systems. New England Journal of Medicine, 385(4), 342-351.
6. Ho, S. S., Gonzalez, R. D., Abelson, J. L., & Liberzon, I. (2012). Neurocircuits underlying cognition–emotion interaction in a social decision making context. NeuroImage, 63(2), 843-857.