Large-Scale Online stimulus Norming and Surveys about Perceptions in Healthcare

National Center For Complementary & Integrative Health

Investigators

Linked publications & trials

Abstract

Pain can be modulated by interpersonal processes that support the patient-provider relationship, and contextual factors related to the treatment environment. In this project, we investigate large-scale norms and beliefs about healthcare in America, and how perceptions of people and other visual cues in the treatment environment influence expectations about pain during treatment and about the pain other people feel. This project uses online survey methodologies to measure large scale normed beliefs and expectations from a geographically, ethnically, and socioeconomically diverse population of Americans. The purpose of this project is to investigate healthcare beliefs that may enhance or diminish pain in the clinic, which can be used to design studies in the laboratory to test if manipulating these beliefs can be advantageous for pain outcomes. In the past year, we published our first set of studies on this project (Necka et al., 2021, Social Science and Medicine). We collected five samples on the online platform Mechanical Turk to examine how first impressions of medical providers influence expectations of pain during a medical procedure and expectations for post-surgical recovery. Past work has demonstrated that perceptions of medical providers competence influence the experience of pain in the clinic, and that perceptions of medical providers similarity to a patient influence the patients experience of pain in a simulated clinical interaction. In domains outside of healthcare, research has demonstrated that even first impressions of traits such as competence can influence expectations and behavior: for instance, more competent looking politicians are more likely to be elected. Therefore, we were interested in measuring whether peoples perceptions of medical providers influenced their expectations about pain they might experience if that medical provider were to conduct a procedure on them. Participants completed surveys in which they viewed images of peoples faces and were told that they should imagine that the faces were those of potential medical providers who could conduct a procedure on them. They first chose medical providers that they would prefer to conduct painful procedures on them, and then they rated how painful they expected those procedures to be and the types of treatment they would expect that they would need in post-surgical recovery. In two of our samples, we used computer generated faces; in three other samples, we used real human faces. In four of our samples, participants rated how similar the stimulus faces were to themselves. In some samples, participants also rated the faces on perceived competence (in other samples, we used stimuli from published stimulus sets on which these characteristics had already been rated). We found that people preferred faces that looked more competent and more similar to them to be their medical providers. Furthermore, they expected to experience less pain and to have less need for prescription-strength pain medication following hypothetical procedures conducted by more competent looking medical providers and following procedures conducted by medical providers that were more similar to them. This indicates that individuals use first impressions to generate pain expectations, and has important implications for telehealth and virtual healthcare tools such as Zocdoc. We also completed data collection on a second series of experiments that evaluates whether individuals perceive pain differently based on the race or sex of a participant. Prior research suggests White individuals need stronger expressions before labeling a face as in pain when viewing the face of a Black actor relative to a White actor. However, these faces differed in identity, were actors, and were all male. We therefore ran a study in which we manipulated the same facial muscle movements (with action units determined based on the pain expression literature) but superimposed these on different identities. Therefore the expressions were identical, it was just the face they were imposed on that differed. We also decided to not only measure potential differences in pain judgments by race, but also by sex, such that female patients and research participants have lower pain tolerance and are less likely to be treated for pain than male patients, similar to the disparities between Black and White individuals regardless of sex. We completed three studies over the course of 2019-2020, including a period of racial unrest that led to increased awareness of racial disparities in the US. Our findings indicate that individuals were less likely to label the same expression as painful when displayed by a computer-generated female face relative to a male face, regardless of race. Participants also tended to label female faces as more fearful than male faces. We did not observe consistent findings as a function of the race of the face across studies, which may reflect the changing context and awareness of racial disparities in the middle of 2020. We have submitted this paper for publication (Dildine, Amir, Atlas, Submitted) and are currently undergoing peer review.