Effect of Physician Race and Gender on Simulated Patients' Ratings

• Sponsor: Yale University (Other)
• Overall Status: Completed
• CT.gov ID: NCT04190901
• Collaborator: (none)
• Enrollment: 3,592
• Location: 1
• Arms: 4
• Actual Duration (Months): 4.7
• Patients Per Site Per Month: 759.2

Study Details

Study Description

Brief Summary

The purpose of the study was to determine whether the race and gender of a simulated doctor affected analog patients' reported confidence and satisfaction in the simulated doctor's diagnosis and treatment plan. The study used two randomized patient analog experiments.

This study is complete and pre-analysis plans (PAPs) for each experiment were published prior to data collection. The PAPs are available at: http://aspredicted.org/blind.php?x=43xj25 (Study 1) and https://aspredicted.org/blind.php?x=369st7 (Study 2).

Condition or Disease	Intervention/Treatment	Phase
Bias, Racial; Bias, Sex; Patient Satisfaction; Physician-Patient Relations	Behavioral: Simulated Black Male Physician; Behavioral: Simulated Black Female Physician; Behavioral: Simulated White Male Physician; Behavioral: Simulated White Female Physician	N/A

Detailed Description

Prior literature is unclear on whether patients display bias in their evaluations of physicians based on their race or gender. The investigators estimated the effects of physician race and gender using an online clinical vignette. Participants played the role of analog patients reporting to the Emergency Department (ED) with symptoms consistent with gastroenteritis. Participants were provided with a diagnosis of gastroenteritis by a simulated ED physician. The race (black or white) and gender (male or female) of the simulated physician was randomly assigned in a 2x2 factorial experiment. Simulated physicians provided a diagnosis of gastroenteritis and contradicted by an Online Symptom Checker. Following the physician's diagnosis and contradiction by the Online Symptom Checker, participants rated the simulated physician on survey measures of satisfaction and confidence in both the treatment plan and diagnosis. The main (null) hypothesis tested was that there were no differences across the four treatment arms (Black Female, Black Male, White Female, White Male). Participants for the first experiment (Study 1) were recruited from Amazon Mechanical Turk (MTurk) and participants for the second experiment (Study 2) were recruited from Lucid.

Primary Aim: To determine whether the race and gender of a simulated physician had a causal effect on participants' confidence and satisfaction in the physician's diagnosis and treatment plan in an ED setting.

Exploratory Aims: To determine whether the race and gender of a simulated physician had a casual effect on participants' perceptions of the warmth and competence of the physician, their willingness to sue or complain about the physician for an incorrect diagnosis, and their perceived fairness of the charge for the visit.

Study Design

• Study Type: Interventional
• Actual Enrollment: 3592 participants
• Allocation: Randomized
• Intervention Model: Factorial Assignment
• Intervention Model Description: This is a 2x2 experimental design. Subjects were randomly assigned to one of four possible conditions: Black Female, Black Male, White Female, White Male. Within each condition, subjects were randomly assigned 1 of 10 possible putative doctors from a total of 40 putative doctors, 10 for each condition. For example, 10 Black Female doctors, etc. The images of putative physicians were selected from actors in the Chicago Face Database and altered to wear a white coat. Given that the vast majority of Emergency Physicians in the United States are white men, the White Males condition served as the "control".This is a 2x2 experimental design. Subjects were randomly assigned to one of four possible conditions: Black Female, Black Male, White Female, White Male. Within each condition, subjects were randomly assigned 1 of 10 possible putative doctors from a total of 40 putative doctors, 10 for each condition. For example, 10 Black Female doctors, etc. The images of putative physicians were selected from actors in the Chicago Face Database and altered to wear a white coat. Given that the vast majority of Emergency Physicians in the United States are white men, the White Males condition served as the "control".
• Masking: None (Open Label)
• Primary Purpose: Health Services Research
• Official Title: Effect of Physician Race and Gender on Simulated Patients' Ratings and Confidence in Their Physicians: A Randomized Trial
• Actual Study Start Date: Mar 9, 2018
• Actual Primary Completion Date: Jul 31, 2018
• Actual Study Completion Date: Jul 31, 2018

Arms and Interventions

Arm	Intervention/Treatment
Experimental: Simulated Black Male Physician Participants are randomized to view the clinical vignette with a simulated Black Male physician.	Behavioral: Simulated Black Male Physician; Participants in this arm of the experiment viewed one of 10 randomly selected possible images of a simulated Black Male physician. This image was paired with a written treatment and diagnosis of gastroenteritis alongside a contradictory diagnosis and treatment plan for appendicitis from an Online Symptom Checker.
Experimental: Simulated Black Female Physician Participants are randomized to view the clinical vignette with a simulated Black Female physician.	Behavioral: Simulated Black Female Physician; Participants in this arm of the experiment viewed one of 10 randomly selected possible images of a simulated Black Female physician. This image was paired with a written treatment and diagnosis of gastroenteritis alongside a contradictory diagnosis and treatment plan for appendicitis from an Online Symptom Checker.
Experimental: Simulated White Male Physician Participants are randomized to view the clinical vignette with a simulated White Male physician.	Behavioral: Simulated White Male Physician; Participants in this arm of the experiment viewed one of 10 randomly selected possible images of a simulated White Male physician. This image was paired with a written treatment and diagnosis of gastroenteritis alongside a contradictory diagnosis and treatment plan for appendicitis from an Online Symptom Checker.
Experimental: Simulated White Female Physician Participants are randomized to view the clinical vignette with a simulated White Female physician.	Behavioral: Simulated White Female Physician; Participants in this arm of the experiment viewed one of 10 randomly selected possible images of a simulated White Female physician. This image was paired with a written treatment and diagnosis of gastroenteritis alongside a contradictory diagnosis and treatment plan for appendicitis from an Online Symptom Checker.

Outcome Measures

Primary Outcome Measures

1. Patient Confidence [Approximately 10 minutes]

   "How confident are you that this doctor made the correct diagnosis?" Study 1: [not at all confident (0) to completely confident (100)] Study 2: [not at all confident (1) to completely confident (5)]* "How confident are you that this doctor made the correct treatment plan?" Study 1: [not at all confident (0) to completely confident (100)] Study 2: [not at all confident (1) to completely confident (5)]* The Patient Confidence outcome for each study participant was the unweighted average of their ratings on questions a and b. In Study 1, this item was measured using 0-100 point scales. In Study 2, this outcome was measured using 5 point scales. For all analyses, these Patient Confidence outcomes from a and b were rescaled to match the 1-5 point range from Study 2.

2. Believed Symptom Checker over Doctor [Approximately 10 minutes]

   "Which diagnosis do you think is more likely to be correct?" [the doctor's diagnosis (0); online symptom checker (1)]

3. Likelihood of Requesting more Tests [Approximately 10 minutes]

   "Would you ask the doctor to perform additional diagnostic tests? (Such as the CT scan recommended by the Symptom Checker)." [definitely not (1); probably not (2); might or might not (3); probably (4); definitely (5)]

4. Patient Satisfaction: scale [Approximately 10 minutes]

   "What number would you use to rate your care during this emergency room visit?" Study 1: [0 (worse possible care) to 100 (best possible care)] Study 2: [0 (worse possible care) to 10 (best possible care)]* *In Study 1, the Patient Satisfaction was measured using a 0-100 point scale. In Study 2, this was measured using a 10 point scale. For all analyses, this Patient Satisfaction outcome from Study 1 was rescaled to match the 0-10 point range in Study 2.

5. Likelihood to Recommend [Approximately 10 minutes]

   "Would you recommend this doctor to your friends and family?" [definitely not (1); probably not (2); might or might not (3); probably (4); definitely (5)]

Secondary Outcome Measures

1. Warmth and Competence [Approximately 10 minutes]

   Study 1: "How do you imagine this doctor would be in a real interaction?". 7-item scale: Tolerant, Warm, Sincere, Good-natured, Intelligent, Competent, Confident. Study 2: "Based on the doctor's diagnosis, to what extent do you find [him/her]". 6-item scale: Kind, Qualified, Intelligent, Competent, Open-minded, Trustworthy. Warmth and Competence were measured in Study 1 (7-item scale) and Study 2 (6-item scale).

2. Willingness to sue or complain [Approximately 10 minutes]

   "You take the doctor's advice and go home. Over the next few days, the pain in your abdomen got worse and you returned to the hospital where you were diagnosed with appendicitis. Your appendix had burst and you developed a serious infection. This required emergency surgery and an extended stay in the hospital's intensive Care Unit"* "Would you file a complaint against this doctor?" [5 = "Definitely"; 4 = "Probably"; 3 = "Might or might not"; 2 = "Probably not"; 1 = "Definitely not"] "Would you consider suing this doctors?" [5 = "Definitely"; 4 = "Probably"; 3 = "Might or might not"; 2 = "Probably not"; 1 = "Definitely not"] Willingness to sue or complain was only measured in Study 2.

3. Fairness of the cost [Approximately 10 minutes]

   "You would be charged about $350 for this emergency department visit. How fair do you think this charge is?" [0 = "Completely unfair" to 100 = "Completely Fair"]* *Fairness of the cost was only measured in Study 1.

Eligibility Criteria

Criteria

• Ages Eligible for Study: 18 Years and Older
• Sexes Eligible for Study: All
• Accepts Healthy Volunteers: Yes

Inclusion Criteria:

• Adults over 18 years old

Exclusion Criteria:

• Participants who reported current pregnancy

• Participants who reported a current or prior diagnosis of cancer

• Participants who reported a history of abdominal surgery

Contacts and Locations

Locations

	Site	City	State	Country	Postal Code
1	Yale University	New Haven	Connecticut	United States	06520

Sponsors and Collaborators

• Yale University

Investigators

None specified.

Study Documents (Full-Text)

• Statistical Analysis Plan - Jul 25, 2018

More Information

Publications

None provided.

• Responsible Party: Yale University
• ClinicalTrials.gov Identifier: NCT04190901
• Other Study ID Numbers: 2000022317; 12916; 9068
• First Posted: Dec 9, 2019
• Last Update Posted: Jul 21, 2020
• Last Verified: Jul 1, 2020

• Individual Participant Data (IPD) Sharing Statement: Yes
• Plan to Share IPD: Yes
• Studies a U.S. FDA-regulated Drug Product: No
• Studies a U.S. FDA-regulated Device Product: No

Study Results

No Results Posted as of Jul 21, 2020