How Race Influences Trust in Online Prostate Cancer Content, Journal Club - Rashid Sayyid & Zachary Klaassen
August 30, 2023
Rashid Sayyid and Zach Klaassen explore a study published in JAMA Network Open about the role of racial concordance in trust levels for online videos about prostate cancer. The randomized clinical trial involved 2,400 U.S. adults, both Black and White, and assessed their trust in videos featuring different presenters—either Black or White, and either a physician or a patient. The study reveals that both Black and White adults trust physicians more than patients and are less trusting of videos about clinical trials compared to those about screening. Notably, Black adults show significantly higher trust when the information is presented by a Black speaker. The findings underscore the need for increased racial diversity and physician involvement in online health information, particularly in the context of prostate cancer, where racial disparities and mistrust have historical roots.
Biographies:
Rashid Sayyid, MD, MSc, Urologic Oncology Fellow, Division of Urology, University of Toronto, Toronto, Ontario
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center, Augusta, GA
Biographies:
Rashid Sayyid, MD, MSc, Urologic Oncology Fellow, Division of Urology, University of Toronto, Toronto, Ontario
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center, Augusta, GA
Read the Full Video Transcript
Rashid Sayyid: Hello everyone, this is Rashid Sayyid. I'm a urologic oncology fellow at the University of Toronto. Along with Zach Klaassen, associate professor and program director at Augusta University, we'll be discussing a recent publication in the JAMA Network Open. This publication looks at the effect of racial concordance on patient trust in online videos about prostate cancer within the context of a randomized clinical trial. As mentioned, this trial was recently published in JAMA Network Open, with Dr. Stacy Loeb as the first author.
Compared to White men, Black men are more likely to be diagnosed with prostate cancer. This has ranged from 1 in 4 to 1 in 6, depending on the reference you quote. Patients diagnosed with prostate cancer have a worse quality of life compared to their White counterparts, and they're also more likely to die from prostate cancer. Clearly, it's a cohort of men who have worse morbidity from this disease. There's been strong evidence in numerous meta-analyses that there's a significant interaction between race and social determinants of health with respect to prostate cancer-specific mortality.
We know that Black adults are more likely to trust online health information compared to White adults. However, Black adults have significantly more medical mistrust and are historically underrepresented in prostate cancer clinical trials and prostate cancer online content. At this point, it's unknown whether this influences decision-making regarding prostate cancer screening and clinical trial participation. There's also strong evidence that racial concordance with clinicians is positively associated with the perceived quality of healthcare for treated Black adults. Whether this same concept extends to online content remains unknown.
The study objective was to assess, within the context of an RCT, whether racial concordance and other source characteristics, such as the speaker being a physician or a patient, impact health consumers' trust in online health videos. The focus was on videos regarding prostate cancer screening and clinical trials. For the purposes of this trial, a collaborative team of clinicians, behavioral scientists, and community stakeholders designed and filmed a video script about prostate cancer screening and clinical trials. Each of the two videos was filmed with four different presenters, all of whom were men: either a Black physician or a White physician and a Black patient or a White patient. This gave a total of eight videos for the purposes of this study.
Videos were embedded into a Qualtrics survey, which included a baseline questionnaire with demographics, including participants' self-reported gender, race, and ethnicity. It also assessed their health literacy using the eHealth Literacy Scale and medical mistrust using the Group-Based Medical Mistrust Scale. After the video, participants were asked whether they trusted the information using a published 4-point Likert scale of trust in clinicians.
Now, for the viewers, this study included either Black or White U.S. adults who are aged 40 years or older. It did include female participants, given that female counterparts are often involved in seeking online information for prostate cancer for a friend, relative, father, son, etc., and are also involved in the shared decision-making process. Given the current representation in the community and across clinics, the authors deemed that a 4:1 ratio, meaning 4 men to 1 female participant, would be a fair ratio. Patients were recruited by Dynata, a large online survey platform used in prior studies across the U.S. The target was 1,200 Black and White adult participants each, for a total of 2,400 participants. Participants were randomized in parallel in eight blocks by a randomized tool to a survey containing one of the eight aforementioned videos. All surveys contained identical questions except for the actual embedded video, allowing for a comparison across these different eight videos.
The primary outcome of differences in trust in videos with a Black versus White presenter, whether the presenter was a physician versus a patient, and whether it mattered whether it was about screening versus a clinical trial topic, was compared using the Chi-square test. Trust was operationalized as dichotomous variables to simplify things. On one end, you either trusted completely or most of the information, which was deemed as "yes, they trusted," versus if there was only some trust or not at all, that was deemed as "no trust" for the purposes of this analysis.
The calculated sample size of 2,400 patients, meaning 1,200 each of Black and White adults, gave a 92% power to detect a 10% difference in trust according to presenter race, qualifications, or the video topic. Multivariable logistic regression analysis was used to evaluate the video features that were associated with trust, with mutual adjustment for speaker race, qualifications, or topic. Conversely, individual factors at the participant level that were associated with trust across all the videos, such as participants' demographics and health literacy, were evaluated as potential predictors of trust versus mistrust.
At this point, I'll turn it over to Zach to go over the results and discussion for this trial.
Zachary Klaassen: Thanks so much, Rashid. This is the flow chart for the selection of the study population. We can see that in terms of Black men and women, there were 1,912 Black adults and 524 Black men, ultimately comprising 1,703 Black adults that were included in the analysis. With regards to White men, 1,161 were recruited for the study, along with 382 White women, comprising 1,201 White adults who were ultimately included in this analysis.
For the demographics of the study population, stratified by Black versus White, we see that the mean age was slightly older for White adults at 63 years of age compared to 55.5 for Black adults. We see almost a 50/50 ratio for men and women in terms of participants with regards to Black participants, which is 75% for men and 25% for women for White adults. When we look at education, it's actually quite equal between these groups, with the most common education level being a college graduate for both groups: 32.2% of Black adults and 35.5% for White adults. With regards to personal cancer history, a majority of these participants did not have previous cancer, roughly 85%, but 9.1% of Black participants had a prior history of prostate cancer compared to 6.7% for White participants. With regards to health literacy, roughly two-thirds of patients had limited health literacy compared to one-third that had adequate health literacy.
This is a summary of the results as we move into some of the further, more advanced analyses. Among Black adults, there was a greater proportion that had high trust in videos with a Black speaker versus a White speaker, at 72.7% versus 64.3%. Secondly, there was a greater proportion that had high trust in videos with a physician versus a patient, at 72.5% versus 64.6%. A marginally higher proportion had high trust in videos about screening versus clinical trials, at 70.7% versus 66.3%. Among White adults, there was a greater proportion that had high trust in videos with a physician versus a patient, at 78.6% versus 72.0%. A greater proportion had high trust in videos about screening versus clinical trials, similar to the Black participants, at 79.1% versus 71.4%. However, there was no difference for Black versus White presenter among White participants in this trial, at 76.8% versus 73.7%.
This table looks at the multivariable logistic regression model for trust in prostate cancer videos according to the video characteristics. Among Black adults, we see there's a significant difference between Black versus White speakers, with an adjusted odds ratio of 1.48 and a 95% confidence interval of 1.21 to 1.82. We see, again, a significant difference between patient versus physician, with an odds ratio of 0.69 and a 95% confidence interval of 0.56 to 0.85. Finally, with regards to Black adults, we see a difference in clinical trials versus screening and trust with these videos, with an adjusted odds ratio of 0.81, which was also statistically significant.
When we look at the White adults, we'll start at the bottom. We see a difference in clinical trials versus screening, with an odds ratio of 0.66 and a P-value of 0.02. We see a difference in patient versus physician, with an odds ratio of 0.70 and a 95% confidence interval of 0.54 to 0.91. What's interesting, though, is we see no difference in Black versus White speakers among White participants, with an adjusted odds ratio of 1.18 and a P-value of 0.21.
This table looks at the univariable and multivariable logistic regression analysis of individual factors from the questionnaire associated with trust across prostate cancer videos. This part of the table is specific to Black adults, and we'll focus on the multivariable regression model, which you can see to the right of this table. I've highlighted with asterisks where there is significance on multivariable analysis. Black versus White speaker had higher trust, with an odds ratio of 1.62 and a 95% confidence interval of 1.28 to 2.05. We see less trust with a patient, with a statistically significant odds ratio of 0.63. We see clinical trials being less trusted than screening trials, with a statistically significant odds ratio of 0.78. Men who were queried were deemed to be more trustworthy, with an odds ratio of 1.34 and a 95% confidence interval of 1.05 to 1.70.
What's interesting is participants who had a personal prostate cancer history also had more trust in these videos, with an odds ratio of 1.79 and a statistically significant P-value of 0.04. Finally, limited health literacy led to decreased trust, with a statistically significant odds ratio of 0.69 and a 95% confidence interval of 0.53 to 0.91. This is among just the Black participants in this trial.
Looking at the White participants, again, we see less trust for patients in the videos, with an odds ratio of 0.71 and a 95% confidence interval of 0.54 to 0.95. Similar to the Black participants, clinical trials had less trust than screening trials, with an odds ratio of 0.57. Finally, as we saw previously among Black participants, those who were White and had a prior prostate cancer history had increased trust, with an odds ratio of 2.29 and a 95% confidence interval of 1.12 to 4.67.
By way of discussion, in this randomized clinical trial of online prostate cancer videos, there were several notable findings. First, for both Black and White adults, physicians were trusted more than patients, and there was less trust about clinical trials versus screening trials. Secondly, Black adults had greater trust in information presented by a Black speaker in this trial. These results highlight actionable opportunities to improve care. First, it's important to increase diversity in medical information for patients. Secondly, it's important for physicians to be deemed as trusted resources, and it's essential for physicians to actively disseminate evidence-based content online. Finally, for online health information platforms, representation certainly does matter. It's beneficial to tailor health communications by perceived racial groups, specifically for prostate cancer, where specific groups such as Black men have worse health outcomes and have sought care in a historical context of mistrust.
In conclusion, this randomized clinical trial of prostate cancer information was considered more trustworthy when delivered by a physician. However, racial concordance was significantly associated with trust only among Black adults. These results highlight the importance of physician participation and increasing racial diversity in the public dissemination of health information and an ongoing need for public education about clinical trials.
We thank you very much for your attention, and we hope you enjoyed this UroToday Journal Club discussion.
Rashid Sayyid: Hello everyone, this is Rashid Sayyid. I'm a urologic oncology fellow at the University of Toronto. Along with Zach Klaassen, associate professor and program director at Augusta University, we'll be discussing a recent publication in the JAMA Network Open. This publication looks at the effect of racial concordance on patient trust in online videos about prostate cancer within the context of a randomized clinical trial. As mentioned, this trial was recently published in JAMA Network Open, with Dr. Stacy Loeb as the first author.
Compared to White men, Black men are more likely to be diagnosed with prostate cancer. This has ranged from 1 in 4 to 1 in 6, depending on the reference you quote. Patients diagnosed with prostate cancer have a worse quality of life compared to their White counterparts, and they're also more likely to die from prostate cancer. Clearly, it's a cohort of men who have worse morbidity from this disease. There's been strong evidence in numerous meta-analyses that there's a significant interaction between race and social determinants of health with respect to prostate cancer-specific mortality.
We know that Black adults are more likely to trust online health information compared to White adults. However, Black adults have significantly more medical mistrust and are historically underrepresented in prostate cancer clinical trials and prostate cancer online content. At this point, it's unknown whether this influences decision-making regarding prostate cancer screening and clinical trial participation. There's also strong evidence that racial concordance with clinicians is positively associated with the perceived quality of healthcare for treated Black adults. Whether this same concept extends to online content remains unknown.
The study objective was to assess, within the context of an RCT, whether racial concordance and other source characteristics, such as the speaker being a physician or a patient, impact health consumers' trust in online health videos. The focus was on videos regarding prostate cancer screening and clinical trials. For the purposes of this trial, a collaborative team of clinicians, behavioral scientists, and community stakeholders designed and filmed a video script about prostate cancer screening and clinical trials. Each of the two videos was filmed with four different presenters, all of whom were men: either a Black physician or a White physician and a Black patient or a White patient. This gave a total of eight videos for the purposes of this study.
Videos were embedded into a Qualtrics survey, which included a baseline questionnaire with demographics, including participants' self-reported gender, race, and ethnicity. It also assessed their health literacy using the eHealth Literacy Scale and medical mistrust using the Group-Based Medical Mistrust Scale. After the video, participants were asked whether they trusted the information using a published 4-point Likert scale of trust in clinicians.
Now, for the viewers, this study included either Black or White U.S. adults who are aged 40 years or older. It did include female participants, given that female counterparts are often involved in seeking online information for prostate cancer for a friend, relative, father, son, etc., and are also involved in the shared decision-making process. Given the current representation in the community and across clinics, the authors deemed that a 4:1 ratio, meaning 4 men to 1 female participant, would be a fair ratio. Patients were recruited by Dynata, a large online survey platform used in prior studies across the U.S. The target was 1,200 Black and White adult participants each, for a total of 2,400 participants. Participants were randomized in parallel in eight blocks by a randomized tool to a survey containing one of the eight aforementioned videos. All surveys contained identical questions except for the actual embedded video, allowing for a comparison across these different eight videos.
The primary outcome of differences in trust in videos with a Black versus White presenter, whether the presenter was a physician versus a patient, and whether it mattered whether it was about screening versus a clinical trial topic, was compared using the Chi-square test. Trust was operationalized as dichotomous variables to simplify things. On one end, you either trusted completely or most of the information, which was deemed as "yes, they trusted," versus if there was only some trust or not at all, that was deemed as "no trust" for the purposes of this analysis.
The calculated sample size of 2,400 patients, meaning 1,200 each of Black and White adults, gave a 92% power to detect a 10% difference in trust according to presenter race, qualifications, or the video topic. Multivariable logistic regression analysis was used to evaluate the video features that were associated with trust, with mutual adjustment for speaker race, qualifications, or topic. Conversely, individual factors at the participant level that were associated with trust across all the videos, such as participants' demographics and health literacy, were evaluated as potential predictors of trust versus mistrust.
At this point, I'll turn it over to Zach to go over the results and discussion for this trial.
Zachary Klaassen: Thanks so much, Rashid. This is the flow chart for the selection of the study population. We can see that in terms of Black men and women, there were 1,912 Black adults and 524 Black men, ultimately comprising 1,703 Black adults that were included in the analysis. With regards to White men, 1,161 were recruited for the study, along with 382 White women, comprising 1,201 White adults who were ultimately included in this analysis.
For the demographics of the study population, stratified by Black versus White, we see that the mean age was slightly older for White adults at 63 years of age compared to 55.5 for Black adults. We see almost a 50/50 ratio for men and women in terms of participants with regards to Black participants, which is 75% for men and 25% for women for White adults. When we look at education, it's actually quite equal between these groups, with the most common education level being a college graduate for both groups: 32.2% of Black adults and 35.5% for White adults. With regards to personal cancer history, a majority of these participants did not have previous cancer, roughly 85%, but 9.1% of Black participants had a prior history of prostate cancer compared to 6.7% for White participants. With regards to health literacy, roughly two-thirds of patients had limited health literacy compared to one-third that had adequate health literacy.
This is a summary of the results as we move into some of the further, more advanced analyses. Among Black adults, there was a greater proportion that had high trust in videos with a Black speaker versus a White speaker, at 72.7% versus 64.3%. Secondly, there was a greater proportion that had high trust in videos with a physician versus a patient, at 72.5% versus 64.6%. A marginally higher proportion had high trust in videos about screening versus clinical trials, at 70.7% versus 66.3%. Among White adults, there was a greater proportion that had high trust in videos with a physician versus a patient, at 78.6% versus 72.0%. A greater proportion had high trust in videos about screening versus clinical trials, similar to the Black participants, at 79.1% versus 71.4%. However, there was no difference for Black versus White presenter among White participants in this trial, at 76.8% versus 73.7%.
This table looks at the multivariable logistic regression model for trust in prostate cancer videos according to the video characteristics. Among Black adults, we see there's a significant difference between Black versus White speakers, with an adjusted odds ratio of 1.48 and a 95% confidence interval of 1.21 to 1.82. We see, again, a significant difference between patient versus physician, with an odds ratio of 0.69 and a 95% confidence interval of 0.56 to 0.85. Finally, with regards to Black adults, we see a difference in clinical trials versus screening and trust with these videos, with an adjusted odds ratio of 0.81, which was also statistically significant.
When we look at the White adults, we'll start at the bottom. We see a difference in clinical trials versus screening, with an odds ratio of 0.66 and a P-value of 0.02. We see a difference in patient versus physician, with an odds ratio of 0.70 and a 95% confidence interval of 0.54 to 0.91. What's interesting, though, is we see no difference in Black versus White speakers among White participants, with an adjusted odds ratio of 1.18 and a P-value of 0.21.
This table looks at the univariable and multivariable logistic regression analysis of individual factors from the questionnaire associated with trust across prostate cancer videos. This part of the table is specific to Black adults, and we'll focus on the multivariable regression model, which you can see to the right of this table. I've highlighted with asterisks where there is significance on multivariable analysis. Black versus White speaker had higher trust, with an odds ratio of 1.62 and a 95% confidence interval of 1.28 to 2.05. We see less trust with a patient, with a statistically significant odds ratio of 0.63. We see clinical trials being less trusted than screening trials, with a statistically significant odds ratio of 0.78. Men who were queried were deemed to be more trustworthy, with an odds ratio of 1.34 and a 95% confidence interval of 1.05 to 1.70.
What's interesting is participants who had a personal prostate cancer history also had more trust in these videos, with an odds ratio of 1.79 and a statistically significant P-value of 0.04. Finally, limited health literacy led to decreased trust, with a statistically significant odds ratio of 0.69 and a 95% confidence interval of 0.53 to 0.91. This is among just the Black participants in this trial.
Looking at the White participants, again, we see less trust for patients in the videos, with an odds ratio of 0.71 and a 95% confidence interval of 0.54 to 0.95. Similar to the Black participants, clinical trials had less trust than screening trials, with an odds ratio of 0.57. Finally, as we saw previously among Black participants, those who were White and had a prior prostate cancer history had increased trust, with an odds ratio of 2.29 and a 95% confidence interval of 1.12 to 4.67.
By way of discussion, in this randomized clinical trial of online prostate cancer videos, there were several notable findings. First, for both Black and White adults, physicians were trusted more than patients, and there was less trust about clinical trials versus screening trials. Secondly, Black adults had greater trust in information presented by a Black speaker in this trial. These results highlight actionable opportunities to improve care. First, it's important to increase diversity in medical information for patients. Secondly, it's important for physicians to be deemed as trusted resources, and it's essential for physicians to actively disseminate evidence-based content online. Finally, for online health information platforms, representation certainly does matter. It's beneficial to tailor health communications by perceived racial groups, specifically for prostate cancer, where specific groups such as Black men have worse health outcomes and have sought care in a historical context of mistrust.
In conclusion, this randomized clinical trial of prostate cancer information was considered more trustworthy when delivered by a physician. However, racial concordance was significantly associated with trust only among Black adults. These results highlight the importance of physician participation and increasing racial diversity in the public dissemination of health information and an ongoing need for public education about clinical trials.
We thank you very much for your attention, and we hope you enjoyed this UroToday Journal Club discussion.