Coronavirus
Conditions
Brief summary
Recent data have shown that covid19 is disproportionately infecting and killing African Americans and Latinx people in the United States. The aim of the study is to determine which messages are most effective at increasing knowledge and changing behaviors that can protect individuals and their communities from the virus. To accomplish this aim, we plan to recruit approximately 20,000 Hispanic and African-American individuals and randomly assign them to videos that vary either the sender or the framing of the message, while providing the relevant public health information.
Detailed description
The following procedures will be followed. The research subjects will participate online. 1. Doctors from MGH who express interest in participating will record messages according to our scripts (US Doctors Script African American, US Doctors Script Latinx). They will be instructed to film several different versions each. 2. We will recruit a sample of Latinx and African American study participants from across the country through Lucid, an online survey firm that has access to a large subject pool. 3. Participants will first read a consent script and give us their informed consent (US Doctors Messaging Consent) 4. Participants will then navigate through the following steps 1. Brief demographics survey questions (US Doctors Messaging Survey) 2. Videos 1 and 2: Introduction, Acknowledgement, Social Distancing (US Doctors Script African American, US Doctors Script Latinx). Individuals in the control group will not see videos. 3. Beliefs survey questions (US Doctors Messaging Survey) 4. Video 3: Masks (US Doctors Script African American, US Doctors Script Latinx). Again, individuals in the control group will not see a video here. 5. Main outcomes survey (US Doctors Messaging Survey) 6. Control group is shown a version of all 3 videos at end (US Doctors Script African American, US Doctors Script Latinx). 5. The video messages each participant in one of the treatment groups sees will be randomized on the following dimensions. The different versions are clearly demarcated in the US Doctors Scripts document. 1. Racial or ethnic identity of the doctor delivering the messages: concordant vs. discordant identity to the subject. 2. Whether the message includes an acknowledgment of elephant in the room issues for each target group: trust in the medical system or fear of deportation. 3. Whether the social distancing component of the message is delivered by Dr. Birx of the CDC or recorded by the MGH physicians. 4. Whether individuals are given information about how representative individuals perceive mask wearers of color. This information comes from results of our nationally representative pilot survey. 5. Some individuals in a control group will only see messages after all surveying has been completed.
Interventions
BEHAVIORALMessaging
o The video messages each participant in one of the treatment groups sees will be randomized on the following dimensions. The different versions are clearly demarcated in the US Doctors Scripts document. * Racial or ethnic identity of the doctor delivering the messages: concordant vs. discordant identity to the subject. * Whether the message includes an acknowledgment of elephant in the room issues for each target group: trust in the medical system or fear of deportation. * Whether the social distancing component of the message is delivered by Dr. Birx of the CDC or recorded by the MGH physicians. * Whether individuals are given information about how representative individuals perceive mask wearers of color. This information comes from results of our nationally representative pilot survey. * Some individuals in a control group will see messages later.
Sponsors
Massachusetts General Hospital
Stanford University
Yale University
Massachusetts Institute of Technology
Harvard University
National Bureau of Economic Research, Inc.
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
Hispanic and African-American adult individuals, men and women - oversample of those with less completed education
Exclusion criteria
children, those who do not identify as above
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Incidence Rate for Knowledge Gaps: Control vs. Any Intervention | The outcome was measured immediately following the intervention in the 15-minute baseline survey. | Incidence rate for knowledge gaps is the count of knowledge gaps divided by the maximum possible count: 7 |
| Incidence Rate for Information-seeking Behavior: Control vs. Any Intervention | The outcome was measured immediately following the intervention in the 15-minute baseline survey. | Incidence rate for Information-seeking behavior is the count of links demanded divided by the maximum possible count: 10 |
| Incidence Rate for Knowledge Gaps by Intervention | The outcome was measured immediately following the intervention in the 15-minute baseline survey. | Incidence rate for knowledge gaps is the count of knowledge gaps divided by the maximum possible count: 7 |
| Incidence Rate for Information-seeking Behavior by Intervention | The outcome was measured immediately following the intervention in the 15-minute baseline survey. | Incidence rate for Information-seeking behavior is the count of links demanded divided by the maximum possible count: 10 |
Countries
United States
Participant flow
Recruitment details
Individuals were recruited throughout the United States by the survey company Lucid from 13 May to 26 May 2020. Only persons aged 18 years or older and self-identifying as Black or Latinx were eligible.
Pre-assignment details
The trial sample was enrolled from 13 May 2020 through 26 May 2020. Of 35,987 persons screened, 20,925 were eligible for participation (15,062 who would have been eligible were excluded because the demographic quotas were met); 5,450 individuals did not consent or failed both basic attention checks, and 1,208 left the survey before randomization. Out of 14,267 participants enrolled and randomized,123 participants did not answer one or more baseline questions and were excluded from analysis.
Participants by arm
| Arm | Count |
|---|---|
| Study Participants - Control Study participants who are assigned to a pure control group | 1,617 |
| Study Participants - Any Intervention Study participants who are assigned to any intervention groups | 12,527 |
| Total | 14,144 |
Baseline characteristics
| Characteristic | Study Participants - Any Intervention | Study Participants - Control | Total |
|---|---|---|---|
| Age, Continuous | 38.9 years STANDARD_DEVIATION 15.2 | 39.16 years STANDARD_DEVIATION 15.6 | 38.93 years STANDARD_DEVIATION 15.3 |
| Education index | 4.57 scores on a scale STANDARD_DEVIATION 1.5 | 4.53 scores on a scale STANDARD_DEVIATION 1.5 | 4.56 scores on a scale STANDARD_DEVIATION 1.5 |
| Hispanic origin Argentina | 74 Participants | 8 Participants | 82 Participants |
| Hispanic origin Colombia | 97 Participants | 8 Participants | 105 Participants |
| Hispanic origin Cuba | 208 Participants | 22 Participants | 230 Participants |
| Hispanic origin Ecuador | 38 Participants | 9 Participants | 47 Participants |
| Hispanic origin El Salvador | 76 Participants | 8 Participants | 84 Participants |
| Hispanic origin Guatemala | 40 Participants | 3 Participants | 43 Participants |
| Hispanic origin Mexican, Mexican American, Chicano | 3271 Participants | 387 Participants | 3658 Participants |
| Hispanic origin Nicaragua | 19 Participants | 5 Participants | 24 Participants |
| Hispanic origin Other country | 608 Participants | 73 Participants | 681 Participants |
| Hispanic origin Panama | 20 Participants | 2 Participants | 22 Participants |
| Hispanic origin Peru | 55 Participants | 8 Participants | 63 Participants |
| Hispanic origin Puerto Rico | 57 Participants | 6 Participants | 63 Participants |
| Hispanic origin Spain | 189 Participants | 27 Participants | 216 Participants |
| Hispanic origin Venezuela | 50 Participants | 9 Participants | 59 Participants |
| Household size | 3.21 person STANDARD_DEVIATION 1.7 | 3.23 person STANDARD_DEVIATION 1.7 | 3.22 person STANDARD_DEVIATION 1.7 |
| Likely to run out of money because of COVID-19 | 6531 Participants | 875 Participants | 7406 Participants |
| Race/Ethnicity, Customized African American | 7725 Participants | 1042 Participants | 8767 Participants |
| Race/Ethnicity, Customized Latinx | 4802 Participants | 575 Participants | 5377 Participants |
| Region of Enrollment United States | 12527 participants | 1617 participants | 14144 participants |
| Sex: Female, Male Female | 7169 Participants | 943 Participants | 8112 Participants |
| Sex: Female, Male Male | 5358 Participants | 674 Participants | 6032 Participants |
| Stratum Female, age ≤44 y, central | 2444 Participants | 327 Participants | 2771 Participants |
| Stratum Female, age ≤44 y, coastal | 1787 Participants | 223 Participants | 2010 Participants |
| Stratum Female, age ≥45 y, central | 1784 Participants | 229 Participants | 2013 Participants |
| Stratum Female, age ≥45 y, coastal | 1154 Participants | 164 Participants | 1318 Participants |
| Stratum Male, age ≤44 y, central | 2009 Participants | 268 Participants | 2277 Participants |
| Stratum Male, age ≤44 y, coastal | 1740 Participants | 199 Participants | 1939 Participants |
| Stratum Male, age ≥45 y, central | 885 Participants | 111 Participants | 996 Participants |
| Stratum Male, age ≥45 y, coastal | 724 Participants | 96 Participants | 820 Participants |
| Trouble paying for food | 4592 Participants | 558 Participants | 5150 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk | EG004 affected / at risk | EG005 affected / at risk | EG006 affected / at risk | EG007 affected / at risk | EG008 affected / at risk | EG009 affected / at risk | EG010 affected / at risk | EG011 affected / at risk | EG012 affected / at risk | EG013 affected / at risk | EG014 affected / at risk | EG015 affected / at risk | EG016 affected / at risk |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 784 | 0 / 791 | 0 / 780 | 0 / 778 | 0 / 777 | 0 / 788 | 0 / 783 | 0 / 788 | 0 / 779 | 0 / 780 | 0 / 782 | 0 / 790 | 0 / 779 | 0 / 786 | 0 / 781 | 0 / 781 | 0 / 1,617 |
| other Total, other adverse events | 0 / 784 | 0 / 791 | 0 / 780 | 0 / 778 | 0 / 777 | 0 / 788 | 0 / 783 | 0 / 788 | 0 / 779 | 0 / 780 | 0 / 782 | 0 / 790 | 0 / 779 | 0 / 786 | 0 / 781 | 0 / 781 | 0 / 1,617 |
| serious Total, serious adverse events | 0 / 784 | 0 / 791 | 0 / 780 | 0 / 778 | 0 / 777 | 0 / 788 | 0 / 783 | 0 / 788 | 0 / 779 | 0 / 780 | 0 / 782 | 0 / 790 | 0 / 779 | 0 / 786 | 0 / 781 | 0 / 781 | 0 / 1,617 |
Outcome results
Primary
Incidence Rate for Information-seeking Behavior by Intervention
Incidence rate for Information-seeking behavior is the count of links demanded divided by the maximum possible count: 10
Time frame: The outcome was measured immediately following the intervention in the 15-minute baseline survey.
Population: Each intervention indicated in Arm/Group Title was indeed an arm that included each sub arm of different types of message that we tested Arm A against Arm B, or vice versa, and pre-specified in our study protocol and analysis plan to combine arms for analysis. Arms A and B are indicated in each Arm/Group Title.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Study Participants - Control | Incidence Rate for Information-seeking Behavior by Intervention | 0.336 proportion of links |
| Study Participants - Any Intervention | Incidence Rate for Information-seeking Behavior by Intervention | 0.353 proportion of links |
| Intervention 2: Video 2 Recorded by MGH Physician/Dr. Birx of the CDC - MGH Physician (Arm A) | Incidence Rate for Information-seeking Behavior by Intervention | 0.343 proportion of links |
| Intervention 2: Video 2 Recorded by MGH Physician/Dr. Birx of the CDC - Dr. Birx of the CDC (Arm B) | Incidence Rate for Information-seeking Behavior by Intervention | 0.348 proportion of links |
| Intervention 3: Received Acknowledgment of Discrimination - No (Arm A) | Incidence Rate for Information-seeking Behavior by Intervention | 0.346 proportion of links |
| Intervention 3: Received Acknowledgment of Discrimination - Yes (Arm B) | Incidence Rate for Information-seeking Behavior by Intervention | 0.341 proportion of links |
| Intervention 4: Received Acknowledgment of Economic Inequalities - No (Arm A) | Incidence Rate for Information-seeking Behavior by Intervention | 0.344 proportion of links |
| Intervention 4: Received Acknowledgment of Economic Inequalities - Yes (Arm B) | Incidence Rate for Information-seeking Behavior by Intervention | 0.345 proportion of links |
| Intervention 5: Received Information About Social Perception - No (Arm A) | Incidence Rate for Information-seeking Behavior by Intervention | 0.345 proportion of links |
| Intervention 5: Received Information About Social Perception - Yes (Arm B) | Incidence Rate for Information-seeking Behavior by Intervention | 0.344 proportion of links |
Primary
Incidence Rate for Information-seeking Behavior: Control vs. Any Intervention
Incidence rate for Information-seeking behavior is the count of links demanded divided by the maximum possible count: 10
Time frame: The outcome was measured immediately following the intervention in the 15-minute baseline survey.
Population: Any intervention was indeed an arm that included each sub arm of different types of message that we tested against a control and pre-specified in our study protocol and analysis plan to combine arms for analysis.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Study Participants - Control | Incidence Rate for Information-seeking Behavior: Control vs. Any Intervention | 0.335 proportion of links |
| Study Participants - Any Intervention | Incidence Rate for Information-seeking Behavior: Control vs. Any Intervention | 0.344 proportion of links |
Primary
Incidence Rate for Knowledge Gaps by Intervention
Incidence rate for knowledge gaps is the count of knowledge gaps divided by the maximum possible count: 7
Time frame: The outcome was measured immediately following the intervention in the 15-minute baseline survey.
Population: Each intervention indicated in Arm/Group Title was indeed an arm that included each sub arm of different types of message that we tested Arm A against Arm B, or vice versa, and pre-specified in our study protocol and analysis plan to combine arms for analysis. Arms A and B are indicated in each Arm/Group Title.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Study Participants - Control | Incidence Rate for Knowledge Gaps by Intervention | 0.066 proportion of knowledge gaps |
| Study Participants - Any Intervention | Incidence Rate for Knowledge Gaps by Intervention | 0.065 proportion of knowledge gaps |
| Intervention 2: Video 2 Recorded by MGH Physician/Dr. Birx of the CDC - MGH Physician (Arm A) | Incidence Rate for Knowledge Gaps by Intervention | 0.065 proportion of knowledge gaps |
| Intervention 2: Video 2 Recorded by MGH Physician/Dr. Birx of the CDC - Dr. Birx of the CDC (Arm B) | Incidence Rate for Knowledge Gaps by Intervention | 0.068 proportion of knowledge gaps |
| Intervention 3: Received Acknowledgment of Discrimination - No (Arm A) | Incidence Rate for Knowledge Gaps by Intervention | 0.066 proportion of knowledge gaps |
| Intervention 3: Received Acknowledgment of Discrimination - Yes (Arm B) | Incidence Rate for Knowledge Gaps by Intervention | 0.063 proportion of knowledge gaps |
| Intervention 4: Received Acknowledgment of Economic Inequalities - No (Arm A) | Incidence Rate for Knowledge Gaps by Intervention | 0.066 proportion of knowledge gaps |
| Intervention 4: Received Acknowledgment of Economic Inequalities - Yes (Arm B) | Incidence Rate for Knowledge Gaps by Intervention | 0.063 proportion of knowledge gaps |
| Intervention 5: Received Information About Social Perception - No (Arm A) | Incidence Rate for Knowledge Gaps by Intervention | 0.066 proportion of knowledge gaps |
| Intervention 5: Received Information About Social Perception - Yes (Arm B) | Incidence Rate for Knowledge Gaps by Intervention | 0.065 proportion of knowledge gaps |
Primary
Incidence Rate for Knowledge Gaps: Control vs. Any Intervention
Incidence rate for knowledge gaps is the count of knowledge gaps divided by the maximum possible count: 7
Time frame: The outcome was measured immediately following the intervention in the 15-minute baseline survey.
Population: Any intervention was indeed an arm that included each sub arm of different types of message that we tested against a control and pre-specified in our study protocol and analysis plan to combine arms for analysis.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Study Participants - Control | Incidence Rate for Knowledge Gaps: Control vs. Any Intervention | 0.085 proportion of knowledge gaps |
| Study Participants - Any Intervention | Incidence Rate for Knowledge Gaps: Control vs. Any Intervention | 0.065 proportion of knowledge gaps |