Effective Community Consultation for Prehospital EFIC Research: Lessons from a 20-Site Pediatric Trial

Abstract

Objectives –

Research in the prehospital setting is frequently conducted with federal Exception from Informed Consent (EFIC). EFIC requires community consultation (CC) and public disclosure (PD) to inform community members about research plans and create opportunities for community feedback and study modification prior to approval. There is limited research on the relative effectiveness of varying CC and PD strategies, particularly with increasing utilization of social media. We sought to characterize the frequency of use and reach of specific CC/PD strategies in the Pediatric Dose Optimization for Seizure in EMS (PediDOSE) trial, assess their relative effectiveness, and offer guidance to future EFIC studies on optimal strategies and potential pitfalls in conducting CC/PD.

Methods –

We completed a retrospective review of the CC/PD plans and summary reports from all 20 sites in the PediDOSE trial. We used descriptive statistics to determine the frequency and reported reach of specific CC/PD strategies. We also described sites' modifications of their initial CC/PD plans.

Results –

All sites completed the required CC/PD activities, which began in October 2021. Sites conducted 52 CC meetings; meeting with community groups had the lowest attendance (median 9, IQR 2-11) and were the most frequently cancelled. Most sites (16/20) utilized social media for CC/PD; however, only 4 sites use paid social media advertisements. Paid advertising had greater reach on Facebook than posts to existing institutional followers (median reach 13992 paid vs 2504 free) at reasonable cost. All sites created PD materials and conducted CC surveys in Spanish; a majority of sites (11/20) reported actively recruiting minoritized populations for inclusion in CC/PD activities with variable success.

Conclusion –

Based on our experience we recommend the following: EFIC studies should utilized paid social media for PD. Studies should establish standardized metrics for success of social media campaigns before beginning CC/PD activities. CC should conduct community meetings with inclusion of all relevant stakeholder groups, and sites should consider using social media to augment CC activities. Investigators of EFIC studies involving children should develop explicit plans to engage children and minoritized populations in CC/PD.

INTRODUCTION

Conducting impactful research on time-critical illnesses requires standard human subjects’ protection review and occasionally a waiver of informed consent. In 1996, the U.S. Food and Drug Administration (FDA) and the U.S. Department of Health & Human Services (HHS) published regulations for ethical conduct of emergency research using the federally regulated exception from informed consent (EFIC) model.^1,2 EFIC is a vital yet rarely used tool to conduct for research on life-threatening conditions. In 2022, a review identified only 110 trials in the U.S. using EFIC. Of these, only 13 were actively recruiting and only 14 were pediatric, with 2 of the pediatric studies actively enrolling at the time of the review.³

Following publication of EFIC guidelines, the FDA produced guidance documents and recommendations on how community consultation (CC) and public disclosure (PD) must be incorporated into the EFIC research approval process.^4,5 These guidance documents encourage discussing the proposed research with groups that are representative of the community in which the research is conducted. For research studies involving children, this includes both disclosure of the proposed project to children’s legally authorized representatives (LARs) and discussion of the proposal with children who are old enough to understand the proposed work. Additionally, when conducting research in EMS systems it is important that EMS clinicians and system leaders be included in the CC/PD process.⁶ This process has been shown to have high acceptability with both participants in CC activities and study subjects.^7-10

The methods by which researchers conduct a CC and PD process are not prescribed by the FDA or HHS. Multiple methods of CC and PD have been previously described and can broadly be grouped into three methods: 1) Uni-directional PD through posting of flyers, news articles, and interviews disseminated through the media;^11,12 2) Bi-directional CC where feedback is obtained through random number dialing, community member interviews, focus groups, and community meetings;^8,11-13 3) Technology driven, bi-directional CC and PD through social media campaigns and virtual meetings that allow for direct public comment.^11,14-18

Despite nearly three decades of EFIC research, there is limited research on the relative effectiveness of varying CC and PD strategies. While multiple studies have described CC and PD activities conducted as part of large multicenter EFIC studies,^11,16,19 these studies did not evaluate the effectiveness of individual strategies. Other studies of smaller EFIC research studies have only examined specific aspects of CC/PD campaigns.^15,20,21 No studies offer practical guidance for future investigators on how to plan and implement a CC/PD strategy.

The Pediatric Dose Optimization for Seizures in EMS (PediDOSE) study is an EFIC study examining the implementation of age-based, standardized medication (midazolam) dosing for pediatric seizures.²² The EFIC plan for PediDOSE allowed sites to create individualized CC and PD plans with guidance for required and optional elements provided by the main study team. The objective of this study was to characterize the frequency of use and reach of specific CC/PD strategies, assess their relative effectiveness, and offer guidance to future EFIC studies on optimal strategies and potential pitfalls in conducting CC/PD.

METHODS

Setting

The PediDOSE study is a large, prospective, multicenter trial conducted through the Pediatric Emergency Care Applied Research Network (PECARN).²³ The study is being conducted in partnership between children’s hospitals and EMS agencies in 20 cities (Supplemental Table). The purpose of the study is to evaluate if a standardized, age-based method for paramedics to administer seizure medication reduces the frequency of ongoing seizure activity on hospital arrival. Because of the emergent nature of seizure treatment, the FDA and the Institutional Review Board (IRB) determined that the study met the requirements for subjects to be enrolled under EFIC.

The study is enrolling patients across 20 communities in which EMS agencies have partnered with academic institutions to conduct the study. The single institutional review board (sIRB) for this study, which approved all CC and PD materials and plans, is at the University of Utah.

Community Consultation and Public Disclosure Plan

The PediDOSE study team developed a CC and PD plan for implementation across all study sites, which included completion of a minimum of 100 online surveys and 10 semi-structured interviews at each site.²² The investigators designed the plan to ensure active consultation of key stakeholder groups in each community, while allowing individual sites flexibility in the methods used for consultation. While certain CC/PD activities were mandatory for all sites, sites were also provided a list of optional activities to complete CC and PD. Both mandatory and optional activities were sub-grouped into interactive CC (Table 1) and passive PD (Table 2) methods. Use of websites and social media accounts was a suggested passive method, while interviews or discussions on social media was a suggested active method. Before conducting CC and PD, sites were required to report their community demographics and encouraged to include minoritized groups in CC/PD activities. All sites were required to create Spanish summary materials for dissemination due to the anticipated prevalence of Spanish as a preferred language in all communities in this study. Planning and execution of the CC/PD activities occurred primarily from December 2021- July 2022 during a phase of the COVID-19 pandemic when variability existed in large in-person gatherings, so plans accounted for this. Individual sites had a budget of $8,000 each to complete EFIC activities before seeking sIRB approval to begin enrollment.

Table 1:

Community Consultation Interactive Methods as Developed by the Study Team and Required by the IRB

Direct Interaction (must do ALL of these)
Surveys of at least 100 patients/families in emergency departments (ED) and/or neurology departments (optional other location(s) Interviews (minimum 10): include bi-directional conversation with a mix of parents or guardians of children with and without pre-existing seizures to solicit feedback and provide education about the study to participants.
Public solicitation (must do ALL of these)
Solicit feedback via handouts/posters in high-risk patient areas (ED, neurology, other areas if feasible) Solicit feedback via study website Solicit feedback from medical community via email (ED, neurology, critical care, general pediatrics if feasible) Solicit feedback from emergency medical services (EMS) Medical Directors and EMS clinicians in participating communities via email
Community Meetings (must do at least 1) *Must be interactive and site must provide recording or transcript for the institutional review board
Present study information at meetings of community, local government, civic, or patient advocacy groups (pre- existing meetings preferred) and gather feedback Interviews or discussions on “talk” radio, television, or interactive social media program Meetings or activities provided by hospitals’ and institutions’ existing community outreach programs Press conferences and briefings Other (please describe):

Table 2:

Public Disclosure Passive Methods as Developed by the Study Team and Required by the IRB

Must do at least 3

Targeted mailings to households in the communities, with information about how to obtain further details. Advertisements and articles in media outlets in the English language, and if appropriate, foreign language (Public outreach documents should be translated into languages that are common in the area served by the facility) Clearly marked links to study information on the participating hospitals’ and emergency medical services agency’s internet web sites/ social media accounts using prompts from social media template (*IRB approved verbiage) Spanish summary materials that are accessible to non-English speaking residents who reside in the community from which research subjects are likely to be drawn (other languages as per local community requirements) Letters/emails to local and regional community leaders (e.g., police or other government entities) Newsletter posts and/or grand rounds presentations to local/regional hospital staff(s) Public service announcements (Newsletter text contains institutional review board approved verbiage) Other (please describe)

CC/PD materials received sIRB approval and were formally made available to all sites on October 19^th, 2021. Sites used these materials to develop their individual CC and PD plans, which were submitted to both local IRBs and the sIRB for approval. Once plans received sIRB approval, sites were allowed to begin CC/PD activities.

Upon completion of CC and PD activities, sites submitted a finalized CC and PD summary to the sIRB and their local IRB. The IRBs determined if the community had been adequately consulted and, if so, approved the study to be implemented at each individual site. After final IRB approval of CC/PD activities, each site was permitted to begin patient enrollment for the PediDOSE study.

Analysis of Community Consultation and Public Disclosure Strategies

To conduct our analysis, we reviewed each site’s CC and PD plan and final summary report. In their finalized CC and PD summary, sites submitted feedback from community meetings, social media impact data, and modifications from their initial CC and PD plans. For this analysis, we presumed all proposed CC/PD activities were completed unless sites explicitly reported plan modifications. We created a spreadsheet of site strategies for each interactive and passive method using Microsoft Excel, allowing us to comparatively review site strategies. We also created spreadsheets describing each community meeting and use of social media by each site.

We reviewed community meeting titles and descriptions to classify intended meeting audiences. We summarized numbers, formats (virtual vs. in-person vs. hybrid), and attendances for community meetings using descriptive statistics and compared across meeting types and audiences. Each site reported descriptions of and feedback from community meetings. The amount of feedback from each meeting was quantified and coded based on discussion topic. We summarized types and frequencies of feedback in aggregate and by meeting type using descriptive statistics. To determine the relative effectiveness of individual meeting types in community consultation, we performed comparisons based on attendance and quantity of feedback received per meeting. We considered meetings unsuccessful if they were either attended by less than 5 people or did not yield questions or feedback. Feedback from CC surveys and interviews have been previously reported,^22,24 and are therefore not described here.

We quantified numbers and types of social media posts using descriptive statistics. We described social media engagement levels using metrics provided by reporting sites, which varied both across and within social media platforms. Specific metrics reported for Twitter (now X) were Impressions (number of individuals viewing a tweet), Retweets (the sharing of a tweet on a user’s own timeline), Likes (user communicated approval of a tweet), and Engagements (individuals retweeting, favoriting, replying to, or clicking on content within a tweet).²⁴ Reported Facebook metrics were Reach (number of users viewing content), Reactions (number of users responding to content with Likes or other emojis), Shares (the placing of a post on a user’s own timeline or sending it to another user), and Likes.²⁵ Likes were the sole Instagram metric reported, while no metrics were reported from LinkedIn or NextDoor. When social media engagement was reported in aggregate across multiple posts, we used the mean engagement for each post to calculate descriptive statistics.

We reviewed all aspects of CC/PD summaries, including plans and meeting descriptions to describe engagement with minoritized populations. Finally, each site reported deviations from their initial CC/PD plans. These were reviewed and characterized using descriptive statistics.

RESULTS

All 20 participating sites completed required CC and PD activities. Two sites were required by the sIRB to conduct additional community meetings to obtain more interactive feedback after submission of their CC/PD summary; all other sites’ reports were approved on first submission. These additional meetings are included in the analysis below. Based on the primarily positive feedback obtained during the CC/PD process, no specific changes to the study protocol were made after the CC/PD process was completed. Sites’ selected optional activities are provided in Table 3.

Table 3.

Sites’ Selection of Optional Community Consultation / Public Disclosure Activities

Activity	Number of Sites (20 total)
Community Consultation Meetings (must do at least 1)
Meetings with community, government, civic or patient advocacy groups	17
Interviews / discussions on radio, television, or social media	6
Meetings of hospitals’ and institutions’ outreach programs	8
Press conferences and briefings	2
Other	4*

Activity	Number of Sites (20 total)
Public Disclosure Passive Methods (must do at least 3)
Targeted mailings to households	4*
Advertisements and articles in media outlets	10
Links to study information on hospitals’ and emergency medical services internet web sites/ social media accounts	13
Accessible Spanish summary materials	19
Letters/emails to local and regional community leaders	10
Newsletter posts and/or grand rounds presentations to hospital staff	14
Public service announcements	1
Other	3**

^*Included engaging epilepsy advocacy group, hospital town hall, and social media posts,

^*6 sites reported attempting to send targeted mailings, 2 were unable to do so

^**Included flyer distribution and letters to stakeholders.

Effectiveness of Variable Community Meeting Strategies

We can make comparisons of the effectiveness of specific meeting types for CC based on attendance, feedback, and frequency of cancellation. Hospital based advisory group and medical meetings both were well attended and provided frequent study feedback. Community based advisory groups meetings had lower attendance, offered less feedback, and were more likely to fall through. Meetings with EMS and government leaders were well attended but provided less feedback. Meetings conducted on social media platforms had much greater attendance than other meetings, though it is unclear the quantity or quality of feedback they elicited. Additionally, no sites explicitly reported on CC meetings with children in their CC/PD summaries.

Community group meetings varied in their ability to facilitate community consultation. Median attendance was 13 individuals (IQR 7-24), variable by meeting type (Table 4). Community interest group meetings were the least attended meeting type (median attendance 9, IQR 2-11), while EMS professional meetings were best attended (median 31, IQR 21-54). 13% (6/48) of meetings with reported attendance had less than 5 attendees. Three were community epilepsy interest groups, two were with a particular hospital’s family advisory council, and one was a community meeting targeting Hispanic and Latino community members.

Table 4.

Community Consultation Meetings by Group Type

Group Type	Number of Meetings	Median Attendance (IQR)
Community Interest Group	11	9 (2 – 11)
Emergency Medical Services Group	14	31 (21 - 54)
Local Government	7	25 (11.8 - 34.5)
Hospital Advisory Group	13	10 (7 - 16)
Hospital Staff Group	7	10 (6.5 - 18.5)
Meeting Type	Number of Meetings	Median Attendance (IQR)
Virtual	46	12 (6 - 20)
In-Person or Hybrid	6	25 (12 – 34.5)

^*Four meetings did not report attendance numbers

Six sites reported not being able to present at nine targeted community meetings, and this was the most common reason for deviation from initial CC/PD plans (Table 5); six of the meetings were with community groups, while three were planned meetings with government agencies.

Table 5 -.

Causes of Modifications to Community Consultation / Public Disclosure Plans

Modification Cause	Sites Reporting
Unable to present at targeted community meeting	6
Did not interview targeted number of Spanish speakers	2
Unable to access mailing lists of neurology patients	2
Unable to present at hospital grand rounds	2
Unable to post on emergency medical services websites / social media accounts	2
Not cost effective to advertise in local newspaper	1
Initial advertising plan deemed to have inadequate reach	1
Unable to engage Hispanic/Latino community group for presentation	1
Not feasible to post study information on hospital-wide computers	1

Quantity and quality of feedback from community consultation also varied by meeting type (Table 6) but elicited common themes (Table 7). Hospital based meetings provided the most community feedback (mean 3.1 question/comments per meeting) while community (0.90 per meeting), government (0.88 per meeting), and EMS (0.93 per meeting) meetings provided the least. Twenty percent (10/50) of meetings did not yield either questions or comments, including 55% (6/11) of meetings with EMS leadership, 30% (3/10) of community meetings, and 14% (1/7) of meetings with government officials.

Table 6 –

Community Meeting Feedback and Questions

Topic	Meeting Type (number of meetings) & Number of Comments/Questions
	Community (10)	EMS (13)	Government (7)	Hospital (13)	Medical (7)	Overall (50)
Medical questions	4	3	1	10	4	22
Study support	2	5	2	7	2	18
Consent concern		1	2	5	1	9
Consent question	1	2		3		6
Portable EEG question				2	3	5
CC/PD question	1			2		3
Community trust Concerns				3		3
Medical concern				3		3
Study design question/concern				2	1	3
EFIC support	1			1		2
EMS liability concern			2			2
Language question/concern			1	1		2
Transport questions				1	1	2
EMS training question		1				1
Total number question or comments	9	12	8	40	12	81

^*Feedback or questions were not received at all meetings

Table 7 –

Community Meeting Feedback Themes

Feedback Themes

Concern for appropriateness of dosing outside of typical weights for ages Questions about why the study age range was selected Questions about community outreach to inform about the study, particularly for underserved populations, recommendations of groups for outreach Medical providers - questions surrounding the use of the a portable EEG device How does consent work if EMS is enrolling in the study Concerns about how EFIC would be received by the community - possible negative reactions

Three sites hosting streaming videos on social media reached more than 9000 individuals. In comparison, 48 community consultation meetings had a total recorded attendance of 1,061 individuals. A Twitter livestream generated 145 impressions and 17 clicks on the study website, but no comments. A Facebook live video yielded 683 views but only 2 comments. A second Facebook video, boosted by $150 in advertisement, reached 7950 people, with 76 individuals clicking the link to the study survey and 2 additional comments.

Effectiveness of Variable Public Disclosure Strategies

Most sites (16) used at least one social media site for CC/PD, including Twitter (now X) (10 sites), Facebook (12 sites), Instagram (6), LinkedIn (2), and NextDoor (1). One site deviated from their initial PD plan as they could not obtain approval from their hospital to post on their Facebook page. Social media posts originated from a variety of sources (Table 8) and had varying levels of engagement (Table 9). However, the four sites that utilized paid advertisements on social media made had greater impact in their social media campaigns than those using free institutional accounts. Three of these sites reported the reach of individual Facebook posts (Figure 1). An additional site using paid social media advertising reported 73,367 impressions across two Facebook and one Instagram post. Only one site reported cost data; $25/day for seven days for their paid Facebook advertisement.

Table 8 –

Social Media Engagement (Excluding Live Events)

a) Twitter
Social Media Audience (n)	Impressions (range, (n))	Likes (range, (n))
Hospital (17)	443-1332 (7)	0-6 (13)
School / Department (4)	27-1518 (3)	2 (1)
Division (6)	50-856 (6)	0-17 (6)
Site Principal Investigator (1)	-	4 (1)
Podcast (1)	-	1 (1)

b) Facebook
Social Media Audience (n)	Reach (range, (n))
Paid Advertisement (3)	8622-15223 (3)
Pre-Existing Audiences
Research Network (1)	24 (1)
Department / Division (2)	336 (1)
Emergency Medical Services (5)	1362-3873 (2)
Foundation (5)	15-26 (2)
Hospital (4)	3720-10183 (4)

• Other reported account holders - Government (1)

Table 9 –

Social Media Engagement by Platform (Excluding Live Streams)

Social Media Platform (number of sites using)	Metric (number of posts with metric reported)	Median (Range)
Twitter (10)	Impressions (16)	461 (27-1518)
	Retweets (24)	1(0-4)
	Likes (22)	2 (0-6)
	Engagement (7)	5 (2-55)
Facebook (12)*	Reach – free (10)	2504 (15-10183)
	Reach – ad (3)	13942 (8622-15223)
	Reactions – free (6)	20 (11-27)
	Reactions – ad (1)	24
	Likes – free (7)	1 (0-22)
	Shares – free (6)	2 (0-6)
Instagram (6)	Likes (4)	11.512(11-14)
LinkedIn (2)	-	-
NextDoor (1)	-	-

^*Likes and Share not reported for Facebook ads

Figure 1:

Facebook Post Reach by Source

Other PD strategies were variably employed and frequently resulted in deviations from initial CC/PD plans. Advertisements tended to be used in smaller metropolitan areas (mean population 2.9 million vs 4.4 million),²⁶ and 2/10 sites planning to advertise either change their initial advertising plan or decided to forego advertising altogether. One third (2/6) of sites planning to send direct mailers to patients’ families were unable to do so (Table 5).

Effectiveness and Variability in Engagement with Minoritized Populations

All sites were required to create Spanish summary materials for dissemination; however, additional outreach to minoritize populations were infrequent and varied by site. Five sites either published or posted PD materials in the community aimed at either Spanish-speaking or minoritized populations. Three sites planned community meetings with minoritized groups. One conducted community meetings focused on Black participants, while a second met their local Immigrant and Refugee Commission. A third site was forced to deviate from their stated aim to meet with a Hispanic community group as they were unable to arrange a meeting to present the study with identified groups. Sites also were also instructed to create summary materials in other languages “as per local community requirements” yet only one site reported creating summary materials in another language.

CC Surveys were available in Spanish at all sites. Three sites discussed actively recruiting Spanish speakers to complete CC surveys, but only one site targeted a specific number of survey completions in Spanish. For CC interviews, six sites reported plans to interview one or more Spanish speakers, yet two of these sites reported deviating from this goal.

DISCUSSION

Strategies for CC/PD activities across 20 sites participating in the PediDOSE trial, a study on pediatric seizures in the EMS setting utilizing EFIC, varied in their effectiveness. We found that while sites frequently utilized social media for PD, few used it for CC, despite its greater reach than traditional CC activities. Few sites paid for social media advertising, greatly reducing the potential reach of PD. Engagement with minoritized populations was variable and engagement with children non-existent. Based on our findings we can make recommendations regarding future CC/PD activities as detailed below.

Social media is being increasingly utilized for PD in EFIC studies in place of traditional media platforms, use of which are costly and time consuming.^19,27 Social media offers a large reach^14,28 but has drawbacks including the short time that viewers engage with content,²⁸ the dynamic nature of trends in social media usage, and difficulty in targeting specific audiences of interest, including minoritized and non-English speaking communities.²⁹ For the PediDOSE study, most sites used social media, while fewer sites utilized traditional media outlets confirming ongoing trends in CC/PD strategies.

Many EFIC studies utilizing social media for CC/PD activities rely on paid advertisements.^14,17,20,28 However in the PediDOSE study, only 5 of 20 sites utilized paid advertisements through social media platforms while 11 others only utilized existing institutional social media accounts of much more limited reach. While only two sites reported cost data for advertisement on social media, these costs ($125-$150) were only a fraction of each site’s $8,000 CC/PD budget. Given the reach and affordability of paid social media campaigns, we recommend EFIC studies develop and implement social media campaign utilizing paid advertisements for PD.

Determining the reach of social media campaigns is complicated by the variable metrics of engagement employed by different social media sites and their different meanings. We recommend EFIC studies require sites to report standardized engagement metrics aligned with the goals of their social media campaign and articulate plans for evaluating their reach and effectiveness. For example, for PD Reach on Facebook and Impressions on X (formerly Twitter) would be appropriate metrics.

The lack of standardization of reporting of social media is consistent with a general lack of standardization in how IRBs consider both CC/PD plans and result reporting. In qualitative interviews of IRB members who review CC/PD plans, no IRB used quantifiable metrics to assess plan adequacy.³⁰ Similarly, in their recollection, IRBs’ review of completed CC/PD activities had never resulted in rejection or modification of an EFIC study.³⁰ To make the evaluation of CC/PD campaigns more rigorous, we recommend studies work with IRBs to identify key metrics to quantify the success of CC/PD campaigns before they begin.

Our study demonstrated the relatively limited reach of community meetings for CC in the PediDOSE study. Meetings of community interest groups had particularly small audiences. Other studies of CC/PD strategies have shown either similar^28,31 or greater but variable³² standing meeting attendance. Sites’ use of social media for CC had greater reach, yet it remains unclear if social media can be an effective means for engagement in meaningful discussion. Smaller meetings have the benefit of allowing each individual participant to express their opinion. Additionally, feedback content in our study varied across meeting types. Sites should strive for diversity in both meeting participants and perspectives. We recommend that sites consider using social media and community meetings for CC, acknowledging that while social media provides significant reach, its effectiveness in facilitating bidirectional communication has not been fully established. We also recommend CC meetings be conducted with all key stakeholder groups, including clinicians, hospital advisory groups, and community groups.

Specific to PediDOSE was the importance of EMS professional engagement. Meetings with frontline EMS clinicians specifically were both well attended and rich in feedback. EMS clinicians are supportive of research in the prehospital setting but often have concerns about their role and liability in these studies.⁶ Many have substantial concerns about enrolling in studies without patient consent.³³ We recommend that EFIC studies in the prehospital setting engage EMS clinicians directly as part of the CC/PD process.

Engagement with minoritized communities during CC/PD activities is important for multiple reasons, including disparate acceptance of EFIC studies in differing populations,³⁴ frequent enrollment of minoritized patients in EFIC studies,³⁵ and increasing emphasis on reducing disparities in and through pediatric research.³⁶ Reliance on the diversity of advisory groups for CC is problematic, as these groups often do not adequately represent the patient populations they serve.³⁷ In the PediDOSE study all sites distributed study materials in Spanish, However, only a subset of sites explicitly targeted Spanish speakers for either interviews or surveys, and a number of these sites were unable to achieve their modest engagement goals. Only three sites reported attempting to conduct community meetings dedicated to minoritized groups, and one of these was ultimately unable to do so. Based on early PediDOSE enrollment data, both interview participants and survey respondents were less likely to be minoritized or Hispanic than either community members or study enrollees.²⁴ This reinforces the need for improved efforts to engage these populations in CC/PD activities.

The CC/PD plan template stated that PD materials translated “as per local community requirements.” However, only one site created summary materials in a language other than English or Spanish. Prior to study enrollment, it was difficult to quantify the extent to which sites had specific language communities large enough to justify translation of study summary materials in non-English languages, since individuals may be multilingual. Preferred language data from actual enrollments in prior EFIC studies in a community may remove ambiguity for sites.

Based on the above, we recommend that all EFIC sites develop explicit plans for engagement with local minoritized communities in CC/PD. These plans should ideally use past enrollment data to help identify languages and minoritized communities for outreach. Studies should establish community language preference thresholds that warrant translation of CC/PD information. When possible, study teams should leverage existing institutional-community partnerships³⁸ to facilitate outreach.

For EFIC studies involving children, the FDA states CC/PD activities should incorporate feedback from children who are old enough to understand the proposed research. However this was not reported on by PediDOSE sites. This is consistent with previous practices as to our knowledge there are no published descriptions of engagement in children in the CC/PD process of pediatric EFIC studies. Given the ethical mandate for child assent in pediatric research³⁹ and published FDA guidance, this is an ongoing challenge for investigators to meaningfully achieve. Future pediatric EFIC studies should prioritize the creation of CC activities that specifically involve minors who are old enough to provide assent, potentially through direct interviews, focus groups, and social media geared towards this age group.

Based on our experience we can highlight the following pitfalls to avoid for sites planning CC/PD activities. When selecting groups for CC meetings, sites should inquire about typical meeting attendance and frequency of meeting cancellations. Community interest group meetings, while offering unique insights and lived experience, were at high risk of being unsuccessful in our study. Sites also need to be aware of institutional policies regarding communication with patients. Multiple sites in our study planned on sending mailers to families of patients with epilepsy but were not ultimately permitted to do so. Similarly, sites should confirm their ability to access institutional resources, such as presentation slots at grand rounds or posts on institutional social media accounts, before submitting CC/PD plans. Finally, sites should be proactive and engage institutional resources to help reach minoritized communities.

The PediDOSE CC/PD campaigns were run by individual study sites based on guidance from the main study team. However, this is not the only model for CC/PD. The Pedi-PART (Pediatric Prehospital Airway Resuscitation Trial) trial ran a centralized social media campaign as part of their CC/PD activities, coordinated by the main study site.⁴⁰ They ran advertisements in all study markets on Facebook and Instagram. They received 23.3 million social media advertisement impressions, reaching 3.4 million unique users and resulting in 13,873 engagements.⁴⁰ By running a social media campaign centrally, this study was able to leverage expertise in social media not necessarily available at all sites as well as economics of scale. Similar strategies have been used for CC/PD in adult trials.²⁰ Future EFIC trials need to either explicitly instruct sites in running social media campaigns or consider centralizing social media efforts to optimize PD through this medium.

Our study was conducted during the COVID-19 pandemic, which profoundly affected the frequency and format of community meetings. While the pandemic severely restricted the ability of individuals to meet in person, wider dissemination of virtual group meeting tools and the flexibility offered by virtual meetings may have reduced barriers to engagement. It remains to be seen whether moving forward, community meetings occur more frequently in-person or virtually, and how this affects the ability of studies to conduct CC activities.

LIMITATIONS

Our study had several limitations. As discussed above, the study was conducted during the COVID-19 pandemic, which severely restricted public gatherings. CC/PD strategies used by the PediDOSE study and levels of participation in these activities may therefore be unique to this period. We also relied on self-reports by sites to the sIRB regarding their CC/PD plans and outcomes. This may have affected our results in different ways. For example, sites may have set internal expectations of enrollment of minoritized patients in CC/PD activities but may not have reported these to the sIRB. Sites were also variable in the level of detail in their reporting of their plans and results, which limited our analysis.

CONCLUSION

Based on our experience as part of PediDOSE study we can make the following recommendations for future studies and sites engaging in CC/PD activities:

• All EFIC studies should engage in paid advertisements across social media platforms to maximize reach of PD. These may either be run centrally or locally with explicit guidance on where and how to advertise.

• EFIC studies should consider conducting hybrid CC campaigns utilizing both social media forums supported by advertising and more traditional community consultation meetings.

• Engagement metrics of interest for social medial campaigns should be identified before CC/PD occurs. Goals for these metrics should be set, and they should be tracked in a standardized fashion across study sites.

• CC meeting plans must identify and target all stakeholder groups for meetings a priori, ensuring a diversity of viewpoints are heard.

• All EFIC sites should develop & implement explicit plans for engagement with local minoritized communities in CC/PD. Standardized community language preference thresholds should be set for translation of CC/PD information. Sites should partner with institutional and community groups to maximize engagement of minoritized and language preference groups.

• Pediatric EFIC studies should develop CC plans that incorporate feedback from age-appropriate children through interviews, focus groups, and social media geared towards minors who are of the age of assent.

By following this guidance and avoiding the pitfalls detailed above, future EFIC study teams may best reach their study communities, ensure consulted communities accurately represent the opinions of study populations, and most positively impact the conduct and outcomes of their studies.

DATA AVAILABILITY –

Deidentified data available on request