Evidence-informed consensus statements to guide COVID-19 patient visitation policies: results from a national stakeholder meeting

We created consensus statements from 36 evidence-informed experiences (i.e., impacts on patients, families, healthcare professionals, and PFCC) and 63 evidence-informed strategies (i.e., ways to improve restricted visitation) identified during a modified Delphi process²⁰ in the restricted visitation multiphase research program, and thematically analyzed these statements into 12 themes (Fig. 1). We emailed a meeting agenda with consensus statements to participants one week prior to the two half-day virtual meeting (held via Zoom videoconferencing [Zoom Video Communications, Inc., San Jose, CA, USA]).

Attendees included 45 individual stakeholders (patients, families, researchers, front-line critical care clinicians, decision-makers) who self-identified as being directly impacted by COVID-19 restricted visitation policies (Table 1). We purposively recruited participants from prior stages of the research program. This meeting received ethical approval from the University of Calgary’s Conjoint Health Research Ethics Board (REB20-0944). Participants provided informed consent prior to participating in the meeting. We reported this study in accordance with the Consolidated Criteria for Reporting Qualitative Research checklist (Electronic Supplementary Material [ESM] eTable 1).²¹

The two half-day meetings included the final two rounds of the Delphi consensus process, which are presented elsewhere²⁰ and two breakout sessions (described in ESM eAppendix 1). Participants were divided into six groups for day 1 and six different groups for day 2 (5 ± 1 participant per group). Where possible, groups included representatives from each key stakeholder group. Each group was assigned one research assistant trained in qualitative methods to facilitate the discussion and one notetaker to record observations. The aims of the meeting were for stakeholders to: 1) discuss and refine consensus statements created from the Delphi consensus process²⁰ (ESM eAppendix 2) (day 1) and 2) discuss implementation of the consensus statements that included evidence-informed strategies to improve restricted visitation (i.e., themes 5–12, Fig. 1). Although stakeholder experiences (i.e., themes 1–4, Fig. 1) are relevant to the context of visitation policies, they are unlikely to inform policies and thus were not included (ESM eAppendix 3).

Each breakout session was audio-recorded, transcribed verbatim, reviewed for accuracy of transcription, deidentified, and imported into NVivo-12 (QSR International, Melbourne, Australia) for data organization and analysis. Data analysis incorporated both deductive (i.e., guided by predetermined objectives to understand implementation of the consensus statements) and data-driven inductive approach to coding. Five researchers (K. K., N. J., L. H., K. M., E. S.) conducted qualitative, thematic content analysis on the transcribed recordings, independently and in duplicate.²² The same researchers met regularly to discuss emerging themes, resolve discrepancies, and develop a list of codes. Coding discrepancies were resolved through discussion and, if necessary, reviewed by a third researcher (K. F.). The researchers kept an audit trail to track the evolving analysis and used a reflexive journal to enhance the study’s methodological rigor. Four of the five researchers then systematically applied the final list of codes to all transcripts.

Analysis from the day 1 discussions identified that consensus statements could be improved in three areas: 1) clarity, 2) accessibility, and 3) feasibility. Electronic Supplementary Material eTable 2 includes the original (from the modified Delphi process) and refined (from the stakeholder meeting) consensus statements. Final prioritized and refined recommendation consensus statements (from the stakeholder meeting) are included in Table 2.

Participants discussed several issues regarding the clarity of the proposed consensus statements. Exemplar quotations and select statements are included in ESM eTable 3. For example, participants were concerned about the ambiguity of the term “family” used in some of the consensus statements. Several participants suggested that “family” be more explicitly defined because it may affect who is permitted to visit (quotation 1 [Q1]). Moreover, participants felt that “family” should be considered distinct from “visitors” and proposed the use of the phrase “essential care partners” instead (Q2). Others also noted the ambiguity in using the word “children” in the consensus statements. Participants suggested ways to clarify the use of the word “children” in the statements, such as “children who are under the age of 16” (Q3) and making exceptions for specific situations (Q4). Lastly, participants agreed that it is important for a statement that included visitation at the end of life to clearly state “while the patient is lucid and able to interact” (Q5), based on their experiences of family only being able to visit after the patient was intubated or confused and delirious (Q6) and not being able to visit while a patient was lucid and not intubated.

Several participants identified that some consensus statements referred to information/personnel/devices that were not accessible to all stakeholders or ICUs (e.g., technology, Wi-Fi, personnel such as an ICU liaison or accessible infection and prevention control expert) (Q7, Q8). Participants recommended addressing these accessibility gaps by amending statements to be more inclusive. For example, one statement referred to using a website to communicate policies. Participants suggested that this statement should include multiple modes of information (e.g., website, pamphlet, notes, in-person) to improve accessibility to the most up-to-date hospital visitation policies (Q9). Moreover, participants suggested amending recommendations to include the word “accessible” such as accessible education information on “infection prevention and control standards” (Q10), accessible supports to debrief among healthcare professionals, and accessible technology literacy support for families and healthcare professionals (Q11).

Participants reworded consensus statements to accurately reflect inequitable access experiences and to use more inclusive language (Q12). Communication barriers between healthcare workers and patients/families or between patients and families (i.e., language, cognitive difficulties, delirium, hearing impairment, translation issues) were added to existing statements regarding the impact of PFCC during restricted visitation (Q13). Geographic considerations (i.e., quarantine for visitors), socioeconomic considerations (i.e., lack of portable devices, Wi-Fi, technical literacy), and cultural and spiritual considerations (i.e., end-of-life practices, ceremonies, familial relationships, last rites) were incorporated as inclusive considerations to consensus statements (Q14).

Time, physical space, coordination efforts, jurisdictional differences, and budgetary considerations were added to make statements more universally feasible since “depending on which hospital you’re at, the policies were different” (Q15). For example, participants agreed with one statement that policy changes should be communicated during regular working hours but added extra context that all staff should be informed about the change before the media (Q16). Participants also agreed that a designated “greeter” would be helpful, though not all agreed it would be feasible (Q17). One statement included designated outdoor spaces for physician-family interaction, which was not deemed feasible in the winter months or with increased healthcare professional workload (Q18). Some participants suggested wording to reflect variable COVID-19 experiences, patient volume in different geographical places, and evolving changes (Q19, Q20) to make statements more generalizable.

Participant discussions of implementation centered on the facilitators and barriers of consensus statements and how the consensus statements should be measured and monitored (Fig. 2).

Participants discussed several barriers to the implementation of consensus statements, many of which related to the dynamic pandemic circumstances. Exemplar quotations are listed in ESM eTable 4. These included consensus statement-related factors such as the number of (and length of) consensus statements (i.e., difficult to implement or communicate) (Q21) or their perceived lack of flexibility that would prevent them from being adapted to local contexts (Q22). Consensus statement-related factors also included external barriers such as changes to public health orders that affect the hospital visitation policy (Q23, Q24) and perceived lack of partnership between the government and hospital staff (Q25). Participants discussed several barriers at the individual level, which included individual values (i.e., healthcare professionals that do not follow the process), capacity (i.e., available time to check changes/e-mail), and emotions (e.g., fear, exhaustion). Participants shared their experience with “change fatigue”, which may affect the implementation of consensus statements (Q26). Participants also expressed that “The staff [are] tired, the administrators are tired because they have a hard time to find staff to take care of patients. So to ask them to implement something very complex, would probably just not work at this point” (Q27). One participant reflected that, at the beginning of the pandemic, fear for safety (i.e., of visitors, staff, and community) “was the motivating factor behind us putting really strict family visitation in” (Q28). Though, with the vast majority of the staff being vaccinated (i.e., facilitator), this fear has dissipated to some degree compared with at the beginning of the pandemic (Q29).

Participants discussed the factors that should be prioritized during strategy implementation. These were: 1) structure of the consensus statements (e.g., content, format), 2) process of implementing the consensus statements, and 3) who to engage in implementation. Participants reiterated that the consensus statements should be structured as “guiding principles rather than strict rules” (Q30). They should be clear, direct, and concise (Q31) and should be adaptable to the local context (Q32) or based on the COVID-19 burden in the region (Q33). Participants liked the idea of these consensus statements being available in different formats, but cautioned that all formats should be version-controlled with the same effective dates to ensure consistency of messaging. Participants agreed that providing rationale for why these consensus statements are being implemented is important for stakeholder buy-in (Q34).

Given the number of consensus statements, participants felt the process for implementing these statements should either involve prioritization of strategies (i.e., those with significant impact up front) or a tiered approach (i.e., a group of statements targeting different stakeholders) (Q35). Several participants liked the option for stakeholders to provide feedback on the consensus statements, with clear (i.e., who the feedback is going to) and meaningful (i.e., how their feedback is used) feedback avenues (Q36). Some participants suggested dissemination may include informing decision-makers, ICU unit managers, and critical care societies about the consensus statements or publishing consensus statements in traditional peer-reviewed publications (Q37). Participants agreed that public awareness of the consensus statement is important but felt that, like most successful campaigns, they should engage an expert in public relations. All participants agreed that multiple stakeholders should be engaged, including patients, families, local patient or family advisory councils, government, healthcare professionals, and critical care societies.

Participants suggested several ways to measure the effect of the implementation of these consensus statements. This should include both quantitative measures (e.g., family satisfaction, surveys but not virtual/in-person visits) and qualitative measures (e.g., asking different stakeholders about their awareness and perceived impact of the consensus statements) (Q38). Participants agreed that evaluation will reveal “which parts of these policies and the way they are implemented are most meaningful and important to keep moving forward, and which might actually not be a good use of resources” (Q39).