Development of a non‐pharmacologic delirium management bundle in paediatric intensive care units

Abstract Background Non‐pharmacologic interventions might be effective to reduce the incidence of delirium in pediatric intensive care units (PICU). Aim To explore expert opinions and generate informed consensus decisions regarding the content of a non‐pharmacologic delirium bundle to manage delirium in PICU patients. Study design A two‐round online Delphi study was conducted from February to April 2021. PICU experts (nurses, physicians, researchers, physical therapists, play specialists, and occupational therapists) located in Europe, North America, South America, Asia, and Australia participated. Results We developed a questionnaire based on the outcomes of a comprehensive literature search in the domains: 1) cognition support; 2) sleep support; and 3) physical activity support. Under these domains, we listed 11 strategies to promote support with 61 interventions. Participants rated the feasibility of each intervention on a 9‐point Likert scale (ranging from 1 strongly disagree to 9 strongly agree). A disagreement index and panel median were calculated to determine the level of agreement among experts. In the second round, participants reassessed the revised statements and ranked the interventions in each domain in order of importance for age groups: 0–2, 3–5, and 6–18 years of age. During the first Delphi round, 53 of 74 (72%) questionnaires were completed, and in the second round 45 of 74 (61%) were completed. Five of the highest ranked interventions across the age groups were: 1) developing a daily routine, 2) adjusting light exposure according to the time of day, 3) scheduling time for sleep, 4) providing eyeglasses and hearing aids if appropriate, 5) encouraging parental presence. Conclusions Based on expert consensus, we developed an age‐specific non‐pharmacologic delirium bundle of interventions to manage delirium in PICU patients. Relevance to Clinical Practice An age‐specific Non‐Pharmacological Delirium bundle is now ready to be tested in the PICU and will hopefully reduce pediatric delirium.

Five of the highest ranked interventions across the age groups were: 1) developing a daily routine, 2) adjusting light exposure according to the time of day, 3) scheduling time for sleep, 4) providing eyeglasses and hearing aids if appropriate, 5) encouraging parental presence.
Conclusions: Based on expert consensus, we developed an age-specific nonpharmacologic delirium bundle of interventions to manage delirium in PICU patients.
Relevance to Clinical Practice: An age-specific Non-Pharmacological Delirium bundle is now ready to be tested in the PICU and will hopefully reduce pediatric delirium.

K E Y W O R D S
delirium, delphi study, management, non-pharmacological, paediatric intensive care unit 1 | BACKGROUND Critically ill children are likely to experience discomfort, pain, distress, withdrawal syndrome, and delirium in the paediatric intensive care unit (PICU). They often undergo unpleasant procedures such as insertion of intravenous lines, drains, catheters, suctioning when on mechanical ventilation, and other necessary, often recurrent, treatment procedures. Noise and light in the PICU environment add to the stressors experienced by critically ill children. 1,2 To reduce their suffering, healthcare providers regularly administer sedatives and analgesics to children, most commonly opioids and benzodiazepines. 3 At the same time, extensive use of benzodiazepines and opioids carries the risk of prolonged mechanical ventilation, prolonged PICU stay, iatrogenic withdrawal syndrome (IWS), paediatric delirium (PD), and delusional memories. [4][5][6][7] Sleep disturbance and PD are frequently treated with additional sedatives, which may lead to a vicious cycle that contributes to increased child morbidity and mortality. 8 Balancing adequate sedation, while avoiding over-and undersedation, is a PICU challenge. 9 Systematic reviews based on hospitalized adult non-ICU patients conclude that multi-component non-pharmacologic interventions reduce the occurrence of delirium compared with usual care. 10,11 However, studies in the adult intensive care unit (ICU) patients show that non-pharmacologic interventions might be efficacious in reducing the incidence and duration of delirium. [12][13][14] In the PICU setting, nonpharmacologic interventions seem promising. 1,15 A recent study described a significant reduction of PD after implementation of a nonpharmacologic bundle in children under 5 years and children after surgery after congenital heart disease surgery. 15 Therefore, we assume that non-pharmacologic interventions, such as promoting orientation, day-night rhythm, and avoiding overstimulation from noise and light, might reduce delirium in critically ill children as well while causing no harm.
The Medical Research Council defines an intervention as being complex if it contains several components. 16 Moreover, the council provides guidelines for the development and evaluation of complex interventions. 16 This framework has four key phases: development, feasibility and piloting, evaluation, and implementation. 16 The present Delphi study represents the development phase. We used the term "management" for both prevention and treatment of delirium, as our assumption is that the same interventions will reduce the incidence, relieve symptoms, and shorten the duration of PD.
What is known about the topic • Critically ill children admitted to the paediatric intensive care unit risk experiencing discomfort, pain, distress, withdrawal syndrome, and delirium.
• Non-pharmacological multi-component bundles for preventing delirium are effective in the adult intensive care unit settings.
• The Delphi method is a systematic, interactive method building on the opinions of a panel of experts within the field of a specific topic. After a number of rounds with questionnaires, consensus on the topic is reached.

What this paper adds
• Consensus of global experts on an age-specific nonpharmacological bundle to manage delirium in paediatric intensive care unit patients is presented.
• A parental and nurse-driven non-pharmacological delirium management bundle consisting of 1) developing a daily routine, 2) adjusting light exposure according to the time of day, 3) scheduling time for sleep, 4) providing eyeglasses and hearing aids if appropriate, 5) encouraging parental presence.
• The first step in development and evaluation of a complex intervention to manage paediatric delirium in the PICU is taken.

| AIMS
We aimed to explore expert opinions and generate informed decisions regarding the content of a "Non-pharmacologic Delirium management Bundle" in PICU patients (NDB-PICU).

| METHODS
We performed a two-round online Delphi study, to achieve agreement on the content of the NDB-PICU among a group of international and interprofessional PICU experts. The strength of a Delphi study is its ability to rapidly obtain consensus opinions from an expert panel based on the assumption that group opinion is more valid than individual opinions on an issue that has no collective comprehension. 17 The online Delphi format allowed us to reach a dispersed and varied group of international experts without compromising anonymity. We expected anonymity would minimize dominating opinions during consensus formation. We used a modified Delphi method, replacing the standard first round presenting the initial open-ended questions and focus group discussions 17 with a comprehensive literature search of non-pharmacologic interventions in both adult and pediatric ICU patients [17][18][19] (see Table S1).

| The questionnaire
Based on the literature search we devised a questionnaire divided into three domains: 1) cognition support, 2) sleep support, and 3) physical activity support. In the three domains, we identified 11 strategies (for example, to promote a structure for the day, reducing noise, and dimming lights). Finally, the strategies were broken down into 61 specific interventions that could be initiated always in accordance with the medical condition and the developmental stage of the child for example, provide bright light during the daytime (See supplementary material, Table S2). To provide an overview of important age-specific items, we subdivided the "child" category in the ranking session into three age groups defined by the author group: 1) 0-2 years (preverbal, neonate to toddler), 2) 3-5 years (preschool age), and 3) 6-18 years (school age). The final questionnaire was tested by the research group to ensure that the wording was understandable, the questions were relevant, and checking that each item contained only one question, and the statements were mutually exclusive.

| Participants
Purposive sampling was used based on predetermined criteria to recruit experts representing clinical PICU practice. In this Delphi study, an expert was a PICU nurse, physician, researcher, physical therapist, play specialist, or occupational therapist. The inclusion criteria were both knowledge and practical experience of paediatric delirium as well as capacity and willingness to contribute. 17 Applying a snowball sampling method, the members of the research group sent e-mails to known colleagues with content knowledge, describing the aim of the Delphi study and asking them to participate and nominate other experts. 17 We sent initial e-mails to colleagues from North America (n = 5), South America (n = 1), Europe (n = 19), Australia (n = 4), and Asia (n = 3) to identify potential participants. The snowball sampling method resulted in 74 potential participants. They were informed about the aim of the study and expected time investment.
We planned to limit the study to two Delphi rounds to prevent a low response rate, and ultimately to reduce attrition bias. 20 We aimed to include at least 20 participants, which has been recommended to produce stable results. 21  Each participant received a personal invitation by e-mail with a link to the questionnaire. In each of the two rounds, the participants were sent an information letter explaining the aim and content of that specific round, a consent form, the estimated time investment, and a deadline for completion. To optimize the response rate, we sent the participants a maximum of three reminders per round. Responses were anonymous both to the panel and the research group. A high response rate was important for the content validity of the results. 18 The participants were asked to rate the feasibility in daily practice of the interventions of the proposed NDB-PICU on a 9 point Likert scale (ranging from "1 strongly disagree" to "9 strongly agree"), and to provide a rationale for their response and including suggestions for improving the description. In addition, they were invited to comment, in their own words, how the suggested interventions could be revised.

| Data collection
In the second round, all 74 participants received a general summary of the first round that presented the overall group results of the feasibility rating. The participants were asked to reconsider or confirm their opinion based on this new information and rank the approved, modified, and newly added interventions that were not part of the first Delphi round for their importance for delirium management, with a rank of 1 being most important. Involving Human Subjects Act. 23 As all data were anonymous, permission to store data was not required.

| Data analysis
For each intervention, we calculated a median rating and a disagreement index (DI) to determine the level of agreement. Median ratings between 7 and 9 were defined as relevant, 4 to 6 as somewhat relevant, and 1 to 3 as irrelevant. To calculate the DI, we divided the interpercentile range (IPR) (IPR 0.3-0.7) into the IPR adjusted for symmetry (IPRAS) 24 (See supplementary material, Table S3). The cutoff for DI was defined as <0. The participants' written comments on the interventions were analysed by two of the researchers. The comments from the participants about the interventions were rephrased into new statements and included in the final round. The participants' overall ranking of the importance for each intervention was calculated as follows: the number of participants who ranked the interventions as "1" multiplied by the total number of interventions, the number who ranked the intervention as "2" multiplied by the number of interventions À1 and so on. The values for all rankings were summed for each intervention.
The highest-ranked intervention for each strategy and age group was included in the final recommendation of the NDB-PICU.

| RESULTS
We enrolled 74 participants to participate in the Delphi study. Participants were predominately from Europe and North America were represented in both rounds, but most participants were nurses and physicians who typically had more than 11 years of PICU experience (Table 1).
For sixty-one interventions in round 1, agreement among the participants was indicated with DI <0.2. Ten interventions with DI ≥0.2 were excluded. Eleven interventions needed to be modified and clarified based on input, and two new interventions were added: "staff should keep identification badge visible" and "use signs on patient door/bed to communicate that the child is sleeping or it is nap time." Forty interventions did not require any modification. All 53 interventions indicated high agreement in round 2. Six interventions were modified and clarified, no new interventions were added, and two interventions were collapsed into one based on expert feedback (Table 2).
After round 2, the expert group agreed on the feasibility of three domains: cognitive support, sleep support, and physical activity support, with 11 strategies covering 52 interventions ( Table 2).
The highest-ranked interventions for the 11 strategies and age groups, are listed in Table 3 and constitute an age-appropriate NDB-PICU. Five of the 11 interventions were similar among the age groups: 1) developing daily structure, 2) adjusting light exposure according to the time of day, 3) scheduling time for sleep, 4) providing eyeglasses and hearing aids if appropriate, 5) encouraging parents to be present.
Furthermore, four interventions were shared by the 0-2 year age group and the 3-5 year age group, and two interventions were shared by the 3-5 year age group and the 6-18 year age group. The two younger age groups each had two specific interventions, and the 6-18 year age group had four (Table 3).
Overall, interventions with parent engagement were highly rated.
Five interventions from each age group included parent collaboration, such as consulting parents about their child's daily structure; scheduling time for sleep; music preferences; bringing familiar objects from home such as comforting sleep objects; and overall encouragement of the parents to be present (see Tables S5, S6, and S7).

Strategies for improving visual or hearing impairment
Ensure that the child uses eyeglasses if appropriate whenever awake and ensure that the glasses are clean. Read aloud or tell a story to the child 9 (0.13) Encourage the parents to read aloud or tell a story to the child -9 (0.00) Provide sleep objects 9 (0.13) Provide sleep objects such as teddy bear, sleeping pillow or cuddle cloth -9 (0.00) Sing for or with the child 8 (0.13) Encourage the parents to sing for or with the child The participants responded that they found it difficult to rank the interventions because of the wide age span within the age groups, particularly in the oldest age group. Also, they commented that they found it difficult to rank equally important interventions (e.g., eyeglasses and hearing aids), and consequently these were combined. Furthermore, some participants found some interventions to be context-specific, and although important, not feasible in their own setting (for example, one expert commented, "I agree with this opinion but quite often the structure and housing conditions of many PICUs do not allow for it"). Interventions with the highest rank from all the strategies.

| DISCUSSION
An international multidisciplinary panel of PICU experts reached consensus on the content of a non-pharmacologic delirium management bundle (NDB-PICU) for three different age groups using a Delphi method. We found that the five interventions general for all age groups preferred by PICU experts to prevent PD were: 1) developing ties. 25 Interventions that support cognition account for more than half of the NDB-PICU. To support these interventions, parents could play an important role in being present, playing, reading, and orienting the child. Encouraging family members to be present has been described previously and may decrease delirium rates. [27][28][29] Interventions such as informing the families of delirium, displaying a daily schedule at the bedside, and letting parents bring familiar items from home such as blankets, pictures, and age-appropriate toys have been positively evaluated. 27,29 Parents feel more engaged and relieved when they can care for their child through such interventions. 27 Although parents might be stressed, we know from adult literature that it looks promising to have a family member delivering a non-pharmacologic intervention to reduce delirium in critically ill adults. 30 To promote sleep in the adult ICU, the use of eye masks and/or earplugs has been tested and significantly decreases the rate of delirium. 31 For children, suggestions in a family-centered toolkit have been made to provide eye masks, dim the light at night, and provide headphones to reduce noise. 27 In our Delphi study, such interventions were rejected. Rather, interventions to modify the environment in the room by avoiding loud talk and dimming lights with curtains or blinds were prioritized. Experts may have rejected the idea of using eye masks and earplugs because such devices may be unfamiliar or disruptive for children, who, unlike adults, are unaccustomed to them. The children may become scared because of the sudden lack of sight and something in their eyes and try to remove it. Sleep intervention support needs to be attuned to the child's age and sleep preferences, while parents should recommend day-night routines that reflect home routines, such as usual bedtime, objects to take to bed, such as cuddle toys, or listening to preferred music.
In the NDB-PICU, strategies to increase mobilization were highly ranked among the participants. The intervention "provide physical therapy daily" was modified to "provide physical therapy as considered appropriate", indicating that the participants found physical therapy to be important for most children, but that level and type should be based on developmental age and criticality. It has previously been established that PICU clinicians find early mobilization important. 32 Despite this, mobilization does not occur in 25% of critically ill children across Europe, for instance, because of tubes or catheters. 33 The most common activity seems to be children being held by family or nurse. 33  This study allowed us to develop the NDB-PICU, which needs to be tested for efficacy on delirium reduction in PICU patients. With step one of the MRC framework completed, the next phase of this complex intervention can be carried out, 16 namely the feasibility, pilot testing, and evaluation.