Singapore's health outcomes after critical illness in kids: A study protocol exploring health outcomes of families 6 months after critical illness

Aim: To explore and understand the impact of paediatric intensive care unit (PICU) admission on longitudinal health outcomes, experiences and support needs of children and their parents in the first 6 months after PICU discharge and to examine the role of ethnicity. Design: This study uses a prospective, longitudinal design. Methods: The sample will include children ( N = 110) and at least one parent ( N = 110) admitted to the PICU


| INTRODUC TI ON
Advances in medical technologies have improved the survival rates of critically ill paediatric patients (Pollack et al., 2014). Every year in the United Kingdom and United States, approximately 20,000 and 230,000 children, respectively, require admission to the paediatric intensive care unit (PICU) (Garber et al., 2003;PICANet, 2017). Most children (95%) will be discharged alive (PICANet, 2017), but with this reduced mortality more children are discharged with increased morbidity (Pollack et al., 2014;Senna et al., 2020). Emotional distress and de-conditioned physical and mental impairment have become increasingly evident in paediatric critical care survivors (Shudy et al., 2006;Watson et al., 2018). In addition, the emotional and social impact of the PICU experience may extend to parents who experience the PICU first-hand by being present (Abela et al., 2020). As a result, the post-intensive care syndrome in paediatrics (PICS-p) framework was conceptualized in recognition of the unique experiences of children and their families after critical illness (Manning et al., 2018). The Indian and others (CMIO) and is therefore a suitable location for the study of these health outcomes. There are two quaternary PICUs providing medical care to critically ill children in Singapore.

| Children
Growing evidence shows that children are being discharged from the PICU with physical and cognitive impairment (Hordijk et al., 2020;Killien et al., 2019) and 30% of PICU child survivors show sustained emotional impact from their critical illness. These patients reported deterioration of their emotional well-being up to 1 year after discharge (Abela et al., 2020;Shudy et al., 2006), and PICU child survivors continue to exhibit negative behaviour and psychological responses in the first year after PICU discharge (Melnyk et al., 2004).
The duration of impairment varies, with some reporting lasting impact (Pollack et al., 2021).
A qualitative systematic review examining the psychosocial impact of a PICU admission revealed that children experienced confusion, relied on their parents' narratives, and were focused on their former selves and the meaning of normality (Manning et al., 2014). These psychological and social impacts were confirmed by two systematic reviews examining quantitative data (Abela et al., 2020;Procter et al., 2021). The sustained negative emotional and social effects on well-being may potentially disrupt childhood as a dynamic state of continuous growth and development after the critical illness.

| Parents
A systematic review by Abela et al. (2020) showed that parents were affected emotionally and socially to a large extent after their child's critical illness. More than half of the parents reported anxiety (60%), up to 50% were depressed and 30% presented with sub-clinical and clinical symptoms of post-traumatic stress disorder (PTSD). Similarly, qualitative studies which examined the experiences of families of PICU survivors found that these families had distressing memories, increased anxiety and were over-protective of the children (Foster et al., 2017). Parenting roles evolved to cope with the behavioural changes of the PICU survivors. Caring for a PICU survivor places considerable financial and time commitment stresses on families (Terp & Sjostrom-Strand, 2017). Alterations in parental role and parenting style after paediatric critical illness were reported as coping responses towards the care of a PICU survivor (Dahav & Sjostrom-Strand, 2018). There is currently limited knowledge about the bidirectional interaction between PICU survivors and their parents' health after a critical illness.

| Ethnicity
Ethnicity is considered a key determinant of health outcomes because of its association with social and economic opportunities and resources. This relationship was well documented in the infant and adult population but is less known in the paediatric population (Mehta et al., 2013). A study examining 12,000 critically ill oncology paediatric patients found that after controlling for severity and cancer type, ethnicity influenced PICU mortality (Leimanis Laurens et al., 2020). Literature on the impact of ethnicity on health outcomes in paediatric critical care is lacking. Most studies do not examine the difference in health outcomes according to ethnicity. The few available studies which examined outcomes by ethnicity were mainly conducted in the United States (Balluffi et al., 2004;Melnyk et al., 2004) (Balluffi et al., 2004;Bronner et al., 2010).
Minority groups also reported higher social care and healthcare utilization. In Singapore, differences in parental satisfaction were noted in the PICU population, with reporting Malay families less satisfied (Sng et al., 2017). Another study showed that Malay families of children with complex health conditions reported higher levels of stress (Chan et al., 2019).
Currently, there is still a gap in the comprehensive understanding of the long-term health outcomes of the PICU survivors and their families following PICU discharge. The association between ethnicity and health outcomes after paediatric critical illness is limited. This proposed study, exploring the trajectory of health outcomes in children and families after PICU discharge in Singapore will contribute to the understanding of the PICS-p in a multiracial context.

| Research question
What are the longitudinal health outcomes, experiences, and support needs of children and their parents up to 6 months after PICU discharge?

| Aim
The aim of this study is to explore and understand the impact of PICU admission on longitudinal health outcomes, experiences, and support needs of children and their parents in the first 6 months after PICU discharge, and to examine the role of ethnicity.
The objectives are the following: 1. To describe the physical, cognitive, emotional, and social health outcomes of children and the trajectory in the first 6 months after PICU discharge.
2. To describe parental emotional and social health outcomes in the first 6 months after PICU discharge.
3. To explore the experiences and support needs of parents after PICU discharge. 4. To understand the role of ethnicity in children and parents' health outcomes after PICU discharge.

| Methodology
A longitudinal, single centre study design will be adopted. This study will have two work streams: (1) a quantitative study exploring prospectively the longitudinal health outcomes of 110 children and 110 parents and (2) an embedded qualitative study exploring the experiences and support needs via semi-structured interviews with at least 12 parents at 1 and 6 months after PICU discharge.

| Theoretical framework
The PICS-p framework ( Figure 1) conceptually organizes health outcomes of children, parents and siblings of a child survivor after PICU discharge (Manning et al., 2018). The framework views childhood as a dynamic state where the critical illness occurs at a time of tremendous growth and development. The framework also accounts for the potential impact of baseline developmental disabilities on quality of life (QOL) after PICU recovery. The PICS-p framework includes physical, cognitive, emotional and social health outcomes of children with various recovery trajectories after PICU discharge.
Unlike the PICS framework used in adults (Needham et al., 2012), the PICS-p includes the family as an interdependent unit. In this framework, parental and siblings' emotional and social health outcomes are integrated within the pathway of the PICU survivor. This integration aids the recognition of potential bidirectional influence of health outcomes, where the PICU survivors are heavily dependent on their family.

| Setting
We will recruit participants from one PICU in a tertiary paediatric hospital, KK Women's and Children's Hospital in Singapore, to capture variation in case mix and patient demographic. The study site is an 800-bed hospital specializing in women's and child (or children's) health, with 400 beds dedicated to paediatric services including a 16 bedded multidisciplinary (medical, surgical and cardiac) quaternary PICU facility. The PICU admits up to 600 patients a year. Following hospital discharge, patients will receive medical follow-up from the respective subspecialist depending on their admission diagnosis to the PICU. Families who require social assistance will continue to receive support from the medical social workers. The suitability of this hospital for the purpose of this study lies in its ability to access a large number of families of children who require critical care, with different socio-demographics and different types of diagnosis. Therefore, this hospital is considered a suitable study site for a representative sample of families of children requiring paediatric intensive care in Singapore.

| Eligibility criteria
Participants for this study include: (1) PICU child survivors and (2) parents/legal guardians. (c) received PICU therapies for at least one organ dysfunction; (d) at least one parent/legal guardian living with the potential study child participant.

| Participants
Quantitative study: We anticipate enrolling 110 children and their parents over an 18-month period. Based on the 20% attrition rate reported in previous PICU studies, we conservatively estimate a 40% attrition over 6 months (Balluffi et al., 2004;Keenan et al., 2006;Rodriguez-Rey et al., 2018). Thus, we expect having complete follow-up (i.e. 6-month outcomes) for 66 participants.
A sample size of 64 will be required for this study, calculated based on the following parameters: mean difference of the Paediatric Quality of Life Inventory (Varni et al., 2001) of 3.0 with estimated standard deviation of paired differences of 9.5 between PICU admission and 6 months post PICU discharge, 80% power and a two-sided alpha of 5%. Paired sample t-test was used to compute sample size. After accounting for a 40% attrition rate over 6 months, we conclude that we will need to recruit 110 patients.
Qualitative study: A stratified purposive sample of at least 12 parents (at least 24 participants if both mother and father agree to participate) will be enrolled into the qualitative interviews (Polit & Beck, 2020). This sample size will capture diverse perspective around experiences from PICU admission up to 6 months after discharge, support needs, care-giving roles and responsibilities (Morse, 2000).
Recruitment will be stratified according to race (CMIO). We will interview at least 12 parents at both 1 and 6 months after PICU discharge.

| Recruitment
We will recruit children and parents during the PICU admission. The study team will screen the medical records of patients admitted to PICU daily to identify eligible patients. We will review clinical information of each patient's record to facilitate identification of eligible participants. We will share information about the study in the participant information sheet. The CONSORT flow diagram will illustrate the recruitment of this study (Moher et al., 2010).
At 1 month post PICU discharge (T2), with parental permission, the study team will invite children (aged 5 years and above) who are competent and able to complete the assessments to take part in the study. The parents will decide if they will ask their child to participate. If the child chooses not to participate, the study team will cease contact with the child. However, parents will still be able to continue their own participation irrespective of their child's involvement. Participants will receive SGD$50 in staggered amounts across the five time points of data collection (T3: SGD$10, T4: SGD$10 and T5: SGD$30).

| Data collection
2.5.1 | Quantitative study Data collection will include baseline and clinical characteristics such as demographics (age, gender, education level, ethnicity and religion) and socioeconomic data, pre-PICU health status and current illness data (PICU admission and discharge diagnoses, co-morbidities, surgical and invasive procedures performed; duration of mechanical ventilation, continuous renal replacement therapy, extracorporeal membrane oxygenation, inotropic support, sedative medications, PICU and hospital length of stay). We will extract data from the electronic medical records by research staff who will enter the data directly into a secure web application (JISC Online Survey) hosted by the University of Plymouth.
Each child (or proxy) and their parent will provide data on health outcomes prospectively over the first 6 months after PICU discharge.
The focus and selection of study measures were informed by the PICS-p framework (Manning et al., 2018). We will collect outcomes at five time points; baseline status (at consent); at PICU discharge; 1, 3 and 6 months after PICU discharge (Figure 2). The outcome measures and time points are presented in Table 1; detailed explanation of the individual measurement instrument is presented in Supporting Information.

| Qualitative study
Qualitative data will be collected virtually with parents of the PICU child survivors. Individual or paired parents' semi-structured interviews lasting approximately 30-60 min will be conducted with parents/legal guardians at 1 (T1) and 6 (T4) months after PICU discharge. An interview guide will be used to explore the experiences and support needs of parents. The interview guide is developed based on semi-structured qualitative research methods (Dicicco-Bloom & Crabtree, 2006), recommendations by research team members and paediatric critical care literatures. Reflective field notes captured during the interviews will be used to highlight gestures and reactions by unveiling important elements such as emotions experienced by participants (Saldana, 2016). Interviews will be conducted in English, audio-recorded and transcribed verbatim.

| Data analysis
Descriptive statistics will be presented for demographic and clinical characteristics. All PICU child survivors' and parents' related outcome measures will be calculated, including means, standard deviation, median and interquartile range for continuous variables.
Frequency counts and percentages will be reported for categorical variables. Normality, outliers and systemic missing data will be assessed. Data transformation will be performed as necessary (Polit & Beck, 2020).

| Quantitative data analysis
Analysis related to specific objectives includes: Objective 1: To describe physical, cognitive, emotional, and social health outcomes of children in the first 6 months after PICU F I G U R E 2 Overview of SHACK workflow in the first 6 months after paediatric intensive care discharge. PICU, paediatric intensive care unit  x indicate the time-points for questionnaires. discharge. Objective 2: To describe parental emotional, social health outcomes and coping in the first 6 months after their child's PICU discharge.
Descriptive statistics such as mean, standard deviation, median and interquartile range will be used to describe participant-reported outcomes in the first 6 months after PICU discharge.
Mixed-model ANOVA for repeated measurements will be used to examine the trajectory in children and their parents across baseline to 6 months after PICU discharge. Differences in the level of QOL, functional status and cognition across five time points will be examined in children. Depression, anxiety, PTSD, family impact and coping strategies will be analysed for parental participants. The assumption for the mixed model will be addressed by analysing the sample for normality, homogeneity of variance, independence, and sphericity. A post hoc test will be used to examine the differences in variables of significant difference with p value < .05.
Objective 4: To understand the role of ethnicity in children's and parents' health outcomes after PICU discharge.
Mixed-model ANOVA for repeated measures will be used to examine the differences in different outcomes and race.

| Qualitative data analysis
Audiotape recordings for each interview session will be transcribed verbatim and double checked for accurate transcription. Transcripts will be imported into the qualitative software package using the QSR International's NVivo 12 software for sorting, coding and categorizing of the data.
The adapted five-stage framework analysis process will be used for qualitative data analysis to achieve Objective 3: To explore the experiences and support needs and coping of parents after their children's PICU discharge. The five stages are: (1) familiarization of the full transcript through immersion (2) development of a theoretical framework by identifying recurrent and important themes; (3) indexing and pilot charting; (4) summarizing data in an analytical framework; and (5) synthesizing data by mapping and interpreting (Ward et al., 2013). Stage 5 will allow for the data to be compared across time-points (1-3 and 3-6 months).

| Ethical considerations
The purpose of the study will be explained to prospective eligible participants as well as the level of involvement required of them for the research. Participants will be informed that the study team will only be able to destroy information that has yet to be anonymized.
We will assign a unique identification number in lieu of names on the questionnaire to maintain confidentiality. All participants will be asked to sign a written consent, with the signed consent form photocopied for the parents' record. Parents will provide consent for children unable to provide consent/assent at time of enrolment.
The patient information sheet and consent, questionnaires and reimbursement schedule were reviewed by three parents of PICU survivors. An incentive of SGD$50 will be given to parents after completing questionnaires across five time points.

| Ethical review
The study protocol was reviewed by the SingHealth Centralised

| Validity and reliability/rigour
The study design creates the opportunity to triangulate and integrate quantitative and qualitative data (Onwuegbuzie & Johnson, 2006).
To promote rigour in the project, we will respect the application of principles specific to each quantitative and qualitative work project.
For the quantitative component, we adopted the principles of validity, reliability and generalizability (Polit & Beck, 2020). Principles of credibility, dependability, confirmability and transferability will be observed for the qualitative component (Lincoln & Guba, 1985).

| Quantitative validity
Outcome measures used in this study were selected for their validity, reliability, ease of use, availability in electronic versions and previous use with PICU child survivors and their families. The selection was further informed by the PICS-p framework and current literature (Manning et al., 2018).

| Qualitative validity
Credibility will be ensured by following the steps in framework analysis and coding the experiences of parents of PICU survivors (Saldana, 2016). Findings will be shared with researchers experienced in qualitative methods with PICU families to refine the final results. Dependability will be enhanced by having two researchers to code all interview transcripts-this is to ensure consistency in coding data across diverse participant groups (Graneheim & Lundman, 2004). An audit trail will be maintained to facilitate evaluation of the consistency of the research process and demonstrate the process of conceptualization (Polit & Beck, 2020). The audit trail will include interview transcripts, coding records, and field notes. Confirmability will be enhanced through consensus during thematic analysisparticipants' narratives will be used to report the study findings.
Participant demographic, description of context, and the audit trail of data source will strengthen the transferability of the study findings (Polit & Beck, 2020).

| DISCUSS ION
Research in long-term health outcomes is limited for children surviving a critical illness. Previous studies focused on specific patient groups (e.g., traumatic brain injuries, congenital heart disease or acute respiratory distress syndrome) using cross-sectional research design, measuring various outcomes (functional health, psychological wellbeing, health related QOL) and often do not account for the child's pre-PICU health status. Understanding of the trajectories and longitudinal health outcomes of children and families after PICU discharge is severely limited. To improve outcomes for children and their families, multinational stakeholders from across 6 continents through consensus determined a set of PICU core outcomes (Fink et al., 2020). The use of longitudinal design and choice of outcome measures in the current study are aligned with the OCEANIC study (Manning et al., 2020). Thereby, our study attempts to create a community of research practice among colleagues in the PICU worldwide and align outcomes measures that enable us to compare robust data. In addition, this study will enhance understanding of the relationship between ethnicity and health outcomes. Collectively, these results can be used to inform practices of health care and other professionals, future research and policy development targeting specific groups.
Dissemination of findings will include presentations at international critical care conferences and the research team will actively engage in dissemination of findings to targeted practitioners, academics, decision-makers and the general public.
Through the Paediatric Acute and Critical Care Medicine Asian Network (PACCMAN), we will have opportunities to disseminate the findings and replicate/scale up this project across Asian countries, reaching an audience much broader than the academic and local communities. This will ensure that information about successful and unsuccessful features of our project is shared.

| Limitations
First, the possibility exists that, because of the child's critical illness, parents are less likely to consent to the study. Therefore, a member from the care team will assess the parents' well-being before inviting them to participate in this study. They will be given time to consider and be reassured that participation is voluntary and that the care of the child will not be affected by their decision to participate.
Second, fathers may be under-represented as shown in previous PICU-related research. This will affect the understanding of the impact of paediatric critical illness in fathers. To improve fathers' participation rate, the study team members will conduct invitations in the presence of both parents, keeping fathers involved throughout the conversation.
Third, due to the nature of the care burden after critical illness, participants are likely to drop out from a longitudinal study.
Participants will be briefed and given an information sheet about their involvement in this study. They will be given a monthly update on the recruitment rates and a pictorial road map to reflect their journey in the study. Based on reported attrition rates of 20% in longitudinal PICU studies, we conservatively estimate a 40% attrition rate.

| CON CLUS ION
Our research outcomes will help improve the delivery of current healthcare services, short term policy, and healthcare innovations for PICU survivors and their families. The results from this study will continue to build a sound foundation for ongoing research and policy development to support the complex needs of the PICU child survivors and their families. Lastly, the results and template of our project are transferable and scalable to other countries across Asia and beyond, with the goal of building understanding in order to improve the health outcomes, experiences and lives of PICU child survivors and their families.

ACK N OWLED G EM ENT
This research is supported though the KK Women's and

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/jan.14911.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.