Outcomes of Children Surviving Pediatric Acute Respiratory Distress Syndrome: ​​from the Second Pediatric Acute Lung Injury Consensus Conference

Abstract

Objective:

To summarize the evidence for the Pediatric Acute Lung Injury Consensus Conference-2 (PALICC-2) recommendations for assessment of outcomes among patients surviving pediatric acute respiratory distress syndrome (PARDS).

Data Sources:

MEDLINE (Ovid), Embase (Elsevier), and CINAHL Complete (EBSCOhost)

Study Selection:

We conducted a scoping review to identify studies evaluating outcomes following PARDS. We included studies of survivors of PARDS, acute respiratory failure with a high proportion of PARDS patients, or other critical illnesses if PARDS-specific outcomes could be extracted.

Data Extraction:

Title/abstract review, full text review, and data extraction using a standardized data collection form.

Data Synthesis:

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used to identify and summarize evidence and develop recommendations. Of 8,037 abstracts screened, we identified 20 articles for inclusion. Morbidity following PARDS was common and affected multiple domains of pulmonary and non-pulmonary function. There was insufficient evidence to generate any evidence-based recommendations. We generated eight good practice statements and five research statements. A panel of 52 experts discussed each proposed good practice statement and research statement, and the agreement rate was measured with an online voting process. Good practice statements describe the approach to clinical outcome assessment, assessment of pulmonary outcomes of children surviving PARDS, and assessment of non-pulmonary outcomes of children surviving PARDS including health-related quality of life and physical, neurocognitive, emotional, family, and social functioning. The five research statements relate to assessment of patient pre-illness status, use of post-discharge endpoints for clinical trials, the association between short-term and longer-term outcomes, the trajectory of recovery following PARDS, and practices to optimize follow-up.

Conclusions:

There is increasing evidence that children are at risk for impairments across a range of pulmonary and non-pulmonary health domains following hospitalization for PARDS. The results of this extensive scoping review and consensus conference involving experts in PARDS research, clinical care, and outcomes assessment provide guidance to clinicians and researchers on post-discharge follow-up to optimize the long-term health of patients surviving PARDS.

There is increasing awareness that survivors of pediatric critical illness are at risk for ongoing morbidities across a range of health domains [1, 2]. Pediatric critical care clinicians, researchers, and families recommend assessment of child overall health, health-related quality of life (HRQL), physical, emotional, and cognitive function, pain, communication, and long-term survival following pediatric intensive care unit (PICU) discharge [3]. It is unknown whether these are also the most important domains to assess among survivors of pediatric acute respiratory distress syndrome (PARDS), who may in particular have additional pulmonary morbidity related to their acute lung injury. Pediatric Acute Lung Injury Consensuses Conference (PALICC) 2015 provided guidelines for outcome assessment following PARDS based on limited pediatric evidence [4]; we aimed to update these guidelines by reviewing new evidence published in the last 10 years. In this article for the Second Pediatric Acute Lung Injury Consensus Conference (PALICC-2), we address Key Question #9 as outlined in the accompanying methodology manuscript [5]: What are the morbidity and long-term outcomes of children with PARDS?

METHODS

Details of the literature search are outlined in the PALICC-2 methodology manuscript in this supplement [5]. We conducted a scoping review based on the PRISMA Extension for Scoping Reviews guidelines [6] to identify relevant studies related to outcomes of children surviving PARDS. We included studies that evaluated either pulmonary or non-pulmonary outcomes at the time of or following hospital discharge among survivors of PARDS, acute respiratory failure with a high proportion of PARDS patients, or other critical illnesses if PARDS-specific outcomes could be extracted. The complete search strategy can be found in the Supplemental Digital Content (Table S1). Details of title/abstract review, full text review, and data extraction, and generation of good practice statements and research statements are outlined in the PALICC-2 methodology manuscript [5].

RESULTS

Of 8,037 abstracts screened, 110 underwent full-text screening and 20 articles were included (Figure S1). From these articles, we developed four categories of recommendations/statements about: a) approach to clinical outcome assessment; b) pulmonary outcomes of patients who survive PARDS; c) non-pulmonary outcomes of patients who survive PARDS; and d) conducting PARDS outcomes research. Complete evidence-to-decision tables supporting the recommendations are provided as Tables S2–8.

Approach to Clinical Outcome Assessment

Post-PICU morbidities among children surviving PARDS and their families may go unrecognized if not intentionally evaluated. The optimal setting for and approach to post-discharge evaluation may differ based on the patient’s medical complexity and available resources, but for many patients may optimally occur in the patient’s existing medical home.

Good practice statement 9.1.1. In patients with PARDS, the patient’s primary care providers should be advised to screen for post-PICU morbidities within three months of discharge from the hospital.

(Ungraded good practice statement, 90% agreement)

Good practice statement 9.1.2. A stepwise approach to clinical evaluation of post-PICU morbidities should be used with initial screening by a primary care provider or electronic/telephonic screen.

(Ungraded good practice statement, 96% agreement)

Remarks: Full assessment, initial management, and serial re-evaluations of impairments should be made by a primary care provider if appropriate and referral to a specialist if deficits persist or are not in the scope of practice of the primary care provider. Location-specific resources should be considered, including availability of dedicated post-PICU follow-up clinics or remote consultation.

Justification

Opportunities for children to be evaluated in formal follow-up clinics remain limited [7]. While follow-up clinics have become increasingly common after adult and neonatal intensive care [8–10], most pediatric-specific clinics are focused on follow-up after pediatric neurocritical care [11–13]. Screening for morbidities following PARDS may alternatively (or even optimally) be performed by primary care providers after PICU discharge for patients without chronic respiratory support needs. For children requiring chronic respiratory support, subspecialty follow-up with pulmonologists or pediatricians with expertise in chronic respiratory care is advised. PICU providers familiar with the patient’s clinical course should advise primary care providers about the patient’s risk for adverse post-discharge outcomes and recommend screening. Primary care providers may elect to subsequently refer their patients to specialty providers including pulmonologists, rehabilitation specialists, physiotherapists, psychologists, or educators depending on the deficits identified.

Benefits

The initial months after PICU admission are resource-intensive for families and patients, and often include multiple healthcare visits, new equipment or medications, or outpatient therapy [14]. These months may represent a sensitive period of recovery for children who have acquired morbidities during their PICU stay, and provision of post-discharge care to optimize recovery is crucial. This resource-intensive period necessitates provision of care in a family-centered manner, which can be accomplished by incorporating post-PICU care into the patient’s medical home.

Harms and burdens

The burden of a one-time electronic or telephone communication from PICU providers to primary care clinicians is minimal but requires incorporation into standard PICU workflow. There is a time and potential financial burden on primary care clinicians to assume responsibility for screening.

Balance of effects

The burden associated with communication between the PICU provider and primary care clinician is minimal, while the risk of failure to communicate a patient’s risk for adverse outcomes and recommended screening for post-discharge impairments is high. The evaluation itself requires additional effort on the part of the clinician, family, and patient, but the associated risk is low while the potential for benefit if impairments are identified is high.

Implementation considerations

Post-PICU assessment and initial management of identified impairments by the primary care provider can minimize burden on patients, families, and specialists. Reserving subspecialty referrals for children with the highest risk of impairments or after impairments are identified may enhance efficiency and lessen the burden on families. The types of evaluation will differ based on local resources with varied settings presenting different opportunities to screen for and treat impairments congruent with resource availability.

Pulmonary Outcomes of Children Who Survive PARDS

While post-discharge pulmonary outcomes have been well-characterized in adult survivors of ARDS, studies of long-term pulmonary sequelae among survivors of PARDS are limited [15,16]. There is increasing evidence, however, that a subset of children surviving PARDS have ongoing pulmonary morbidities following discharge [17–20].

Good practice statement 9.2.1. Patients with PARDS should be screened for pulmonary function abnormalities within the first three months after hospital discharge.

(Ungraded good practice statement, 90% agreement)

Remarks: The screening should include a minimum of a respiratory symptom questionnaire, a respiratory exam, and pulse oximetry.

Good practice statement 9.2.2. Patients with PARDS who are of sufficient developmental age and capabilities should also be assessed by spirometry to screen for pulmonary function abnormalities within the first three months after discharge. (Ungraded good practice statement, 94% agreement)

Remarks: A follow-up assessment within the first year should be added if spirometry is abnormal.

Good practice statement 9.2.3. Patients should be referred to a specialist (pediatrician or pediatric pulmonologist) when deficits in pulmonary function are identified for further assessment, treatment, and long-term pulmonary follow-up.

(Ungraded good practice statement, 94% agreement)

Justification

The 2015 PALICC recommendations/statements were based on adult data and six small studies (n=2 to 11) reporting pulmonary function abnormalities in children following PARDS at variable time points after hospital discharge [4]. While still limited by small sample sizes, single-center cohorts, and loss to follow-up, four recent observational studies (n=29 to 47) provide additional evidence that children with PARDS are at risk for developing or having persistent respiratory symptoms and pulmonary function abnormalities post-discharge [17–20] (Table 1, Table S9). One study included children meeting the American European Consensus Conference (AECC) ARDS definition [20], while the other three included children meeting the PALICC definition of PARDS [17–19]. Abnormal respiratory symptoms including cough, breathing difficulty, and wheezing were reported on non-standardized questionnaires in approximately one-third of children with PARDS at three months post-discharge [17,18], with most symptoms resolving by 12 months [18]. A minority had an abnormal respiratory examination [18]. In one retrospective study conducted using insurance claims data, over 80% of the 47 children with severe PARDS were evaluated by a pulmonologist in the year following their PICU admission and 57.4% were treated with pulmonary medication(s) [19] (Figure 1).

Table 1:

Summary of evidence related to pulmonary outcomes following acute respiratory distress syndrome in children

Ref (Year)	Centers	Population	Country	Outcome Domains	Timepoints	No.a	Results
17 (2020)	Single center	General ARDS	Canada	Respiratory symptoms	3 months	38	36.8% any symptoms; 18.4% new symptoms
				Pulmonary function	3 months	10	30% abnormal spirometry; 25% decreased inspiratory pressure
18 (2018)	Single center	General ARDS	India	Respiratory symptoms	3 months 9–12 months	29 27	37.9% any symptoms 0% any symptoms
				Respiratory exam	3 months 9–12 months	29 27	3.4% abnormal exam 0% abnormal exam
				Pulmonary function	3 months 9–12 months	29 27	82.7% abnormal spirometry 18.5% abnormal spirometry
19 (2022)	Single center	Severe ARDS	US	Pulmonary care	1 year	47	80.9% ≥1 pulmonary outpatient visit; 57.4% pulmonary medication use
20 (2017)	Single center	General ARDS	US	Pulmonary function	11 monthsb	17	35.3% abnormal spirometry

ARDS, acute respiratory distress syndrome.

^aNumber of patients evaluated at each follow-up timepoint

^bMean time to follow-up

Figure 1.

Proportion of patients with impairments in pulmonary outcomes based on time of assessment and domain. Studies listed by population evaluated and reference number. ARDS, acute respiratory distress syndrome.

In addition to abnormal respiratory symptoms, 30–80% of children surviving PARDS had pulmonary function abnormalities on lung function testing at three months post-discharge [17,18,20], with 18.5% - 35% having persistent abnormalities at one year [18,20]. Abnormalities were mixed, including obstructive and restrictive disease, and some were not accompanied by respiratory symptoms. Spirometry, a technique that is often limited to cooperative school-aged or older children, was the most commonly used assessment method. We are unable to recommend other methods of pulmonary function testing as only a small number of patients underwent alternative methods including oscillometry [17], plethysmography [20], maximal inspiratory and expiratory pressure measurements [17], or infant pulmonary function testing [20].

Benefits

Post-discharge screening for respiratory symptoms, examination, and pulmonary function testing can identify at-risk children who may benefit from targeted management and follow-up, including reassessment within the first year post-discharge. These results should be interpreted considering the child’s pulmonary status prior to PARDS. While the initial assessment may be performed by a primary care provider, we recommend referral to a pediatric specialist if persistent respiratory symptoms or impairments in pulmonary function are identified. Additional testing (e.g., chest imaging, more extensive pulmonary function testing) may be recommended by the specialist to further evaluate the impairment and initiate targeted management, if appropriate.

Harms and burdens

A symptom questionnaire, clinical examination, and pulse oximetry are often performed in routine medical visits, and thus the associated burden is minimal. Although spirometry is a routine, non-invasive procedure that can be safely performed in age-appropriate, cooperative children (generally five years and older) [21], it can be time-consuming for patients and families and may necessitate traveling to another facility where the test is available. Additionally, in-depth assessments and referral to specialists may be a burden for families, generate additional anxiety, and increase costs to families and the health care system.

Balance of effects

Taken together, these findings suggest that children surviving PARDS are at risk of pulmonary morbidity. Failure to manage and follow these deficits, particularly when severe, may lead to long-term morbidity in affected children.

Implementation considerations

The proposed evaluations differ based on the resources available. Spirometry, particularly when performed in younger children, requires trained technicians, and may have limited availability in rural or low-resource settings [21]. Additionally, accessibility to specialist care may be limited, and the roles of various health providers may differ from one setting to another. The application of these good practice statements will need to be adapted to specific health care environments.

Non-Pulmonary Outcomes of Children Who Survive PARDS

Increasingly, pediatric critical care researchers have attempted to better understand the range of morbidities that children may experience following PICU hospitalization [2]. While studies of outcomes specific to PARDS survivors remain limited, in recent years there has been substantially more research into non-pulmonary outcomes following PARDS. Therefore, we have been able to augment the previous 2015 PALICC recommendations/statements regarding neurocognitive outcomes, which were based on the adult literature and only a single pediatric study demonstrating decline in Pediatric Overall Performance Category (POPC) and Pediatric Cerebral Performance Category (PCPC) at hospital discharge following acute lung injury [22].

Good practice statement 9.3.1. For patients who survive PARDS, health-related quality of life (HRQL), physical, neurocognitive, emotional, family, and social function should be evaluated within three months of hospital discharge.

(Ungraded good practice statement, 100% agreement)

Good practice statement 9.3.2. For infants and toddlers who survive PARDS, additional evaluation of HRQL, physical, neurocognitive, emotional, family, and social function should be performed prior to entering school, for example, at 4–6 years of age.

(Ungraded good practice statement, 90% agreement)

Good practice statement 9.3.3. Patients with identified abnormalities should be treated or referred for more in-depth assessment and treatment by appropriate subspecialists and educators.

(Ungraded good practice statement, 98% agreement)

Justification

Recent studies demonstrated impairments in HRQL [20,23–25], functional status [23–30], neurocognitive status [31,32], emotional health [24], and school participation [18,33] in the months and years after hospitalization with PARDS, as well as high rates of ongoing care needs after hospital discharge [28,29,34], post-discharge health resource use [14,17,19], late mortality [14,18,23,31,35,36], and work absences for parents [33] (Table 2, Table S9).

Table 2:

Summary of evidence related to non-pulmonary outcomes following acute respiratory distress syndrome in children

Ref (Year)	Centers	Population	Country	Outcome Domains	Timepoints	No.a	Results
14(2022)	Single center	Moderate-severe ARDS	US	Health resource use	1 year	241	15.8% high resource use
				Late mortality	1 year	241	4.6% post-discharge mortality
17 (2020)	Single center	General ARDS	Canada	Health resource use	3 months	38	23.7% ED visit, 10.5% readmission, 18.4% specialist referral
18 (2018)	Single center	General ARDS	India	Late mortality	3 months 9–12 months	29 27	0% post-discharge mortality 7.4% post-discharge mortality
				Return to school	3 months 9–12 months	29 27	44.8% not in school 14.8% not in school
19 (2022)	Single center	Severe ARDS	US	Health resource use	1 year	47	66.0% ED visit; 25.5% readmission
20 (2017)	Single center	General ARDS	US	HRQL	11 monthsb	23	Health perception 46.5/100; general health 63/100; physical functioning 77/100; behavior 75/100
23 (2022)	Multicenter	General ARDS	US	HRQL	28 days	95	53% >10% decline from baseline
					3 months	72	25% >10% decline from baseline
					9 months	61	31% >10% decline from baseline
				Functional status	28 days	95	48% FSS change ≥3
					3 months	72	31% FSS change ≥3
					9 months	61	23% FSS change ≥3
24 (2018)	Multicenter	Acute respiratory failurec	US	Functional status	6 months	949	19% POPC decline; 11% PCPC decline
				HRQL	6 months	838	27% impaired HRQL
				Emotional health	6 months	102	30.4% positive PTSD screen
25 (2019)	Multicenter	Moderate-severe ARDS	US	Functional status	6 months	689	22.4% decline in POPC
				HRQL	6 months	448	21.0% impaired HRQL
26 (2021)	Multicenter	Acute respiratory failurec	US	Functional status	Discharge	427	9.4% decline in POPC
27 (2022)	Single center	ARDS requiring ECMO	US	Functional status	3–4 yearsb	59	Mean POPC 2.3; mean PCPC 1.9
28 (2018)	Single center	General ARDS	US	Post-discharge care	Discharge	274	24.4% to post-discharge care
				Functional status	Discharge	274	23.0% FSS change ≥3
				Late mortality	1 year 3 years	274 259	5.5% post-discharge mortality 8.0% post-discharge mortality
29 (2020)	Single center	Trauma patients with ARDS	US	Post-discharge care	Discharge	68	76.9% to post-discharge care
				Functional status	Discharge	68	77.1% abnormal FIM
30 (2021)	Single center	General ARDS	Singapore	Functional status	Discharge	90	15.6% FSS change ≥3; 38.9% any change; 27.8% new feeding support; 11.1% new respiratory support
31 (2017)	Single center	Immunocompromised patients with ARDS	Germany	Late mortality	3 years	15	40.0% post-discharge mortality
				Neurologic outcome	3 years	15	26.7% learning difficulties; 6.7% cognitive impairment
32 (2022)	Multicenter	Moderate-severe ARDS	US	Neurocognitive status	3–8 years	74	18.9% IQ ≥15 points below sibling
33 (2021)	Multicenter	Acute respiratory failurec	US	School absenteeism	6 months	399	69.9% missed school days; 38.3% chronic absenteeism
				Work absenteeism	6 months	506	54.7% missed work days
34 (2019)	National database	Trauma patients with ARDS	US	Post-discharge care	Discharge	2073	55.8% to post-discharge care
				Tracheostomy	Discharge	2073	18.4% discharged with tracheostomy
35 (2008)	Single center	ARDS requiring ECMO	UK	Late mortality	1 year	77	5.2% post-discharge mortality
36 (2020)	Single center	ARDS requiring ECMO	France	Late mortality	6 months	111	3.0% post-discharge mortality
39 (2020)	Single center	ARDS with NMB	India	Neuropathy/myopathy	3 months	23	0% neuropathy or myopathy

ARDS, acute respiratory distress syndrome; ECMO, extracorporeal membrane oxygenation; ED, Emergency Department; FIM, Functional Independence Measure; FSS, Functional Status Scale; HRQL, health-related quality of life; IQ, intelligence quotient; NMB, neuromuscular blockade; PCPC, Pediatric Cerebral Performance Category; POPC, Pediatric Overall Performance Category; PTSD, post-traumatic stress disorder.

^aNumber of patients evaluated at each follow-up timepoint

^bMean time to follow-up

^c73% of cohort with moderate-severe ARDS.

A small, single-center cohort study demonstrated that at a mean of 10.7 months post-admission, PARDS survivors had significantly worse HRQL than U.S. norms and children with asthma in the domains of general health perceptions and physical functioning; general health perception scores were particularly low at a mean of 46.5/100 [20]. In a larger multicenter study, 53% of PARDS survivors experienced a decline of >10% from their baseline HRQL at the earlier of 28 days post-admission or ICU discharge with 25% having persistent HRQL decline of >10% three months post-admission, a prevalence that rose to 31% by nine months post-admission [23]. The large multicenter Randomized Evaluation of Sedation Titration for Respiratory Failure (RESTORE) trial included HRQL assessment six months after ICU discharge in 838 survivors of acute respiratory failure, of whom 73% had moderate to severe PARDS; 27% of children and teens had impaired HRQL (at least one standard deviation below the population mean), with lower scores than healthy population norms across all HRQL domains. Scores for infants and toddlers were similarly lower than population norms for multiple HRQL domains [24]. Among the subset of participants who met criteria for moderate to severe PARDS and had normal pre-admission function, 21% of children and teens had impaired HRQL, and 22% of infants and toddlers had impaired growth and development [25] (Figure 2).

Figure 2.

Proportion of patients with impairments in non-pulmonary outcomes based on time of assessment and domain. Studies listed by population evaluated and reference number. ARDS, acute respiratory distress syndrome; ARF, acute respiratory failure; HRQL, health-related quality of life; Immuno, immunocompromised; NMB, neuromuscular blockade; PTSD, post-traumatic stress disorder. *Assessed neuropathy/myopathy.

In addition to the impaired physical functioning observed within domains of HRQL, multiple studies specifically evaluated functional status following PARDS. In a study of pediatric trauma patients with PARDS, 77.1% of were discharged with an abnormal Functional Independence Measure [37] score compared to 30.7% of trauma patients without PARDS [29], and 9.4% of these trauma patients with acute respiratory failure had a decline in POPC at discharge [26]. In two cohorts of general PICU patients with PARDS, 16–23% experienced new morbidity at hospital discharge based on increase in Functional Status Scale (FSS) [38] score by ≥3 points [28,30] with 39% experiencing any FSS increase, primarily manifested by 28% requiring new feeding support and 11% requiring new respiratory support [30]. Following discharge, 31% of PARDS patients had a new morbidity defined by FSS increase ≥3 points at three months post-admission relative to pre-illness baseline, and 23% had new morbidity at nine months [23]. Among patients with moderate to severe PARDS in the RESTORE trial, 22% experienced a decline in POPC and/or PCPC from baseline to six-month follow-up [25]. In a separate study, patients evaluated three to four years following PARDS requiring extracorporeal membrane oxygenation (ECMO) had ongoing impairments as demonstrated by mean POPC score of 2.3 and mean PCPC score of 1.9 [27]. In a study specifically evaluating neuromuscular outcomes following neuromuscular blockade for treatment of PARDS, neither myopathy nor neuropathy was identified at three-month follow-up [39].

Several studies report impairments in neurocognition, school participation, and emotional functioning following PARDS. In a small cohort of immunocompromised patients evaluated three years after admission with PARDS, 27% had learning difficulties and 6.7% had cognitive impairment [31], while among patients with moderate-severe PARDS in the RESTORE trial without a history of cognitive impairment, 19% had an estimated intelligence quotient at least one standard deviation below that of their sibling when measured three to eight years following their PARDS hospitalization [32]. A small cohort study in India found that only 55% of PARDS survivors had returned to school by three months, and only 85% had returned to school by nine to twelve months following hospitalization [18]. Among nearly 400 school-eligible patients in the RESTORE trial, 70% missed school in the six months post-admission, with 38% experiencing chronic absenteeism. Parents also struggled with absenteeism following their child’s PARDS admission, with 55% experiencing missed days of work during the six months post-admission [33]. Finally, 30% of patients in the RESTORE trial screened positive for post-traumatic stress disorder at six months after discharge [25].

Patients surviving PARDS also have high rates of ongoing care needs following hospital discharge, with 24% of general PICU patients with PARDS discharged to inpatient rehabilitation and 3.3% to chronic care facilities [28]. Among pediatric trauma patients with ARDS, nearly 50% required post-discharge care compared to 12–16% without PARDS [29,34], and 18.4% underwent tracheostomy placement [34]. In the three months following hospitalization, 18% of general PICU patients with PARDS were seen by a specialty care provider, 24% presented to an Emergency Department, and 11% were readmitted to the hospital [17]. Over the course of the first year, 66% of patients with severe PARDS presented to an Emergency Department and 26% were readmitted to the hospital [19]. A study using machine learning techniques applied to post-discharge insurance claims to identify phenotypes of outcomes after acute respiratory failure requiring at least three days of invasive ventilation identified that nearly one in five patients with moderate-to-severe PARDS demonstrated the high morbidity phenotype identified as having the highest health resource use in the year following hospitalization [14].

Post-discharge mortality varied by population and time point evaluated; among PARDS patients requiring ECMO who survived to hospital discharge, post-discharge mortality was 3.0% at six months [36] and 5.2% at one year [35], while among general PICU patients with PARDS, post-discharge mortality was 4.6–5.5% at one year [14,28] and 8.0% at three years [28]. Among 15 immunocompromised patients with PARDS, three-year post-discharge mortality was 40% [31] (Figure 3).

Figure 3.

Proportion of patients with post-discharge mortality based on time of assessment. Studies listed by population evaluated and reference number. ARDS, acute respiratory distress syndrome; ECMO, extracorporeal membrane oxygenation; Immuno, immunocompromised.

Benefits

These findings suggest that children surviving PARDS are at high risk for morbidities across a range of health domains. Early identification of impairments within three months following hospital discharge may facilitate treatment to optimize long-term outcomes. Given that impairments may not be as easily measured or recognized in young children due to inability to participate in evaluation, their developmental capacity, or inadequacy of standard measures to detect impairments in young children [40], re-evaluation prior to beginning school may identify previously undetected morbidities that could negatively impact development.

Harms and burdens

Screening for a wide range of morbidities may be time-consuming for patients, families, and clinicians, especially early after hospital discharge when families may be experiencing high rates of stress, anxiety, and healthcare encounters [41–44]. In-depth assessments and referrals to subspecialists may place an additional time and financial burden on patients and families. Interventions may also be costly and time-consuming.

Balance of effects

Identifying and treating deficits in HRQL, physical, neurocognitive, emotional, family, and social function among children surviving PARDS is important to optimize recovery. Identification of abnormalities in any of the domains recommended for screening warrants further evaluation to confirm and elucidate areas of impairments to develop appropriate treatment plans. The decision to provide more comprehensive assessments and/or treatment by a primary care provider versus a subspecialist will depend on the type and magnitude of the impairment, the experience and expertise of the primary care provider, and the availability of subspecialty providers. Failure to optimize recovery through identification and treatment of impairments may result in long-term harm to the patient or their family. Impairments could also progressively worsen over time if not treated. Shared decision-making among the provider, family, and patient, when appropriate, to discuss the risks and benefits of further assessment and treatment are likely to be valuable.

Implementation considerations

The type of evaluation will differ based on local resources. Some sites may have the capacity to conduct comprehensive in-person evaluations for patients through a post-PICU follow-up clinic, while other sites might conduct screening remotely via questionnaires, in primary care clinics, or through school-based assessments. We do not have evidence that non-pulmonary morbidities following PARDS are different than morbidities identified in other critically ill pediatric cohorts, and thus providers may refer to the pediatric critical care Core Outcome Set and Core Outcome Measures Set [2,3,45] for essential elements, domains, and instruments for screening, with acknowledgment that providers should consider where an instrument was validated and whether the validation is generalizable to their patient(s). Some locations have resources to provide additional in-depth assessments and availability of subspecialists to evaluate and treat post-discharge deficits among PARDS survivors. Others may have limited or no availability of subspecialty providers, and the responsibility for further assessment and treatment may fall to the primary care providers who may have variability in experience and resources to provide these services. By approaching the evaluation and treatment of post-PICU morbidities in a stepwise manner as recommended, the aspects of assessment and care that are most challenging to implement should be limited to the children at highest risk for ongoing deficits who have the greatest opportunity to benefit from treatment.

Conducting PARDS Outcomes Research

While the literature on outcomes among PARDS survivors is growing, our understanding of the associations between patient and clinical factors with poor long-term outcomes, and how post-discharge outcomes can be incorporated into clinical trials, remains limited.

Research statement 9.4.1. When feasible, pre-PICU baseline status for each outcome measure should be established or estimated when evaluation of post-PICU morbidity is anticipated.

(Ungraded research statement, 98% agreement)

Research statement 9.4.2. We cannot make a recommendation regarding the use of alternative post-discharge endpoints. Given declining mortality among children with PARDS, additional studies are needed to evaluate potential alternative post-discharge endpoints for clinical trials.

(Ungraded research statement, 96% agreement)

Remarks: Potential post-discharge endpoints to evaluate should include hospital and PICU readmissions (e.g., within 30 days of discharge), unplanned health resource use, HRQL, and physical, pulmonary, neurocognitive, emotional, family, and social function.

Research statement 9.4.3. Clinical studies should be designed to evaluate the association between short-term outcomes (e.g., new or progressive organ dysfunction) and longer-term post-discharge outcomes for patients with PARDS.

(Ungraded research statement, 100% agreement)

Research statement 9.4.4. Further studies are needed to determine factors that may affect the trajectory of recovery following PARDS.

(Ungraded research statement, 100% agreement)

Remarks: Additional research should include demographic characteristics, clinical factors, PICU exposures, social determinants of health, and access to care.

Research statement 9.4.5. Practices to optimize follow-up (e.g., incentives, multimodal evaluations) should be used to minimize bias due to differential loss-to-follow-up when conducting post-PICU outcomes research.

(Ungraded research statement, 96% agreement)

Justification

The heterogeneity of patient and hospitalization characteristics in children who develop PARDS limits our ability to clearly delineate the impact of severe lung injury and associated life-sustaining measures (e.g., immobility, sedation) likely to have negative side effects. To better delineate the impact of PARDS, it is imperative to understand the child’s functioning after resolution of PARDS relative to their pre-illness baseline, which is frequently omitted in pediatric critical illness research [2,46]. Additionally, understanding the long-term impact of lung injury and the associated critical care support on a PARDS survivor necessitates a breadth of outcome assessments including multiple health domains. The use of patient-centered endpoints is a central component of clinical outcomes research, and includes: quantifying the rate of subsequent hospital and PICU readmissions (e.g., within 30 days of discharge); recording unplanned health resource use; measuring pulmonary outcomes; and evaluating the domains of the PICU Core Outcome Set including overall health (inclusive of health-related quality of life) and physical, cognitive, emotional, and family functioning [3]. Additionally, the relationship between short- and long-term outcomes will need to be more clearly delineated to facilitate future targeting of key inpatient outcomes that may impact long-term outcomes. Lastly, outcomes research is frequently restricted to a limited population due to exclusion criteria that enhance study feasibility (e.g., language, need for internet access) at the expense of generalizability. High attrition after enrollment is also a substantial concern in many studies; loss-to-follow-up in the included prospective studies in this review ranged from 22–46% for survey data collection [23–25,32,33] and as high as 54% for in-person pulmonary evaluations [20]. There is also differential follow-up based on patient and family demographic factors; patients with follow-up data were more commonly of non-Hispanic White race [20,23,24] with higher parental level of education and greater household income than patients without follow-up data [24]. Future studies will need to facilitate participation by all patients, not only to ensure equity in access but also because results based on those who are currently followed-up may not be generalizable to those excluded or systematically lost to follow-up.

DISCUSSION

Literature characterizing outcomes of survivors of PARDS has grown substantially since PALICC 2015 [4]. We now have information about recovery after PICU admission for PARDS in single-center, multi-center, and population-based studies [14,17–20,23–36,39]. Unfortunately, it is increasingly clear that morbidity after hospital discharge from PARDS is common and affects all domains of function, HRQL, and patients’ families. The relationship of morbidity to illness severity is inconsistent among domains and populations studied. Thus, children with PARDS, regardless of severity, are at risk of new morbidity after the acute illness has resolved, and we advise screening for impairments in all patients surviving PARDS.

Many of the health domains identified in the literature as affecting PARDS survivors overlap with the domains recommended in the PICU Core Outcome Set [3] for survivors of pediatric critical illness. There were no studies, however, that directly compared outcomes among children surviving PARDS to children surviving other etiologies of critical illness. It thus remains unknown whether the prevalence and severity of impairments are similar for patients with and without PARDS. A domain of particular relevance to survivors of PARDS that was not identified as a core domain in the PICU Core Outcome Set is pulmonary morbidity. Pulmonary abnormalities have been identified in nearly one in three PARDS survivors and over half of PARDS survivors were prescribed pulmonary medications in the months following discharge [17–20]. While sample sizes for studies evaluating pulmonary function were small, the high prevalence of impairments across the range of PARDS severity suggest that all patients surviving PARDS may benefit from ongoing evaluation of their pulmonary function, including those without abnormal respiratory symptoms as some impairments may be subclinical.

To optimize their recovery, patients surviving PARDS warrant vigilant follow-up after discharge to identify and treat impairments in respiratory function as well as physical, emotional, cognitive, social, and family functioning. We generated eight good practice statements for the approach to clinical outcome assessment, assessment of pulmonary outcomes of children surviving PARDS, and assessment of non-pulmonary outcomes of children surviving PARDS including HRQL and physical, neurocognitive, emotional, family, and social functioning. There was no evidence for restricting follow-up evaluation only to patients with severe PARDS, and thus these good practice statements apply to all children surviving PARDS regardless of disease severity.

To minimize the burden to providers and families of conducting comprehensive post-discharge evaluations on all patients, we advise a stepwise approach to evaluation beginning with an initial screening followed by full assessment if deficits are identified on screening, with subsequent treatment and serial re-evaluations of impairments as appropriate. As most patients will not have access to dedicated PICU follow-up clinics, we suggest that follow-up care begin in the patient’s medical home with their primary care provider, with subsequent referral to subspecialty providers as needed. For this model to be effective, PICU providers must communicate with primary care providers of patients surviving PARDS to advise them of the post-discharge morbidities that the patient may face and suggest resources for screening, including the PICU Core Outcome Measures Set [45] that provides recommendations for screening measures and more in-depth assessment tools across the outcome domains recommended in the PICU Core Outcome Set [3]. More work is needed, in collaboration with primary care providers, to identify best practices for transitioning care of the PARDS patient from the PICU to the general hospital wards to the outpatient setting to ensure appropriate communication and adequate resources for ongoing evaluation. Additionally, further research is needed to determine how to optimize conducting post-discharge pulmonary evaluation, as we were unable to recommend specific screening tools for respiratory symptom screening based on the existing literature, and studies that have incorporated post-discharge spirometry assessments have been limited by small sample sizes and high rates of loss-to-follow-up.

We also generated five research statements reflecting the need to better delineate trajectories and mechanisms of impairment and recovery following PARDS, relationships between short- and long-term outcomes, and methods to optimize the study of post-discharge outcomes to ensure results are relevant to families from all backgrounds. Existing studies of PARDS outcomes have predominantly been conducted in North America and Western Europe, and multiple prospective studies have identified differential loss-to-follow-up based on patient and family demographic characteristics, with non-Hispanic White patients with higher socioeconomic status more commonly represented in follow-up cohorts.

This work has several limitations. While we employed an extensive search strategy, we may not have identified all relevant studies. Studies were primarily conducted in high-income countries and thus the statements we generated may not be relevant for patients in other geographic regions; however, we did develop these statements with experts from resource-limited settings to ensure broad applicability. No studies contained high quality evidence and thus we were only able to make ungraded good practice and research statements rather than recommendations for outcome assessment following PARDS.

CONCLUSION

Many children who survive PARDS suffer declines in physical, cognitive, and emotional function and diminished HRQL, and their families may also suffer. Their trajectory of recovery has not been well studied, but morbidities have been shown to persist for months or years after the episode of PARDS. The good practice statements developed through this extensive scoping review process and consensus conference involving experts in PARDS research, clinical care, and outcomes assessment provide guidance to clinicians caring for children surviving PARDS across the scope of pediatric care to identify and treat impairments to optimize recovery. The accompanying research statements address the limitations of existing PARDS outcomes research and aim to promote future research that advances our understanding of long-term outcomes among children surviving PARDS. This vulnerable population of children and their families warrant assiduous long-term care and additional rigorous research to inform efforts that will optimize their health and well-being.