Continuing Care For Critically Ill Children Beyond Hospital Discharge: Current State of Follow-up

A brief, anonymous, adaptive Qualtrics survey link was distributed to the PNCRG and the Pediatric Acute Lung Injury and Sepsis Investigators (PALISI) outcomes group. This sampling frame represents international multiprofessional providers with interest in PICU research, primarily within large US children’s hospitals. Recipients were encouraged to forward links to colleagues, and the survey was administered February 15, 2021 to April 12, 2021 with no incentives. Multiprofessional health care providers were targeted, and no geographic limitations were set. This study was approved by the institutional review board.

The goals of the authors of this article were: (1) to identify and characterize PFUPs, (2) elicit stakeholder opinions on PFUP essential components and benefits, and (3) to identify barriers to establishing and maintaining programs. The survey was created utilizing established processes of literature review and expert consensus during PNCRG outcomes group meetings to determine essential content.²⁶  The sampling frame was based on study team consensus for likelihood of response from the majority of existing PFUPs and targeting participants with relevant PICU expertise. Specific survey question creation and revision followed well-established survey design principles during multiple expert team meetings, including measure development expertise (author T.A.H.), and was piloted before distribution.²⁷  The survey (Supplemental Fig 5) queried respondent demographics and roles, and the presence of PFUPs at their institution, which dictated further question exposure. Respondents with PFUPs provided program details if able. All were asked opinions regarding PFUPs in general, and additional optional open-ended questions regarding logistics of PFUPs, barriers, and attitudes. Depending on responses, the survey took approximately 5 to 15 minutes.

Five institutions had multiple respondents with discrepant PFUP details (funding sources, diagnoses included, and ages treated), all resolved by contacting PNCRG members at the institutions to clarify. Some responses were collapsed for reporting and statistical analysis. Descriptive statistics characterized the full sample and PFUP details. Likert responses on a 5-point scale (1 = strongly agree to 5 = strongly disagree) summarized cohort responses. Thirteen items focused on barriers to establishing and maintaining a PFUP. We used a consensus process to group the items into 4 composite variables for analyses (Supplemental Table 4). t test, one-way ANOVA, or χ² analysis compared groups by clinical role and presence of a PFUP. We aimed to test differences in responses regarding attitudes and barriers between these groups to provide important context to results as we hypothesized responses would differ by provider perspective and PFUP experience. Quantitative analyses were conducted using SPSS (v26).

Qualitative analysis used the 6-step deductive thematic analysis methodology outlined by Kiger and Varpio²⁸  with 2 independent reviewers (CF and JK). Results did not reach thematic saturation due to limited responses. Results are presented in relevant sections of quantitative data. Variability in qualitative responses was explored by presence of a PFUP.

Among 116 respondents, 97 completed the survey, and 14 completed >50%. Five were excluded as incomplete. The 111 analyzed respondents represented 60 different institutions located in the US (n = 55), Canada (n = 3), Australia (n = 1), and United Kingdom (n = 1). Most (87%) institutions were academic medical centers. US geographic region varied: Northeast (15%), Southeast (18%), Midwest (38%), Southwest (7%), Mountain West (3%), and West (13%). Most participants were physicians (80%), followed by neuropsychologists (9%), nurses (4%), nurse practitioners (3%), psychologists (2%), program managers (2%), and education specialists (1%). Among physicians, specialties included pediatric critical care (65%), pediatric neurology (15%), general pediatrics (16%), physiatry (2%), and developmental pediatrics (2%). Most reported practicing for <10 years (61%), with the remaining practicing ≥10 years (33%) or in-training (6%). Forty-eight (43%) respondents from 21 individual programs reported an active PFUP. Sixty-three (57%) respondents from 39 individual programs reported no PFUP. Most without PFUPs indicated interest (70%) or actively working to start a PFUP (11%), and few (19%) had “no interest.”

Twenty-one PFUPs were identified, located in the US (n = 18), Canada (n = 2), and Australia (n = 1). Seventeen PFUPs provided program details (Tables 1–3). PFUP funding included clinical revenue (59%), institutional or departmental support (53%), research funding (24%), and philanthropy (12%). Five (29%) PFUPs reported all PICU patients were eligible or included broad PICU populations, but most focused on narrower populations, including neurologic (53%) and respiratory (12%) diagnoses. One PFUP limited inclusion to children under age 5 years. Eight programs expanded inclusion to populations over age 18 years. All 17 detailed programs included at least 1 outpatient visit within 6 months of discharge. Eight (47%) programs reported inpatient consultation, and 10 programs (59%) reported longitudinal outpatient follow-up. Five (29%) outpatient PFUPs were physician only programs. Physicians from institutions with PFUPs reported comfort managing physical morbidity (85%) more often than cognitive (55%), emotional (50%), or social or family (40%) morbidities. Eleven (65%) PFUPs reported outpatient neuropsychology providers. All neuropsychologists from institutions with PFUPs reported comfort managing cognitive, emotional, and family or social morbidities. Among respondents with PFUPs, 51% reported using data for quality improvement projects, and 19% reported using data for advocacy projects. Additional PFUP activities and benefits are found in Fig 1.

Participants with a PFUP listed specific measures, both for research and clinical applications, including the “Bayley,” “WeeFIM,” “PEDICAT,” “PEDSTBI,” and “PedsQL” assessments, and general responses such as “neuropsychologic testing” or “depression screening.” Five active PFUPs reported a developmental screen, 5 reported functional outcome screens, and 3 reported quality of life-focused tools as key measures.

All respondents rated essential members, components, and benefits of a PFUP. Most responses were rated in the ”strongly agree” to ”neutral” categories. No statistically significant differences were found between groups by clinical role or presence of a PFUP for any of these analyses. Respondents rated multidisciplinary providers as all essential to a PICU follow-up team, but only nonphysician members received >50% strongly agree responses (Fig 2). Evaluation of all PICS domains (physical, cognitive, emotional, and psychosocial), providing family education and support, and plans for school reintegration were rated as essential by ≥90% of respondents (Supplemental Fig 6).

More than 80% of respondents agreed that PFUPs provide direct benefits to patients and families, improve the identification, evaluation, and management of PICS, and are essential to advance knowledge and research on long-term PICU outcomes (Supplemental Fig 7). Less than 20% of respondents indicated general pediatricians alone were equipped to identify and manage PICS morbidities. More than 85% agreed that PFUPs are necessary to advance knowledge, but only 9 (43%) institutions with an active PFUP reported collecting their outcomes data, and only 2 programs reported collaborating with at least 1 other institution on data collection. Most respondents (76%) indicated willingness to contribute to a multicenter outcomes database of longitudinal PICU outcomes.

Qualitative responses regarding benefits of PFUPs described the ability of clinics to identify and manage specific issues associated with PICS, potential to improve care coordination, access to care, and to reduce disparities. Respondents with PFUPs focused on the programs’ ability to provide education and support for families as they navigate the health care system and reintegrate into the school setting. Respondents without PFUPs expressed a need for further research in this area or uncertainty about the benefits of a program.

Respondents identified Lack of Support as the most important barrier overall. The least important barrier was Unclear Benefit, with 76% respondents rating as “not a barrier” or a “small barrier.” As shown in Fig 3, programs with a PFUP rated Lack of Support (P = .01) and Getting Started (P = .005) barriers significantly lower than those without a PFUP. Differences were not statistically significant for Lack of Measures or Unclear Benefit. Thematic analysis showed the most common responses centered on the availability of funding for a program and the difficulty in obtaining appropriate remuneration, followed by lack of institutional or departmental support. Patient access was a less commonly reported but included concerns about “no shows” and difficulties with large catchment areas and travel distances.

This survey, representing 60 unique institutions, identified a critical gap between belief and capability of providing optimal care for children recovering from critical illness. Despite stakeholder agreement on the importance and benefits of PFUPs, only one-third indicated presence of a PFUP, although many expressed interest or active work to start a program. Among current PFUPs, variability reflects the importance of adapting to meet the needs of patients and families within the local context. However, this local adaptation should be balanced with interinstitutional standardization to ease collaboration and advance clinical knowledge, as well as provide comprehensive support for PICS morbidities. Summary recommendations are found in Fig 4.

Providing patients and families with coordinated follow-up and providers experienced with post-ICU morbidities were highly agreed upon benefits of PFUPs. Most PICU survivors with brain injury or physical impairments lack access to coordinated care programs outside of inpatient rehabilitation facilities.²⁹  Childhood cancer survivor clinics^30–32  and NICU follow-up programs¹⁴  have become a patient-valued standard of care. However, unlike their high-risk counterparts surviving childhood cancer or prematurity, most patients struggling after PICU discharge do not receive the support of a longitudinal care program. Even without a primary brain injury diagnosis, PICU patients develop PICS due to secondary effects related to ICU procedures, delirium, sedative and anesthetic exposures, inflammation, sleep disruption, and hypoxia.^1,33,34  PICU survivors often have ongoing needs for multisystem specialty medical care including chronic pain, sedative dependence, dysautonomia, sleep disturbances, physical deconditioning, hair loss, and wound care that may not be routinely cared for by pediatric providers outside those experienced in post-PICU populations.³⁴  Furthermore, these same patients are at risk for cognitive impairments that require specific expert assessments to diagnose and intervene.³⁵  PICU survivors (and their parents) develop psychological distress, which impacts social functioning and school or work performance, but access to mental health services is a growing crisis in the US.^8,36,37  While PFUPs are not mental health clinics, they can normalize patients’ symptoms, serve as referral hubs after discharge, and establish programmatic access to services. PFUPs can also help families summarize the outcome of their admission, identify gaps in care for ongoing PICS morbidities, and articulate family goals moving forward, while adjusting to the realities of a sometimes alarming “new normal.”

A definitive minority of centers providing pediatric critical care offer a formalized PFUP, and considerable variation exists in the resources offered. This likely reflects the unique challenges of addressing the diverse needs of a highly heterogeneous population. Only 5 PFUPs detailed in this study were broadly inclusive of PICU survivors, while the majority focused on neurologic populations that are among the highest risk groups for PICS morbidities.²⁵  Responses highlighted the need for integrated care as nonphysician providers were rated highly essential. Unease in a physician-only model was reflected in both respondents’ impressions of general pediatricians’ capacity to diagnose and manage the spectrum of PICS morbidity in isolation, as well as reported discomfort for nonphysical symptom management by physician respondents. Managing morbidities beyond the physical, particularly in cognitive, emotional, and social domains, faced by PICU survivors, is often outside the normal physician scope. However, the complementary expertise of other providers, such as neuropsychologists, can fill gaps in physician-only models of care to assess and treat these important morbidities. Data show benefit and high patient satisfaction in post-PICU care with integration of neuropsychology,²²  consistent with our responses. Well-resourced programs are also able to provide both therapeutic intervention through integration of psychologists and rehabilitation providers (eg, physical and occupational therapy), and social and family support through social work and school liaison services. Respondents agreed these additional services should be considered essential but were not often available in active PFUPs. Data are lacking on quantifying direct benefits of these additional services within PFUPs but is actively being studied.

An area with less consensus was whether PFUPs decrease health care disparities or improve access to specialty care. Poverty contributes to poorer health outcomes for children, including but not limited to PICU diagnoses.^38–40  Children from impoverished neighborhoods are more likely to require PICU admission.⁴¹  Additionally, for rural populations, there continues to be significant regional differences with many families receiving care in PICUs >50 miles from home,⁴²  and discharging to rural communities that lack pediatric specialists.²⁴  Respondents noted some institutions serve a large geographical area, which places a significant financial and time burden on families needing care after discharge. PFUPs with systematic referral programs, specific caregiver education about the need for postdischarge follow-up, and integrated care with multiple subspecialists within the same appointment could overcome some challenges, but providers should account for disparity when planning and implementing a PFUP. Notably, some elements identified as essential by respondents have been deployed successfully using telehealth, which may reduce support barriers, as well as family burden. Recent pandemic-related health care needs resulted in a surge in telehealth and virtual technologies from a baseline of <10% of visits to a new proportion of 43% to 51%.⁴³  This offers nascent or resource-limited PFUPs the opportunity to develop relatively cost-effective and time-efficient telehealth programs. While the ability to fully assess and treat PICS morbidities is limited in the telehealth setting, studies show utility of telehealth in some aspects of medical care, mental health support, and limited neuropsychological assessments.^44–48  Further exploration of telehealth to supplement components of PICS care is warranted given the significant financial, employment, and time burdens placed on many families by PICU admission and ongoing morbidity.²⁴ 

Reasons for the relative lack of PFUPs were highlighted in responses regarding barriers, which largely centered around insufficient support and funding. Institutional “buy in” generally requires compelling evidence for need and efficacy, in addition to financial justification. This challenge is exacerbated by the lack of guidance for programmatic necessities and current paucity of clinical data demonstrating cost-effectiveness. The survey provides areas of consensus for PFUP components, but more data are needed to demonstrate financial benefits. Responses from active programs show multiple potential sources of funding to help stakeholders overcome initial challenges related to financing and reimbursement. Some programs started with philanthropic or research-related funds, but careful planning and data capture regarding provider reimbursement, show rates, referrals for procedures (eg, imaging or audiology) or specialty care (eg, ophthalmology or physical therapy) within the institution, and patient or family satisfaction are important to sustainable institutional support. Additionally, building the program into existing clinic infrastructure (eg, use clinic space and staff already allocated to pediatric neurology, physiatry, or neuropsychology providers) helps overcome space and staffing challenges, and reduces clinical overhead for new programs.

Despite the burgeoning PICS literature, there remains a wide variety of constraints to our understanding of the scope, including the paucity of validated diagnostic outcomes measures across PICS domains. It is vital for the pediatric community to push the field of outcomes research forward through standardization and collaboration to advocate for widespread systematic PICU follow-up. Recently, stakeholders participated in multidisciplinary projects seeking to standardize outcomes assessments through the creation of the Core Outcomes Set (COS)⁴⁹  and the forthcoming Core Outcomes Measurement Set (COMS). These projects are a step forward in the quantification of post-PICU morbidity by standardizing assessments across PICS domains, and the field now needs to focus on collaborative data collection of longitudinal outcomes. Despite reported willingness, our study showed very few PFUPs are collecting data to fulfill this role and document the vital data needed to advance clinical care and research.

Not all potentially relevant stakeholders were afforded participation given our chosen sampling frame and distribution methods despite intent of inclusivity. Additionally, due to ability to forward the links individually, response rate is unknown. The participants likely had an interest in PFUPs, and likely differ from members of the organizations as a whole, indicating risk of response bias. The sampling frame primarily targeted large academic children’s hospitals in the US, with private practice and smaller institutions underrepresented. Additionally, the majority of respondents were specialty physicians, and represented few primary care, rehabilitation, or hospitalist providers, though additional perspectives of these groups regarding PFUPs and post-PICU needs are essential to advancing knowledge in future research. The study team relied on measure development and content expertise; however, the survey was not designed as a psychometrically anchored measurement tool, no formal approaches to measure validation were completed, and responses are subject to interpretation bias.

We identified few PFUPs despite strong agreement around the importance of systematic multispecialty care, and the direct benefit to patients and families. Stakeholders involved in care of these children and their families need to recognize morbidities across PICS domains and need to offer systematic support rather than assuming services can be accessed after discharge. A combination of providers, such as physicians and neuropsychologists, as well as screening instruments, are needed to ensure the breadth of PICS morbidities are identified and treated. Even without a complete arsenal of therapeutic resources, PFUPs can provide a safety net by identifying new morbidity and appropriate referral resources that can be accessed in a timely manner, including mental health, school, and family support services, and rehabilitation therapies through institution or community partnerships. Most successful programs identify a longitudinal approach to care with integrated multidisciplinary providers, and financially supported by a variety of sources. Addressing the needs of PICU survivors requires greater investment and collaboration at the provider, institution, and health-policy levels than is currently available.