1. Introduction

Parenting is among the most demanding multi-domain capacities adults perform, integrating emotional availability and attunement (S. H. Goodman et al., 2011), executive functions of planning, inhibition, and working memory (Crandall et al., 2015), physical caregiving labour, and a coherent self-narrative that can be communicated to a child (Rolland, 1999). When parenting is going well it is largely invisible: children’s routines, regulation, learning and security rest on capacities the parent recruits.

Surviving a critical illness disturbs several of these capacities simultaneously. Needham et al. (2012) formalised the survivor phenotype as post-intensive care syndrome (PICS): new or worsened impairments in physical, cognitive, or mental-health domains arising after critical illness and persisting beyond acute hospitalisation, a structure reaffirmed in the 2020 SCCM international consensus update (Mikkelsen et al., 2020). PICS is therefore a constellation of impairments that tracks closely onto the domains parenting recruits. A parallel construct, PICS-Family (PICS-F), describes psychological sequelae (anxiety, depression, complicated grief, posttraumatic stress) in family members exposed to a loved one’s critical illness (Davidson et al., 2012).

The scale of parental disturbance after ICU admission is therefore predictably large. A meta-analysis of 19 studies involving 10 179 ICU survivors estimated a pooled PICS prevalence of 54.35%, with substantial burden across physical (46.0%), cognitive (32.1%) and mental health (32.4%) domains. Heterogeneity was extreme (I² = 98.5% overall, 95.9–96.8% by domain), and trim-and-fill adjustment for funnel asymmetry lowered the pooled estimate to 46.6%, so these figures index magnitude rather than a precise rate (Ayenew et al., 2025). In the BRAIN-ICU follow-up cohort, 64% of 406 survivors had at least one PICS problem at three months and 56% at twelve months, with 25% and 21% respectively affected in two or more domains (Marra et al., 2018). Multi-domain impairment matters because concurrent physical, cognitive, and emotional impairment is more likely to overwhelm caregiving capacity, emotional availability, and routine maintenance.

Despite the volume of PICS research and the clear conceptual mapping onto parenting, parenting after adult ICU survivorship is essentially unstudied. PICS-F has been operationalised primarily for spouses and adult next-of-kin (Davidson et al., 2012; Johnson et al., 2019); the paediatric counterpart (PICS-p; Manning et al., 2018) addresses children who themselves were patients. Parenting after PICS, with its own sensitivity, executive, physical, and identity dimensions, has not been systematically studied, and the adult cohort studies constituting the PICS and PICS-F evidence base do not routinely record whether patients have dependent children.

This review covers three directly affected populations as analytically distinct even though they overlap in any household. The first is the survivor-parent: an adult ICU survivor carrying PICS impairments (cognitive, psychological, physical, and biographical) whose attempt to resume the parenting role is shaped by them. The second is the well co-parent, whose role combines witnessing trauma, surrogate decision-making, and compensatory caregiving load, and whose psychological sequelae fall within the established PICS-F construct. ‘Well’ denotes availability to the parenting role rather than absence of distress; this person is frequently anxious, traumatised, socially unsupported, financially strained, or themselves medically vulnerable, and may be the more impaired parent on any given measure. The third is the child: directly exposed to parental absence and altered functioning, indirectly exposed via the co-parent’s distress, and not routinely captured by either PICS or PICS-F as operationalised. We use two-parent language for analytic convenience, but the configurations that matter clinically are wider: single-parent households, separated and re-partnered parents, same-sex parents, blended families, grandparents as primary caregivers, kinship care, and culturally organised caregiving norms. Where the survivor is the sole resident parent, the compensatory function falls to extended kin, to services, or to the children themselves.

The review asks two interlocking questions: how parenting is reshaped when a parent survives the ICU and develops PICS, and what happens to the children of these parents during admission and across recovery. The analysis draws on family systems theory (Minuchin, 1974), attachment theory (Bowlby, 1969), trauma transmission models (Yehuda & Lehrner, 2018), Bronfenbrenner’s bioecological framework (Bronfenbrenner & Morris, 2006), and Rolland’s (1999) biopsychosocial conceptualisation of illness within families. Parenting capacity is the dependent variable, PICS the proximal disruptor, and the family system the unit of clinical concern.

2. Methods

We conducted a narrative review with a structured literature search and hierarchical evidence categorisation. Because direct evidence was expected to be scarce, the review integrates adjacent literatures, ranked by proximity to the target phenomenon rather than by study design. First-line evidence is direct adult ICU and PICS work. Second-line evidence is adult cardiac arrest survivorship: these survivors typically pass through the ICU and experience PICS-relevant sequelae, while additionally carrying a phenotype shaped by hypoxic-ischaemic brain injury, sudden onset, and frequent family witnessing of the index event (Moulaert et al., 2009; Sawyer et al., 2020). These features sharpen rather than displace the PICS frame, and this is one of the few areas within ICU survivorship measuring family outcomes dyadically (Rojas et al., 2024). Third-line evidence is drawn from analogous parental serious illness (cancer, traumatic brain injury, stroke, posttraumatic stress disorder, chronic illness) and from the young-carer and parentification literatures. The hierarchy is conceptual rather than evidentiary in the systematic-review sense, and direct evidence is distinguished from inference throughout.

PubMed/MEDLINE was searched in April 2026 and updated in July 2026 using four strategies mapped to the evidence hierarchy: adult ICU and PICS survivorship combined with family, caregiver dyadic and offspring terms (D1); a gap-confirmation search pairing survivorship terms with named parenting constructs and commonly reported instruments (D2); cardiac arrest survivorship with family, caregiver or witness outcomes (D3); and review-level evidence on analogous parental serious illness, young caring and parentification (D4). Ovid APA PsycINFO was searched against the same parenting constructs, Google Scholar was used for supplementary discovery, and key reviews were citation-tracked (Mikkelsen et al., 2020; Sawyer et al., 2020; Southern et al., 2024). Eligible sources were English-language peer-reviewed empirical studies, systematic reviews, scientific statements and conceptual frameworks concerning adult survivors of critical illness and their families or children, without date restriction; paediatric ICU survivor literature was excluded except where it informed conceptual frameworks (e.g. PICS-p). Of 852 deduplicated records, screened on title and abstract by the first author, 168 were then read in full by two authors, with full-text exclusion resting on informational redundancy alone. Foundational theory, measurement and service-system sources, and peer-review suggestions entered through supplementary routes. Full search strategies with strings, database yields, deduplication, screening flow and source accounting are reported in Supplementary Methods.

3. Findings

3.1 Parenting capacity after PICS

Figure 1 sets out the conceptual model that organises this section.

3.1.1 Emotional availability and responsiveness to the child

Emotional availability, the parent’s capacity to read, register, and respond contingently to a child’s emotional signals, is among the parenting capacities most consistently linked with child mental health, with maternal depression a particularly well-documented disruptor (S. H. Goodman et al., 2011), and is readily eroded by parental psychopathology (Zitzmann et al., 2024). The PICS evidence base suggests it is at risk after ICU admission. Across PICS cohorts, anxiety occurs in approximately 30% (Nikayin et al., 2016), depression in 30% (Rabiee et al., 2016), and PTSD in approximately one fifth of survivors at 12 months (Parker et al., 2015). The cardiac arrest subset shows comparable morbidity (PTSD 19–27%, anxiety 24–26%, depression approximately 19%; X. Chen et al., 2023; Wilder Schaaf et al., 2013; Yaow et al., 2022) and provides the only ICU-survivor evidence to test dyadic concordance: Rojas et al. (2024) found survivor post-traumatic stress symptoms, but not physical dependence or cognitive impairment, independently associated with family-member distress at one month, in a cross-sectional design that cannot establish direction.

The implication for parenting is clinically plausible but not directly demonstrated. Parental PTSD has been associated with parenting stress, less optimal parent–child relationships, and more frequent negative parenting practices across trauma-exposed populations (Christie et al., 2019). Similarly, maternal depression is meta-analytically associated with child internalising, externalising, and general psychopathology (S. H. Goodman et al., 2011), with the parenting mechanisms including reduced sensitivity and increased withdrawal or hostility (Lovejoy et al., 2000). Although parental PTSD is most studied in combat veteran, refugee, and Holocaust survivor populations, the mechanisms implicated (intrusive symptomatology, emotional numbing, hyperarousal, avoidance) interfere with attunement and predictability in ways that are not trauma-type specific (Lambert et al., 2014; Leen-Feldner et al., 2013). PICS survivors with clinically significant PTSD may therefore exhibit a similar parenting profile, modulated by the medical rather than interpersonal nature of the index trauma.

3.1.2 The cognitive work of parenting

Parenting may be the most executive-function-loaded everyday role adults perform. Crandall et al. (2015) articulate executive function as a foundational substrate of parenting: planning routines, inhibiting reactive responses, holding multiple children’s needs in working memory, and shifting flexibly between caregiving and competing demands. These demands overlap with the domains commonly impaired after ICU admission (Marra et al., 2018).

Pandharipande et al. (2013), in the BRAIN-ICU study, found that one in three medical and surgical ICU survivors had cognitive impairment one year after discharge equivalent to moderate traumatic brain injury, with a quarter scoring at levels consistent with mild Alzheimer’s disease; delirium duration during ICU stay was a robust predictor. Cardiac arrest survivors are a particularly cognitively affected subset, reflecting the impact of hypoxic-ischaemic injury (Boyce van der Wal et al., 2015; Moulaert et al., 2009). The mapping onto parenting is clinically plausible: dysexecutive sequelae may compromise the planning, working memory, inhibition, and problem-solving that routines, supervision, school communication, and consistent discipline require. Parental cognitive sequelae have been associated with worsened child behavioural outcomes in the analogous parental TBI literature (Pessar et al., 1993; Uysal et al., 1998).

3.1.3 The physical caregiving role

Parenting is bodily work: lifting, carrying, bathing and comforting young children; driving, attending events and sustaining household function for older ones. The physical domain of PICS, encompassing ICU-acquired weakness, deconditioning, neuromuscular impairment, dyspnoea, and chronic pain, directly limits this work (Herridge et al., 2011; Needham et al., 2012; Ohtake et al., 2018). Return-to-work rates fall substantially below pre-illness baselines (Kamdar et al., 2020), roughly two in five survivors report serious financial stress after discharge (Khandelwal et al., 2018), and family members face high rates of financial hardship (Khandelwal et al., 2025), all of which interact with impaired physical capacity and the well co-parent’s compensatory load. Qualitative work describes family caregivers, most often but not only spouses, assuming routine domestic and personal care after home discharge (Choi et al., 2018); parenting-specific data remain essentially absent.

3.1.4 Parental identity and biographical disruption

Parenting is in part the ongoing performance of an identity that is meaningful to the parent and embedded in parent–child functioning (Fadjukoff et al., 2016). Critical illness has been conceptualised as a biographical disruption encompassing loss of self and altered identity after ICU admission (Tembo, 2017), extending Bury’s account of chronic illness as a rupture in biography and self-concept (Bury, 1982).

Parental identity is particularly vulnerable. Mothers, who commonly perform a disproportionate share of childcare and routine caregiving labour (Craig & Mullan, 2011), may experience inability to perform that labour as a disruption to maternal identity rather than only as a medical sequela (Vallido et al., 2010). Fathers may experience identity threats organised around provider and protector roles; direct empirical work is scarce, though qualitative synthesis of fathers caring for a child with a life-limiting illness describes assumed responsibility to lead the family, performance of strength, and suppression of painful emotions (Postavaru et al., 2021), plausibly transferable in part despite the reversed caregiving direction. Gender-stratified parenting data are scarce in the ICU literature; the parental cancer literature suggests heterogeneous gendered patterns in guilt, communication style, and role contraction (Semple & McCance, 2010). Within the cardiac arrest subset, recovery extends beyond physical survival into psychological adaptation, altered identity, and family-supported role reconstruction (Dainty et al., 2021; Harrod et al., 2021).

3.1.5 Parenting self-efficacy and parent–child relationship quality

Parenting self-efficacy, the parent’s belief in their capacity to influence child outcomes, is a robust predictor of warmth, consistency, and child adjustment (Jones & Prinz, 2005). PICS plausibly reduces self-efficacy directly via cognitive and physical limitation, indirectly via mood symptoms, and reflectively via the felt discrepancy between pre- and post-illness self. Family caregivers, most often spouses, commonly absorb the domestic and caregiving work the survivor can no longer do (Choi et al., 2018; Howard et al., 2026). Whether such compensation, protective in the short term, entrenches a ‘diminished parent’ identity is untested: neither study collected data on parenting or dependent children.

Direct relationship-quality data after PICS are essentially absent. The closest evidence comes from the cardiac arrest dyadic subliterature: Rojas et al. (2024) found survivor post-traumatic stress symptoms associated with family distress at one month, and Jensen et al. (2020), in a narrative analysis of older survivor–spouse couples, identified a dyadic mismatch in which survivors emphasised getting back to life as it was while spouses foregrounded anxiety and changed life trajectories (‘we don’t talk about his heart’). Jensen et al. sampled spouses exclusively; Rojas et al. sampled mixed adult family members, 43% of them spouses or partners. Extending an adult dyadic concordance signal to parent–child relationships is an inference rather than a finding.

3.2 Children of PICS survivors

3.2.1 Children visiting the adult ICU: the most direct evidence base

The most direct child-focused empirical evidence concerns children visiting an adult ICU, and it sits within the PICS frame. Lamiani et al. (2021), in a systematic review of five studies involving 141 children and adolescents aged 4–17, concluded that visits could be traumatic but also helped children understand what was happening and preserve family relationships; anxiety and depression symptoms were common regardless of whether children visited. An important qualification runs through this literature: the hospitalised relative is variously a parent, grandparent or sibling, and most studies do not disaggregate. Those that do report the same skew. Hauw-Berlemont et al. (2023) found that of 39 patients with a child aged 1–15 in the close family, 11 were the child’s parent and 22 the grandparent, and that 87% of patients or families found the admission difficult to explain to children. Lamiani et al. (2025), across three Italian adult ICUs, recorded the patient as grandparent for 14 of 25 children aged 7–17, parent for 8, and sibling for 3. That study also supplies the first quantitative test of the visit itself: children who visited after preparation with age-banded booklets showed reduced separation anxiety and no rise in post-traumatic stress symptoms, while separation anxiety among non-visitors was unchanged. Allocation followed family choice rather than randomisation, so this supports rather than settles the inference that exclusion does not itself protect children.

MacEachnie et al. (2018), in a qualitative meta-synthesis of children’s and young people’s experiences of a relative’s critical illness, identified four organising themes, including the parent–child bond and the experience of being overlooked as close family members. Its six included papers, published between 2006 and 2016, comprise three pairs of companion or secondary analyses from the same cohorts, so the effective base is three participant samples. However, those primary studies also overlap those synthesised by Lamiani et al. (2021), so the two reviews are better read as concordant readings of one small literature than as independent confirmation. The authors note that no included study focused solely on parental illness, which is itself informative about the state of the evidence. Belser et al. (2024), interviewing six adults who visited a critically ill family member accompanied by their own children aged five months to 17 years, described shock affecting the whole family and a perception that prepared children cope better. However, the critically ill relative was usually the children’s grandparent, so the findings speak to exposure to adult critical illness generally rather than to parental critical illness. Knutsson et al. (2021) supply the rare counter-case: seven Swedish children aged 6–18, in every instance the patient’s own children, described the visit as making the seriousness comprehensible, and judged the information they were given superficial. Laurent et al. (2019), synthesising seven studies of children visiting adult ICUs, concluded that preparation and support determine whether the visit goes well. The convergent message is that these children need inclusion and explanation rather than silence. That message concerns the admission itself. One study follows these young people past discharge: Fergé et al. (2021), assessing 46 adolescent relatives of 32 French ICU patients one year on, found probable post-traumatic stress in 33%, with a past sense of threat and concurrent anxiety and depressive symptoms the independent correlates. Its participants were first- to third-degree relatives rather than the patient’s children, so it documents persistence after adult critical illness rather than strictly after parental critical illness.

3.2.2 Trauma exposure: secondary traumatic stress and the witnessed-event extreme

Even where the ICU admission was not directly witnessed, children are exposed to its sequelae: the sudden absence of the parent, hospital environments, parental physical changes (tracheostomy, weight loss, scars), and the cascading anxiety of the well co-parent. Secondary traumatic stress is well documented in children of war veterans, first responders, and refugees with PTSD (Lambert et al., 2014; Yehuda & Lehrner, 2018), yet the adult-focused PICS-F construct has no established paediatric counterpart.

At the extreme end of exposure, children of cardiac arrest survivors may witness the arrest itself, since out-of-hospital cardiac arrest most often occurs at home (Sawyer et al., 2020). Approximately one in three adult bystanders show signs of pathological psychological processing weeks afterwards (Brinkrolf et al., 2021), and Dainty et al. (2025) documented a ‘base trauma’ phenomenology in 33 family witnesses, whose accounts include household hypervigilance extending to children even though children were not sampled. A child who directly witnesses a parent’s cardiac arrest meets the DSM-5-TR Criterion A definition through witnessing threatened death (American Psychiatric Association, 2022), and perceived life threat, poor family functioning, and low social support, each plausibly present here, are among the strongest predictors of PTSD in trauma-exposed children (Trickey et al., 2012).

3.2.3 Attachment disruption

Attachment theory (Ainsworth, 1979; Bowlby, 1969) predicts that prolonged separation, parental unresponsiveness, or frightening parental behaviour disrupts the attachment system and shapes the internal working model of relationships. Each is plausibly produced by PICS: hospitalisation entails separation, post-discharge psychopathology and dysexecutive symptoms may reduce sensitivity, and frightening parental behaviour in the sense of Main and Hesse (1990) may arise from PTSD-related dissociation or hyperarousal. Unresolved parental loss or trauma is a meta-analytically established correlate of disorganised attachment (r = .31), whereas frightening or frightened caregiver behaviour rested on only two observational studies and was not pooled, and parental depression as a source of unpredictable parental inaccessibility yielded only r = .06 (van Ijzendoorn et al., 1999). Prolonged parental absence and post-recovery emotional unavailability were not examined as precursors, and remain plausible but untested risks here. For infants, the window is acute (Ainsworth et al., 1978). Nevertheless, direct empirical work in infants of PICS survivors was not identified in our literature review.

3.2.4 Behavioural and emotional outcomes (third-line analogous evidence)

In parental stroke, Visser-Meily et al. (2005) reported elevated internalising or externalising problems in 29% of children one year after the parent’s stroke (54% at the start of inpatient rehabilitation, 23% at two months), with younger child age predicting more externalising behaviour. In the same cohort at three years, daughters reported more stress than sons, and the ill parent’s depressive symptoms were the most consistent correlate and the sole independent predictor of child stress (Sieh, Meijer, & Visser-Meily, 2010). In parental cancer, a systematic review of 28 studies found children’s misadjustment was most consistently predicted by poor family functioning and parental depressive mood, with intrusive symptoms elevated where the parent’s disease was recurrent or advanced; children’s age was not itself a predictive factor (Krattenmacher et al., 2012). In parental TBI, the closest analogue to severe PICS with cognitive sequelae, children show elevated behavioural and emotional difficulties tracking changes in parenting, co-parent distress, and family functioning (Pessar et al., 1993; Uysal et al., 1998). In parental chronic illness more broadly, a meta-analysis of 19 studies found a small elevation in children’s internalising problems (d = 0.23) and a weaker externalising effect that did not survive random-effects modelling (d = 0.09; Sieh, Meijer, Oort, et al., 2010). These stroke and chronic-illness sources are not fully independent, since the Visser-Meily cohort contributes to both Sieh analyses.

These analogues are not interchangeable, and their heterogeneity is itself informative about which illness characteristics matter. At the opposite pole from critical illness, a systematic review of offspring of parents with type 1 and type 2 diabetes found that most studies detected no difference from offspring of healthy parents, and that offspring of parents with diabetes fared better than offspring of parents with metastatic cancer (Landi et al., 2020). This strengthens the case for studying critical illness directly: onset speed, prognostic uncertainty, and visible physical change plausibly drive offspring impact, and PICS sits at the extreme end of each dimension.

3.2.5 Developmental considerations

Effects vary by child age. Infants are vulnerable in the attachment system; maternal post-ICU depression is of particular concern given established effects on child psychopathology (S. H. Goodman et al., 2011). Preschool and school-age children differ in illness understanding and in how distress presents, though age-specific patterns after parental critical illness have not been documented and are offered here as clinical expectation rather than evidence. Adolescents face the developmental demand of individuation against the contradictory pull to remain home as caregiver; Kim et al. (2022), in a Korean national survey of 266 adolescents with a parent diagnosed with cancer and 3,163 controls, found no overall association with depressive mood (adjusted odds ratio 1.29, 95% CI 0.83–2.01) but a significant one where the mother was the diagnosed parent (aOR 1.73, 95% CI 1.10–2.73), without testing the child’s sex and the parent’s sex jointly. Christensen, Frivold, et al. (2026), interviewing 16 relatives aged 12–19 of ICU patients in Norway, eleven of whom were the patient’s child, described independent navigation of adolescence during critical illness, changed relational bonds and adaptation to a new everyday life, and a need for structured adult support rather than treatment as peripheral observers.

A terminological distinction is necessary here, because the evidence base does not observe it. Dependent children, adolescents, young carers, young adult carers, and independent adult children of survivors are different populations with different developmental tasks, and the parenting-relevant construct applies only to the first three. Much of the PICS-F literature that appears to concern ‘children of the patient’ in fact concerns independent adults: in the systematic review most often cited for this risk category, being an adult child of the patient was among the more consistent predictors of caregiver depression, anxiety, and PTSD (Johnson et al., 2019), and the constituent cohorts recruited adult family members. Adult children experience a real caregiver-burden cascade and three-generation family stress, but their situation is not one in which the survivor’s parenting capacity is the outcome of interest. Inferences from that literature to dependent children cross a developmental boundary and are treated as such throughout.

3.2.6 Ordinary helping, young caring, and parentification

Children’s involvement in family recovery should not be read as uniformly harmful. Becker (2007) situates children’s caring on a continuum of caregiving load, running from routine help that is age and culturally appropriate, undertaken by most children, to substantial, regular, and significant caregiving that is age and culturally inappropriate, undertaken by few; the ordering variables are amount, regularity, complexity, time involved, intimacy, and duration. Only about 2–4% of children in the United Kingdom, Australia, and the United States meet the threshold of a young carer, defined as a person under 18 who carries out, often on a regular basis, significant or substantial caring tasks and assumes a level of responsibility that would usually be associated with an adult (Becker, 2007). Becker treats intimacy as one of six ordering dimensions rather than a threshold in itself, but the entry of intimate and personal care into the child’s role is the marker most often held to distinguish young caring from ordinary household help.

Two distinctions matter. First, young adult caring (conventionally 18–24, and to 25 in Australian usage) is an age-defined category rather than a further degree of severity; young adult carers frequently care for partners or their own children, a configuration different from that of children who care (Becker, 2007). Second, parentification is better understood as a relational configuration that can arise from caring than as the severe end of the load continuum: the alteration or removal of boundaries within family structures that occurs when children take on the roles and responsibilities of the adult (Hendricks et al., 2021). Its outcomes are bimodal rather than uniformly pathological, and are conditioned by perceived fairness, felt obligation, family resources and support, cultural normativity, and the ill parent’s capacity to remain aware of the child’s needs. A child may therefore carry a heavy caring load without role reversal, or be parentified at a modest objective load. The relevance of this configuration to trauma-exposed families is not merely theoretical: in offspring of Holocaust survivors, parental PTSD was associated with poorer adult outcomes, with parent–child role reversal, secondary traumatisation, and depressive symptoms acting as mediators (Hoffman & Shrira, 2019). Self-reported childhood parentification is meta-analytically associated with adult psychopathology (Hooper et al., 2011). Theoretical accounts distinguish instrumental parentification, the assumption of adult tasks, from emotional parentification, in which the child becomes the parent’s confidant, the latter generally held to carry greater developmental risk. That differential has quantitative support: in 132 adolescents from 47 families with a chronically ill parent, emotional parentification predicted distress in multilevel models whereas instrumental parentification did not (C. Y. Chen & Panebianco, 2020); qualitative work with adolescents of seriously but not critically ill parents finds both forms present, the emotional the more pervasive (Barr et al., 2025).

Exposure and impact are separable empirically as well as conceptually. The Multidimensional Assessment of Caring Activities (MACA-YC18) and the Positive and Negative Outcomes of Caring Questionnaire (PANOC-YC20) measure amount and effect independently, and comparable caring loads predict divergent outcomes by gender and developmental stage (Joseph et al., 2009). Systematic review documents both negative outcomes in young carers (anxiety, guilt, shame surrounding intimate care, school absence and bullying, internalising and externalising problems) and positive ones (self-esteem, early maturity, closer parent–child relationships), with harm concentrating where caring is unchosen, unshared, unacknowledged, and displaces age-appropriate activity (Chikhradze et al., 2017). Young adult carers also report higher adverse childhood experience burden than non-carer peers, particularly emotional neglect and abuse. Moreover, this burden predicts depression, anxiety, and reduced wellbeing. In turn, caregiving intensity itself was not associated with adversity exposure, and standard instruments do not capture carer-specific events such as responding to a life-threatening health crisis (Landi, Bowman Grangel, et al., 2025).

For children of PICS survivors, this framing generates a specific proposition: risk is unlikely to track survivor impairment severity alone. It should concentrate where sustained caring load coincides with role reversal, obligation, concealment, absent adult alternatives, and unrecognised effort, and where the survivor’s own psychological sequelae reduce their capacity to notice the child’s needs. Concealment is already documented in this setting: among 16 Norwegian adolescent relatives of ICU patients, eleven of them were the patient’s child, every participant withheld their own feelings and needs from their parents on the grounds that the parents had enough to manage (Christensen, Rohde, et al., 2026). Hidden caring may be a particular concern after critical illness, because household reorganisation occurs abruptly, without the gradual visibility of chronic illness, and outside any service configured to look for it.

3.3 Co-parenting and the family system

3.3.1 The well co-parent: PICS-F and the redistribution of parenting labour

Parenting in two-parent families is co-regulated. When one parent is impaired, the family’s parenting capacity depends not only on that parent’s residual function but on the well co-parent’s ability to absorb additional load while remaining functional. Howard et al. (2026) identified five family-caregiver roles in recovery at home (physical care, emotional support, information seeking, monitoring and resource advocacy) that overlap directly with parenting labour and are layered on pre-existing parenting responsibilities; several caregivers described managing a relative’s recovery alongside care of their own children without additional support. Choi et al. (2018) described the abrupt transition from family visitor to active caregiver, with emotional needs that often persisted even when patients regained physical independence. Johnson et al. (2019) documented anxiety, depression, and PTSD prevalences of approximately 20–40% in this population, the canonical PICS-F profile. In turn, Ahn et al. (2025), in a post hoc dyadic analysis of 148 survivor–caregiver pairs, found caregiver burden persisting at 3 and 12 months and correlating cross-sectionally with survivor PICS impairment.

More recent adult ICU data extend this dyadic evidence. In the Australian multicentre PRICE cohort, clinically significant PTSD, depression, or anxiety symptoms were present in 42% of family members of ICU survivors at 3 months and 34% at 12 months, and family members were more likely to report these problems when the paired survivor reported them, supporting a dyadic model of post-ICU morbidity (Rai et al., 2025). Rousseau et al. (2025) showed that survivors and caregivers may not appraise post-ICU impairment in the same way, and that post-ICU consultation may partially realign divergent understandings of recovery. For survivor-parent families, the well co-parent is not only a caregiver but the interpreter of the survivor’s recovery for the children.

The cardiac arrest subset has provided the most quantitative dyadic data. In a prospective cohort of 195 family caregivers, van Wijnen et al. (2017) found high caregiver strain in 15% at twelve months, anxiety in 25%, and depression in 14%. Bohm et al. (2021) showed substantially higher burden in caregivers of cognitively impaired survivors. The DANCAS national survey found 24% of relatives experienced caregiver strain one to five years after the index event, with younger relatives more affected (Hermansen et al., 2024). A scoping review documented persistent uncertainty and unmet needs across discharge and bereavement (Rojas et al., 2023). The implication for parenting is direct: the well co-parent’s distress is itself a determinant of child outcomes, and family-systems theory (Minuchin, 1974) predicts that strain in the parental dyad propagates to parent–child and sibling subsystems via boundary distortion, scapegoating, or coalition formation.

3.3.2 Family communication, the shared narrative, and negotiating a new normal

How families construct a shared narrative of the illness is conceptualised as central to adjustment in adult and child members alike. Rolland’s (1999) family-systems-illness model identifies that narrative as a central family task. ICU events are particularly difficult to narrate: the patient often has fragmentary memory due to delirium and sedation, the well co-parent has emotionally over-saturated memory, and the child has variably accurate, sometimes phantasmatic, perception. ICU diaries were developed partly to address this, though a multicentre trial across 35 French ICUs found no significant reduction in patient PTSD symptoms at three months (risk difference −4%, 95% CI − 15% to 6%; Garrouste-Orgeas et al., 2019) and the evidence base is mixed. Their utility as a child-facing artefact is less examined but theoretically promising.

Less open and more problem-laden family communication has been associated with worse child adjustment in the parental cancer literature, though on a small and mixed evidence base (Krattenmacher et al., 2012). Within ICU survivorship, the cardiac arrest meta-ethnography by Southern et al. (2024), synthesising 32 qualitative studies, captured the shared survivor and key-supporter process as negotiating a new normal, with key supporters specifically experiencing emotional turmoil, becoming a carer, and engaging with a new and unknown world. Similar themes are likely to characterise the wider PICS population.

3.3.3 Sibling effects

Parental attention is unevenly allocated in families with multiple children, the youngest or most demanding often receiving residual capacity while older children absorb more household and emotional labour and may perform second-parent functions, with both protective and risk implications. The PICS-p framework treats siblings as part of the affected family system (Manning et al., 2018); empirical sibling-outcome data after adult critical illness remain limited.

3.3.4 A counterpoint: not all families fragment

The evidence should not be read in a deficit-only direction. Ahlberg et al. (2023), in a Swedish study of 60 former ICU patients and 85 family members, found most families rated functioning and hardiness as healthy, with small variation overall though greater within families than between them. The study concluded that families nonetheless need continuing information, communication support, and help developing coping strategies. Critical illness does not mechanically produce family dysfunction; it produces a renegotiation in which prior functioning, resources, communication patterns and the survivor’s impairment mix all moderate outcomes. Which PICS-survivor families need targeted intervention is not yet operationalised.

3.4 Where parenting falls between services

Institutional support for parenting after PICS varies markedly across the recovery trajectory, and the central observation is structural: parenting is too relational for medical follow-up, too medically determined for ordinary parenting support, too child-focused for adult rehabilitation, and too adult-linked for child mental-health services. It therefore falls between four systems, none of which is fully accountable for it. Table 1 sets out what exists at each phase across three anglophone health systems and where the parenting-specific gap sits.

During acute admission, whether a survivor family is recognised as having a parenting situation requiring structured response, including household composition and dependent children’s care arrangements, depends in practice on whether a bedside biopsychosocial assessment by social work or allied health is completed and acted on (Cagle & Bunting, 2018; Hartman-Shea et al., 2011). Visiting policies, which set the conditions under which children encounter their hospitalised parent, are typically determined at unit rather than national level (Intensive Care Foundation, n.d..). Where dependent children are present, statutory child-safeguarding frameworks overlay clinical practice, and screening conducted with the well co-parent can generate child-protective pathways the medical team alone is not positioned to initiate (Australian Institute of Family Studies, 2026; Department of Communities and Justice, n.d.). Practical supports bearing on parenting include emergency care for children when the patient is the primary caregiver, and interpretation of medical communication on the child’s behalf; for culturally and linguistically diverse families that brokerage runs through interpreter and cultural liaison services, since a co-parent who has not understood the information cannot transmit it to a child. Practices vary by hospital, state, and funding pathway: Table 1 is illustrative rather than a description of uniform national practice. Family members during admission carry substantial psychological risk (Davidson et al., 2012; Johnson et al., 2019; Smith et al., 2025), but the dependent-child subset is rarely identified separately.

Post-ICU support is markedly less integrated. International models include ICU recovery clinics and structured screening for PICS domains (Mikkelsen et al., 2020; Sevin et al., 2018). Cardiac arrest survivorship guidance goes further on family inclusion, with the American Heart Association scientific statement recognising co-survivors (Sawyer et al., 2020) and naming key supporters as part of the recovery population (Resuscitation Council UK [UK], 2024). In turn, survivors and families endorse a need for earlier, family-inclusive follow-up (Mion et al., 2021; Rojas et al., 2023). For PICS more broadly, family-inclusive follow-up is less standardised and parenting-specific screening is not well described. Dedicated ICU follow-up clinics remain uncommon in Australia, with only two of 107 responding ICUs operating one and a single service dedicatedly funded (Cook et al., 2020). The discontinuity is predictable: families are visible during admission, when the survivor is not expected to parent, and invisible after discharge, when parenting demands intensify.

The three systems in Table 1 each create a different opportunity for parenting-specific work. Patient-led charity infrastructure (ICUsteps, n.d.) is well placed to normalise the survivor-parent experience and surface hidden young caring; a professional collaborative of recovery centres (Society of Critical Care Medicine, n.d..) to standardise screening instruments and referral thresholds; and publicly funded allied health accessed through general-practice care plans, alongside federal carer supports (Australia, 2023; Gateway, n.d.), to deliver family intervention at low marginal cost to the household. We did not identify parenting-specific screening as standard in any of them. On this illustrative comparison, Australia appears comparatively well placed to deliver such a pathway through its allied-health infrastructure, yet less likely to identify eligible families, because dedicated ICU follow-up remains uncommon.

3.5 Selected studies summarising the evidence base

Tables 2 and 3 summarise the primary PICS and adult ICU evidence and the complementary cardiac arrest studies, each reporting family-member relationship, patient sex, ages of any children included, and whether parental status or dependent children were recorded.

The pattern is the most direct evidence in this review for the gap it claims. Of the twenty-three studies, fourteen are adult ICU, PICS or cardiac arrest survivorship studies, and not one of these recorded the parental status of the ill adult or the presence, number, or ages of dependent children at home. Patient sex is reported in most of them, but in no case is a patient identified as a parent. Where family members were sampled, spouses and partners dominate, with adult children and parents of the patient the other substantial categories. In one study the relationship was not recorded at all. The remaining nine studies were designed around children, and even there parental status is recorded unevenly: three record the patient’s relationship to each child, one admitted only children of the patient by eligibility, one records the parental status of the accompanying adult rather than of the patient, and four report only broad relationship categories without counts. Where the relationship is recorded, the patient is frequently a grandparent rather than a parent. In one national survey, the only significant sociodemographic predictor of caregiver strain was younger relative age, attributed in discussion to relatives with children at home, a variable the survey did not collect (Hermansen et al., 2024).

4. Discussion

4.1 Integration across frameworks

Triangulating these literatures yields a multilevel account. PICS produces mood, anxiety, trauma, cognitive and physical sequelae that pressure the foundations of parenting capacity; at the dyadic level these alter sensitivity, predictability and physical availability, where the only direct evidence remains the small cardiac arrest dyadic subliterature, which sampled adult family members, predominantly but not exclusively spouses (Jensen et al., 2020; Rojas et al., 2024). At the family level the well co-parent’s PICS-F and the redistribution of caregiving labour reorganise the household (Howard et al., 2026; Southern et al., 2024), though neither study measured child outcomes. Ecological conditions moderate whether that reorganisation is sustainable, and transmission is likely to run through behavioural and relational pathways rather than the biological mechanisms emphasised in intergenerational-trauma models (Yehuda & Lehrner, 2018), with cognitive impairment distinguishing this population from interpersonal-trauma populations.

4.2 Theoretical implications

Three theoretical implications follow. First, PICS-F as operationalised does not address dependent children as a distinct affected group, and its evidence base concerns adults. Extending it requires a developmentally specific construct, we therefore propose PICS-F-Children as a preliminary organising framework, structurally analogous to PICS-p, for research, screening, and clinical formulation. We emphasise that it is a heuristic construct and not a validated syndrome or diagnostic entity: it has not been operationalised, its boundaries have not been tested against existing paediatric trauma and adjustment constructs, and its incremental validity over generic measures of child distress is unknown. Its proposed content, offered as hypotheses rather than criteria, includes intrusive imagery of the medical scene, hypervigilance to parental health cues, separation anxiety, sleep disturbance, somatic complaints, avoidance of hospitals, and role reorganisation within the household. Table 4 consolidates the domains for assessment.

Second, parenting should be considered an outcome domain in PICS research alongside the existing physical, cognitive, and psychological domains, with validated tools for parenting self-efficacy, parenting stress, and parent–child relationship quality added at low burden to existing ICU follow-up cohorts. Third, the cardiac arrest subset demonstrates that, for some PICS aetiologies, the index event itself is a relevant unit of analysis for the family: witnessed arrest is a discrete child-trauma event not adequately captured by post-discharge frameworks.

4.3 Clinical implications

The clinical task is to make parenting visible at every contact point. ICU follow-up clinics should consider screening for parental status and offering family-component pathways: brief parenting-focused screening for the survivor, PICS-F screening for the well co-parent, and routed paediatric mental-health screening for children. Mion et al. (2021) found that 95% of family members in their cardiac arrest follow-up survey wanted dedicated family follow-up, a signal that likely extends to the broader PICS population. For psychologists in critical-care liaison or rehabilitation, useful prompts operationalise family-systems thinking: ‘What parenting tasks feel easy now, and what feel hard?’; ‘How has your partner’s role changed?’; ‘What do your children know about what happened?’; ‘Are any children taking on more responsibility than feels right for their age, and does anyone thank them for it?’ The last operationalises Section 3.2.6 (Ordinary helping, young caring, and parentification): what matters is not whether a child helps but whether the help is chosen, shared, acknowledged, and age-appropriate.

The intervention literature for offspring of seriously ill parents provides both a warrant and a caution. A meta-analysis of 27 randomised trials involving 3,590 young offspring of parents with serious physical or mental illness found small but reliable benefits of psychosocial intervention on overall psychological adjustment (d = 0.17) and internalising problems (d = 0.13) at post-intervention, with larger effects on positive adjustment (d = 0.36) that strengthened at short-term follow-up (d = 0.62). Effects on externalising behaviour, family functioning, and social functioning were null, and gains were not sustained beyond fifteen months (Landi, Pakenham, et al., 2025). Group formats outperformed individual delivery at follow-up and shorter interventions outperformed longer ones. Type of parental illness did not moderate efficacy, although that test was available only at post-intervention and was underpowered, so it licenses transfer to a new illness context only weakly. Two features matter for PICS. First, only six of the 27 trials addressed parental physical illness, and those covered only HIV and cancer, both chronic and neither acute in onset; none addressed offspring of a parent with sudden, high-uncertainty critical illness. Second, the authors’ recommendations, a whole-family approach coordinating assessment for patient and family together and booster sessions to sustain gains, together with their observation that the family structure of adult patients is often not collected in clinical settings, transpose directly to ICU follow-up. A further boundary is that the trials enrolled offspring aged 9 to 25, so this evidence speaks to school-age children, adolescents, and young adults, not to infants or preschoolers.

Plausible starting components are therefore psychoeducation about predictable PICS sequelae, developmentally appropriate family communication, coping-skills work, and practical routine repair, staged early; trauma-focused intervention where survivor PTSD or family-member trauma is prominent, since survivor post-traumatic stress symptoms specifically, rather than impairment severity, were associated with family distress (Rojas et al., 2024); and neuropsychological input where cognitive impairment limits compensatory strategies. On present evidence, however, the expected effect of any such package is modest, and its durability beyond a year is unproven.

4.4 Limitations of the evidence base

The largest limitation is the scarcity of direct evidence, documented in Section 3.5: parental sensitivity, attachment security, parental reflective functioning, discipline style, and parent–child relationship quality were not measured in any identified adult ICU survivorship cohort, and ICU databases do not routinely record whether patients have dependent children. Most claims rest on extrapolation from the cardiac arrest dyadic subset, itself small, and from third-line analogues. This last analogue body, however, differs in important ways: parental cancer typically involves a longer trajectory, parental TBI a more focal cognitive sequela, parental stroke skews older, parental diabetes is chronic and predictable, and the parental PTSD literature is dominated by interpersonal trauma. Two further inferential boundaries deserve emphasis. The PICS-F evidence concerns spouses and independent adult children, so its extension to dependent children crosses a developmental boundary; and the studies of children visiting adult ICUs frequently sample children whose hospitalised relative was a grandparent, so they establish exposure to adult critical illness rather than to parental critical illness.

Methodological limitations of this review include its narrative, non-systematic nature; English-language and high-income-country bias; limited representation of fathers as survivors, and of single-parent, same-sex-parent, blended, kinship-care, and culturally diverse families; and inconsistent timing across studies. The evidence hierarchy used here ranks sources by proximity to the target phenomenon and does not substitute for prospective PICS-parenting cohorts.

4.5 Future research directions

Parental status (presence, number and ages of dependent children at home) must become a routine ICU data element, which no cohort in Tables 2 and 3 currently records. Three designs would then address the core gaps.

The priority design is a prospective dyadic cohort co-located within an existing ICU follow-up clinic (Mikkelsen et al., 2020; Sevin et al., 2018), recruiting adult ICU survivors with at least one dependent child at home and assessing the survivor-parent, the well co-parent, and each child at ICU discharge and at 3, 12, and 24 months. Low-burden instruments include the Parenting Sense of Competence Scale (Johnston & Mash, 1989) and the Parenting Stress Index–Short Form (Abidin, 2012) for the survivor-parent, alongside the SCCM consensus screening recommendations (Mikkelsen et al., 2020), and the Hospital Anxiety and Depression Scale (Zigmond & Snaith, 1983) and PTSD Checklist for DSM-5 (Blevins et al., 2015) for the co-parent. For the child, age-stratified measurement is essential: developmental screening and attachment observation in infants and toddlers; the Strengths and Difficulties Questionnaire (R. Goodman, 1997) and Child Behavior Checklist (Achenbach & Rescorla, 2001) in preschool- and school-age children; and the Children’s Revised Impact of Event Scale (Perrin et al., 2005) in adolescents, indexed to a witnessed-index-event subgroup. Caring activity and its outcomes should be measured separately, using the MACA-YC18 and PANOC-YC20 (Joseph et al., 2009) with an explicit parentification measure, so that ordinary helping, young caring, and role reversal are distinguished empirically. This design would also test whether the PICS-F-Children profile is distinguishable from generic child distress.

The second design is a cross-sectional dyadic study at 6–12 months post-discharge, sized for actor–partner interdependence modelling, to test survivor–family symptom concordance across the broader PICS population as Rojas et al. (2024) did for the cardiac arrest subset, and whether survivor PTSD specifically predicts child outcomes. The third is a pilot intervention trial adapting a family-meaning-making protocol into the ICU follow-up clinic, informed by the format findings of Landi, Pakenham, et al. (2025): group delivery, short duration, and a planned booster to address the observed decay of effects, with the child Strengths and Difficulties Questionnaire at six months as the primary outcome. Cross-cutting questions include gender, culture, and socioeconomic moderators; the contribution of co-parent PICS-F net of survivor impairment; and biological transmission mechanisms. Together, these studies would operationalise the PICS-F-Children construct and shift parenting from an inferred to a measured outcome of adult critical illness.

5. Conclusion

Parenting is a multi-domain capacity, and PICS is a multi-domain disruptor. The mapping between the two is close enough that the absence of any identified PICS-parenting outcome literature is critical. None of the adult ICU, PICS or cardiac arrest cohorts tabulated here records whether its patients are parents. The most informative signals available, the cardiac arrest dyadic subliterature and the adult-ICU children-visiting studies, document distress on each side of the survivor–family relationship and the value of prepared inclusion, but measure no reciprocal parent-to-child effects. Convergent third-line evidence suggests that children of PICS survivors are an under-recognised at-risk population, with risk concentrated where family reorganisation shifts sustained, unacknowledged caring onto them. Treating parenting disruption as a substantive outcome of critical illness is the conceptual move this field still needs.