It’s a typical Friday night in emergency department (ED) — the kind where the corridor feels more like a game of Tetris than a functioning workspace. You’ve already reviewed three wheezers, two toddlers with gastro, and a teenager with “abdo pain, ?appendix.” You’re halfway through your coffee when you get asked to see Sophie, a 2-year-old brought in with a fever and reduced intake.
The triage note reads: “Mum worried — child not herself. Low-grade fever. Soft abdomen. Wet nappies. Alert.” You steel yourself for what might be the tenth viral illness of the shift.
But as you enter the cubicle, Sophie doesn’t bounce into your lap or reach for your stethoscope like most toddlers. She sits slumped in Mum’s arms, quiet, clingy, avoiding eye contact. Her obs are within normal limits — just on the edge of tachycardic, but not quite enough to trigger escalation.
You chat to Mum, expecting to reassure her. But instead, she leans forward and says, “I don’t know what it is, but she’s just not right. She’s never like this. I feel like something bad’s going to happen.”
You pause. You’ve heard that before. Sometimes it’s nothing — but sometimes, it isn’t.
The nurse asks the standard new question, now part of routine obs: “Are you worried your child is getting worse?” Mum doesn’t hesitate. “Yes.”
It’s documented. But what does that “yes” really mean? Does it hold clinical weight? Or is it just another checkbox in an overcrowded notes section?
A few hours later, Sophie’s heart rate climbs, her oxygen saturations dip, and her GCS drops a point. She ends up being admitted to the Paediatric Intensive Care Unit (PICU). You’re left wondering: could you have predicted this based on Mum’s gut feeling alone?
Background
Ask any paediatric trainee what to do when a parent says, “I’m worried”, and the answer is instinctive: take it seriously. But despite this deeply ingrained principle — that parents know their children best — formal recognition of caregiver concern as a measurable, actionable clinical sign has lagged behind.
Delayed recognition of deterioration remains a leading cause of preventable morbidity and mortality in children. In the UK, Why Children Die (CEMACH, 2008) highlighted the failure to identify early signs of serious illness as a key theme in avoidable child deaths. Reviews from the National Confidential Enquiry into Patient Outcome and Death (NCEPOD) have reinforced this, calling for better early warning systems and improved communication with families.
Australian coronial reports echo these concerns. In one well-known case, a mother voiced increasing anxiety about her child’s condition repeatedly — documented but unacted upon — before the child died of undiagnosed sepsis. These are not isolated tragedies. They reveal a system that often hears but doesn’t listen.
In response, initiatives like Ryan’s Rule in Queensland and Martha’s Rule in the UK were introduced, empowering families to escalate concerns when they feel ignored. However, these systems are reactive, relying on families to assert themselves, often in intimidating clinical environments where escalation pathways may not be clear. These rules are necessary but not always sufficient.
Proactively incorporating parental concerns into clinical reviews is not a new idea — many paediatric early warning scores (PEWS) already include a checkbox for parental or staff concerns. However, how this is measured, used, and responded to varies widely.
A 2022 systematic review and meta-analysis by Chong et al. examined whether PEWS systems, as implemented across inpatient paediatric care, actually reduced critical events. Across 15 studies and over 580,000 admissions, PEWS use was associated with reduced mortality and fewer unplanned code events — but there was wide variability in outcomes and systems. Crucially, very few studies isolated caregiver concern as an independent factor.
This study does exactly that.
Erin Mills, Phyllis Lin, Mohammad Asghari-Jafarabadi, Adam West, Simon Craig,
Association between caregiver concern for clinical deterioration and critical illness in children presenting to hospital: a prospective cohort study, The Lancet Child & Adolescent Health,
2025.
https://doi.org/10.1016/S2352-4642(25)00098-7.
Summary of the Study
This Australian prospective cohort study asked a simple question:
Over 26 months, clinicians at a tertiary paediatric ED in Melbourne embedded one proactive question into routine vital sign observations:
This wasn’t a vague chat at the bedside. It was a structured, repeated prompt — asked before vital signs were recorded, translated when needed, and documented in the electronic medical record (EMR). If caregivers said “yes,” the response was clarified to ensure it related to clinical deterioration.
Rather than relying on families to initiate escalation, this approach brought caregiver concern into the clinical record by design.
Population
Children under 19 years presenting to the ED between November 2020 and December 2022.
Inclusion required at least one complete set of post-triage vital signs and a documented response to the caregiver concern question.
Inclusion criteria
- ≥1 set of post-triage vital signs (at least three of HR, RR, BP, SpO₂, or conscious state)
- A recorded response to: “Are you worried your child is getting worse?”
Exclusion criteria
- Transfer to another hospital (outcome data not available)
- No post-triage observations
- No caregiver present
Intervention
A documented “Yes” response to the concern question, recorded as part of routine obs.
Comparator
Children with a “No” response or no documented concern question.
Outcomes
- Primary: PICU admission
- Secondary: Mechanical ventilation, death during admission
Adjustments were made for key confounders, including age, triage category, abnormal vitals (as per Medical Emergency Team (MET) criteria), socioeconomic status (Index of Relative Socio-economic Advantage or Disadvantage (IRSAD)), language, and Indigenous status.
Results and Key Findings
From 73,845 eligible presentations:
- 24,239 (33%) had at least one documented caregiver concern response
- 4,721 caregivers (19.5%) expressed concern for clinical deterioration
Children whose caregivers expressed concern were:
- Four times more likely to be admitted to PICU (Unadjusted OR 4.04; Adjusted OR 1.72, 95% CI 1.40–2.11)
- Nearly five times more likely to require mechanical ventilation (Unadjusted OR 4.60; Adjusted OR 1.85, 95% CI 1.36–3.15)
- Slightly more likely to die during admission, though this was not statistically significant after adjustment (Adjusted OR 1.02, 95% CI 0.18–5.81)
In 1,898 cases where concern and abnormal vitals were present, caregiver concern preceded the abnormal vital signs in 19% — with a median lead time of 7.3 hours.
Only 4.7% of all caregiver concern responses were “Yes,” suggesting that families were selective, not alarmist. Concern had a positive predictive value of 6.9% for PICU admission and a negative predictive value of 98.2%.
This wasn’t just correlation. It was a clinically meaningful signal — and often an early one.
Why This Matters
Asking one structured question — “Are you worried your child is getting worse?” — could add hours of clinical foresight.
Critical Appraisal (Using the CASP Checklist for Cohort Studies)
This study stands up well to critical appraisal using the Critical Appraisal Skills Programme (CASP) framework. Here’s how it performs against each of the 12 questions:
1. Did the study address a clearly focused issue?
Yes. The researchers asked whether caregiver concern — when proactively documented — was associated with critical illness in paediatric patients and whether this association was independent of traditional predictors like abnormal vital signs. The clinical question is directly relevant to front-line paediatrics and has clear implications for escalation systems and early warning tools.
2. Was the cohort recruited in an acceptable way?
Yes. The cohort included all eligible children presenting to a single tertiary paediatric emergency department over two years. Inclusion was based on routine clinical data collection: patients needed at least one set of post-triage vital signs and a documented caregiver concern response. This method was pragmatic and reduced selection bias. However, the requirement for vital signs after triage means that healthier children (discharged quickly or never had repeat observations) were excluded — potentially skewing the cohort toward a more unwell population.
3. Was the exposure accurately measured?
Yes. Caregiver concern was assessed using a standardised, repeated question: “Are you worried your child is getting worse?” This was asked by clinical staff as part of routine vital sign collection before measuring physiological parameters. If caregivers said yes, staff asked scripted follow-up questions to determine whether the concern was clinical. These responses were then recorded in a structured EMR field, minimising variability in interpretation. This is a significant strength of the study — caregiver concern was no longer anecdotal but documented and analysable.
4. Was the outcome accurately measured?
Yes. All outcomes were objective and obtained from the electronic medical record. These included PICU admission, mechanical ventilation, and death during the same admission. The hospital’s Medical Emergency Team (MET) criteria were used to define abnormal vital signs based on age-specific thresholds. These outcomes are clinically meaningful and not open to interpretation, increasing the reliability of the findings.
5. Have the authors identified all important confounding factors?
Yes. The multivariable analysis included adjustments for key demographic and clinical factors associated with deterioration. These included:
- Age
- Gender
- Socioeconomic status (IRSAD quintile)
- Indigenous status
- Primary language
- Country of birth
- Mode of arrival (e.g. ambulance)
- Australasian Triage Scale (ATS) category
- Abnormalities in heart rate, respiratory rate, oxygen saturation, blood pressure, and conscious state
While the study did not explicitly account for medical complexity or underlying chronic illness, the wide range of adjustments makes it unlikely that the observed association is purely confounded.
6. Was the follow-up of subjects complete and long enough?
Yes. The entire hospital admission was used to track outcomes. Patients transferred to another hospital were excluded, as their outcome data couldn’t be verified. This exclusion was clearly reported and appropriate. The length of follow-up was sufficient to capture all PICU admissions, ventilation episodes, and in-hospital deaths.
7. What are the results?
Caregiver concern for clinical deterioration was strongly associated with serious clinical outcomes:
- PICU admission: Adjusted OR 1.72 (95% CI 1.40–2.11)
- Mechanical ventilation: Adjusted OR 1.85 (95% CI 1.36–3.15)
- Death: No significant association after adjustment (Adjusted OR 1.02, 95% CI 0.18–5.81)
In 19% of cases where caregiver concern and abnormal vital signs were present, the concern came first — with a median lead time of over 7 hours. This highlights caregiver concern as a potential early signal for deterioration, not just a response to observed clinical change.
Only 4.7% of all caregiver concern responses were “Yes,” suggesting that families were selective and appropriate in voicing concern — not simply anxious.
8. How precise are the results?
The results are precise and statistically robust. Confidence intervals for PICU admission and ventilation were narrow and did not cross 1, indicating a significant association. The models used cluster-robust standard errors to account for repeated presentations, and sensitivity analyses supported the findings.
9. Do you believe the results?
Yes. The results are consistent with clinical intuition and prior qualitative literature on parental concern. The methodology was strong, with a large sample size, objective outcome measures, and appropriate statistical adjustment. The finding that caregiver concern can precede abnormal vitals adds credibility and makes physiological sense — parents are often the first to notice subtle changes in behaviour, feeding, or energy levels.
10. Can the results be applied to your local population?
Yes, likely. The study was conducted in a high-income, urban tertiary paediatric hospital—a setting similar to many large UK, Irish, and other international paediatric centres. While some variation in patient mix and documentation systems exists, the core concept—proactively seeking and acting on caregiver concern—is broadly applicable and potentially scalable. This may also have utility in resource-limited settings.
11. Do the results of this study fit with other available evidence?
Yes. Previous studies, particularly in sepsis and paediatric oncology, have demonstrated the value of caregiver concern as a diagnostic red flag. What this study adds is generalisability — it shows that concern predicts deterioration across a wide range of presentations, not just within one disease group. It also moves beyond feasibility or acceptability, providing evidence of predictive power.
12. What are the implications for practice?
The implications are significant. Caregiver concern is more than anecdote — it’s data. This study supports embedding a proactive question about concern into routine observations, not as a courtesy but as a clinical tool. Escalation systems should respond to this input in the same way they do to abnormal vitals or triage scores. If families are to be genuine partners in care, we need to listen in a structured, systematised way — and act when they speak up.
Implications for PEWS and Practice
This study reinforces what many paediatric clinicians already suspect — that caregiver concern isn’t just a “soft sign,” but a measurable, actionable predictor of serious illness. The next challenge is incorporating this insight into how we monitor and respond to deteriorating children.
Most PEWS tools used in Australia, the UK, and internationally already include some version of “parental concern” — but how that concern is elicited, recorded, and responded to varies widely.
This study suggests several changes worth considering:
Ultimately, embedding caregiver concern into escalation pathways isn’t just about listening — it’s about operationalising trust in families as part of the safety system.
Limitations
As strong as this study is, a few limitations are worth noting:
- Single-centre setting: The study was conducted at one large tertiary paediatric hospital in Australia. While the findings are compelling, generalisability to smaller, regional, or international settings may vary.
- Selection bias: Inclusion required at least one full set of post-triage observations. Children with minor illness or very short stays (who may not have had obs recorded) were excluded, potentially skewing the cohort toward more unwell patients.
- Caregiver concern may not always be recorded: Although over 24,000 presentations had a documented response, it’s unclear how often the question was skipped or inconsistently applied — especially in time-pressured or low-acuity cases.
- No adjustment for medical complexity: The study adjusted for many demographic and clinical factors but did not include underlying comorbidities, which could influence caregiver perception and illness trajectory.
- Documentation ≠ escalation: Documenting a caregiver’s concern doesn’t mean it triggered a clinical response. We don’t yet know how often concern led to meaningful action—or how outcomes may have differed if it had.
Despite these limitations, the study provides robust, real-world data and opens the door to refining how we incorporate caregiver insight into safe, responsive systems.
Clinical Bottom Line
This study shows that caregiver concern is a reliable, independent predictor of critical illness in children. It often comes before the numbers, and it matters.
If we’re serious about early recognition of deterioration, we must start treating caregiver concern as signal, not noise.
Infographic
Authors’ Comments – Simon Craig and Erin Mills
During our paediatric training we are frequently reminded: ‘If the parent is worried, you should be worried’ and ‘A high level of parental concern is a red flag for serious illness’. With that in mind, we set out to systematically include the voice of parents and caregivers into paediatric patient monitoring systems. How do you distinguish the highly worried parent from the many who are concerned enough to bring their child into the emergency department? Despite the intuitive logic, there was surprisingly little evidence supporting the idea that a parent’s concern was associated with serious illness in the child.
The question ‘Are you worried your child is getting worse?’ was co-designed with parents and carers to specifically identify concerns about clinical deterioration. While all concerns warrant attention, our results demonstrate that concerns are strongly and independently associated with clinically important outcomes: intensive care admission, mechanical ventilation and death. Therefore, parental concern about deterioration requires a clear escalation pathway including timely clinical review. That review should prompt clinicians to re-evaluate the diagnosis, consider possible complications, and identify whether there is something else that has been overlooked.
This paper reinforces what many paediatricians and parents already instinctively understand: parents are the experts in their child, and their intuition – or ‘gut feeling’ – is often accurate. As clinicians, we should actively harness this expertise. By explaining differential diagnoses, red flags and the expected course of illness, we equip families with the tools they need to partner with clinicians in decision making and assist with rapid identification of deterioration.
References
- Confidential Enquiry into Maternal and Child Health (CEMACH). Why Children Die: A Pilot Study 2006. London: CEMACH; 2008.
- National Confidential Enquiry into Patient Outcome and Death (NCEPOD). Time to Intervene? A Review of Patients Who Needed Advanced Airway Management. London: NCEPOD; 2020.
- Queensland Health. Ryan’s Rule. Available from: https://www.health.qld.gov.au/ryansrule
- NHS England. Martha’s Rule. NHS England; 2024. Available from: https://www.england.nhs.uk/patient-safety/marthas-rule/
- Van den Bruel A, Haj-Hassan T, Thompson M, Buntinx F, Mant D. Diagnostic value of clinical features at presentation to identify serious infection in children in developed countries: a systematic review. Lancet. 2010;375(9717):834–845.
- Van den Bruel A, Bartholomeeusen S, Aertgeerts B, Bruyninckx R, Goossens M, Vermeire E, et al. Serious infections in children: an incidence study in family practice. BMC Fam Pract. 2006;7:23.
- Chong SL, Goh MS, Ong GY, Acworth J, Sultana R, Yao SH, et al. Do paediatric early warning systems reduce mortality and critical deterioration events among children? A systematic review and meta-analysis. Resusc Plus. 2022;11:100262.
- Gill FJ, Leslie GD, Marshall AP. The impact of implementation of Family‐Initiated escalation of care for the deteriorating patient in hospital: A systematic review. Worldviews Evid Based Nurs. 2016;13(4):303–13.
