An Early Warning Model and Nursing Care Bundle for Sepsis‐Induced Myocardial Injury in the Paediatric Intensive Care Unit (PICU): A Study Protocol

The management of paediatric sepsis and its complications represents a significant challenge in the PICU [1]. SIMI, in particular, is a life-threatening condition associated with high mortality [2]. Septic cardiomyopathy, characterised by systolic and diastolic dysfunction induced by the systemic inflammatory response (which involves complex mechanisms such as mitochondrial dysfunction, calcium handling impairment and excessive inflammatory signalling) [3], places the heart at considerable risk [4]. In this context, as highlighted in recent literature [4, 5], the role of the PICU nurse evolves beyond protocol-based care to encompass continuous physiological synthesis, anticipatory intervention and meticulous clinical decision-making. To systematically address this challenge, we have designed a three-phase study protocol aimed at developing an early warning model and a standardised nursing care bundle for SIMI. This article details the protocol and uses an illustrative case to demonstrate the potential application and feasibility of the proposed care principles in clinical practice. It is important to clarify that the target population for the main study is all septic patients in the PICU, not limited to those with trauma. The case of a child with polytrauma is used solely to vividly illustrate the severity and management complexity of SIMI. This aligns with research highlighting the role of inflammatory response in traumatic injury outcomes [6]. This protocol outlines a three-phase, mixed methods study designed to develop and preliminarily evaluate an early warning model and a corresponding nursing care bundle for SIMI in the PICU (see Figure 1 for the overall study flowchart). A single-centre retrospective cohort study will be conducted. Data will be extracted from the electronic health record (EHR) system of the PICU at Guiyang Maternal and Child Health Care Hospital covering a 5-year period from January 1, 2020 to December 31, 2024. (1) Aged between 1 month and 18 years, (2) admitted to the PICU with a diagnosis meeting the Sepsis-3.0 international consensus criteria [1], (3) PICU length of stay > 24 h. (1) Known congenital heart disease or primary cardiomyopathy, (2) inability to access complete medical records, (3) missing > 30% of key predictor variables (e.g., troponin, BNP). Based on historical data from our centre, the incidence of SIMI is approximately 20%. Adhering to the rule of thumb in predictive model development requiring at least 10 outcome events per predictor variable, and planning for up to 10 predictors, at least 100 SIMI events are needed. Therefore, the target total sample size is 500 patients (expected to yield ~100 SIMI events), which is expected to cover eligible cases over the 5-year period. Four categories of variables will be collected: (1) demographics, (2) vital signs and laboratory parameters (e.g., lactate, troponin I, NT-proBNP, inflammatory markers) at PICU admission and serially within the first 24 h, (3) therapeutic interventions (e.g., vasoactive-inotropic score), (4) outcome measures (SIMI diagnosis, 28-day mortality). Data will be extracted by two independent researchers. Variables with > 30% missing data will be excluded from analysis. For variables with ≤ 30% missing, multiple imputation will be used. SIMI diagnosis will be based on specific criteria (e.g., cardiac troponin I elevation > 99th percentile upper reference limit). Following univariate analysis, variable selection will be performed using LASSO regression, and the final predictive model will be constructed using multivariate logistic regression. Model performance will be evaluated via the Area Under the Curve (AUC), calibration curves, and decision curve analysis. Internal validation will be performed using 10-fold cross-validation. Such approaches are aligned with methodologies used to validate paediatric early warning systems in clinical practice [7]. This phase has been approved by the Ethics Committee of Guiyang Maternal and Child Health Care Hospital (Approval No: [2025-89], Date of Approval: 2025 November 17). The use of retrospective, de-identified data qualifies for a waiver of informed consent. This phase aims to develop a structured, evidence-based nursing care bundle for SIMI through a scoping review and expert consensus process. The overall framework for bundle development and implementation is shown in Figure 2. We will first conduct a scoping review of existing guidelines, systematic reviews and high-quality primary studies on SIMI management to identify potential evidence-based nursing interventions. In this study, 15–20 experts will be invited, including paediatric intensivists (5–6), PICU specialist nurses (5–6), paediatric cardiologists (2–3) and clinical pharmacists (1–2). A draft bundle comprising approximately 15–20 items, based on Phase 1 results and the scoping review, will be presented. Experts will rate each item on a 5-point Likert scale for ‘necessity’ and ‘feasibility’ and provide comments. The mean score, standard deviation and level of consensus (percentage of experts rating ≥ 4) will be calculated for each item. The pre-defined consensus criterion is ≥ 80% agreement. The results of the first round will be fed back to the panel, and items not reaching consensus or those modified will be rated again. The research team will determine the final SIMI nursing care bundle based on the results of the two Delphi rounds. A non-randomised, pre–post pilot study design will be used. The final bundle will be embedded into the hospital's electronic health record system as a structured checklist, triggered when a patient is identified as high-risk for SIMI by the Phase 1 model. All PICU nurses will receive a dedicated 2-week training programme, including theoretical sessions and case simulations. Bedside prompt cards and quick reference guides will be developed and distributed. Overall bundle adherence rate (defined as the proportion of applicable items that are fully executed), with a target of > 80%. Comparison of SIMI incidence, duration of vasoactive drug use, and PICU length of stay between the pre-implementation (6 months) and post-implementation (6 months) periods. Focus group interviews will be conducted with nursing staff to understand the acceptability of the bundle and to identify perceived barriers and facilitators to its use. The following case of a child with polytrauma and sepsis is presented solely to illustrate the clinical challenge of SIMI and to provide a concrete context for the proposed nursing care principles. The target population for the main study is all septic patients meeting the inclusion criteria described in Section 2.1.2. The subject was a 3-year, 7-month-old female admitted to the PICU after a high-impact road traffic accident. The initial presentation was consistent with traumatic haemorrhagic shock and respiratory failure. The initial assessment revealed an intubated child with a GCS of 3T. Haemodynamic instability was evident, with tachycardia (HR 164 bpm), borderline hypotension (BP 104/56 mmHg), and cool extremities. A nasogastric tube drained coffee-ground material. Examinations confirmed extensive facial trauma, multiple fractures, a right adrenal haematoma and bilateral pulmonary contusions. Initial laboratory results were significant for a haemoglobin of 68 g/L, markedly elevated inflammatory markers (IL-6 > 1000 pg/mL), and evidence of severe myocardial injury with a troponin T level of 766.8 pg/mL and an elevated NT-proBNP [6, 7]. These findings confirmed the diagnosis of septic cardiomyopathy, establishing cardiac protection as the primary clinical focus (Table S1). The nursing care plan was dynamic, centred on the primary goal of myocardial protection while coordinating support for other compromised systems (Table S2). A primary nursing challenge was optimising ventilatory support without compromising cardiac function. High intrathoracic pressure from mechanical ventilation can reduce venous return and increase right ventricular afterload, and changes that are detrimental to an injured heart is detrimental to an injured heart [8]. Our nursing interventions were therefore guided by close monitoring of dynamic lung compliance (Crs) and oxygenation indices. We adopted a conservative oxygenation strategy that targeted lower SpO2 limits to minimise oxygen toxicity while ensuring adequate systemic oxygen delivery [9]. A declining Crs was immediately communicated to guide adjustments in PEEP and fluid balance. We maintained a meticulous record of intake and output, shifting to a precision-based, restrictive fluid approach to prevent volume overload on the compromised myocardium [10]. The confirmed myocardial injury necessitated a nursing-driven strategy to minimise cardiac workload and oxygen consumption [11]. We instituted a strict ‘cardiac rest’ protocol, clustering nursing care activities to provide long, uninterrupted periods of rest. Sedation and analgesia were titrated specifically to blunt the stress response and reduce catecholamine-driven increases in myocardial oxygen demand. The administration of vasoactive infusions was managed with a dual objective: to maintain a minimum acceptable perfusion pressure while using the lowest effective dose to limit cardiac afterload. Trends in cardiac biomarkers were treated as a direct report on the efficacy of our nursing care [5]. A rising or plateauing troponin level prompted an immediate bedside reassessment. Advanced imaging modalities like speckle-tracking echocardiography further enhance the assessment of myocardial dysfunction in sepsis [12]. In the context of a known intracranial haemorrhage, we carefully balanced the use of vasopressors to support cerebral perfusion pressure against the risk of increasing cardiac strain. The initiation of total parenteral nutrition (TPN) and subsequent continuous renal replacement therapy (CRRT) demanded flawless nursing management to maintain electrolyte and metabolic homeostasis, which is critical for stable cardiac rhythm and function. The patient's recovery was gradual but sustained. Nursing-sensitive indicators reflecting positive progress included the successful weaning and discontinuation of vasoactive support, consistent improvement in peripheral perfusion, and a steady downward trend in troponin T and NT-proBNP levels [13, 14] (Figure S2). The patient was eventually successfully weaned from mechanical ventilation and transferred out of the PICU. The detailed clinical course is provided in Figure S1. This case highlighted that in septic shock, the heart can be a silent victim. We learned that vigilant serial tracking of troponin and BNP was crucial for revealing the true severity of myocardial injury, beyond conventional haemodynamic parameters [2, 13, 15]. This experience underscores the value of biomarker-guided, personalised management, a principle supported by recent sepsis research [16]. Furthermore, this case led us to reframe fundamental nursing tasks as direct therapeutic interventions. Clustering care became a strategy for myocardial rest; titrating sedation was an intervention to lower oxygen demand; and every millilitre of fluid administered was a deliberate choice. This shift from task-completion to physiology-management represents a profound change in practice. The implementation plan for the care bundle (Phase 3) is informed by established implementation science frameworks (see Figure 2 for the implementation framework). We recognise key enablers such as integration into the electronic health record, structured training and leadership support. Potential barriers including workflow disruption and resource constraints will be proactively addressed through nurse champions and iterative feedback cycles. This study protocol has several limitations. Firstly, Phase 1 is a single-centre retrospective study, carrying risks of confounding and information bias; the developed model requires further validation in prospective multi-centre studies. Secondly, the Delphi expert panel was constituted from a single centre, and thus the consensus results may be influenced by our local geographical context and the specific professional backgrounds of the involved experts. Finally, Phase 3, as a pilot study, has a small sample size and lacks randomisation; its primary aim is to assess feasibility rather than confirm efficacy, and results should be generalised with caution. The successful navigation of this child's critical illness was contingent upon a nursing philosophy rooted in integrating physiology, anticipating deterioration and applying clinical reasoning [1, 4]. By moving beyond task completion to become proactive managers of a complex physiological system, the PICU nursing team played a definitive role. This experience validates the role of the critical care nurse as a knowledge worker and informs development of the standardised protocol outlined in this study. We are deeply indebted to our nursing colleagues in the PICU, whose vigilant monitoring and expert execution of the care plan were indispensable. Their frontline insights provided the crucial data for clinical decision-making. We also extend our thanks to the paediatric intensivists, respiratory therapists and nutritionists who formed our collaborative team. Finally, we acknowledge the patient's family for their resilience and partnership in care. The authors have nothing to report. This study was approved by the Institutional Review Board of Guiyang Maternal and Child Health Care Hospital (Approval No: 2025-89; Date of Approval: 2025 November 17). This approval covers the use of all clinical data, including the illustrative case with its laboratory results (Figures S1 and S2). Informed consent for the publication of the case details was obtained from the patient's legal guardian. The need for informed consent for the retrospective phase (Phase 1) was waived due to the use of de-identified data. Informed consent was obtained from the legal guardian of the patient. The authors declare no conflicts of interest. The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Figure S1: Clinical course, major interventions and corresponding physiological trends of the illustrative case. Timeline (x-axis) shows days since PICU admission. Key clinical events (e.g., initiation/weaning of vasoactive support, mechanical ventilation, CRRT) are annotated. Vital signs (e.g., heart rate, mean arterial pressure) and key laboratory trends (e.g., lactate, inflammatory markers) are plotted on the y-axes to demonstrate temporal correlations with interventions. Figure S2: Serial trends in myocardial injury and haemodynamic stress biomarkers for the illustrative case. Biomarker levels (Troponin T, NT-proBNP) are plotted against time (days). The shaded area indicates the period of vasoactive support. Dashed horizontal lines represent institutional upper reference limits. A steady decline in both biomarkers corresponds with clinical improvement and weaning of support. Table S1: Key diagnostic and laboratory data on admission for the illustrative case. CRRT, continuous renal replacement therapy; IL-6, interleukin-6; NT-proBNP, N-terminal pro-B-type natriuretic peptide. Table S2: Dynamic nursing care plan for the illustrative case. This table summarises the problem-focused nursing approach, linking identified clinical problems to core interventions and monitored outcomes. ABP, arterial blood pressure; CVP, central venous pressure; MAP, mean arterial pressure; VAP, ventilator-associated pneumonia; CLABSI, central line-associated bloodstream infection; CAUTI, catheter-associated urinary tract infection; GI, gastrointestinal. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.