Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Dec 17.
Published in final edited form as: J Intensive Care Med. 2025 Apr 21;40(9):964–972. doi: 10.1177/08850666251329902

Strain in the Pediatric Intensive Care Unit and its Association with Patient Outcomes

Matthew S Baker 1,2, Kristen R Miller 1, Jaime S LaVelle 2, Julia A Heneghan 3, Michele M Loi 1,2, Cameron F Gunville 1,2,*, Aline B Maddux 1,2,*
PMCID: PMC12706758  NIHMSID: NIHMS2122699  PMID: 40255081

Abstract

Objective:

To characterize measures of pediatric intensive care unit (PICU) strain and test for associations between strain and patient outcomes. We hypothesized that periods of increased strain would be associated with increased odds of experiencing a post-ICU floor escalation event.

Design:

Retrospective cohort study.

Setting:

Quaternary care children’s hospital PICU.

Study Population:

PICU admissions (2014–2023).

Interventions:

None.

Measurements and Main Results:

We measured PICU strain metrics daily including percent occupancy, percent turnover, and acuity. Percent occupancy and turnover were higher during annual peak viral season versus non-peak season (percent occupancy: median 84.4 [interquartile range (IQR) 72.9, 91.7] versus 68.8 [IQR 56.3, 82.1], P < .001; percent turnover: median 25.0 [IQR 20.0, 31.3] versus 21.9 [IQR 15.6, 28.1], P < .001). Acuity metrics did not differ during these periods. In patients admitted for respiratory or neurologic illness, we used multivariable logistic regression to test for associations between strain metrics on the day of transfer and odds of a floor escalation event defined as an unplanned PICU readmission or rapid response or code blue activation within 48 h of transfer. Of 12 832 patient transfers, 429 (3.3%) experienced a floor escalation event. After controlling for patient and clinical characteristics, percent occupancy and turnover were independently associated with an increase in floor escalation events. This risk was not linear for occupancy: above 87.5% (95% confidence interval (CI): 80.4, 94.6%), per 5% increase in occupancy patients experienced more floor escalation events (odds ratio [OR] 1.29 [95% CI: 1.07, 1.57]). For every 5% increase in turnover, patients experienced more floor escalation events (OR 1.06 [95% CI: 1.01, 1.12]).

Conclusions:

Occupancy and turnover characterized recognized periods of PICU strain. High occupancy and turnover were associated with more floor escalation events. Multicenter studies are needed to evaluate the generalizability of these findings across other PICUs.

Keywords: intensive care unit, pediatric intensive care, critical care, resource utilization, outcomes, pediatrics

Introduction

Pediatric intensive care units (PICUs) frequently face surges brought about by seasonal illnesses, endemics, and pandemics. These surges lead to strain on PICUs, hospitals, and the healthcare system.1–4 While strain can generally be understood as a discrepancy between resources necessary to provide optimal care and demand brought about by increased patient volume and/or acuity, strain metrics in critical care units are often difficult to define and measure due to constant fluctuations throughout each day.5–7

In adult critical care, studies suggest that strain is associated with adverse patient outcomes such as increases in mortality, medication errors, pressure injuries, and readmission rates.8–12 Many of these adult studies were published in response to the COVID-19 pandemic and described the association between high patient volumes and increased mortality.5,9,13–15 The association between PICU strain and outcomes is likely to differ from adult populations due to differences in distribution and types of comorbidities, etiologies of critical illnesses, mortality rates, and lengths of hospitalizations. Additionally, pediatric critical illness is disproportionately cared for in large pediatric referral centers with these patients also having higher rates of underlying medical complexity and subsequent resource intensive needs potentially making these centers more vulnerable to the impact of strain.16–20 Furthermore, the hospital environment surrounding the PICU can also impact the outcomes of PICU patients prior to and after their PICU admission. In this study, we focused on developing a better understanding of PICU strain and its association with patient outcomes immediately after transfer out of the PICU. However, it is key to acknowledge that post-PICU outcomes could be impacted by strain on PICU or inpatient floor resources.

In this study, we aimed to characterize PICU strain based on metrics specific to the pediatric setting and evaluate for an association between strain metrics and patient outcomes. We hypothesized that PICU strain metrics of occupancy, turnover, and acuity would quantify varying aspects of pediatric critical care resource strain and that increased strain would be associated with increased risk of experiencing a floor escalation event defined as an unplanned PICU readmission, rapid response, or code blue activation within 48 h of transfer from the PICU.

Materials and Methods

This study was approved by Colorado’s Multiple Institutional Review Board (#23-1447; Sep. 22, 2023; Associations Between Strain and Outcomes in a Quaternary PICU) as an exempt study and was conducted under the standards of the 1975 Helsinki Declaration. We conducted a retrospective, single center study at a quaternary care pediatric referral center with a 48-bed maximum capacity PICU and up to 3500 PICU admissions annually. We characterized strain metrics using all patients cared for within our PICU between September 1, 2014 and March 31, 2023.

Patient-level data were collected from our site’s VPS (Virtual Pediatric Systems, LLC) database including age, sex, primary and all active diagnoses, date and time of admission, admission source, Pediatric Risk of Mortality III (PRISM III) score,21 support with and duration of invasive mechanical ventilation, support with extracorporeal membrane oxygenation (ECMO) and/or continuous renal replacement therapy (CRRT) and associated dates, placement of central venous catheters and arterial lines while in the PICU including dates of placement and removal, PICU length of stay, time of transfer/discharge from PICU, unplanned readmission to PICU within 48 h of transfer, and hospital length of stay (Supplemental Digital Content [SDC], eTable 1). Operational PICU data were also collected from our site’s VPS database including number of admissions, transfers, and discharges in each 24-h period (0700 to 0659), as well as overall PICU census and bed capacity measured at 0700 daily. Rapid response team and code blue activations were collected from our hospital’s database. All patient and unit operational data collected from VPS and the hospital code blue and rapid response team database were complete.

Daily PICU strain metrics were calculated on a 24-h basis from 0700 to 0659. PICU bed capacity was defined as maximum physical PICU patient beds available each day. PICU bed capacity varied throughout the observational study with it changing from the lowest of a static 32 bed capacity throughout 2014 up to the highest capacity of 32–48 beds in 2023 as a result of increases in inpatient admissions at the institution over time and the addition of seasonal and surge bed flexing. PICU strain metrics were categorized as daily percent occupancy, percent turnover and acuity. Percent occupancy was defined as number of patients in the PICU measured at 0700 daily divided by PICU bed capacity on that day. Percent turnover was defined as number of PICU admissions, transfers, and discharges from the PICU divided by the day’s PICU bed capacity. Acuity was quantified as 2 distinct metrics, both quantified for each day during the study period: 1) average acuity score: number of patients receiving invasive mechanical ventilation, continuous renal replacement therapy, and/or with an arterial line in place divided by total PICU bed capacity and 2) average PRISM III risk of mortality of all newly patients admitted on the day.

Our primary outcome was experiencing a floor escalation event, a composite of an unplanned PICU readmission, rapid response or code blue activation within 48 h of transfer from the PICU.

Statistical Analysis

Median and interquartile range [IQR] were reported for continuous variables and frequency and percentage for categorical variables. We created run charts for each strain metric with 30-day moving averages. We calculated strain metrics and their variability by annual quarter and during known illness surges including the annual respiratory viral peak season (December through February annually), the enterovirus D-68 outbreak (September 2014), and the poly-viral illness surge following the COVID-19 pandemic (October 2022 through April 2023). Coefficient of variation (CV; ratio of standard deviation to the mean) was reported for each period to reflect the amount of relative variability in the metric and was used to compare the dispersion of measurements. Comparisons of medians and CVs across period were performed using Wilcoxon rank sum tests and asymptotic tests for the equality of coefficients of variation, respectively.22

In patients admitted with a primary neurologic or respiratory diagnosis, we evaluated the association between PICU strain metrics measured on each patient’s transfer day and a floor escalation event within 48 h after transfer from the PICU. An early unplanned floor escalation event was defined as occurring during the first 48 h following transfer from the PICU. We selected this timeframe because it has been utilized by previous studies describing unplanned ICU readmissions and reflects the society of critical care medicine’s quality care metric definition of an unplanned ICU readmission.23–26 We selected to perform these analyses in patients with neurologic and respiratory primary admission diagnoses as these patients were known to experience the highest rates of readmission at our institution, allowing for prognostic enrichment of the cohort given the low event rate of our primary outcome. Patients who died while in the PICU or were discharged directly home or transferred to another hospital or another ICU within our institution were excluded as they were not at risk of experiencing a post-ICU floor escalation event. Patient data were collected at the hospitalization level; only the first PICU admission in a hospitalization was evaluated in the patient-specific analyses. We used univariate logistic regression to test for an association between each independent variable and the primary outcome. We assessed for non-linear relationships between each strain metric and the primary outcome using natural cubic splines with 1 (linear) to 5 knots, evenly spaced across the distribution of the independent variable. We chose the functional form of the strain metric by selecting the number of knots that resulted in the lowest Akaike information criterion (AIC) from the natural cubic spline models. For strain metrics that demonstrated a non-linear association, we used the R package segmented to further evaluate for a threshold using the algorithm developed by Muggeo.27 We tested for pairwise interactions between age and primary diagnosis and combinations of strain metrics, decided upon a priori (Supplemental Table 1). A separate multivariable logistic regression model was built for each strain metric. In each model, we included age, chronic complex condition (CCC) status, and PRISM III based on their known association with PICU readmission.28–30 Additional variables with P < .20 in univariable analyses were included in the multivariable model. As a sensitivity analysis, we fit a generalized estimating equation (GEE) with an exchangeable correlation structure for the final multivariable models to evaluate the impact of patients with more than one hospitalization. We used R version 4.3.1 for the analyses (R Foundation for Statistical Computing; Vienna, Austria).

Results

Strain Metrics

The strain metrics demonstrated notable variability throughout the study period (Figure 1; Supplemental Figure 1). Daily occupancy was higher during the annual respiratory viral peak season compared with periods outside of the annual respiratory viral peak season (median 84.4% [IQR 72.9%, 91.7%] vs 68.8% [IQR 56.3%, 82.1%], P < .001) and variability was lower (CV 18.8% vs 26.2%, P < .001) (Figure 2; Supplemental Tables 2, 3). Occupancy demonstrated a similar pattern during the post-COVID poly-viral surge compared with study days outside of the post-COVID-19 poly-viral surge (median 81.3% [IQR 66.7%, 91.7%] vs 71.9% [IQR 56.3%, 84.4%], P < .001 and CV 19.1% vs 25.8%, P < .001) (Figure 2; Supplemental Tables 2, 4). Turnover was also higher during annual respiratory viral peak season compared with other periods (median 25.0% [IQR 20.0%, 31.3%] vs 21.9% [IQR 15.6%, 28.1%], P < .001) and variability was lower (CV 31.1% vs 42.1%, P < .001) (Figure 2; Supplemental Tables 2, 3). Turnover demonstrated a similar pattern during the post-COVID-19 poly-viral surge compared with study days outside of the post-COVID-19 poly-viral surge (median 27.1% [IQR 22.1%, 33.3%] vs 21.9% [IQR 15.6%, 28.1%], P < .001 and CV 32.6% vs 40.5%, P = .002) (Figure 2; Supplemental Tables 2, 4). The enterovirus D-68 surge that occurred in September 2014 demonstrated a similar increase in occupancy compared with study days outside of the enterovirus D-68 surge (median 79.7% [IQR 75.0%, 86.7%] vs 71.9% [IQR 56.3%, 85.0%], P < .01), however percent turnover was not significantly different (Figure 2; Supplemental Tables 2, 5). Acuity strain metrics did not demonstrate the same patterns of variability in comparing respiratory virus surge and non-surge periods (Supplemental Figure 2, eTable 2).

Figure 1.

Figure 1.

Open in a new tab

Volume-based strain metrics (percent occupancy and percent turnover) demonstrated variability during the study period. Shaded bars represent various illness surges including the enterovirus D-68 outbreak (September 2014), annual respiratory viral peak season (December through February annually), and the post-COVID 19 viral surge (October 2022 through April 2023). Dark grey lines represent daily rates, and the black solid line represents a 30-day moving average.

Figure 2.

Figure 2.

Open in a new tab

Volume-based strain metrics (percent occupancy and percent turnover) demonstrated higher strain during known illness surges including annual respiratory viral peak season (Dec-Feb annually), enterovirus D-68 outbreak (Sep 2014) and the post-COVID-19 poly-viral surge (Oct 2022 and April 2023).

Patient-Specific Risk of Floor Escalation Event

For the patient-specific analyses, we excluded 1268 patients including 458 patients who died while in the PICU, 532 patients who were discharged directly home from the PICU, and 278 patients who were transferred to another hospital or another ICU within our institution (Supplemental Figure 3). During the study period, 12 832 patients with a primary respiratory (n = 10 264; 80%) or neurologic (n = 2568; 20%) diagnosis were transferred from the PICU to a general medical floor (Table 1). Of these, 429 (3.3%) had a floor escalation event within 48 h of PICU transfer including 335 (3.3%) and 94 (3.7%) with a primary respiratory or neurologic diagnosis, respectively. Of these 429 patients with escalation events, 50 (11.7%) experienced a readmission only, 148 (34.5%) a rapid response activation only, 4 (0.93%) a code blue activation only, and 227 (52.9%) a combination of escalation events (Supplemental Table 6).

Table 1.

Clinical Characteristics of Patients Admitted with a Primary Respiratory or Neurologic Diagnosis.

Characteristics All (n = 12 832) No floor escalation event (n = 12 403) Floor escalation event (n = 429)
Age at admission (years), median (IQR) 2.38 (1.13, 7.11) 2.37 (1.14, 7.04) 2.79 (0.90, 10.6)
Male sex, n (%) 7372 (57.5) 7119 (57.4) 253 (59.0)
Presence of a chronic complex condition, n (%) 5906 (46.0) 5643 (45.5) 263 (61.3)
Patient Origin, n (%)
Inpatient 3138 (24.5) 2998 (24.2) 140 (32.6)
OR 1121 (8.7) 1098 (8.9) 23 (5.4)
Outpatient 8573 (66.8) 8307 (67.0) 266 (62.0)
Pediatric Risk of Mortality (PRISM) III score, median (IQR) 0.49 (0.34, 0.71) 0.49 (0.34, 0.71) 0.51 (0.34, 0.90)
Invasive mechanical ventilation, n (%) 1787 (13.9) 1711 (13.8) 76 (17.7)
Cardiopulmonary resuscitation, n (%) 35 (0.3) 34 (0.3) 1 (0.2)
Extracorporeal membrane oxygenation, n (%) 22 (0.2) 22 (0.2) 0 (0)
Continuous renal replacement therapy, n (%) 7 (0.1) 7 (0.1) 0 (0)
Arterial line while in PICU, n (%) 971 (7.6) 945 (7.6) 26 (6.1)
Central line placed in PICU, n (%) 627 (4.9) 595 (4.8) 32 (7.5)
PICU length of stay, median (IQR) 1.69 (0.97, 2.99) 1.69 (0.97, 2.97) 1.85 (0.99, 3.65)
Nighttime transfer from PICU, n (%) 1802 (14.0) 1728 (13.9) 74 (17.2)
Open in a new tab

In univariate modeling, older age, presence of a CCC, higher PRISM III score, inpatient origin (compared to operating room), longer PICU stay, support with invasive mechanical ventilation, and placement of a central line in the PICU were associated with having a floor escalation event after discharge (Supplemental Table 7). We identified an interaction between age and primary diagnosis in that older patients admitted with a primary neurologic diagnosis were at increased risk of experiencing a floor escalation event compared with younger patients whereas age was less important in children with a primary respiratory illness (Supplemental Figure 4).

Higher occupancy and turnover on the day of transfer were associated with increased odds of having a floor escalation event in univariate analyses. Natural cubic spline modeling identified that the relationship between occupancy and odds of experiencing a floor escalation event was best modeled in a non-linear fashion using natural cubic splines with 3 knots (Figure 3). The best fitting model for all other strain metrics was linear (Supplemental Table 8). We tested for pairwise interactions between any combination of strain metrics, but none were identified (Supplemental Table 9).

Figure 3.

Figure 3.

Open in a new tab

Probability of experiencing the primary outcome by PICU percent occupancy on the day of transfer for a patient with covariates equal to the reference group (for categorical predictors) and the average values (for continuous predictors).

After adjusting for patient-level confounders and utilizing the best-fitting 3 spline model, we identified that the odds of experiencing a floor escalation event increased significantly above 87.5% occupancy (95% CI: 80.4, 94.6%) (Figure 3). For every five percent occupancy above 87.5%, the odds of experiencing a floor escalation event increased 1.29-fold (95% CI: 1.07, 1.57) (Figure 4). At occupancy levels below 87.5%, there was not an increased risk of experiencing a floor escalation event (OR 1, 95% CI: 0.95, 1.04). In the multivariable model including the same patient-level confounders, turnover was associated with an increased odds of having a floor escalation event (OR 1.06 per 5% increase in turnover [95% CI: 1.01, 1.12]) (Figure 5). For both occupancy and turnover, findings were similar when tested using a generalized estimating equation model accounting for patients with multiple hospitalizations (Supplemental Tables 10, 11). The acuity strain metrics were not associated with an increased odds of experiencing a floor escalation event in multivariable modeling (Supplemental Figures 5, 6).

Figure 4.

Figure 4.

Open in a new tab

Multivariable model evaluating percent occupancy of the PICU on the day of a patient’s PICU transfer and its association with having a floor escalation event within 48 h of transfer.

Figure 5.

Figure 5.

Open in a new tab

Multivariable model evaluating percent turnover of the PICU on the day of a patient’s PICU transfer and its association with having a floor escalation event within 48 h of transfer.

Discussion

PICU strain can be measured using quantifiable metrics including percent occupancy and turnover. The results of this study demonstrated that PICU strain, as measured by volume-based metrics, reflect the surges experienced during known periods of high PICU volume. Percent occupancy captured known periods of strain and is a metric that has demonstrated importance in adult critical care studies evaluating strain. After accounting for patient-level confounders, we found that higher percent occupancy and turnover were associated with an increased odds of a patient experiencing a floor escalation event within 48 h after transfer from the PICU. Notably, the data highlight a critical threshold in percent occupancy at 87.5%, above which there is a dramatic increase in the odds of post-transfer floor escalation events. Overall, these results demonstrated that markers of strain are an important metric to monitor and address to improve patient outcomes.

Strain metrics, including occupancy and turnover, as well as novel scoring systems that account for aspects of patient acuity and volume-based unit characteristics have been studied, primarily in adult critical care settings.8–15 These adult studies demonstrated notable associations between higher strain and adverse patient outcomes including higher mortality and readmission rates. In a multicenter adult ICU study published in 2023, Pilcher et al9 developed the Activity Index, a strain metric which simultaneously accounted for nurse to patient ratios, patient acuity, patients with COVID-19, and unit occupancy. This study identified that an increased ICU Activity Index was associated with increased odds of mortality and readmission. Our study had similar findings in that periods of increased strain as demonstrated by percent occupancy and turnover on the day of transfer were associated with increased odds of floor escalation events, including early PICU readmission. However, most adult studies including this study by Pilcher have focused on periods of known strain, such as the COVID-19 pandemic, rather than evaluating for annual variability in strain which appears to have less significant of an impact in adult critical care units.8–15 In our study, we evaluated seasonal variation due to annual viral surges that are prominent in the PICU setting.31 Overall, our study demonstrated similar but additive results as have been found in recent adult ICU literature, that increased unit occupancy, as well as turnover, are important unit operations metrics that require attention due to their association with adverse patient outcomes.

The volume-based strain metrics including percent occupancy and turnover demonstrated variability throughout the study period with the highest levels occurring during viral illness surges. Prior studies have demonstrated that pediatric hospitals have implemented minimal acute operational changes in response to fluctuations in occupancy.32,33 With PICU admissions increasing annually and large pediatric referral centers receiving a larger percentage of these admissions, volume-based strain metrics will likely continue to fluctuate but at consistently higher levels highlighting the need for enhanced surge response planning and utilization of hospital census predictive models.20,34,35 Our study suggests that when the PICU is busiest from either high patient occupancy or turnover, patients being transferred out of the PICU have higher odds of experiencing a PICU readmission, a floor rapid response or a code blue event within the first 48 h. These patient events are quality metrics by which ICUs are evaluated both internally and externally.36–38 Our study identifies strain based factors, periods of high occupancy and turnover, which may contribute to early, unplanned PICU readmissions and therefore could be targets for future quality improvement initiatives aimed at reducing preventable readmissions.

Our results identified a nonlinear association of PICU percent occupancy and increased odds of experiencing a floor escalation event. This threshold of increased risk is likely the result of several factors. The provider workload at this occupancy may impair the ability to accurately assess a patient’s floor readiness at the time of transfer decision. In addition, pending admissions may be boarding elsewhere due to lack of PICU bed availability prompting PICU providers to transfer patients earlier than in lower occupancy times. There also may be a similar reflection in hospital-wide census during these times leading to strain across the hospital system decreasing resources available for individual patients. While each of these factors may have contributed to this critical threshold in percent occupancy, our results provide an operational warning indicator to inform staffing and resource utilization decisions.

Mitigating the negative impacts of resource strain in the ICU setting during surge periods must be as multifaceted as strain itself. Hospitals should consider multiple operational strategies to buffer against strain’s impact including utilizing predictive modeling to proactively prepare inpatient facilities by temporarily increasing bed capacity and staffing, standardizing emergency preparedness planning, developing critical care outreach programs that support the non-ICU medical teams during high strain periods, and identifying ways to visibly communicate strained periods to staff and providers to promote a shared-mental model of the challenges faced by all involved. In addition, various institutions should not work in isolation but instead should collaborate to discuss shared obstacles and troubleshoot solutions.

The study has notable limitations. As a single center study, generalizability may be limited, particularly due to differences in patient populations, PICU and hospital practices, and illness surge characteristics. For prognostic enrichment purposes, we chose to limit the population for the patient-specific analyses to patients admitted with primary respiratory and neurologic admit diagnoses, however this may further limit generalizability. In this study, we were unable to quantify the hospital-wide census and associated strain and their potential impact on floor escalation events. Despite the large sample size, the event rate was low and replication of study findings in a larger cohort inclusive of a more generalizable cohort of patients would be beneficial. Finally, while this study identified multiple contributing factors to resource strain, it did not capture all extrinsic factors such as equipment and staffing limitations, including nurse/respiratory therapist to patient ratios, or intrapersonal factors such as clinical experience and wellbeing.

Conclusions

PICU strain metrics demonstrated notable variability throughout the 10-year study period with pediatric illness surges reflecting the highest strain with the least variability. At times of high occupancy and turnover, patients in our study more frequently experienced a floor escalation event within 48 h after transfer from the PICU with an inflection point noted above 87.5% occupancy. Multicenter studies are needed to evaluate the generalizability of these findings.

Supplementary Material

Supplementary Digital Material

Supplemental material for this article is available online.

Acknowledgements

Research support was provided to Dr Baker through a University of Colorado Pediatric Critical Care Research Award. Virtual Pediatric Systems data was provided by our site’s VPS database.

Funding

Research support was provided to Dr Matthew Baker through a University of Colorado Pediatric Critical Care Research Award. Dr Aline Maddux receives funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (K23HD096018).

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Considerations

Not applicable.

Consent to Participate

This study was approved by Colorado’s Multiple Institutional Review Board (#23-1447; Sep. 22, 2023; Associations Between Strain and Outcomes in a Quaternary PICU) as an exempt study and was conducted under the standards of the 1975 Helsinki Declaration. Informed consent was waived by Colorado’s Multiple Institutional Review Board.

Consent for Publication

Not applicable.

Data Availability

Deidentified data are available through collaboration with the study investigators through a data use agreement.

References

  • 1.Campbell C The benefits of designing a stratification system for New York city pediatric intensive care units for use in regional surge capacity planning and management. J Community Health. 2010;35(4):337–347. [DOI] [PubMed] [Google Scholar]
  • 2.Cushing AM, Bucholz EM, Chien AT, Rauch DA, Michelson KA. Availability of pediatric inpatient services in the United States. Pediatrics. 2021;148(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Kanter RK, Cooper A. Mass critical care: pediatric considerations in extending and rationing care in public health emergencies. Disaster Med Public Health Prep. 2009;3(Suppl 2):S166–S171. [DOI] [PubMed] [Google Scholar]
  • 4.Markovitz BP. Pediatric critical care surge capacity. J Trauma. 2009;67(2 Suppl):S140–S142. [DOI] [PubMed] [Google Scholar]
  • 5.Anesi GL, Kerlin MP. The impact of resource limitations on care delivery and outcomes: routine variation, the coronavirus disease 2019 pandemic, and persistent shortage. Curr Opin Crit Care. 2021;27(5):513–519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Halpern SD. ICU Capacity strain and the quality and allocation of critical care. Curr Opin Crit Care. 2011;17(6):648–657. [DOI] [PubMed] [Google Scholar]
  • 7.Rewa OG, Stelfox HT, Ingolfsson A, et al. Indicators of intensive care unit capacity strain: a systematic review. Crit Care. 2018;22(1):86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Bagshaw SM, Wang X, Zygun DA, et al. Association between strained capacity and mortality among patients admitted to intensive care: a path-analysis modeling strategy. J Crit Care. 2018;43:81–87. [DOI] [PubMed] [Google Scholar]
  • 9.Pilcher DV, Hensman T, Bihari S, et al. Measuring the impact of ICU strain on mortality, after-hours discharge, discharge delay, interhospital transfer, and readmission in Australia with the activity index. Crit Care Med. 2023;51(12):1623–1637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Wagner J, Gabler NB, Ratcliffe SJ, Brown SE, Strom BL, Halpern SD. Outcomes among patients discharged from busy intensive care units. Ann Intern Med. 2013;159(7):447–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Weissman GE, Gabler NB, Brown SE, Halpern SD. Intensive care unit capacity strain and adherence to prophylaxis guidelines. J Crit Care. 2015;30(6):1303–1309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Weissman JS, Rothschild JM, Bendavid E, et al. Hospital workload and adverse events. Med Care. 2007;45(5):448–455. [DOI] [PubMed] [Google Scholar]
  • 13.Duggal A, Mathews KS. Impact of ICU strain on outcomes. Curr Opin Crit Care. 2022;28(6):667–673. [DOI] [PubMed] [Google Scholar]
  • 14.Rubinson L Intensive care unit strain and mortality risk among critically ill patients with COVID-19-there is no “me” in COVID. JAMA Netw Open. 2021;4(1):e2035041. [DOI] [PubMed] [Google Scholar]
  • 15.Wilcox ME, Rowan KM, Harrison DA, Doidge JC. Does unprecedented ICU capacity strain, as experienced during the COVID-19 pandemic, impact patient outcome? Crit Care Med. 2022;50(6):e548–ee56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Chan T, Rodean J, Richardson T, et al. Pediatric critical care resource use by children with medical complexity. J Pediatr. 2016;177:197–203.e1. [DOI] [PubMed] [Google Scholar]
  • 17.Edwards JD, Goodman DM. The child with severe chronic illness in the ICU: a concise review. Crit Care Med. 2022;50(5):848–859. [DOI] [PubMed] [Google Scholar]
  • 18.Horak RV, Griffin JF, Brown AM, et al. Growth and changing characteristics of pediatric intensive care 2001–2016. Crit Care Med. 2019;47(8):1135–1142. [DOI] [PubMed] [Google Scholar]
  • 19.Murphy Salem S, Graham RJ. Chronic illness in pediatric critical care. Front Pediatr. 2021;9:686206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Killien EY, Keller MR, Watson RS, Hartman ME. Epidemiology of intensive care admissions for children in the US from 2001 to 2019. JAMA Pediatr. 2023;177(5):506–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Pollack MM, Patel KM, Ruttimann UE. PRISM III: an updated pediatric risk of mortality score. Crit Care Med. 1996;24(5):743–752. [DOI] [PubMed] [Google Scholar]
  • 22.Feltz CJ, Miller GE. An asymptotic test for the equality of coefficients of variation from k populations. Stat Med. 1996;15(6):647–6458. [DOI] [PubMed] [Google Scholar]
  • 23.Candidate Critical Care Quality Indicators. Society of Critical Care Medicine Quality Indicators Committee: 1995. Society of Critical Care Medicine; 1995. [Google Scholar]
  • 24.Christiansen CF, Moller MH, Nielsen H, Christensen S. The Danish intensive care database. Clin Epidemiol. 2016;8:525–530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Mandell IM, Bynum F, Marshall L, Bart R, Gold JI, Rubin S. Pediatric Early Warning Score and unplanned readmission to the pediatric intensive care unit. J Crit Care. 2015;30(5):1090–1095. [DOI] [PubMed] [Google Scholar]
  • 26.Su L, Ma X, Gao S, et al. Evaluation of ICUs and weight of quality control indicators: an exploratory study based on Chinese ICU quality data from 2015 to 2020. Front Med. 2023;17(4):675–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Muggeo VM. Estimating regression models with unknown breakpoints. Stat Med. 2003;22(19):3055–3071. [DOI] [PubMed] [Google Scholar]
  • 28.Czaja AS, Hosokawa PW, Henderson WG. Unscheduled readmissions to the PICU: epidemiology, risk factors, and variation among centers. Pediatr Crit Care Med. 2013;14(6):571–579. [DOI] [PubMed] [Google Scholar]
  • 29.Kotsakis A, Stevens D, Frndova H, et al. Description of PICU unplanned readmission. Pediatr Crit Care Med. 2016;17(6):558–562. [DOI] [PubMed] [Google Scholar]
  • 30.Odetola FO, Clark SJ, Dechert RE, Shanley TP. Going back for more: an evaluation of clinical outcomes and characteristics of readmissions to a pediatric intensive care unit. Pediatr Crit Care Med. 2007;8(4):343–347; CEU quiz 57. [DOI] [PubMed] [Google Scholar]
  • 31.Wunsch H, Wagner J, Herlim M, Chong DH, Kramer AA, Halpern SD. ICU Occupancy and mechanical ventilator use in the United States. Crit Care Med. 2013;41(12):2712–2719. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Fieldston E, Ragavan M, Jayaraman B, Metlay J, Pati S. Traditional measures of hospital utilization may not accurately reflect dynamic patient demand: findings from a children’s hospital. Hosp Pediatr. 2012;2(1):10–18. [DOI] [PubMed] [Google Scholar]
  • 33.Fieldston ES, Hall M, Sills MR, et al. Children’s hospitals do not acutely respond to high occupancy. Pediatrics. 2010;125(5):974–981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Epstein RH, Dexter F. A predictive model for patient census and ventilator requirements at individual hospitals during the coronavirus disease 2019 (COVID-19) pandemic: a preliminary technical report. Cureus. 2020;12(6):e8501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Nguyen HM, Turk PJ, McWilliams AD. Forecasting COVID-19 hospital census: a multivariate time-series model based on local infection incidence. JMIR Public Health Surveill. 2021;7(8):e28195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Hosein FS, Roberts DJ, Turin TC, Zygun D, Ghali WA, Stelfox HT. A meta-analysis to derive literature-based benchmarks for readmission and hospital mortality after patient discharge from intensive care. Crit Care. 2014;18(6):715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Pleacher KM, Bratton SL. Pediatric intensive care unit bounce-backs: are we evaluating them? Pediatr Crit Care Med. 2007;8(4):401–402. [DOI] [PubMed] [Google Scholar]
  • 38.Scanlon MC, Mistry KP, Jeffries HE. Determining pediatric intensive care unit quality indicators for measuring pediatric intensive care unit safety. Pediatr Crit Care Med. 2007;8(2 Suppl):S3–S10. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Digital Material
NIHMS2122699-supplement-Supplementary_Digital_Material.docx (3.6MB, docx)

Data Availability Statement

Deidentified data are available through collaboration with the study investigators through a data use agreement.

RESOURCES