Time to Clinical Stability in Children With Community-Acquired Pneumonia

The Catalyzing Ambulatory Research in Pneumonia Etiology and Diagnostic Innovations in Emergency Medicine study was a prospective study of children 3 months to 18 years old who presented to the emergency department (ED) with signs and symptoms of lower respiratory tract infection and received a chest radiograph for clinical suspicion of CAP. Enrollment occurred from July 2013 to December 2017. The study was approved by the Cincinnati Children’s Hospital Medical Center Institutional Review Board. Written informed consent was obtained from all legal guardians, and assent was obtained from children ≥11 years of age. Among this cohort of children with clinically suspected CAP, we performed a sub-analysis to evaluate only patients with definitive radiographic or equivocal (atelectasis versus pneumonia) pneumonia. Further details of the study have been described previously.^13 –15 

We excluded children hospitalized in the 14 days before the index ED visit, those with a history of aspiration, and those with immunocompromising or chronic medical conditions that predispose to severe or recurrent pneumonia (eg, immunodeficiency, chronic corticosteroid use, chronic lung disease, malignancy, sickle cell disease, congenital heart disease, tracheostomy use, and neuromuscular disorders impacting respiration). Patients enrolled within 30 days before the ED visit were excluded to ensure a distinct infection episode.

Potential participants were identified by trained research coordinators. After informed consent, demographic, comorbidity, and historical information were obtained from the patient or parent by using a structured questionnaire. Clinicians completed a case report form assessing clinical signs. Clinical data, including vital signs, LOS, medications, and interventions (eg, oxygen use), were extracted from the electronic medical record.

Baseline data, including demographics, illness history, and physical examination findings, were collected from the parent and ED clinician at the time of the study ED visit. Data for vital signs, including temperature, HR, RR, oxygen saturation, and use of supplemental oxygen, were obtained from the electronic medical record. For inpatients, these parameters were generally measured at least every 4 hours from admission through discharge as part of routine care.

Stable temperature was defined as a recorded temperature of 36.0 to 37.9°C. RR and HR were considered stable if they were <99th percentile for age.¹⁶  Oxygenation was stable when the oxygen saturation was >90% and there was no administration of supplemental oxygen. If the last recorded value for a given parameter during the hospitalization was abnormal, that parameter was considered unstable at discharge. Otherwise, the time and date of the last abnormal value for each parameter were subtracted from the admission time and date to determine the TCS for that parameter in hours.

We first evaluated each individual variable as being stable or unstable at the time of discharge. We then assessed TCS using various combinations of the 4 stability parameters, including combinations of 2 (eg, HR and RR), 3 (eg, HR, RR, and temperature), and all 4 variables (HR, RR, temperature, and oxygenation). Stability for each combination was defined as the normalization of all TCS variables in question. LOS was calculated as the time of discharge from the hospital minus the time of admission order.

The TCS measure including all 4 parameters was compared with other outcomes, including hospital LOS and illness severity. Disease severity was defined by using a 3-tiered categorical variable (mild, moderate, severe) previously used to define clinical outcomes in children with CAP.¹³  Mild was defined as hospitalization for <24 hours with no inpatient receipt of supplemental oxygen or IV fluids (maintenance or boluses). Moderate was defined as hospitalization lasting >24 hours, or hospitalization <24 hours with the use of supplemental oxygen or IV hydration. Severe disease was defined as requiring ICU admission for >24 hours, having a diagnosis of severe sepsis or septic shock, receipt of vasoactive infusions, positive-pressure ventilation (ie, continuous positive airway pressure, bilevel positive airway pressure, or intubation with mechanical ventilation), chest drainage procedures for empyema, extracorporeal membrane oxygenation, or death. To evaluate if patients who were discharged from the ED or hospital with unstable parameters had clinical deterioration after discharge, we evaluated readmission rates to the ED and to the hospital. Lastly, we evaluated whether various clinical factors present at the time of hospital admission from the ED were associated with early TCS, which was defined as reaching stability within 24 hours after admission.

We calculated summary statistics, including counts and percentages for categorical variables and medians and quartiles for quantitative variables. We compared patient characteristics across age groups using χ² or Fisher’s exact tests for categorical variables and the Wilcoxon rank test for continuous variables. We compared TCS with LOS using the Wilcoxon signed rank test. We examined factors associated with (1) the TCS and (2) the achievement of early clinical stability with Cox proportional hazards and logistic regression models, respectively. We derived hazard and odds ratios with 95% confidence intervals from these models. All analyses were performed in the open-source R software environment.¹⁷ 

Overall, 571 patients were hospitalized for CAP and included in this analysis. The median patient age was 3.1 years (interquartile range [IQR], 1.3–4.5), and 308 (54%) were boys (Table 1). The median LOS was 35.5 hours (IQR 20.4–49.7). Children <1 year old had longer hospitalizations (40.6 hours [23–72.6]) compared with older children (Table 1). There were 152 (27%) children classified as mild severity, 359 (64%) classified as moderate, and 47 (8%) classified as severe.

Of the 571 hospitalized children, 187 (32.7%) children had at least 1 abnormal parameter at the time of hospital discharge, 42 (7%) had 2 abnormal parameters, and no child had 3 or 4 abnormal parameters (Table 1). Although most children <12 months old reached stability in all 4 parameters, this occurred in approximately half of older children.

Of the individual parameters, RR was least likely to reach stability at discharge (n = 282 [65%]), followed by HR (n = 316 [84%]), temperature (n = 218 [95%]), and oxygenation (n = 270 [97%]; Table 2). The median TCS for each of the 4 parameters was <24 hours, with RR and oxygenation having a longer TCS (median 16.1 and 17.5 hours, respectively), compared with HR (12.8 hours) and temperature (12.1 hours). When stratified by age, the youngest and oldest children had the longest TCS for individual parameters (Supplemental Table 4). TCS for the 4 parameters statistically increased with disease severity (4.1–7.9 hours for mild, 15.6–19.3 for moderate, and 24.6–63.2 for severe), as did LOS (18.2, 40.8, and 69.8 hours for mild, moderate, and severe groups; Fig 1, Supplemental Table 5). In all severity groups, the TCS was statistically less than the LOS.

We also examined combinations of the 4 parameters. Most (83.4%) patients in our cohort had >1 unstable parameter at presentation. Because of limited sample size, we are unable to draw conclusions about the small number of patients with isolated unstable parameters. In general, the more parameters that were unstable at the time of hospitalization, the longer the TCS (with 12.1–17.5 hours for single parameters and 47.9 hours when all 4 parameters were unstable). In patients hospitalized with at least 2 unstable parameters on admission, most reached stability before discharge (Table 2). The youngest and oldest age groups had the longest TCS when exploring the combinations of parameters (Supplemental Tables 4 and 6). For both individual parameters and combinations of parameters, results were similar for patients with radiographic pneumonia (Supplemental Table 6, Supplemental Fig 2).

Older age (adjusted odds ratio 0.96 [95% confidence interval, 0.93–0.99]), vomiting (0.77 [0.63–0.95]), and prolonged capillary refill (0.77 [0.60–0.98]) were associated with decreased odds of early TCS (within 24 hours), whereas diffusely decreased breath sounds (versus no decreased breath sounds; 1.94 [1.02, 3.80]) were associated with reaching stability within the first 24 hours after admission (Table 3). Similar results were found in the subset with radiographic CAP, with the addition of grunting associated with decreased odds of early TCS and focally decreased breath sounds and hyperinflation associated with increased odds of early TCS (Supplemental Table 7).

Of the 571 children, 46 (8.1%) revisited the ED. Overall, there were no statistical differences in the odds of revisit in those who were stable versus unstable for each parameter (Supplemental Table 8), although analyses are limited because of smaller sample sizes.

In this prospective cohort study of children with CAP, a TCS outcome using routinely collected physiologic data can serve as an objective measure of clinical response and discharge readiness. Of those patients who were unstable on admission to the hospital, most reached stability for all or most of their parameters in <24 hours. LOS was significantly longer than TCS, which suggests that many hospitalized patients could potentially be discharged sooner if a TCS measure was incorporated into discharge decisions or in establishing standardized discharge readiness criteria.¹⁸  TCS could aid in the identification of those who might be nearing readiness for discharge versus those more likely to require a prolonged LOS. Additionally, factors associated with an earlier TCS can be used to anticipate hospital courses and facilitate management decisions.

Although TCS has been used for adults with CAP, only 1 previous study by Wolf et al was conducted to assess its utility in children.^{7 ,8 ,12 ,19 ,20}  We have expanded on this work in several ways. Wolf et al defined oxygenation as the use of supplemental oxygen rather than the measurement of oxygen saturation, introducing uncertainty about whether the patient was truly hypoxemic and requiring oxygen. We attempted to increase granularity by adding measured hypoxemia to the oxygenation stability parameter. In addition, Wolf et al used Pediatric Advanced Life Support vital sign thresholds, which lack specificity and include wide age ranges.²¹  We attempted to increase granularity and specificity in vital sign parameters by using age-based percentiles.¹⁶  More patients in our cohort were discharged with unstable parameters than in the cohort reported by Wolf et al (33% vs 21%).

Our results were also consistent with the previous pediatric study by Wolf et al in several ways. Oxygenation and RR took the longest to reach clinical stability, and RR was most likely to be abnormal at discharge (13.7% in the Wolf et al study and 26.1% in this study). But despite the higher percentages of “unstable” discharges in our sample, there was not a statistical increase in revisits. When combinations of study parameters were considered together, any combination that included oxygenation took longer to reach stability. However, unlike the study by Wolf et al, in which RR and O₂ had similar LOS to more complicated combinations of parameters, the combination of RR and O₂ did not compare similarly to more complicated combinations (RR and oxygenation TCS 25.2 hours, all 4 parameters TCS 47.9 hours) in our study, although it was unclear if all parameters were unstable on admission (as opposed to any individual parameter) in the previous study. Both studies revealed that TCS increased with disease severity.¹² 

The objective stability parameters collected for this study consisted of routinely collected data on all patients hospitalized for CAP, making their use easy and transferable across most institutions. Objective data alone do not independently determine the overall LOS, given patient and family characteristics (eg, access to transportation, social determinants of health) and health care team characteristics (eg, provider rounding, efficiency of responding to TCS, time of day).^22 –25  However, given the differences in TCS and LOS, our results suggest that using stability parameters to determine readiness for discharge may help to facilitate safe and earlier discharges for children in the hospital.

Older children were more likely to have 1 to 2 parameters that were unstable at discharge, as opposed to infants who were more typically discharged with all parameters stable. Children <12 months of age had the longest LOS, which was possibly influenced by the time taken for all parameters to become “stable.” This finding suggests that current practice includes all 4 parameters being considered together in determining readiness for discharge in infants. This may reflect the nature of viral respiratory infections in this younger age group or a more conservative approach to discharge decisions in infants. No patients in our study were discharged with >2 unstable parameters, with HR and RR being the most likely to have not reached stability. However, despite patients being discharged with unstable parameters, these patients were not more likely to revisit and require admission to the hospital within 7 days. Our HR and RR findings also may reflect the variability in the way these signs are measured, the way in which the signs were categorized as unstable, or the relative lack of prognostic value of HR and RR compared with oxygenation and temperature. Given that being discharged with unstable parameters did not lead to readmission, potentially in an older population, oxygenation and temperature alone may be sufficient to determine readiness for discharge.

We found that several patient characteristics (older age, presence of vomiting, and initial prolonged capillary refill) were associated with prolonged TCS (>24 hours). This combination suggests that older children and those admitted for IV fluids and potential dehydration may improve slower than other patients. Patients without these factors might be candidates for management on an observation or short-stay unit. More research may be needed to understand why children in our cohort with diffusely decreased breath sounds reached clinical stability earlier than those without. Possibly these patients had a component of reactive airway disease or asthma and therefore improved because of the use of alternative medications, such as steroids and β agonists. Given that our study is among the first to evaluate factors associated with TCS, we support further work to better understand if this data could lead to less inpatient care for a defined population.

Our study has several limitations. We tested 4 objective physiologic parameters; however, there may be other nonclinical factors (parent presence, transportation, pharmacy availability, etc) that contribute to time in the hospital for children with CAP. In addition, in adults, the ability to ambulate, a back-to-baseline mental status, and return of toleration of normal oral intake are markers used to determine readiness for discharge.⁸  These data were not available in the current study. Given that dehydration may be an important reason for the admission of children with CAP, some combinations of these other markers may add value. Finally, although our revisit data suggest that being discharged with unstable parameters did not lead to revisits and readmission, our study was not powered a priori for this uncommon safety outcome. Because this study was performed at a single center, it is unknown if these data are generalizable. However, the Cincinnati Children’s Hospital Medical Center sees 99.6% of children hospitalized for CAP in Hamilton County, making it less likely that children are returning to other hospitals after discharge.²⁶ 

This prospective study provides further insight into potential objective measures of TCS that could be used in assessing disease recovery in children hospitalized for CAP and potentially targeting ways to predict or shorten the LOS (eg, such as notifying the primary team when the patient no longer requires oxygen). Given that patient factors on admission were associated with early TCS, these patients might benefit from observation units to allow for expedited discharges to eliminate unnecessary further inpatient usage. In addition, given that patients being discharged with nonnormalized parameters (ie, unstable HR or RR) do not seem to require readmission, a less complex model, potentially the combination of oxygenation and temperature, might be sufficient to determine medical readiness for discharge. Overall, our findings have several important implications regarding clinical utility, including the potential to decrease LOS in hospitalized children with CAP. In addition, factors associated with early TCS could be used by ED providers to guide disposition decisions, including determining which patients are most appropriate for boarding versus observation versus inpatient admission during times when inpatient beds are limited. However, before any implementation of TCS in discharge readiness checklists, research is needed to further understand the residual factors impacting LOS after TCS has been achieved. Future work to validate these is an important next step in understanding the role of incorporating a TCS measure into the clinical environment and its effect on clinical and process outcomes of CAP in children.

CAP

community-acquired pneumonia

ED

emergency department

HR

heart rate

IQR

interquartile range

IV

intravenous

LOS

length of stay

Oxy

oxygenation

RR

respiratory rate

TCS

time to clinical stability

Temp

temperature