Rebecca Steuart, MD

Children with neurologic impairment (NI; eg, hypoxic-­ischemic encephalopathy, muscular dystrophy) are characterized by functional and/or intellectual impairments resulting from a variety of neurologic diseases.¹ These children commonly have respiratory comorbidities, including central hypoventilation, impaired cough, and oromotor dysfunction, that may lead to chronic respiratory insufficiency and a need for chronic respiratory support at baseline.^2,3 Baseline respiratory support modalities can include supplemental oxygen, noninvasive positive pressure ventilation, or invasive mechanical ventilation.

Acute respiratory infections (ARI; eg, pneumonia, bronchiolitis) are the most common cause of hospitalization, intensive care unit (ICU) admission, and death for children with NI.^1,3 Discharge criteria for otherwise healthy children admitted to the hospital with ARI often include return to respiratory baseline.⁴ Children with complex chronic conditions have longer hospitalizations when hospitalized with respiratory infections,^5-7 because, in part, comorbidities and complications prolong the time to return to baseline. This prolonged return to respiratory baseline in combination with family knowledge, comfort, and skill in managing respiratory support and other complexities at home may alter discharge practices in the population of children with NI. In our clinical experience, discharge before return to baseline respiratory support occurs more frequently in children with NI than in otherwise healthy children when hospitalized with ARI. However, the consequences of discharging children with NI prior to return to respiratory baseline are unknown.

In this study, we sought to determine if discharge prior to return to baseline respiratory support is associated with reutilization among children with NI hospitalized with ARI. We hypothesized that patients discharged prior to return to respiratory baseline would have higher rates of 30-day hospital reutilization.

Study Design and Data Source

This single-center, retrospective cohort study of children hospitalized at Cincinnati Children’s Hospital Medical Center (CCHMC) used data from the Pediatric Health Information System (PHIS) and the electronic medical record (EMR). PHIS, an administrative database of 45 not-for-profit, tertiary care, US pediatric hospitals managed by Children’s Hospital Association (Lenexa, Kansas), was used to identify eligible children, examine demographic and clinical variables, and define outcomes. PHIS contains data regarding patient demographics, inpatient resource utilization, and diagnoses. Encrypted medical record numbers in PHIS allowed for local identification of patients’ medical records to complete EMR review to confirm eligibility and obtain detailed patient-level clinical information (eg, respiratory support needs) not available in PHIS.

Pilot medical record reviews allowed for standardized study definitions and procedures. All study staff underwent training with the primary investigator, including detailed review of 10 initial abstractions. Two investigators (K.M. and S.C.) performed repeat abstractions from 40 randomly selected records to enable assessment of interrater reliability. Average reliability, indicated by the κ statistic, indicated substantial to near-perfect reliability⁸ (κ = 0.97, 95% CI 0.90-1.00) for the primary exposure. EMR data were managed using Research Electronic Data Capture (REDCap, Nashville, Tennessee)⁹ and subsequently merged with PHIS data.

Study Population

Hospitalizations of children with NI aged 1 to 18 years at CCHMC between January 2010 and September 2015 were eligible for inclusion if they had a principal discharge diagnosis indicative of ARI and required increased respiratory support from baseline during hospitalization. NI was defined as a high-intensity, chronic neurological diagnosis with substantial functional impairments according to previously defined diagnosis codes.^1,10 ARI was identified using codes in the Clinical Classification Software (Agency for Healthcare Research and Quality, Rockville, MD) respiratory group indicative of ARI (eg, pneumonia, bronchiolitis, influenza; Appendix Table).

Children transferred to CCHMC were excluded because records from their initial illness presentation and management were not available. Because of expected differences in management and outcomes, children with a known diagnosis of tuberculosis or human immunodeficiency virus were excluded. Because exposure criteria were dependent on hospital discharge status, hospitalizations for children who died during admission (4 of 632 hospitalizations, 0.63%) were excluded from the final cohort (Appendix Figure).

Study Definitions

Baseline respiratory support (ie, “respiratory baseline”) was defined as the child’s highest level of respiratory support needed prior to admission when well (ie, no support, supplemental oxygen, continuous positive airway pressure [CPAP] or bilevel positive airway pressure [BiPAP], or ventilator support), and further characterized by night or day/night requirement. Respiratory baseline was identified using EMR documentation of home respiratory support use at the time of index admission. Return to respiratory baseline was defined as the date on which the child achieved documented home respiratory support settings, regardless of clinical symptoms.

Children may have required increased respiratory support from baseline at any time during hospitalization. Maximum respiratory support required was categorized as one of the following: (1) initiation of supplemental oxygen or increase in oxygen flow or duration; (2) initiation of CPAP or BiPAP; (3) increase in pressure settings or duration of pressure support for those with baseline CPAP, BiPAP, or ventilator use; and (4) initiation of full mechanical ventilation. Respiratory support categories were mutually exclusive: children requiring multiple types of increased respiratory support were classified for analysis by the most invasive form of respiratory support used (eg, a child requiring increase in both oxygen flow and pressure settings was categorized as an increase in pressure settings). Children who received heated high-flow nasal cannula therapy as maximum support were categorized as initiation or increase in oxygen support.

Time to return to respiratory baseline was defined as the difference in days between date of return to respiratory baseline and date of admission. Time to return to respiratory baseline was determined only for children who were discharged at respiratory baseline.

Primary Exposure and Outcome Measures

The primary exposure was hospital discharge before return to respiratory baseline (ie, discharge on higher respiratory support than at baseline settings). At our institution, standardized discharge criteria for children with NI do not exist. The primary outcome was all-cause, 30-day hospital reutilization, including hospital readmissions and emergency department (ED) revisits. Secondary outcomes included 30-day reutilization for ARI and hospital length of stay (LOS) in days.

Patient Demographics and Clinical Characteristics

Demographic and patient characteristics that might influence hospital discharge before return to respiratory baseline or readmission were obtained from PHIS (eg, demographic information, age, insurance type, measures of clinical complexity, illness severity) and by EMR review (eg, baseline respiratory support needs, maximum respiratory support during hospitalization). Measures of clinical complexity included comorbid complex chronic conditions (CCCs)^11-14 and technology dependence^14-16 using previously defined diagnostic codes. Measures of illness severity included sepsis¹⁷ and ICU-level care. At our institution, children with baseline ventilator use do not require admission to the ICU unless they are clinically unstable.

Statistical Analysis

Continuous variables were described using medians and interquartile ranges (IQR). Categorical variables were described using counts and percentages. Patient characteristics and outcomes were stratified by primary exposure and compared using chi-square test or Fisher exact test for categorical variables and Wilcoxon rank sum test for continuous variables.

To examine the independent association between discharge before return to respiratory baseline and hospital reutilization, a generalized estimating equation was used that included potential confounders while accounting for within-patient clustering. Patient demographics included age, race, ethnicity, and insurance type; measures of clinical complexity included number of CCCs, technology dependence, and baseline respiratory support; and measures of acute illness severity included ARI diagnosis, degree of increase in respiratory support during hospitalization, and ICU-level care. LOS was also included in the model as a covariate because of its expected association with both exposure and outcome.

Secondary analyses were conducted using the outcome of 30-day reutilization for ARI. Subgroup analysis excluding hospitalizations of children lost to follow-up (ie, no encounters in the 6 months after hospital discharge) was also conducted. All analyses were performed with SAS v9.3 (SAS Institute, Cary, North Carolina). P values < .05 were considered statistically significant. This study was approved by the Institutional Review Board.

Study Cohort

A total of 632 hospitalizations experienced by 366 children with NI who were admitted with ARI were included (Appendix Figure). Most children (66.4%) in the cohort experienced only one hospitalization, 17.5% had two hospitalizations, 7.9% had three hospitalizations, and 8.2% had four or more hospitalizations. The median age at hospitalization was 5.0 years (IQR 2.8-10.5) and most hospitalizations were for children who were male (56.6%), white (78.3%), non-Hispanic (96.0%), and publicly insured (51.7%; Table 1). More than one-quarter (28.6%) of hospitalizations were for children with four or more CCCs, and in 73.4% of hospitalizations, children were technology dependent (Table 1). Baseline respiratory support was common (46.8%), including home mechanical ventilation in 11.1% of hospitalizations (Table 1). Bacterial pneumonia, including aspiration pneumonia, was the most common discharge diagnosis (50.5%, Table 1).

Demographic and Clinical Characteristics

Children were discharged before return to respiratory baseline in 30.4% of hospitalizations (Appendix Figure). Children discharged before return to respiratory baseline were older (median age 5.7 years, IQR 3.1-11.0, vs 4.9 years, IQR 2.6-9.7; P = .04) and more likely to be privately insured (54.2% vs 43.4%; P = .04), compared with children discharged at respiratory baseline (Table 1). Children discharged before return to respiratory baseline were also more likely to have a respiratory CCC (59.9% vs 30.9%; P < .001), have a respiratory technology dependence diagnosis code (44.8% vs 24.1%; P < .001), and have baseline respiratory support needs on EMR review (67.7% vs 37.7%; P < .001), compared with children discharged at baseline (Table 1).

Children discharged before return to respiratory baseline required significantly greater escalation in respiratory support during hospitalization, compared with children discharged at respiratory baseline, including higher rates of initiation of CPAP or BiPAP, increased pressure settings from baseline (for home CPAP, BiPAP, or ventilator users), and initiation of full mechanical ventilation (Table 1). Hospitalizations in which children were discharged before return to respiratory baseline were also more likely to include ICU care than were those for children discharged at baseline (52.1% vs 35.2%; P < .001; Table 1).

Clinical Outcomes and Utilization

Reutilization within 30 days occurred after 32.1% of hospitalizations, with 26.1% requiring hospital readmission and 6.0% requiring ED revisit (Table 2). There was no statistical association in either unadjusted (Table 2) or adjusted (Table 3) analysis between children discharged before return to respiratory baseline and 30-day all-cause hospital reutilizations, hospital readmissions, or ED revisits.

In analysis of secondary outcomes, 30-day reutilization because of ARI occurred after 21.5% of hospitalizations, with 19.0% requiring hospital readmission and 2.5% requiring ED revisit. Median hospital LOS for the cohort was 4 days (IQR 2-8; Table 2). Hospitalizations in which children were discharged before return to respiratory baseline were longer than in those discharged at baseline (median 6 days, IQR 3-11, vs 4 days, IQR 2-7; P < .001; Table 2).

For hospitalizations of children discharged at respiratory baseline, the median time to return to respiratory baseline was 3 days (IQR 1-5, complete range 0-80). In these encounters, discharge occurred soon after return to respiratory baseline (median 1 day, IQR 0-1.5, complete range 0-54).

In subgroup analysis excluding the 18 hospitalizations in which children were lost to follow-up (2.8% of the total cohort), discharge before return to respiratory baseline was not associated with 30-day all-cause hospital reutilization (Table 4).

In this retrospective cohort study, children with NI hospitalized with ARI were frequently discharged using increased respiratory support from baseline. However, those discharged before return to respiratory baseline, despite their greater clinical complexity and acute illness severity, did not have increased hospital reutilization, compared with children discharged at respiratory baseline. Our findings suggest that discharge before return to baseline respiratory support after ARI may be clinically appropriate in some children with NI.

With the growing emphasis on decreasing hospital costs, concern exists that patients are being discharged from hospitals “quicker and sicker,”^18,19 with shortening lengths of stay and higher patient instability at discharge. Clinical instability at discharge has been associated with adverse postdischarge outcomes in adults with pneumonia^20-23; however, studies evaluating discharge readiness have not examined the population of children with NI. Our findings of no difference in reutilization for children with NI discharged before return to respiratory baseline, which would be expected to approximate one or more clinical instabilities, contrast these concerns.

Clinicians caring for children with NI hospitalized with ARI may find it difficult to determine a child’s discharge readiness, in part because many children with NI have longer times to return to respiratory baseline and some never return to their pre-­illness baseline.²⁴ In otherwise healthy children hospitalized with respiratory infections such as pneumonia, discharge criteria typically include complete wean from respiratory support prior to discharge.^4,25 In our study’s more complex children, whose parents already manage respiratory support at home, we hypothesize that discharging providers may be comfortable with discharge when the child has certain types of increased respiratory support compatible with home equipment, a parent skilled with monitoring the child’s respiratory status, and the support of an experienced outpatient provider and home nursing providers. At our institution, outpatient respiratory support weans are primarily performed by pediatric pulmonologists and, for isolated weaning of supplemental oxygen or time using support, by parents and outpatient pediatricians.

Another important factor in determining a child’s discharge readiness is the perspective of the child’s parent. Berry et al found that children whose parents believe they are not healthy enough for discharge are more likely to experience unplanned hospital readmissions,²⁴ signaling the role of child- and family-­specific factors in safe discharge decisions. Therefore, parents of children with NI should be proactively involved throughout the multidisciplinary discharge process,^26,27 including the decision to discharge before return to respiratory baseline. Parents have identified ongoing provider support, opportunities to practice home care skills, and written instructions with contingency plans as important components of discharge readiness.²⁸ Further work to create partnerships with these highly skilled caregivers in discharge decision making and transition planning are needed to promote safe discharge practices in this complex population.

In our study, children discharged before return to respiratory baseline were more likely to be older and privately insured compared with children discharged at respiratory baseline. Prior studies have found that social factors including low socioeconomic status influence ED provider admissions decisions for children with pneumonia.^29,30 However, the role of socioeconomic factors in provider discharge decisions for children with NI has not been assessed. These traits may also be proxy markers of other sociodemographic factors, such as parent education level, financial hardship influencing ability to participate in a child’s care at the bedside, access to comprehensive outpatient primary care, and availability of private home nursing. We hypothesize that these related characteristics directly and indirectly influence provider discharge decisions.

Discharging providers are likely more comfortable with discharge prior to return to respiratory baseline when the family has private duty nursing in the home. Home nurses can assist families in providing increased respiratory support and recognizing respiratory problems that may arise following discharge. However, home nursing shortages are common nationwide.^31,32 Low-income children, children with respiratory technology use, and children without Medicaid have been found to have larger gaps in home nursing availability.^31,32 Further studies are needed to understand the role of home nursing availability in provider discharge decisions in this population.

This study has several limitations. The retrospective design of this study creates the potential for residual confounding; there may be other clinical or demographic factors influencing hospital discharge decisions that we are unable to capture using EMR review, including parental knowledge and comfort managing illness, quality of discharge instructions, frequency of follow-up visits, and presence of skilled home nursing services. Categorization of children based on respiratory support status at discharge lends potential for misclassification of exposure; however, our substantial interrater reliability suggests that misclassification bias is small. This study’s primary finding indicated no difference between exposure groups; although we may be unable to detect small differences, we had sufficient power with our sample size to detect meaningful differences in reutilization outcomes.

This study was not designed to capture outpatient time to return to respiratory baseline; prospective studies are needed to identify rates of return to respiratory baseline following ARI in children with NI. We did not measure the level of respiratory support used by children at the time of discharge and, therefore, are unable to estimate the amount of respiratory support weaning needed following discharge or the compatibility of support with home equipment using our data. In addition, this study focused on respiratory support modalities and, thus, did not measure inpatient utilization of mucociliary clearance technologies that might be hypothesized to decrease the time to return to baseline respiratory support. Next steps in evaluating treatment of ARI include investigating the effect of mucociliary clearance on both exposure and outcome in this population.

There may be other clinically meaningful outcomes for this population apart from reutilization that we have not assessed in this study, including increased respiratory support required following discharge, primary care reutilization, healthcare costs, or parent satisfaction with timing of and outcomes after discharge. Finally, although our hospital has reutilization rates for children with NI that are similar to other institutions in the United States,³³ our results may not be generalizable to children with NI hospitalized at other institutions because local discharge processes and systems of care may be different. Prospective, multicenter investigation is needed to evaluate the clinical consequences of discharge before return to respiratory baseline more broadly.

At our institution, approximately one-quarter of children with NI hospitalized with ARI were discharged before return to respiratory baseline, but these children were not at increased risk of reutilization, compared with children discharged at respiratory baseline. Our findings suggest that return to baseline respiratory support might not be a necessary component of hospital discharge criteria. In otherwise clinically stable children with NI, discharge before return to respiratory baseline may be reasonable if their parents are comfortable managing respiratory support at home.

The authors thank Jonathan Rodean of the Children’s Hospital Association for his assistance with abstraction of PHIS data.

Children with neurologic impairment (NI; eg, hypoxic-­ischemic encephalopathy, muscular dystrophy) are characterized by functional and/or intellectual impairments resulting from a variety of neurologic diseases.¹ These children commonly have respiratory comorbidities, including central hypoventilation, impaired cough, and oromotor dysfunction, that may lead to chronic respiratory insufficiency and a need for chronic respiratory support at baseline.^2,3 Baseline respiratory support modalities can include supplemental oxygen, noninvasive positive pressure ventilation, or invasive mechanical ventilation.

Acute respiratory infections (ARI; eg, pneumonia, bronchiolitis) are the most common cause of hospitalization, intensive care unit (ICU) admission, and death for children with NI.^1,3 Discharge criteria for otherwise healthy children admitted to the hospital with ARI often include return to respiratory baseline.⁴ Children with complex chronic conditions have longer hospitalizations when hospitalized with respiratory infections,^5-7 because, in part, comorbidities and complications prolong the time to return to baseline. This prolonged return to respiratory baseline in combination with family knowledge, comfort, and skill in managing respiratory support and other complexities at home may alter discharge practices in the population of children with NI. In our clinical experience, discharge before return to baseline respiratory support occurs more frequently in children with NI than in otherwise healthy children when hospitalized with ARI. However, the consequences of discharging children with NI prior to return to respiratory baseline are unknown.

In this study, we sought to determine if discharge prior to return to baseline respiratory support is associated with reutilization among children with NI hospitalized with ARI. We hypothesized that patients discharged prior to return to respiratory baseline would have higher rates of 30-day hospital reutilization.

Study Design and Data Source

This single-center, retrospective cohort study of children hospitalized at Cincinnati Children’s Hospital Medical Center (CCHMC) used data from the Pediatric Health Information System (PHIS) and the electronic medical record (EMR). PHIS, an administrative database of 45 not-for-profit, tertiary care, US pediatric hospitals managed by Children’s Hospital Association (Lenexa, Kansas), was used to identify eligible children, examine demographic and clinical variables, and define outcomes. PHIS contains data regarding patient demographics, inpatient resource utilization, and diagnoses. Encrypted medical record numbers in PHIS allowed for local identification of patients’ medical records to complete EMR review to confirm eligibility and obtain detailed patient-level clinical information (eg, respiratory support needs) not available in PHIS.

Pilot medical record reviews allowed for standardized study definitions and procedures. All study staff underwent training with the primary investigator, including detailed review of 10 initial abstractions. Two investigators (K.M. and S.C.) performed repeat abstractions from 40 randomly selected records to enable assessment of interrater reliability. Average reliability, indicated by the κ statistic, indicated substantial to near-perfect reliability⁸ (κ = 0.97, 95% CI 0.90-1.00) for the primary exposure. EMR data were managed using Research Electronic Data Capture (REDCap, Nashville, Tennessee)⁹ and subsequently merged with PHIS data.

Study Population

Hospitalizations of children with NI aged 1 to 18 years at CCHMC between January 2010 and September 2015 were eligible for inclusion if they had a principal discharge diagnosis indicative of ARI and required increased respiratory support from baseline during hospitalization. NI was defined as a high-intensity, chronic neurological diagnosis with substantial functional impairments according to previously defined diagnosis codes.^1,10 ARI was identified using codes in the Clinical Classification Software (Agency for Healthcare Research and Quality, Rockville, MD) respiratory group indicative of ARI (eg, pneumonia, bronchiolitis, influenza; Appendix Table).

Children transferred to CCHMC were excluded because records from their initial illness presentation and management were not available. Because of expected differences in management and outcomes, children with a known diagnosis of tuberculosis or human immunodeficiency virus were excluded. Because exposure criteria were dependent on hospital discharge status, hospitalizations for children who died during admission (4 of 632 hospitalizations, 0.63%) were excluded from the final cohort (Appendix Figure).

Study Definitions

Baseline respiratory support (ie, “respiratory baseline”) was defined as the child’s highest level of respiratory support needed prior to admission when well (ie, no support, supplemental oxygen, continuous positive airway pressure [CPAP] or bilevel positive airway pressure [BiPAP], or ventilator support), and further characterized by night or day/night requirement. Respiratory baseline was identified using EMR documentation of home respiratory support use at the time of index admission. Return to respiratory baseline was defined as the date on which the child achieved documented home respiratory support settings, regardless of clinical symptoms.

Children may have required increased respiratory support from baseline at any time during hospitalization. Maximum respiratory support required was categorized as one of the following: (1) initiation of supplemental oxygen or increase in oxygen flow or duration; (2) initiation of CPAP or BiPAP; (3) increase in pressure settings or duration of pressure support for those with baseline CPAP, BiPAP, or ventilator use; and (4) initiation of full mechanical ventilation. Respiratory support categories were mutually exclusive: children requiring multiple types of increased respiratory support were classified for analysis by the most invasive form of respiratory support used (eg, a child requiring increase in both oxygen flow and pressure settings was categorized as an increase in pressure settings). Children who received heated high-flow nasal cannula therapy as maximum support were categorized as initiation or increase in oxygen support.

Time to return to respiratory baseline was defined as the difference in days between date of return to respiratory baseline and date of admission. Time to return to respiratory baseline was determined only for children who were discharged at respiratory baseline.

Primary Exposure and Outcome Measures

The primary exposure was hospital discharge before return to respiratory baseline (ie, discharge on higher respiratory support than at baseline settings). At our institution, standardized discharge criteria for children with NI do not exist. The primary outcome was all-cause, 30-day hospital reutilization, including hospital readmissions and emergency department (ED) revisits. Secondary outcomes included 30-day reutilization for ARI and hospital length of stay (LOS) in days.

Patient Demographics and Clinical Characteristics

Demographic and patient characteristics that might influence hospital discharge before return to respiratory baseline or readmission were obtained from PHIS (eg, demographic information, age, insurance type, measures of clinical complexity, illness severity) and by EMR review (eg, baseline respiratory support needs, maximum respiratory support during hospitalization). Measures of clinical complexity included comorbid complex chronic conditions (CCCs)^11-14 and technology dependence^14-16 using previously defined diagnostic codes. Measures of illness severity included sepsis¹⁷ and ICU-level care. At our institution, children with baseline ventilator use do not require admission to the ICU unless they are clinically unstable.

Statistical Analysis

Continuous variables were described using medians and interquartile ranges (IQR). Categorical variables were described using counts and percentages. Patient characteristics and outcomes were stratified by primary exposure and compared using chi-square test or Fisher exact test for categorical variables and Wilcoxon rank sum test for continuous variables.

To examine the independent association between discharge before return to respiratory baseline and hospital reutilization, a generalized estimating equation was used that included potential confounders while accounting for within-patient clustering. Patient demographics included age, race, ethnicity, and insurance type; measures of clinical complexity included number of CCCs, technology dependence, and baseline respiratory support; and measures of acute illness severity included ARI diagnosis, degree of increase in respiratory support during hospitalization, and ICU-level care. LOS was also included in the model as a covariate because of its expected association with both exposure and outcome.

Secondary analyses were conducted using the outcome of 30-day reutilization for ARI. Subgroup analysis excluding hospitalizations of children lost to follow-up (ie, no encounters in the 6 months after hospital discharge) was also conducted. All analyses were performed with SAS v9.3 (SAS Institute, Cary, North Carolina). P values < .05 were considered statistically significant. This study was approved by the Institutional Review Board.

Study Cohort

A total of 632 hospitalizations experienced by 366 children with NI who were admitted with ARI were included (Appendix Figure). Most children (66.4%) in the cohort experienced only one hospitalization, 17.5% had two hospitalizations, 7.9% had three hospitalizations, and 8.2% had four or more hospitalizations. The median age at hospitalization was 5.0 years (IQR 2.8-10.5) and most hospitalizations were for children who were male (56.6%), white (78.3%), non-Hispanic (96.0%), and publicly insured (51.7%; Table 1). More than one-quarter (28.6%) of hospitalizations were for children with four or more CCCs, and in 73.4% of hospitalizations, children were technology dependent (Table 1). Baseline respiratory support was common (46.8%), including home mechanical ventilation in 11.1% of hospitalizations (Table 1). Bacterial pneumonia, including aspiration pneumonia, was the most common discharge diagnosis (50.5%, Table 1).

Demographic and Clinical Characteristics

Children were discharged before return to respiratory baseline in 30.4% of hospitalizations (Appendix Figure). Children discharged before return to respiratory baseline were older (median age 5.7 years, IQR 3.1-11.0, vs 4.9 years, IQR 2.6-9.7; P = .04) and more likely to be privately insured (54.2% vs 43.4%; P = .04), compared with children discharged at respiratory baseline (Table 1). Children discharged before return to respiratory baseline were also more likely to have a respiratory CCC (59.9% vs 30.9%; P < .001), have a respiratory technology dependence diagnosis code (44.8% vs 24.1%; P < .001), and have baseline respiratory support needs on EMR review (67.7% vs 37.7%; P < .001), compared with children discharged at baseline (Table 1).

Children discharged before return to respiratory baseline required significantly greater escalation in respiratory support during hospitalization, compared with children discharged at respiratory baseline, including higher rates of initiation of CPAP or BiPAP, increased pressure settings from baseline (for home CPAP, BiPAP, or ventilator users), and initiation of full mechanical ventilation (Table 1). Hospitalizations in which children were discharged before return to respiratory baseline were also more likely to include ICU care than were those for children discharged at baseline (52.1% vs 35.2%; P < .001; Table 1).

Clinical Outcomes and Utilization

Reutilization within 30 days occurred after 32.1% of hospitalizations, with 26.1% requiring hospital readmission and 6.0% requiring ED revisit (Table 2). There was no statistical association in either unadjusted (Table 2) or adjusted (Table 3) analysis between children discharged before return to respiratory baseline and 30-day all-cause hospital reutilizations, hospital readmissions, or ED revisits.

In analysis of secondary outcomes, 30-day reutilization because of ARI occurred after 21.5% of hospitalizations, with 19.0% requiring hospital readmission and 2.5% requiring ED revisit. Median hospital LOS for the cohort was 4 days (IQR 2-8; Table 2). Hospitalizations in which children were discharged before return to respiratory baseline were longer than in those discharged at baseline (median 6 days, IQR 3-11, vs 4 days, IQR 2-7; P < .001; Table 2).

For hospitalizations of children discharged at respiratory baseline, the median time to return to respiratory baseline was 3 days (IQR 1-5, complete range 0-80). In these encounters, discharge occurred soon after return to respiratory baseline (median 1 day, IQR 0-1.5, complete range 0-54).

In subgroup analysis excluding the 18 hospitalizations in which children were lost to follow-up (2.8% of the total cohort), discharge before return to respiratory baseline was not associated with 30-day all-cause hospital reutilization (Table 4).

In this retrospective cohort study, children with NI hospitalized with ARI were frequently discharged using increased respiratory support from baseline. However, those discharged before return to respiratory baseline, despite their greater clinical complexity and acute illness severity, did not have increased hospital reutilization, compared with children discharged at respiratory baseline. Our findings suggest that discharge before return to baseline respiratory support after ARI may be clinically appropriate in some children with NI.

With the growing emphasis on decreasing hospital costs, concern exists that patients are being discharged from hospitals “quicker and sicker,”^18,19 with shortening lengths of stay and higher patient instability at discharge. Clinical instability at discharge has been associated with adverse postdischarge outcomes in adults with pneumonia^20-23; however, studies evaluating discharge readiness have not examined the population of children with NI. Our findings of no difference in reutilization for children with NI discharged before return to respiratory baseline, which would be expected to approximate one or more clinical instabilities, contrast these concerns.

Clinicians caring for children with NI hospitalized with ARI may find it difficult to determine a child’s discharge readiness, in part because many children with NI have longer times to return to respiratory baseline and some never return to their pre-­illness baseline.²⁴ In otherwise healthy children hospitalized with respiratory infections such as pneumonia, discharge criteria typically include complete wean from respiratory support prior to discharge.^4,25 In our study’s more complex children, whose parents already manage respiratory support at home, we hypothesize that discharging providers may be comfortable with discharge when the child has certain types of increased respiratory support compatible with home equipment, a parent skilled with monitoring the child’s respiratory status, and the support of an experienced outpatient provider and home nursing providers. At our institution, outpatient respiratory support weans are primarily performed by pediatric pulmonologists and, for isolated weaning of supplemental oxygen or time using support, by parents and outpatient pediatricians.

Another important factor in determining a child’s discharge readiness is the perspective of the child’s parent. Berry et al found that children whose parents believe they are not healthy enough for discharge are more likely to experience unplanned hospital readmissions,²⁴ signaling the role of child- and family-­specific factors in safe discharge decisions. Therefore, parents of children with NI should be proactively involved throughout the multidisciplinary discharge process,^26,27 including the decision to discharge before return to respiratory baseline. Parents have identified ongoing provider support, opportunities to practice home care skills, and written instructions with contingency plans as important components of discharge readiness.²⁸ Further work to create partnerships with these highly skilled caregivers in discharge decision making and transition planning are needed to promote safe discharge practices in this complex population.

In our study, children discharged before return to respiratory baseline were more likely to be older and privately insured compared with children discharged at respiratory baseline. Prior studies have found that social factors including low socioeconomic status influence ED provider admissions decisions for children with pneumonia.^29,30 However, the role of socioeconomic factors in provider discharge decisions for children with NI has not been assessed. These traits may also be proxy markers of other sociodemographic factors, such as parent education level, financial hardship influencing ability to participate in a child’s care at the bedside, access to comprehensive outpatient primary care, and availability of private home nursing. We hypothesize that these related characteristics directly and indirectly influence provider discharge decisions.

Discharging providers are likely more comfortable with discharge prior to return to respiratory baseline when the family has private duty nursing in the home. Home nurses can assist families in providing increased respiratory support and recognizing respiratory problems that may arise following discharge. However, home nursing shortages are common nationwide.^31,32 Low-income children, children with respiratory technology use, and children without Medicaid have been found to have larger gaps in home nursing availability.^31,32 Further studies are needed to understand the role of home nursing availability in provider discharge decisions in this population.

This study has several limitations. The retrospective design of this study creates the potential for residual confounding; there may be other clinical or demographic factors influencing hospital discharge decisions that we are unable to capture using EMR review, including parental knowledge and comfort managing illness, quality of discharge instructions, frequency of follow-up visits, and presence of skilled home nursing services. Categorization of children based on respiratory support status at discharge lends potential for misclassification of exposure; however, our substantial interrater reliability suggests that misclassification bias is small. This study’s primary finding indicated no difference between exposure groups; although we may be unable to detect small differences, we had sufficient power with our sample size to detect meaningful differences in reutilization outcomes.

This study was not designed to capture outpatient time to return to respiratory baseline; prospective studies are needed to identify rates of return to respiratory baseline following ARI in children with NI. We did not measure the level of respiratory support used by children at the time of discharge and, therefore, are unable to estimate the amount of respiratory support weaning needed following discharge or the compatibility of support with home equipment using our data. In addition, this study focused on respiratory support modalities and, thus, did not measure inpatient utilization of mucociliary clearance technologies that might be hypothesized to decrease the time to return to baseline respiratory support. Next steps in evaluating treatment of ARI include investigating the effect of mucociliary clearance on both exposure and outcome in this population.

There may be other clinically meaningful outcomes for this population apart from reutilization that we have not assessed in this study, including increased respiratory support required following discharge, primary care reutilization, healthcare costs, or parent satisfaction with timing of and outcomes after discharge. Finally, although our hospital has reutilization rates for children with NI that are similar to other institutions in the United States,³³ our results may not be generalizable to children with NI hospitalized at other institutions because local discharge processes and systems of care may be different. Prospective, multicenter investigation is needed to evaluate the clinical consequences of discharge before return to respiratory baseline more broadly.

At our institution, approximately one-quarter of children with NI hospitalized with ARI were discharged before return to respiratory baseline, but these children were not at increased risk of reutilization, compared with children discharged at respiratory baseline. Our findings suggest that return to baseline respiratory support might not be a necessary component of hospital discharge criteria. In otherwise clinically stable children with NI, discharge before return to respiratory baseline may be reasonable if their parents are comfortable managing respiratory support at home.

The authors thank Jonathan Rodean of the Children’s Hospital Association for his assistance with abstraction of PHIS data.

Children with neurologic impairment (NI; eg, hypoxic-­ischemic encephalopathy, muscular dystrophy) are characterized by functional and/or intellectual impairments resulting from a variety of neurologic diseases.¹ These children commonly have respiratory comorbidities, including central hypoventilation, impaired cough, and oromotor dysfunction, that may lead to chronic respiratory insufficiency and a need for chronic respiratory support at baseline.^2,3 Baseline respiratory support modalities can include supplemental oxygen, noninvasive positive pressure ventilation, or invasive mechanical ventilation.

Acute respiratory infections (ARI; eg, pneumonia, bronchiolitis) are the most common cause of hospitalization, intensive care unit (ICU) admission, and death for children with NI.^1,3 Discharge criteria for otherwise healthy children admitted to the hospital with ARI often include return to respiratory baseline.⁴ Children with complex chronic conditions have longer hospitalizations when hospitalized with respiratory infections,^5-7 because, in part, comorbidities and complications prolong the time to return to baseline. This prolonged return to respiratory baseline in combination with family knowledge, comfort, and skill in managing respiratory support and other complexities at home may alter discharge practices in the population of children with NI. In our clinical experience, discharge before return to baseline respiratory support occurs more frequently in children with NI than in otherwise healthy children when hospitalized with ARI. However, the consequences of discharging children with NI prior to return to respiratory baseline are unknown.

In this study, we sought to determine if discharge prior to return to baseline respiratory support is associated with reutilization among children with NI hospitalized with ARI. We hypothesized that patients discharged prior to return to respiratory baseline would have higher rates of 30-day hospital reutilization.

Study Design and Data Source

This single-center, retrospective cohort study of children hospitalized at Cincinnati Children’s Hospital Medical Center (CCHMC) used data from the Pediatric Health Information System (PHIS) and the electronic medical record (EMR). PHIS, an administrative database of 45 not-for-profit, tertiary care, US pediatric hospitals managed by Children’s Hospital Association (Lenexa, Kansas), was used to identify eligible children, examine demographic and clinical variables, and define outcomes. PHIS contains data regarding patient demographics, inpatient resource utilization, and diagnoses. Encrypted medical record numbers in PHIS allowed for local identification of patients’ medical records to complete EMR review to confirm eligibility and obtain detailed patient-level clinical information (eg, respiratory support needs) not available in PHIS.

Pilot medical record reviews allowed for standardized study definitions and procedures. All study staff underwent training with the primary investigator, including detailed review of 10 initial abstractions. Two investigators (K.M. and S.C.) performed repeat abstractions from 40 randomly selected records to enable assessment of interrater reliability. Average reliability, indicated by the κ statistic, indicated substantial to near-perfect reliability⁸ (κ = 0.97, 95% CI 0.90-1.00) for the primary exposure. EMR data were managed using Research Electronic Data Capture (REDCap, Nashville, Tennessee)⁹ and subsequently merged with PHIS data.

Study Population

Hospitalizations of children with NI aged 1 to 18 years at CCHMC between January 2010 and September 2015 were eligible for inclusion if they had a principal discharge diagnosis indicative of ARI and required increased respiratory support from baseline during hospitalization. NI was defined as a high-intensity, chronic neurological diagnosis with substantial functional impairments according to previously defined diagnosis codes.^1,10 ARI was identified using codes in the Clinical Classification Software (Agency for Healthcare Research and Quality, Rockville, MD) respiratory group indicative of ARI (eg, pneumonia, bronchiolitis, influenza; Appendix Table).

Children transferred to CCHMC were excluded because records from their initial illness presentation and management were not available. Because of expected differences in management and outcomes, children with a known diagnosis of tuberculosis or human immunodeficiency virus were excluded. Because exposure criteria were dependent on hospital discharge status, hospitalizations for children who died during admission (4 of 632 hospitalizations, 0.63%) were excluded from the final cohort (Appendix Figure).

Study Definitions

Baseline respiratory support (ie, “respiratory baseline”) was defined as the child’s highest level of respiratory support needed prior to admission when well (ie, no support, supplemental oxygen, continuous positive airway pressure [CPAP] or bilevel positive airway pressure [BiPAP], or ventilator support), and further characterized by night or day/night requirement. Respiratory baseline was identified using EMR documentation of home respiratory support use at the time of index admission. Return to respiratory baseline was defined as the date on which the child achieved documented home respiratory support settings, regardless of clinical symptoms.

Children may have required increased respiratory support from baseline at any time during hospitalization. Maximum respiratory support required was categorized as one of the following: (1) initiation of supplemental oxygen or increase in oxygen flow or duration; (2) initiation of CPAP or BiPAP; (3) increase in pressure settings or duration of pressure support for those with baseline CPAP, BiPAP, or ventilator use; and (4) initiation of full mechanical ventilation. Respiratory support categories were mutually exclusive: children requiring multiple types of increased respiratory support were classified for analysis by the most invasive form of respiratory support used (eg, a child requiring increase in both oxygen flow and pressure settings was categorized as an increase in pressure settings). Children who received heated high-flow nasal cannula therapy as maximum support were categorized as initiation or increase in oxygen support.

Time to return to respiratory baseline was defined as the difference in days between date of return to respiratory baseline and date of admission. Time to return to respiratory baseline was determined only for children who were discharged at respiratory baseline.

Primary Exposure and Outcome Measures

The primary exposure was hospital discharge before return to respiratory baseline (ie, discharge on higher respiratory support than at baseline settings). At our institution, standardized discharge criteria for children with NI do not exist. The primary outcome was all-cause, 30-day hospital reutilization, including hospital readmissions and emergency department (ED) revisits. Secondary outcomes included 30-day reutilization for ARI and hospital length of stay (LOS) in days.

Patient Demographics and Clinical Characteristics

Demographic and patient characteristics that might influence hospital discharge before return to respiratory baseline or readmission were obtained from PHIS (eg, demographic information, age, insurance type, measures of clinical complexity, illness severity) and by EMR review (eg, baseline respiratory support needs, maximum respiratory support during hospitalization). Measures of clinical complexity included comorbid complex chronic conditions (CCCs)^11-14 and technology dependence^14-16 using previously defined diagnostic codes. Measures of illness severity included sepsis¹⁷ and ICU-level care. At our institution, children with baseline ventilator use do not require admission to the ICU unless they are clinically unstable.

Statistical Analysis

Continuous variables were described using medians and interquartile ranges (IQR). Categorical variables were described using counts and percentages. Patient characteristics and outcomes were stratified by primary exposure and compared using chi-square test or Fisher exact test for categorical variables and Wilcoxon rank sum test for continuous variables.

To examine the independent association between discharge before return to respiratory baseline and hospital reutilization, a generalized estimating equation was used that included potential confounders while accounting for within-patient clustering. Patient demographics included age, race, ethnicity, and insurance type; measures of clinical complexity included number of CCCs, technology dependence, and baseline respiratory support; and measures of acute illness severity included ARI diagnosis, degree of increase in respiratory support during hospitalization, and ICU-level care. LOS was also included in the model as a covariate because of its expected association with both exposure and outcome.

Secondary analyses were conducted using the outcome of 30-day reutilization for ARI. Subgroup analysis excluding hospitalizations of children lost to follow-up (ie, no encounters in the 6 months after hospital discharge) was also conducted. All analyses were performed with SAS v9.3 (SAS Institute, Cary, North Carolina). P values < .05 were considered statistically significant. This study was approved by the Institutional Review Board.

Study Cohort

A total of 632 hospitalizations experienced by 366 children with NI who were admitted with ARI were included (Appendix Figure). Most children (66.4%) in the cohort experienced only one hospitalization, 17.5% had two hospitalizations, 7.9% had three hospitalizations, and 8.2% had four or more hospitalizations. The median age at hospitalization was 5.0 years (IQR 2.8-10.5) and most hospitalizations were for children who were male (56.6%), white (78.3%), non-Hispanic (96.0%), and publicly insured (51.7%; Table 1). More than one-quarter (28.6%) of hospitalizations were for children with four or more CCCs, and in 73.4% of hospitalizations, children were technology dependent (Table 1). Baseline respiratory support was common (46.8%), including home mechanical ventilation in 11.1% of hospitalizations (Table 1). Bacterial pneumonia, including aspiration pneumonia, was the most common discharge diagnosis (50.5%, Table 1).

Demographic and Clinical Characteristics

Children were discharged before return to respiratory baseline in 30.4% of hospitalizations (Appendix Figure). Children discharged before return to respiratory baseline were older (median age 5.7 years, IQR 3.1-11.0, vs 4.9 years, IQR 2.6-9.7; P = .04) and more likely to be privately insured (54.2% vs 43.4%; P = .04), compared with children discharged at respiratory baseline (Table 1). Children discharged before return to respiratory baseline were also more likely to have a respiratory CCC (59.9% vs 30.9%; P < .001), have a respiratory technology dependence diagnosis code (44.8% vs 24.1%; P < .001), and have baseline respiratory support needs on EMR review (67.7% vs 37.7%; P < .001), compared with children discharged at baseline (Table 1).

Children discharged before return to respiratory baseline required significantly greater escalation in respiratory support during hospitalization, compared with children discharged at respiratory baseline, including higher rates of initiation of CPAP or BiPAP, increased pressure settings from baseline (for home CPAP, BiPAP, or ventilator users), and initiation of full mechanical ventilation (Table 1). Hospitalizations in which children were discharged before return to respiratory baseline were also more likely to include ICU care than were those for children discharged at baseline (52.1% vs 35.2%; P < .001; Table 1).

Clinical Outcomes and Utilization

Reutilization within 30 days occurred after 32.1% of hospitalizations, with 26.1% requiring hospital readmission and 6.0% requiring ED revisit (Table 2). There was no statistical association in either unadjusted (Table 2) or adjusted (Table 3) analysis between children discharged before return to respiratory baseline and 30-day all-cause hospital reutilizations, hospital readmissions, or ED revisits.

In analysis of secondary outcomes, 30-day reutilization because of ARI occurred after 21.5% of hospitalizations, with 19.0% requiring hospital readmission and 2.5% requiring ED revisit. Median hospital LOS for the cohort was 4 days (IQR 2-8; Table 2). Hospitalizations in which children were discharged before return to respiratory baseline were longer than in those discharged at baseline (median 6 days, IQR 3-11, vs 4 days, IQR 2-7; P < .001; Table 2).

For hospitalizations of children discharged at respiratory baseline, the median time to return to respiratory baseline was 3 days (IQR 1-5, complete range 0-80). In these encounters, discharge occurred soon after return to respiratory baseline (median 1 day, IQR 0-1.5, complete range 0-54).

In subgroup analysis excluding the 18 hospitalizations in which children were lost to follow-up (2.8% of the total cohort), discharge before return to respiratory baseline was not associated with 30-day all-cause hospital reutilization (Table 4).

In this retrospective cohort study, children with NI hospitalized with ARI were frequently discharged using increased respiratory support from baseline. However, those discharged before return to respiratory baseline, despite their greater clinical complexity and acute illness severity, did not have increased hospital reutilization, compared with children discharged at respiratory baseline. Our findings suggest that discharge before return to baseline respiratory support after ARI may be clinically appropriate in some children with NI.

With the growing emphasis on decreasing hospital costs, concern exists that patients are being discharged from hospitals “quicker and sicker,”^18,19 with shortening lengths of stay and higher patient instability at discharge. Clinical instability at discharge has been associated with adverse postdischarge outcomes in adults with pneumonia^20-23; however, studies evaluating discharge readiness have not examined the population of children with NI. Our findings of no difference in reutilization for children with NI discharged before return to respiratory baseline, which would be expected to approximate one or more clinical instabilities, contrast these concerns.

Clinicians caring for children with NI hospitalized with ARI may find it difficult to determine a child’s discharge readiness, in part because many children with NI have longer times to return to respiratory baseline and some never return to their pre-­illness baseline.²⁴ In otherwise healthy children hospitalized with respiratory infections such as pneumonia, discharge criteria typically include complete wean from respiratory support prior to discharge.^4,25 In our study’s more complex children, whose parents already manage respiratory support at home, we hypothesize that discharging providers may be comfortable with discharge when the child has certain types of increased respiratory support compatible with home equipment, a parent skilled with monitoring the child’s respiratory status, and the support of an experienced outpatient provider and home nursing providers. At our institution, outpatient respiratory support weans are primarily performed by pediatric pulmonologists and, for isolated weaning of supplemental oxygen or time using support, by parents and outpatient pediatricians.

Another important factor in determining a child’s discharge readiness is the perspective of the child’s parent. Berry et al found that children whose parents believe they are not healthy enough for discharge are more likely to experience unplanned hospital readmissions,²⁴ signaling the role of child- and family-­specific factors in safe discharge decisions. Therefore, parents of children with NI should be proactively involved throughout the multidisciplinary discharge process,^26,27 including the decision to discharge before return to respiratory baseline. Parents have identified ongoing provider support, opportunities to practice home care skills, and written instructions with contingency plans as important components of discharge readiness.²⁸ Further work to create partnerships with these highly skilled caregivers in discharge decision making and transition planning are needed to promote safe discharge practices in this complex population.

In our study, children discharged before return to respiratory baseline were more likely to be older and privately insured compared with children discharged at respiratory baseline. Prior studies have found that social factors including low socioeconomic status influence ED provider admissions decisions for children with pneumonia.^29,30 However, the role of socioeconomic factors in provider discharge decisions for children with NI has not been assessed. These traits may also be proxy markers of other sociodemographic factors, such as parent education level, financial hardship influencing ability to participate in a child’s care at the bedside, access to comprehensive outpatient primary care, and availability of private home nursing. We hypothesize that these related characteristics directly and indirectly influence provider discharge decisions.

Discharging providers are likely more comfortable with discharge prior to return to respiratory baseline when the family has private duty nursing in the home. Home nurses can assist families in providing increased respiratory support and recognizing respiratory problems that may arise following discharge. However, home nursing shortages are common nationwide.^31,32 Low-income children, children with respiratory technology use, and children without Medicaid have been found to have larger gaps in home nursing availability.^31,32 Further studies are needed to understand the role of home nursing availability in provider discharge decisions in this population.

This study has several limitations. The retrospective design of this study creates the potential for residual confounding; there may be other clinical or demographic factors influencing hospital discharge decisions that we are unable to capture using EMR review, including parental knowledge and comfort managing illness, quality of discharge instructions, frequency of follow-up visits, and presence of skilled home nursing services. Categorization of children based on respiratory support status at discharge lends potential for misclassification of exposure; however, our substantial interrater reliability suggests that misclassification bias is small. This study’s primary finding indicated no difference between exposure groups; although we may be unable to detect small differences, we had sufficient power with our sample size to detect meaningful differences in reutilization outcomes.

This study was not designed to capture outpatient time to return to respiratory baseline; prospective studies are needed to identify rates of return to respiratory baseline following ARI in children with NI. We did not measure the level of respiratory support used by children at the time of discharge and, therefore, are unable to estimate the amount of respiratory support weaning needed following discharge or the compatibility of support with home equipment using our data. In addition, this study focused on respiratory support modalities and, thus, did not measure inpatient utilization of mucociliary clearance technologies that might be hypothesized to decrease the time to return to baseline respiratory support. Next steps in evaluating treatment of ARI include investigating the effect of mucociliary clearance on both exposure and outcome in this population.

There may be other clinically meaningful outcomes for this population apart from reutilization that we have not assessed in this study, including increased respiratory support required following discharge, primary care reutilization, healthcare costs, or parent satisfaction with timing of and outcomes after discharge. Finally, although our hospital has reutilization rates for children with NI that are similar to other institutions in the United States,³³ our results may not be generalizable to children with NI hospitalized at other institutions because local discharge processes and systems of care may be different. Prospective, multicenter investigation is needed to evaluate the clinical consequences of discharge before return to respiratory baseline more broadly.

At our institution, approximately one-quarter of children with NI hospitalized with ARI were discharged before return to respiratory baseline, but these children were not at increased risk of reutilization, compared with children discharged at respiratory baseline. Our findings suggest that return to baseline respiratory support might not be a necessary component of hospital discharge criteria. In otherwise clinically stable children with NI, discharge before return to respiratory baseline may be reasonable if their parents are comfortable managing respiratory support at home.

The authors thank Jonathan Rodean of the Children’s Hospital Association for his assistance with abstraction of PHIS data.

Due to the pandemic spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the cause of COVID-19 infection, there is significant disruption to the global supply of PPE.² Order volumes of PPE have increased, prices have surged, and distributors are experiencing challenges meeting order demands.³ With decreased overseas exports, suppliers have placed hospitals on PPE allocations, and many hospitals’ orders for PPE have been only partially filled.^3,4 Unless hospitals have established stockpiles, most only have supplies for 3-7 days of routine use, leaving them vulnerable to exhausting PPE supplies. At the onset of the pandemic, 86% of United States hospitals reported concerns about their PPE supply.⁴

The potential for PPE shortages has led both the Centers for Disease Control and Prevention (CDC) and the World Health Organization to call for the rational and appropriate use of PPE in order to conserve supplies.^2,3 By the time COVID-19 was declared a pandemic, 54% of hospitals had imposed PPE conservation protocols,⁴ with more expected to follow in the weeks and months to come. Innovative protocols have been conceptualized and used to conserve PPE in hospitals (Table).

Yet these conservation protocols often fail to identify missed opportunities to improve the value of PPE that already exist in hospital care. By defining the value of inpatient PPE, hospitals can identify opportunities for value improvement. Changes implemented now will maximize PPE value and preserve supply during this pandemic and beyond.

In order to conserve PPE supply, hospitals might consider limiting PPE to cases in which clear evidence exists to support its use. However, evidence for PPE use can be challenging to interpret because the impact of preventing nosocomial infections (an outcome that did not occur) is inherently problematic to measure. This makes assessing the value of PPE in preventing nosocomial transmission in specific situations difficult.

The basis of using PPE is its effectiveness in controlling outbreaks.¹ A meta-analysis of 6 case-control studies from the SARS outbreak of 2003, which disproportionately infected healthcare workers, suggested that handwashing and PPE were effective in preventing disease transmission. Handwashing alone reduced transmission by 55%, wearing gloves by 57%, and wearing facemasks by 68%; the cumulative effect of handwashing, masks, gloves, and gowns reduced transmission by 91%.⁵ A cohort study of healthcare workers exposed to H1N1 influenza A in 2009 found that use of a facemask or an N95 respirator was associated with negative viral serology suggesting noninfected status.⁶ With respiratory syncytial virus (RSV) outbreaks, a narrative synthesis of 4 studies examining transmission also suggested gowns, facemasks, and eye protection are effective, with eye protection perhaps more effective than gowns and masks.⁷ Yet these studies’ conclusions are limited by study design differences and small sample sizes.

The evidence supporting PPE use for routine hospital conditions is more challenging to interpret. One pediatric study of seasonal respiratory viruses showed that adding droplet precautions to an existing policy of contact precautions alone decreased nosocomial infections for most viruses evaluated.⁸ Yet this study, like many of PPE use, is limited by sample size and possible misclassification of exposure and outcome biases. Because PPE is always utilized in conjunction with other preventive measures, isolating the impact of PPE is challenging, let alone isolating the individual effects of PPE components. In the absence of strong empirical evidence, hospitals must rely on the inherent rationale of PPE use for patient and healthcare worker safety in assessing its value.

In order to protect patients from disease transmission during a pandemic, hospitals might also reconsider whether to use PPE in cases in which evidence is absent, such as routine prevention for colonized but noninfected patients. However, evidence of the possible patient harms of PPE are emerging. Healthcare providers spend less time with isolated patients^9,10 and document fewer vital signs.¹¹ Patients in PPE may experience delays in admission¹² and discharge,¹³ and have higher rates of falls, pressure ulcers, and medication errors.^14,15 They may also experience higher rates of anxiety and depression.¹⁶ Yet no evidence suggests PPE use for noninfected patients prevents transmission to patients or to healthcare workers. Using PPE when it is not indicated deemphasizes the value of other preventative precautions (eg, handwashing), unnecessarily depletes PPE supply, and may create patient harm without added benefit. High-value PPE, both during a pandemic and beyond, is defined by a system designed so that healthcare workers use PPE when they need it, and do not use PPE when not indicated.

While all hospitalized patients are admitted using standard precautions, decisions surrounding PPE can be nuanced for even experienced clinicians. Although the CDC does provide guidance for PPE use based on symptoms that correlate with potential for transmission (eg, patients with cough should be placed in at least droplet precautions),¹ guidelines must rely on provider evaluation and interpretation. For instance, three etiologies of cough—pneumococcal pneumonia, RSV bronchiolitis, and pulmonary tuberculosis—would all require different PPE. The clinician must weigh the probabilities of each pathogen and assess the harm of not protecting against certain pathogens in his or her decision.

Amidst the stress and cognitive burdens placed on clinicians, accuracy in PPE decisions is easily deprioritized. Clinicians may not completely consider patient-specific indications for PPE, implications for patients and staff, and supply shortages. Although the CDC and many hospitals have PPE initiation and discontinuation criteria, clinicians may favor educated guesswork and reliance on past experience when guidelines are poorly accessible or poorly searchable. Such individual, nonstandardized decisions likely lead to variability in practice patterns, inaccuracies in PPE decisions, and ultimately waste of PPE resources.

At our institution, the inconveniences, cognitive burden, and perceived benefit of routine PPE interventions have created a system in which PPE is regularly overused. On our hospital medicine wards, we found that PPE was both over-ordered upon admission (eg, contact/droplet precautions ordered for influenza when droplet precautions only would have sufficed) and unnecessarily continued even after children met discontinuation criteria.

On discharge review from our general pediatric ward in 2019, 18% of children discharged with PPE orders no longer met criteria for PPE. Two conditions—community-acquired bacterial pneumonia and skin and soft-tissue infections—accounted for 47% of discharges with unnecessary PPE orders. At an estimated cost of $0.13-$0.53 for droplet precautions per use, $0.69 for contact precautions, and $0.82-$1.22 for both, the absolute cost of continuing PPE without indication could be as high as $61/day per patient when estimating 50 uses per day. This direct cost represents healthcare spending without added value when PPE are not necessary. Furthermore, the additional emotional cost to the patient and family in their hospitalization experience, the cost of clinician time donning and doffing, the environmental cost of PPE waste, and the cost to the limited PPE supply are not considered in these calculations.

During a pandemic characterized by PPE shortages nationwide, allowing missed opportunities for PPE discontinuation to persist is not only wasteful, but inattentive to public health.

For individual clinicians, opportunities exist to improve PPE usage in daily patient care. Clinicians should not overlook PPE decisions; instead they should make it a practice to review PPE orders daily during rounds as they would lab orders. Clinicians and nursing staff should work together to identify PPE discontinuation opportunities, leveraging the electronic medical record when possible. For the benefit of patients and families, clinicians and bedside staff should recognize and assist in managing patient expectations of PPE.

Hospitals should work to make PPE references easily accessible and interpretable by frontline clinicians. To minimize variability of use, PPE ordering for routine conditions should be standardized and streamlined, including discontinuation criteria. Hospitals should invest in behavioral health programs to support patients with conditions necessitating PPE and develop policies to ensure ancillary services are equally available to all patients. To alleviate concerns about limited clinician time spent with isolated patients, hospitals should assign clinician workloads while accounting for the known increased time needed to care for patients with PPE.

For hospitals with extreme supply shortages, conservation might include decreased use of PPE for conditions in which its use is controversial (eg, patients colonized with methicillin-resistant Staphylococcus aureus or multidrug resistant organisms) as has been trialed in institutions prior to this pandemic.^17,18 Such PPE policy changes might occur in addition to, or in conjunction with, the conservation strategies suggested by other institutions (Table).

Healthcare systems should continually reassess the value of PPE for their hospitals and make changes accordingly. In the midst of difficulties directly demonstrating PPE’s value, hospitals must rely on the inherent rationale of PPE use in assessing value decisions weighed against harms while balancing healthcare worker protection regulations. Decisions should always occur while continuing other sensible infection-control procedures, such as handwashing and environmental hygiene measures.

To effect maximal change, healthcare systems should invest in redesigning PPE ordering systems at the highest level. This should include harnessing existing technologies to streamline PPE ordering decisions to meet clinicians’ cognitive needs. Decision support and auditing technologies could automate and monitor PPE orders efficiently. Likely to be most effective, an investment in creating and maintaining centralized PPE expert management teams to assess, order, and discontinue PPE would minimize individual ordering variation, minimize cost, and maximize value to patients, staff, and hospitals.

In this pandemic, we have the opportunity to rethink how we understand and use PPE in hospitalized patients. It is vitally important now more than ever to consciously conserve the limited supply of PPE resources. As we seek to increase healthcare value while limiting overuse and waste, PPE is a prime target for value improvement efforts as the effective but also burdensome tool that it is. Hospitalists are well-positioned to lead the redesign of how, when, and why PPE is used and to create a more optimized, lasting system that provides maximal value to patients, families, and healthcare workers during this current crisis and beyond.

Due to the pandemic spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the cause of COVID-19 infection, there is significant disruption to the global supply of PPE.² Order volumes of PPE have increased, prices have surged, and distributors are experiencing challenges meeting order demands.³ With decreased overseas exports, suppliers have placed hospitals on PPE allocations, and many hospitals’ orders for PPE have been only partially filled.^3,4 Unless hospitals have established stockpiles, most only have supplies for 3-7 days of routine use, leaving them vulnerable to exhausting PPE supplies. At the onset of the pandemic, 86% of United States hospitals reported concerns about their PPE supply.⁴

The potential for PPE shortages has led both the Centers for Disease Control and Prevention (CDC) and the World Health Organization to call for the rational and appropriate use of PPE in order to conserve supplies.^2,3 By the time COVID-19 was declared a pandemic, 54% of hospitals had imposed PPE conservation protocols,⁴ with more expected to follow in the weeks and months to come. Innovative protocols have been conceptualized and used to conserve PPE in hospitals (Table).

Yet these conservation protocols often fail to identify missed opportunities to improve the value of PPE that already exist in hospital care. By defining the value of inpatient PPE, hospitals can identify opportunities for value improvement. Changes implemented now will maximize PPE value and preserve supply during this pandemic and beyond.

In order to conserve PPE supply, hospitals might consider limiting PPE to cases in which clear evidence exists to support its use. However, evidence for PPE use can be challenging to interpret because the impact of preventing nosocomial infections (an outcome that did not occur) is inherently problematic to measure. This makes assessing the value of PPE in preventing nosocomial transmission in specific situations difficult.

The basis of using PPE is its effectiveness in controlling outbreaks.¹ A meta-analysis of 6 case-control studies from the SARS outbreak of 2003, which disproportionately infected healthcare workers, suggested that handwashing and PPE were effective in preventing disease transmission. Handwashing alone reduced transmission by 55%, wearing gloves by 57%, and wearing facemasks by 68%; the cumulative effect of handwashing, masks, gloves, and gowns reduced transmission by 91%.⁵ A cohort study of healthcare workers exposed to H1N1 influenza A in 2009 found that use of a facemask or an N95 respirator was associated with negative viral serology suggesting noninfected status.⁶ With respiratory syncytial virus (RSV) outbreaks, a narrative synthesis of 4 studies examining transmission also suggested gowns, facemasks, and eye protection are effective, with eye protection perhaps more effective than gowns and masks.⁷ Yet these studies’ conclusions are limited by study design differences and small sample sizes.

The evidence supporting PPE use for routine hospital conditions is more challenging to interpret. One pediatric study of seasonal respiratory viruses showed that adding droplet precautions to an existing policy of contact precautions alone decreased nosocomial infections for most viruses evaluated.⁸ Yet this study, like many of PPE use, is limited by sample size and possible misclassification of exposure and outcome biases. Because PPE is always utilized in conjunction with other preventive measures, isolating the impact of PPE is challenging, let alone isolating the individual effects of PPE components. In the absence of strong empirical evidence, hospitals must rely on the inherent rationale of PPE use for patient and healthcare worker safety in assessing its value.

In order to protect patients from disease transmission during a pandemic, hospitals might also reconsider whether to use PPE in cases in which evidence is absent, such as routine prevention for colonized but noninfected patients. However, evidence of the possible patient harms of PPE are emerging. Healthcare providers spend less time with isolated patients^9,10 and document fewer vital signs.¹¹ Patients in PPE may experience delays in admission¹² and discharge,¹³ and have higher rates of falls, pressure ulcers, and medication errors.^14,15 They may also experience higher rates of anxiety and depression.¹⁶ Yet no evidence suggests PPE use for noninfected patients prevents transmission to patients or to healthcare workers. Using PPE when it is not indicated deemphasizes the value of other preventative precautions (eg, handwashing), unnecessarily depletes PPE supply, and may create patient harm without added benefit. High-value PPE, both during a pandemic and beyond, is defined by a system designed so that healthcare workers use PPE when they need it, and do not use PPE when not indicated.

While all hospitalized patients are admitted using standard precautions, decisions surrounding PPE can be nuanced for even experienced clinicians. Although the CDC does provide guidance for PPE use based on symptoms that correlate with potential for transmission (eg, patients with cough should be placed in at least droplet precautions),¹ guidelines must rely on provider evaluation and interpretation. For instance, three etiologies of cough—pneumococcal pneumonia, RSV bronchiolitis, and pulmonary tuberculosis—would all require different PPE. The clinician must weigh the probabilities of each pathogen and assess the harm of not protecting against certain pathogens in his or her decision.

Amidst the stress and cognitive burdens placed on clinicians, accuracy in PPE decisions is easily deprioritized. Clinicians may not completely consider patient-specific indications for PPE, implications for patients and staff, and supply shortages. Although the CDC and many hospitals have PPE initiation and discontinuation criteria, clinicians may favor educated guesswork and reliance on past experience when guidelines are poorly accessible or poorly searchable. Such individual, nonstandardized decisions likely lead to variability in practice patterns, inaccuracies in PPE decisions, and ultimately waste of PPE resources.

At our institution, the inconveniences, cognitive burden, and perceived benefit of routine PPE interventions have created a system in which PPE is regularly overused. On our hospital medicine wards, we found that PPE was both over-ordered upon admission (eg, contact/droplet precautions ordered for influenza when droplet precautions only would have sufficed) and unnecessarily continued even after children met discontinuation criteria.

On discharge review from our general pediatric ward in 2019, 18% of children discharged with PPE orders no longer met criteria for PPE. Two conditions—community-acquired bacterial pneumonia and skin and soft-tissue infections—accounted for 47% of discharges with unnecessary PPE orders. At an estimated cost of $0.13-$0.53 for droplet precautions per use, $0.69 for contact precautions, and $0.82-$1.22 for both, the absolute cost of continuing PPE without indication could be as high as $61/day per patient when estimating 50 uses per day. This direct cost represents healthcare spending without added value when PPE are not necessary. Furthermore, the additional emotional cost to the patient and family in their hospitalization experience, the cost of clinician time donning and doffing, the environmental cost of PPE waste, and the cost to the limited PPE supply are not considered in these calculations.

During a pandemic characterized by PPE shortages nationwide, allowing missed opportunities for PPE discontinuation to persist is not only wasteful, but inattentive to public health.

For individual clinicians, opportunities exist to improve PPE usage in daily patient care. Clinicians should not overlook PPE decisions; instead they should make it a practice to review PPE orders daily during rounds as they would lab orders. Clinicians and nursing staff should work together to identify PPE discontinuation opportunities, leveraging the electronic medical record when possible. For the benefit of patients and families, clinicians and bedside staff should recognize and assist in managing patient expectations of PPE.

Hospitals should work to make PPE references easily accessible and interpretable by frontline clinicians. To minimize variability of use, PPE ordering for routine conditions should be standardized and streamlined, including discontinuation criteria. Hospitals should invest in behavioral health programs to support patients with conditions necessitating PPE and develop policies to ensure ancillary services are equally available to all patients. To alleviate concerns about limited clinician time spent with isolated patients, hospitals should assign clinician workloads while accounting for the known increased time needed to care for patients with PPE.

For hospitals with extreme supply shortages, conservation might include decreased use of PPE for conditions in which its use is controversial (eg, patients colonized with methicillin-resistant Staphylococcus aureus or multidrug resistant organisms) as has been trialed in institutions prior to this pandemic.^17,18 Such PPE policy changes might occur in addition to, or in conjunction with, the conservation strategies suggested by other institutions (Table).

Healthcare systems should continually reassess the value of PPE for their hospitals and make changes accordingly. In the midst of difficulties directly demonstrating PPE’s value, hospitals must rely on the inherent rationale of PPE use in assessing value decisions weighed against harms while balancing healthcare worker protection regulations. Decisions should always occur while continuing other sensible infection-control procedures, such as handwashing and environmental hygiene measures.

To effect maximal change, healthcare systems should invest in redesigning PPE ordering systems at the highest level. This should include harnessing existing technologies to streamline PPE ordering decisions to meet clinicians’ cognitive needs. Decision support and auditing technologies could automate and monitor PPE orders efficiently. Likely to be most effective, an investment in creating and maintaining centralized PPE expert management teams to assess, order, and discontinue PPE would minimize individual ordering variation, minimize cost, and maximize value to patients, staff, and hospitals.

In this pandemic, we have the opportunity to rethink how we understand and use PPE in hospitalized patients. It is vitally important now more than ever to consciously conserve the limited supply of PPE resources. As we seek to increase healthcare value while limiting overuse and waste, PPE is a prime target for value improvement efforts as the effective but also burdensome tool that it is. Hospitalists are well-positioned to lead the redesign of how, when, and why PPE is used and to create a more optimized, lasting system that provides maximal value to patients, families, and healthcare workers during this current crisis and beyond.

Due to the pandemic spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the cause of COVID-19 infection, there is significant disruption to the global supply of PPE.² Order volumes of PPE have increased, prices have surged, and distributors are experiencing challenges meeting order demands.³ With decreased overseas exports, suppliers have placed hospitals on PPE allocations, and many hospitals’ orders for PPE have been only partially filled.^3,4 Unless hospitals have established stockpiles, most only have supplies for 3-7 days of routine use, leaving them vulnerable to exhausting PPE supplies. At the onset of the pandemic, 86% of United States hospitals reported concerns about their PPE supply.⁴

The potential for PPE shortages has led both the Centers for Disease Control and Prevention (CDC) and the World Health Organization to call for the rational and appropriate use of PPE in order to conserve supplies.^2,3 By the time COVID-19 was declared a pandemic, 54% of hospitals had imposed PPE conservation protocols,⁴ with more expected to follow in the weeks and months to come. Innovative protocols have been conceptualized and used to conserve PPE in hospitals (Table).

Yet these conservation protocols often fail to identify missed opportunities to improve the value of PPE that already exist in hospital care. By defining the value of inpatient PPE, hospitals can identify opportunities for value improvement. Changes implemented now will maximize PPE value and preserve supply during this pandemic and beyond.

In order to conserve PPE supply, hospitals might consider limiting PPE to cases in which clear evidence exists to support its use. However, evidence for PPE use can be challenging to interpret because the impact of preventing nosocomial infections (an outcome that did not occur) is inherently problematic to measure. This makes assessing the value of PPE in preventing nosocomial transmission in specific situations difficult.

The basis of using PPE is its effectiveness in controlling outbreaks.¹ A meta-analysis of 6 case-control studies from the SARS outbreak of 2003, which disproportionately infected healthcare workers, suggested that handwashing and PPE were effective in preventing disease transmission. Handwashing alone reduced transmission by 55%, wearing gloves by 57%, and wearing facemasks by 68%; the cumulative effect of handwashing, masks, gloves, and gowns reduced transmission by 91%.⁵ A cohort study of healthcare workers exposed to H1N1 influenza A in 2009 found that use of a facemask or an N95 respirator was associated with negative viral serology suggesting noninfected status.⁶ With respiratory syncytial virus (RSV) outbreaks, a narrative synthesis of 4 studies examining transmission also suggested gowns, facemasks, and eye protection are effective, with eye protection perhaps more effective than gowns and masks.⁷ Yet these studies’ conclusions are limited by study design differences and small sample sizes.

The evidence supporting PPE use for routine hospital conditions is more challenging to interpret. One pediatric study of seasonal respiratory viruses showed that adding droplet precautions to an existing policy of contact precautions alone decreased nosocomial infections for most viruses evaluated.⁸ Yet this study, like many of PPE use, is limited by sample size and possible misclassification of exposure and outcome biases. Because PPE is always utilized in conjunction with other preventive measures, isolating the impact of PPE is challenging, let alone isolating the individual effects of PPE components. In the absence of strong empirical evidence, hospitals must rely on the inherent rationale of PPE use for patient and healthcare worker safety in assessing its value.

In order to protect patients from disease transmission during a pandemic, hospitals might also reconsider whether to use PPE in cases in which evidence is absent, such as routine prevention for colonized but noninfected patients. However, evidence of the possible patient harms of PPE are emerging. Healthcare providers spend less time with isolated patients^9,10 and document fewer vital signs.¹¹ Patients in PPE may experience delays in admission¹² and discharge,¹³ and have higher rates of falls, pressure ulcers, and medication errors.^14,15 They may also experience higher rates of anxiety and depression.¹⁶ Yet no evidence suggests PPE use for noninfected patients prevents transmission to patients or to healthcare workers. Using PPE when it is not indicated deemphasizes the value of other preventative precautions (eg, handwashing), unnecessarily depletes PPE supply, and may create patient harm without added benefit. High-value PPE, both during a pandemic and beyond, is defined by a system designed so that healthcare workers use PPE when they need it, and do not use PPE when not indicated.

While all hospitalized patients are admitted using standard precautions, decisions surrounding PPE can be nuanced for even experienced clinicians. Although the CDC does provide guidance for PPE use based on symptoms that correlate with potential for transmission (eg, patients with cough should be placed in at least droplet precautions),¹ guidelines must rely on provider evaluation and interpretation. For instance, three etiologies of cough—pneumococcal pneumonia, RSV bronchiolitis, and pulmonary tuberculosis—would all require different PPE. The clinician must weigh the probabilities of each pathogen and assess the harm of not protecting against certain pathogens in his or her decision.

Amidst the stress and cognitive burdens placed on clinicians, accuracy in PPE decisions is easily deprioritized. Clinicians may not completely consider patient-specific indications for PPE, implications for patients and staff, and supply shortages. Although the CDC and many hospitals have PPE initiation and discontinuation criteria, clinicians may favor educated guesswork and reliance on past experience when guidelines are poorly accessible or poorly searchable. Such individual, nonstandardized decisions likely lead to variability in practice patterns, inaccuracies in PPE decisions, and ultimately waste of PPE resources.

At our institution, the inconveniences, cognitive burden, and perceived benefit of routine PPE interventions have created a system in which PPE is regularly overused. On our hospital medicine wards, we found that PPE was both over-ordered upon admission (eg, contact/droplet precautions ordered for influenza when droplet precautions only would have sufficed) and unnecessarily continued even after children met discontinuation criteria.

On discharge review from our general pediatric ward in 2019, 18% of children discharged with PPE orders no longer met criteria for PPE. Two conditions—community-acquired bacterial pneumonia and skin and soft-tissue infections—accounted for 47% of discharges with unnecessary PPE orders. At an estimated cost of $0.13-$0.53 for droplet precautions per use, $0.69 for contact precautions, and $0.82-$1.22 for both, the absolute cost of continuing PPE without indication could be as high as $61/day per patient when estimating 50 uses per day. This direct cost represents healthcare spending without added value when PPE are not necessary. Furthermore, the additional emotional cost to the patient and family in their hospitalization experience, the cost of clinician time donning and doffing, the environmental cost of PPE waste, and the cost to the limited PPE supply are not considered in these calculations.

During a pandemic characterized by PPE shortages nationwide, allowing missed opportunities for PPE discontinuation to persist is not only wasteful, but inattentive to public health.

For individual clinicians, opportunities exist to improve PPE usage in daily patient care. Clinicians should not overlook PPE decisions; instead they should make it a practice to review PPE orders daily during rounds as they would lab orders. Clinicians and nursing staff should work together to identify PPE discontinuation opportunities, leveraging the electronic medical record when possible. For the benefit of patients and families, clinicians and bedside staff should recognize and assist in managing patient expectations of PPE.

Hospitals should work to make PPE references easily accessible and interpretable by frontline clinicians. To minimize variability of use, PPE ordering for routine conditions should be standardized and streamlined, including discontinuation criteria. Hospitals should invest in behavioral health programs to support patients with conditions necessitating PPE and develop policies to ensure ancillary services are equally available to all patients. To alleviate concerns about limited clinician time spent with isolated patients, hospitals should assign clinician workloads while accounting for the known increased time needed to care for patients with PPE.

For hospitals with extreme supply shortages, conservation might include decreased use of PPE for conditions in which its use is controversial (eg, patients colonized with methicillin-resistant Staphylococcus aureus or multidrug resistant organisms) as has been trialed in institutions prior to this pandemic.^17,18 Such PPE policy changes might occur in addition to, or in conjunction with, the conservation strategies suggested by other institutions (Table).

Healthcare systems should continually reassess the value of PPE for their hospitals and make changes accordingly. In the midst of difficulties directly demonstrating PPE’s value, hospitals must rely on the inherent rationale of PPE use in assessing value decisions weighed against harms while balancing healthcare worker protection regulations. Decisions should always occur while continuing other sensible infection-control procedures, such as handwashing and environmental hygiene measures.

To effect maximal change, healthcare systems should invest in redesigning PPE ordering systems at the highest level. This should include harnessing existing technologies to streamline PPE ordering decisions to meet clinicians’ cognitive needs. Decision support and auditing technologies could automate and monitor PPE orders efficiently. Likely to be most effective, an investment in creating and maintaining centralized PPE expert management teams to assess, order, and discontinue PPE would minimize individual ordering variation, minimize cost, and maximize value to patients, staff, and hospitals.

In this pandemic, we have the opportunity to rethink how we understand and use PPE in hospitalized patients. It is vitally important now more than ever to consciously conserve the limited supply of PPE resources. As we seek to increase healthcare value while limiting overuse and waste, PPE is a prime target for value improvement efforts as the effective but also burdensome tool that it is. Hospitalists are well-positioned to lead the redesign of how, when, and why PPE is used and to create a more optimized, lasting system that provides maximal value to patients, families, and healthcare workers during this current crisis and beyond.