Video Abstract

Video Abstract

BACKGROUND:

Improvement initiatives promote safe and efficient care for hospitalized children. However, these may be associated with limited cost savings. In this article, we sought to understand the potential financial benefit yielded by improvement initiatives by describing the inpatient allocation of costs for common pediatric diagnoses.

METHODS:

This study is a retrospective cross-sectional analysis of pediatric patients aged 0 to 21 years from 48 children’s hospitals included in the Pediatric Health Information System database from January 1, 2017, to December 31, 2017. We included hospitalizations for 8 common inpatient pediatric diagnoses (seizure, bronchiolitis, asthma, pneumonia, acute gastroenteritis, upper respiratory tract infection, other gastrointestinal diagnoses, and skin and soft tissue infection) and categorized the distribution of hospitalization costs (room, clinical, laboratory, imaging, pharmacy, supplies, and other). We summarized our findings with mean percentages and percent of total costs and used mixed-effects models to account for disease severity and to describe hospital-level variation.

RESULTS:

For 195 436 hospitalizations, room costs accounted for 52.5% to 70.3% of total hospitalization costs. We observed wide hospital-level variation in nonroom costs for the same diagnoses (25%–81% for seizure, 12%–51% for bronchiolitis, 19%–63% for asthma, 19%–62% for pneumonia, 21%–78% for acute gastroenteritis, 21%–63% for upper respiratory tract infection, 28%–69% for other gastrointestinal diagnoses, and 21%–71% for skin and soft tissue infection). However, to achieve a cost reduction equal to 10% of room costs, large, often unattainable reductions (>100%) in nonroom cost categories are needed.

CONCLUSIONS:

Inconsistencies in nonroom costs for similar diagnoses suggest hospital-level treatment variation and improvement opportunities. However, individual improvement initiatives may not result in significant cost savings without specifically addressing room costs.

What’s Known on This Subject:

Many quality improvement projects have improved care for patients, but few have shown significant cost savings for overall hospitalization costs.

What This Study Adds:

This study describes the proportion of hospitalization costs across cost categories, disease processes, and hospitals. The majority of costs across all diagnosis were related to room costs; improvement projects that address room costs will have the most financial impact.

Wide variability in patient care among inpatient hospitalizations has been shown to lead to higher costs without an improvement in the quality of care.1  At a minimum, 20% of national health care costs may be attributed to waste.2  To improve health care delivery, quality improvement (QI) science refines processes within the clinical setting to foster safe, efficient clinical decision-making and to improve patient experiences using the application of evidence-based medicine. Clinical targets for QI initiatives include the application of evidence-based medicine to “safely do less” in an effort to provide only the care necessary to achieve optimal patient outcomes.3,4  This includes limiting single interventions that have no clinical benefit (eg, decreasing unneeded testing) and defining clinical pathways that increase therapeutic efficacy along the continuum of an illness.57  By using this broad perspective, QI projects have successfully streamlined and improved care for hospitalized patients.

Although maximizing patient outcomes, QI interventions often reveal minimal cost savings for hospitalizations. For example, a common target for QI projects is decreasing the use of unnecessary laboratory testing.5,814  Although the cost of obtaining individual studies may seem substantial, when compared to the total cost of hospitalization, the cost accrued by unnecessary laboratory testing often becomes negligible. Another strategy to lower hospitalization costs is addressing provider- and hospital-level variation. Although provider practice patterns may lead to significant hospital-level practice variation, the proportion of modifiable costs attributed to provider- or hospital-level variation compared to the cost of the room is unknown.15 

Hospital charges are often classified into categories (room, clinical, laboratory, imaging, pharmacy, supply, and other). The room cost broadly encompasses the cost of providing infrastructure and staffing during the duration of a patient’s stay. The other, nonroom cost categories reflect charges incurred after a provider’s decision to proceed with a specific treatment, laboratory, or imaging service. Whereas these categories are directly impacted by provider-driven treatment plans, the room charge for that stay is mostly impacted by the length of stay of the patient. Classification of hospital costs into those associated with the room and those associated with nonroom services may help inform future QI projects that aim for both substantial cost savings and improvement of patient outcomes.16,17  In these terms, our objectives were to do the following: (1) describe the allocation of both room and nonroom hospital costs by cost category, (2) compare the contribution of each category to the overall cost of hospitalization, and (3) describe variation among cost categories in pediatric hospitals for children hospitalized with common pediatric diagnoses.

We conducted a retrospective, cross-sectional descriptive analysis of inpatient and observation encounters from January 1, 2017, to December 31, 2017, for children and adolescents aged 0 to 21 years using the Pediatric Health Information System (PHIS) (Children’s Hospital Association, Lenexa, KS). The PHIS is an administrative database containing patient data from 48 tertiary pediatric hospitals. For all encounters, the PHIS contains demographic characteristics and up to 41 International Classification of Diseases, 10th Revision, Clinical Modification diagnosis and procedure codes. These are used to determine the primary reason and severity (minor, moderate, major, or extreme) for hospitalization by using the All Patient Refined Diagnosis Related Group (APR-DRG) software by 3M. We focused our analysis on the 8 most frequent APR-DRGs typically managed by pediatric hospitalists, which included seizure (APR-DRG 53), bronchiolitis (APR-DRG 138), asthma (APR-DRG 141), pneumonia (APR-DRG 139), acute gastroenteritis (AGE) (APR-DRG 249), upper respiratory tract infection (URI) (APR-DRG 113), other gastrointestinal (GI) diagnoses (APR-DRG 254), and skin and soft tissue infection (SSTI) (APR-DRG 383). These diagnoses accounted for approximately one-quarter of all discharges in the PHIS.

Daily, itemized billing service codes along with charge information are submitted to the PHIS. Hospital charges are converted to estimated costs by using 29 department-specific cost/charge ratios collected from each hospital annually. We classified costs from each hospitalization into the following cost categories: room, clinical, laboratory, imaging, pharmacy, supply, and other. Room costs encompass overhead and staffing costs, including nursing and other general facility costs (building, electricity, and administration) necessary to care for patients. Room costs are primarily impacted by the days of hospitalizaiton.1618  The remaining categories fall under nonroom costs and encompass costs associated with the delivery of medical interventions to the patient. Clinical costs include procedures or administration of therapy such as conducting electroencephalography or administering intravenous fluids. Laboratory testing, radiographic imaging, and medications were included in the laboratory, imaging, and pharmacy categories, respectively. Supplies included miscellaneous items such as equipment associated with nasal cannula or nebulization machines. The “other” category included costs not contained within other categories, such as operating-room time or emergency-department costs included in the hospitalization.

We used frequencies and percentages to summarize categorical variables and used medians with interquartile ranges for continuous variables. For each cost category within each APR-DRG, we calculated mean percentages of cost within each category with 95% confidence intervals (CIs). Next, for each APR-DRG, we compared the mean percentages of cost across levels of severity using mixed-effects models, controlling for hospital clustering with a random intercept for each hospital. We calculated the percent of cost from the nonroom cost categories necessary to achieve a cost reduction equivalent to a 10% reduction in room costs. Finally, to identify areas of high cost and hospital-level variation, we calculated the adjusted percent of cost from nonroom sources for each hospital in each APR-DRG, using a generalized linear mixed-effects model adjusted for severity, age, number of complex chronic conditions,19  patient type (inpatient versus observation), sex, race, and payer. All analyses were performed by using SAS version 9.4 (SAS Institute, Inc, Cary, NC), and P < .05 was considered statistically significant. Because we used deidentified data, the Children’s Mercy Kansas City Institutional Review Board deemed the study exempt.

A total of 195 436 hospital discharges met inclusion criteria, which represented 24.6% of all discharges across the hospitals studied (Table 1). Age varied among diagnoses. Male patients represented >50% of discharges for all diagnosis, most notably with the diagnosis of asthma (61.3%). Patients hospitalized with asthma were more often of non-Hispanic African American race/ethnicity (42.7%); all other diagnoses were more often in those of non-Hispanic white race and/or ethnicity (45.1%–54.5%). Most patients did not have a complex chronic condition (73.8%) and were covered by public insurance (59.8%). All patients had a short length of stay, with a median of 1 to 2 days for all diagnoses.

TABLE 1

Demographic and Clinical Characteristics of Included Cohort

SeizureURIBronchiolitisPneumoniaAsthmaAGEOther GI DiagnosisSSTI
N 39 780 21 522 36 323 18 129 31 186 19 607 17 450 11 439 
Age, y, n (%)         
 <1 4829 (12.1) 5889 (27.4) 24 555 (67.6) 2115 (11.7) 483 (1.5) 4869 (24.8) 3024 (17.3) 1631 (14.7) 
 1–4 12 582 (31.6) 9609 (44.6) 11 390 (31.4) 3682 (47.9) 13 112 (42.0) 7575 (38.6) 4453 (25.5) 3622 (31.7) 
 5–9 9377 (23.6) 3150 (14.6) 235 (0.8) 4151 (22.9) 10 627 (34.1) 3321 (16.9) 3848 (22.1) 2269 (19.8) 
 10–18 12 992 (32.7) 2874 (13.4) 93 (0.3) 3181 (17.5) 6964 (22.3) 3342 (19.6) 6125 (35.1) 3867 (33.8) 
Female sex, n (%) 18 519 (46.6) 8681 (40.3) 14 858 (40.9) 8538 (47.1) 12 060 (38.7) 9195 (46.9) 8325 (47.7) 5450 (47.6) 
Race, n (%)         
 Non-Hispanic white 21 150 (53.2) 9781 (45.4) 16 374 (45.1) 8229 (45.4) 8507 (27.3) 9960 (50.8) 9504 (54.5) 5232 (45.7) 
 Non-Hispanic African American 6870 (17.3) 4019 (18.7) 7253 (20.0) 3216 (17.7) 13 327 (42.7) 3115 (15.9) 2942 (16.9) 2185 (19.1) 
 Hispanic 6695 (16.8) 4990 (23.2) 7966 (21.9) 4169 (23.0) 6050 (19.4) 4253 (21.7) 3131 (17.9) 2635 (23.0) 
 Other 5065 (12.7) 2732 (12.7) 4730 (13.0) 2515 (13.9) 3302 (10.6) 2279 (11.6) 1873 (10.7) 1387 (12.1) 
Payer, n (%)         
 Public 22 360 (56.2) 12 571 (58.4) 22 936 (63.1) 10 237 (56.5) 20 590 (66.0) 11 426 (58.3) 9990 (57.2) 6677 (58.4) 
 Private 15 531 (39.0) 8075 (37.5) 12 187 (33.6) 7183 (39.6) 9309 (29.8) 7397 (37.7) 6319 (39.1) 4277 (37.4) 
 Other 1889 (4.7) 876 (4.1) 1200 (3.3) 709 (3.9) 1287 (4.1) 784 (4.0) 641 (3.7) 485 (4.2) 
CCC count, n (%)         
 0 22 329 (56.1) 15 649 (72.7) 30 912 (85.1) 12 697 (70.0) 28 921 (92.7) 12 870 (65.6) 10 883 (62.4) 9934 (86.8) 
 1 11 508 (23.9) 3397 (15.8) 3528 (9.7) 2669 (14.7) 1788 (5.7) 4009 (20.4) 3724 (21.3) 971 (8.5) 
 ≥2 5943 (14.9) 2476 (11.5) 1883 (5.2) 2763 (15.2) 477 (1.5) 2728 (13.9) 2843 (16.3) 534 (4.7) 
Patient type, n (%)         
 Inpatient 25 313 (63.6) 123 52 (57.4) 25 760 (70.9) 14 395 (79.4) 20 098 (64.4) 10 319 (55.2) 9734 (55.8) 7715 (67.4) 
 Observation 14 467 (36.4) 9170 (42.6) 10 563 (29.1) 3734 (20.6) 11 088 (35.6) 8788 (44.8) 7716 (44.2) 3724 (32.6) 
LOS, d, median (IQR) 1 (1–2) 1 (1–2) 2 (1–3) 2 (1–3) 1 (1–2) 2 (1–2) 2 (1–3) 2 (1–3) 
SeizureURIBronchiolitisPneumoniaAsthmaAGEOther GI DiagnosisSSTI
N 39 780 21 522 36 323 18 129 31 186 19 607 17 450 11 439 
Age, y, n (%)         
 <1 4829 (12.1) 5889 (27.4) 24 555 (67.6) 2115 (11.7) 483 (1.5) 4869 (24.8) 3024 (17.3) 1631 (14.7) 
 1–4 12 582 (31.6) 9609 (44.6) 11 390 (31.4) 3682 (47.9) 13 112 (42.0) 7575 (38.6) 4453 (25.5) 3622 (31.7) 
 5–9 9377 (23.6) 3150 (14.6) 235 (0.8) 4151 (22.9) 10 627 (34.1) 3321 (16.9) 3848 (22.1) 2269 (19.8) 
 10–18 12 992 (32.7) 2874 (13.4) 93 (0.3) 3181 (17.5) 6964 (22.3) 3342 (19.6) 6125 (35.1) 3867 (33.8) 
Female sex, n (%) 18 519 (46.6) 8681 (40.3) 14 858 (40.9) 8538 (47.1) 12 060 (38.7) 9195 (46.9) 8325 (47.7) 5450 (47.6) 
Race, n (%)         
 Non-Hispanic white 21 150 (53.2) 9781 (45.4) 16 374 (45.1) 8229 (45.4) 8507 (27.3) 9960 (50.8) 9504 (54.5) 5232 (45.7) 
 Non-Hispanic African American 6870 (17.3) 4019 (18.7) 7253 (20.0) 3216 (17.7) 13 327 (42.7) 3115 (15.9) 2942 (16.9) 2185 (19.1) 
 Hispanic 6695 (16.8) 4990 (23.2) 7966 (21.9) 4169 (23.0) 6050 (19.4) 4253 (21.7) 3131 (17.9) 2635 (23.0) 
 Other 5065 (12.7) 2732 (12.7) 4730 (13.0) 2515 (13.9) 3302 (10.6) 2279 (11.6) 1873 (10.7) 1387 (12.1) 
Payer, n (%)         
 Public 22 360 (56.2) 12 571 (58.4) 22 936 (63.1) 10 237 (56.5) 20 590 (66.0) 11 426 (58.3) 9990 (57.2) 6677 (58.4) 
 Private 15 531 (39.0) 8075 (37.5) 12 187 (33.6) 7183 (39.6) 9309 (29.8) 7397 (37.7) 6319 (39.1) 4277 (37.4) 
 Other 1889 (4.7) 876 (4.1) 1200 (3.3) 709 (3.9) 1287 (4.1) 784 (4.0) 641 (3.7) 485 (4.2) 
CCC count, n (%)         
 0 22 329 (56.1) 15 649 (72.7) 30 912 (85.1) 12 697 (70.0) 28 921 (92.7) 12 870 (65.6) 10 883 (62.4) 9934 (86.8) 
 1 11 508 (23.9) 3397 (15.8) 3528 (9.7) 2669 (14.7) 1788 (5.7) 4009 (20.4) 3724 (21.3) 971 (8.5) 
 ≥2 5943 (14.9) 2476 (11.5) 1883 (5.2) 2763 (15.2) 477 (1.5) 2728 (13.9) 2843 (16.3) 534 (4.7) 
Patient type, n (%)         
 Inpatient 25 313 (63.6) 123 52 (57.4) 25 760 (70.9) 14 395 (79.4) 20 098 (64.4) 10 319 (55.2) 9734 (55.8) 7715 (67.4) 
 Observation 14 467 (36.4) 9170 (42.6) 10 563 (29.1) 3734 (20.6) 11 088 (35.6) 8788 (44.8) 7716 (44.2) 3724 (32.6) 
LOS, d, median (IQR) 1 (1–2) 1 (1–2) 2 (1–3) 2 (1–3) 1 (1–2) 2 (1–2) 2 (1–3) 2 (1–3) 

Diagnostic groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (APB-DRG 138), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383). CCC, complex chronic condition; IQR, interquartile range; LOS, length of stay.

For each hospitalization, we calculated the mean percentage of cost within each cost category (room, clinical, laboratory, imaging, pharmacy, supply, and other; Fig 1, Supplemental Table 2). The majority of total hospital costs were related to room costs, ranging from 52.5% of total hospital costs for seizure to 70.3% for bronchiolitis. Clinical costs were highest for seizure and asthma (29.1% and 22.6%, respectively) related to EEG procedures and administration of intravenous fluids. Diagnoses of SSTI and other GI diagnosis had the highest percentage of “other” cost (16% and 17%, respectively) secondary to emergency-department critical-care time and operating-room services. The remainder of nonroom costs (ie, laboratory, imaging, pharmacy, and supply) were marginal compared to the room and clinical charges for all studied diagnoses.

FIGURE 1

Mean percent of cost in each cost category by diagnosis group. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (DRG 13S), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383).

FIGURE 1

Mean percent of cost in each cost category by diagnosis group. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (DRG 13S), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383).

To determine differences in cost related to illness severity, we calculated the distribution of hospitalization cost across cost categories at each level of APR-DRG severity of illness for each diagnosis (Supplemental Tables 3 and 4). For the diagnosis of URI, bronchiolitis, pneumonia, and SSTI, patients with increased illness severity required more nonroom resources, such as laboratory and pharmacy. For those hospitalized with a seizure, asthma, viral gastroenteritis, or other GI diagnosis, those with higher illness severity had a higher proportion of total cost from the room. Regardless of diagnosis or severity, room costs represented >50% of total costs.

We calculated the percentage of nonroom costs needed to equal a modest (10%) reduction in room costs (Fig 2). For example, to equal a 10% reduction in room costs for a child hospitalized with bronchiolitis, a hospital would need to reduce clinical costs by 60%, laboratory costs by 201%, imaging costs by 586%, or pharmacy costs by 370%. This pattern holds true for every diagnosis: The cost categories of laboratory, imaging, pharmacy, and supply would need to be significantly reduced (most by >100%) to equal a 10% reduction in room costs. The clinical and other categories showed somewhat more attainable reductions to equal a 10% reduction in room costs (18%–60% and 31%–77%, respectively).

FIGURE 2

Percent cost reduction in each service group to equal a 10% reduction in room cost. Cells from 0% to 75% are white, from 76% to 150% are light gray, and >150% are dark gray. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (APR-DRG 138), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383).

FIGURE 2

Percent cost reduction in each service group to equal a 10% reduction in room cost. Cells from 0% to 75% are white, from 76% to 150% are light gray, and >150% are dark gray. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (APR-DRG 138), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383).

To compare hospital-level variation by severity-adjusted percent of nonroom costs, we created a heat map for each of the 8 common pediatric diagnoses (Fig 3). Notable hospital-level variation in nonroom costs was seen across all diagnosis (seizure 25%–81%, bronchiolitis 12%–51%, asthma 19%–63%, pneumonia 19%–62%, AGE 21%–78%, URI 21%–63%, other GI diagnosis 28%–69%, and SSTI 21%–71%). The asthma, seizure, and other GI diagnoses had higher nonroom costs on average, perhaps related to a greater number of common interventions during hospitalization, such as albuterol administration for asthma. Hospitals with a higher percentage of nonroom costs in one diagnosis tended to have a higher percentage of nonroom costs in other APR-DRG categories as well.

FIGURE 3

Severity-adjusted percent of cost from nonroom sources for each PHIS hospital. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (APR-DRG 138), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383). a Greater adjusted percent of nonroom costs (lower percent room costs) appear red or orange. Higher adjusted percent of room costs (lower nonroom costs) appear purple or blue.

FIGURE 3

Severity-adjusted percent of cost from nonroom sources for each PHIS hospital. Diagnosis groups are as follows: seizure (APR-DRG 53), URI (APR-DRG 113), bronchiolitis (APR-DRG 138), pneumonia (APR-DRG 139), asthma (APR-DRG 141), AGE (APR-DRG 249), other GI diagnosis (APR-DRG 254), and SSTI (APR-DRG 383). a Greater adjusted percent of nonroom costs (lower percent room costs) appear red or orange. Higher adjusted percent of room costs (lower nonroom costs) appear purple or blue.

In this study of 195 436 common pediatric hospitalizations, room costs alone accounted for >50% of total hospitalization costs even when accounting for illness severity. Overall, we found that to equal a modest reduction (10%) in room costs, reductions within individual nonroom cost categories would often be unrealistically high (often >100%). Room charges can encompass building, utilities, staffing, and administrative costs. As such, they can be more difficult to directly affect from a provider standpoint, whereas nonroom costs can be more immediately impacted by provider decisions. Despite this, initiatives addressing reduction in individual nonroom costs are unlikely to lead to large cost savings when viewed in the context of the total cost of hospitalization. Initiatives targeting variability in care delivery7  (eg, clinical pathways or guidelines that reduce diagnostics or unneeded treatments) across a hospital system, however, may create widespread changes in practice that result in notable cost savings. Furthermore, there are likely additional benefits of standardizing care in nonroom cost categories beyond cost reduction, including reduced variability and improved timeliness of care.

This study supports previous work suggesting QI projects focused on reducing individual unnecessary tests or treatments for specific diagnoses result in minimal overall cost savings.5,8,9,12,20,21  However, a system-wide initiative focusing on redundant or unnecessary laboratory testing may result in significant savings. Redundant testing in the United States has been estimated to cost upward of $8 billion per year.22  Several large health care systems have shown substantial reduction in cost by applying system-wide test-reduction initiatives. Examples include reducing routine daily testing, implementing focused testing (eg, ordering a hemoglobin count instead of the complete blood count), obtaining tests based on predictive value (eg, obtaining blood cultures only in those with high risk of bacteremia),23  and monitoring the use of high-cost genetic testing. Although the proportion of unneeded tests related to each individual disease may be small, the global impact on the health system of small changes can result in substantial savings.

We found substantial hospital-level variation in nonroom costs, ranging from 20% to 80% of total costs of hospitalization regardless of diagnosis. This hospital-level variation outlines opportunities for improvement in standardization of practices. Instead of projects focused on a single disease, several institutions have implemented clinical practice guidelines and care process models to reduce variation for multiple diagnoses, leading to fewer unnecessary tests and interventions and resulting in improved patient-level outcomes.6,2426  For example, a large-scale improvement project aimed at implementation of practice pathways within a freestanding tertiary children’s hospital system showed significant reduction in costs and decreased length of stay for many of the 15 clinical pathways studied.27  However, the development and maintenance of these pathways requires substantial cost and resources, which may reduce overall cost savings resulting from their implementation.

In this study, we describe cost allocation related to room and nonroom costs of hospitalized children with common childhood diagnoses. The costs we report reflect operating expenses for room and nonroom costs. It is likely that QI initiatives aimed at reducing nonroom costs may have more of a substantial financial impact if they are also able to reduce room costs in the long-term. For example, avoidance of unnecessary laboratories may reduce length of stay by decreasing follow-up testing or allowing for more expedited discharge. Future studies examining the direct and indirect impact of an initiative may demonstrate that the downstream effects of such an intervention provide meaningful cost savings.

Our study has several limitations. The study population includes only patients admitted to tertiary pediatric hospitals. As such, our results may not be generalizable to pediatric hospitalizations at large because practice styles may vary on the basis of hospital type (tertiary-care, urban hospital versus community hospital). However, our sample included 48 hospitals with reasonable variation in practice, and all demonstrated a predominance of costs associated with room costs. Second, patient-level characteristics and hospital clinical practice patterns influence length of stay, which in turn affect the percent of total charges attributed to room charge. We accounted for these effects by adjusting modeling for patient-level (including demographics, severity, number of complex chronic conditions, and patient type [inpatient versus observation]) and hospital-level covariates (including clustering on hospital and state). We also estimated costs from charges instead of using cost accounting systems, which may not accurately reflect actual costs. However, because we are looking at proportions of costs, differences in actual costs should be minimal. Finally, we examined the 8 most frequent APR-DRGs, which represented 24.6% of the total number of hospitalized patients during the study period. It is likely that different patient populations, including those with chronic conditions requiring frequent hospitalization (eg, cancer, diabetes, and renal failure), have different proportions of charges, in which a QI initiative may be more successful at reducing overall hospitalization costs.

Room costs account for a large proportion of total hospitalization costs. Improvement initiatives directly or indirectly affecting determinants of room costs, such as length of stay, or broadly addressing nonroom costs, such as system-wide reductions of unnecessary testing and reducing variability in care through standardization, may promote safe and effective care while also providing measurable cost savings.

Dr Synhorst participated in the study design, analysis, and interpretation of data and was the primary author of the manuscript; Drs Johnson, Bettenhausen, and Hall participated in the study design, analysis, and interpretation of data and were authors of the manuscript; Dr Kyler participated in the study design and interpretation of the data and was an author of the manuscript; Dr Richardson participated in the study design and interpretation of the data; Dr Mann participated in the study design, analysis, and interpretation of data; and all authors provided critical intellectual content in the revision of the manuscript and approved the final version of the manuscript as submitted.

This article and the data it contains have not been published previously in print or electronic format and are not under consideration by another publication or electronic medium.

FUNDING: No external funding.

COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2020-0619.

     
  • AGE

    acute gastroenteritis

  •  
  • APR-DRG

    All Patient Refined Diagnosis Related Group

  •  
  • CI

    confidence interval

  •  
  • GI

    gastrointestinal

  •  
  • PHIS

    Pediatric Health Information System

  •  
  • QI

    quality improvement

  •  
  • SSTI

    skin and soft tissue infection

  •  
  • URI

    upper respiratory tract infection

1
Yasaitis
L
,
Fisher
ES
,
Skinner
JS
,
Chandra
A
.
Hospital quality and intensity of spending: is there an association?
Health Aff (Millwood)
.
2009
;
28
(
4
):
w566
w572
2
Berwick
DM
,
Hackbarth
AD
.
Eliminating waste in US health care
.
JAMA
.
2012
;
307
(
14
):
1513
1516
3
Quinonez
RA
,
Garber
MD
,
Schroeder
AR
, et al
.
Choosing wisely in pediatric hospital medicine: five opportunities for improved healthcare value
.
J Hosp Med
.
2013
;
8
(
9
):
479
485
4
Agency for Healthcare Research and Quality
.
Six domains of health care quality.
2015
. Available at: www.ahrq.gov/professionals/quality-patient-safety/talkingquality/create/sixdomains.html. Accessed August 11, 2019
5
Johnson
DP
,
Lind
C
,
Parker
SES
, et al
.
Toward high-value care: a quality improvement initiative to reduce unnecessary repeat complete blood counts and basic metabolic panels on a pediatric hospitalist service
.
Hosp Pediatr
.
2016
;
6
(
1
):
1
8
6
Rotter
T
,
Kinsman
L
,
James
E
, et al
.
Clinical pathways: effects on professional practice, patient outcomes, length of stay and hospital costs
.
Cochrane Database Syst Rev
.
2010
;(
3
):
CD006632
7
Bryan
MA
,
Desai
AD
,
Wilson
L
,
Wright
DR
,
Mangione-Smith
R
.
Association of bronchiolitis clinical pathway adherence with length of stay and costs
.
Pediatrics
.
2017
;
139
(
3
):
e20163432
8
Corson
AH
,
Fan
VS
,
White
T
, et al
.
A multifaceted hospitalist quality improvement intervention: decreased frequency of common labs
.
J Hosp Med
.
2015
;
10
(
6
):
390
395
9
Krasowski
MD
,
Davis
SR
,
Drees
D
, et al
.
Autoverification in a core clinical chemistry laboratory at an academic medical center
.
J Pathol Inform
.
2014
;
5
(
1
):
13
10
Procop
GW
,
Yerian
LM
,
Wyllie
R
,
Harrison
AM
,
Kottke-Marchant
K
.
Duplicate laboratory test reduction using a clinical decision support tool
.
Am J Clin Pathol
.
2014
;
141
(
5
):
718
723
11
Konger
RL
,
Ndekwe
P
,
Jones
G
, et al
.
Reduction in unnecessary clinical laboratory testing through utilization management at a US government veterans affairs hospital
.
Am J Clin Pathol
.
2016
;
145
(
3
):
355
364
12
May
TA
,
Clancy
M
,
Critchfield
J
, et al
.
Reducing unnecessary inpatient laboratory testing in a teaching hospital
.
Am J Clin Pathol
.
2006
;
126
(
2
):
200
206
13
Ducatman
AM
,
Tacker
DH
,
Ducatman
BS
, et al
.
Quality improvement intervention for reduction of redundant testing
.
Acad Pathol
.
2017
;
4
:
2374289517707506
14
Tchou
MJ
,
Tang Girdwood
S
,
Wormser
B
, et al
.
Reducing electrolyte testing in hospitalized children by using quality improvement methods
.
Pediatrics
.
2018
;
141
(
5
):
e20173187
15
Jonas
JA
,
Shah
SS
,
Zaniletti
I
, et al
.
Regional variation in standardized costs of care at children’s hospitals
.
J Hosp Med
.
2017
;
12
(
10
):
818
825
16
Rauh
SS
,
Wadsworth
E
,
Weeks
WB
.
The fixed-cost dilemma: what counts when counting cost-reduction efforts?
Healthc Financ Manage
.
2010
;
64
(
3
):
60
63
17
Roberts
RR
,
Frutos
PW
,
Ciavarella
GG
, et al
.
Distribution of variable vs fixed costs of hospital care
.
JAMA
.
1999
;
281
(
7
):
644
649
18
Rauh
SS
,
Wadsworth
EB
,
Weeks
WB
,
Weinstein
JN
.
The savings illusion–why clinical quality improvement fails to deliver bottom-line results
.
N Engl J Med
.
2011
;
365
(
26
):
e48
19
Feudtner
C
,
Feinstein
JA
,
Zhong
W
,
Hall
M
,
Dai
D
.
Pediatric complex chronic conditions classification system version 2: updated for ICD-10 and complex medical technology dependence and transplantation
.
BMC Pediatr
.
2014
;
14
:
199
20
McCulloh
RJ
,
Koster
MP
,
Yin
DE
, et al
.
Evaluating the use of blood cultures in the management of children hospitalized for community-acquired pneumonia
.
PLoS One
.
2015
;
10
(
2
):
e0117462
21
Auerbach
AD
,
Wachter
RM
.
Focusing on value: this time is different
.
J Hosp Med
.
2013
;
8
(
9
):
543
544
22
Jha
AK
,
Chan
DC
,
Ridgway
AB
,
Franz
C
,
Bates
DW
.
Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals
.
Health Aff (Millwood)
.
2009
;
28
(
5
):
1475
1484
23
Zwang
O
,
Albert
RK
.
Analysis of strategies to improve cost effectiveness of blood cultures
.
J Hosp Med
.
2006
;
1
(
5
):
272
276
24
Neuman
MI
,
Hall
M
,
Hersh
AL
, et al
.
Influence of hospital guidelines on management of children hospitalized with pneumonia
.
Pediatrics
.
2012
;
130
(
5
):
e823
e830
25
Newman
RE
,
Hedican
EB
,
Herigon
JC
,
Williams
DD
,
Williams
AR
,
Newland
JG
.
Impact of a guideline on management of children hospitalized with community-acquired pneumonia
.
Pediatrics
.
2012
;
129
(
3
):
e597
e604
26
Ralston
SL
,
Garber
MD
,
Rice-Conboy
E
, et al;
Value in Inpatient Pediatrics Network Quality Collaborative for Improving Hospital Compliance With AAP Bronchiolitis Guideline (BQIP)
.
A multicenter collaborative to reduce unnecessary care in inpatient bronchiolitis
.
Pediatrics
.
2016
;
137
(
1
):
e20150851
27
Lion
KC
,
Wright
DR
,
Spencer
S
,
Zhou
C
,
Del Beccaro
M
,
Mangione-Smith
R
.
Standardized clinical pathways for hospitalized children and outcomes
.
Pediatrics
.
2016
;
137
(
4
):
e20151202

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data