Introduction
Pulmonary hypertension (PH) has been defined as pulmonary arterial pressures greater than 25 mmHg at rest, as per the first World Symposium on Pulmonary Hypertension (WSPH) organized by the World Health Organization (WHO) in Geneva in 1973 [1]. Pediatric PH differs in terms of pathophysiology and etiology as compared to adults [2]. Knowledge acquired over the last decade about pediatric PH is now included in the clinical classification of PH [3]. PH in pediatric patients is associated with significant morbidity and mortality [4]. These patients have various comorbidities, with the most common being congenital heart disease (CHD), bronchopulmonary dysplasia (BPD), genetic anomalies affecting the lung, and pathological insults on growing lungs [5,6]. While the number of hospitalizations for PH has increased over time, there has been a decrease in mortality of these patients, accompanied by a considerable increase in hospital charges [4].
Hospital readmission is one of the quality indicators tracked by the Centers for Medicare & Medicaid Services (CMS) and a target for hospital cost saving as per the Patient Protection and Affordable Care Act [7]. Many states have Medicaid readmission penalties in the pediatric population that focus on overall readmission rates [8,9]. PH is associated with increased hospital readmissions in both children and adults [10,11]. According to Awerbach et al. (2018), the 30-day readmission rate in pediatric patients with PH was 26.3% and was associated with increased odds in those with public insurance [11]. This study highlights readmission rates and other covariates using Pediatric Health Information System (PHIS) among patients in tertiary care hospitals.
Current evidence suggests that readmission rates can be improved if the patients who are at higher risk for readmission are identified in a timely manner, and necessary measures are taken to reduce the deleted readmission. These measures include proper discharge planning, patient education, and timely outpatient appointments [12].
Materials and methods
Study design
A retrospective study was performed using an all-patient publicly available database for the year 2017. Diagnosis codes pertaining to PH were used to identify index admissions. Our study was exempted from the Institutional Review Board (IRB) at the University of South Alabama as we used a publicly available database.
Data source
We obtained data using the National Readmission Database (NRD) from January 1, 2017, to December 1, 2017. NRD is the largest all-patient, all-payer inpatient database available in the United States. The NRD is developed and maintained by the Agency for Healthcare Research and Quality for the Healthcare Cost and Utilization Project (HCUP) and includes publicly available hospitalization data from nonfederal hospitals. The 2017 NRD includes data from 28 states and approximately five million pediatric discharges. It represents 58.2% of all hospitalizations in the year 2017 and 60% of the population. Unique patient identifiers are allotted to each admission to allow for tracking readmissions within a state in a given year. Moreover, data regarding patient and hospital demographics as well as diagnostic and procedural codes assigned at the time of discharge are also available.
Inclusion criteria included patients less than 18 years of age, who had a primary diagnosis of PH using the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) codes (Appendix A).
Exclusion criteria included the following: (1) age greater than 18 years, (2) initial visits in December 2017 because they could not be followed beyond 30 days, (3) admitted electively on the readmission visit, (4) no patient encounters within 30 days from the primary visit, (5) transferred out for rehabilitation services, and (6) died on the index admission.
Covariates
Patient demographics for the index admission including age, gender, insurance status, median household income, loss of function, and discharge disposition were collected. Comorbidities were also included using ICD-10-CM codes (Appendices B-F). Hospital demographics such as bed size, ownership (government versus private), designation (large, small, micropolitan, non-urban), and teaching status were also obtained. Various procedures done during the hospitalization were identified using ICD-10 procedure codes (Appendices G-I). The severity of illness was estimated using All Patients Refined Diagnosis Related Group (APR-DRG) (Appendix J).
Statistical analysis
All statistical analysis was performed using IBM SPSS Statistics for Windows, version 1.0.0.1327 (IBM Corp., Armonk, USA). Baseline characteristics of participants such as age, gender, weekend versus weekday admissions, household income, payer status, loss of function, and the likelihood of dying were tested for statistical differences using the Pearson chi-square test for categorical variables and Mann-Whitney U-Test for continuous variables with no readmission as the reference group. Clinical predictors for 30-day readmission were analyzed using multivariable logistic regression. Furthermore, multiple logistic regressions were performed to independently determine the predictors of readmission within each clinical variable with values presented as Odds Ratios (OR) with 95% confidence intervals (CI).
Results
Of 5,529,389 pediatric hospital encounters in NRD, 16,590 encounters were for PH. Among patients with PH, 748 (7.1%) patients died during the index admission and were excluded from the study. Infants comprised the majority of deaths with 577 (77.1%). Comparison of severity of illness using APR-DRG severity score showed that among infants who died, 475 (82.4%) were classified as extreme loss of function compared to 66.3% among infants who survived (p < 0.001).
A total of 10,501 unique patients met the selection criteria of which 14.4% were readmitted within 30 days. The 30-day readmission rate for PH was higher than that from all other causes combined {Odds Ratio (OR) 4.02 (3.84-4.20), p < 0.001}. Among all PH patients, 56.6% were infants, who were found to have decreased odds of 30-day readmission as compared to the other age groups {OR 0.72 (0.65-0.80) p < 0.001}. For patients readmitted in less than 30 days, the average time to readmission was 17 days (±7 days).
Overall, lengths of stay for readmissions were significantly shorter than those for the initial admissions (median four days vs. nine days, p < 0.001). As a result, readmissions were significantly less expensive than the initial visits (median $36,313 vs. $85,121, p < 0.001). In 2017, total charges for all pediatric PH patient encounters in the United States amounted to $1.63 billion, while charges for 30-day unplanned readmissions for PH were $328 million in 2017. Medicaid remained the largest payer (63.9%) among all PH admissions (Table 1).
Table 1
Demographic characteristics of the sample
Comparison of various demographic characteristics was done using chi-square analysis.
IQR, interquartile range.
No Readmission | 30-Day Readmissions | p-Value | |
Number of Patients | 8985 (85.56%) | 1516 (14.43%) | <0.001 |
Age {IQR} | <1 year {0-3} | 1 year {0-3} | |
Age < 1 (%) | 5189 (57.8) | 755 (49.8) | |
Age 1-7 years (%) | 2391 (26.6) | 501 (33) | |
Age 8-13 years (%) | 600 (6.7) | 140 (9.2) | |
Age 14-18 (%) | 804 (8.9) | 120 (7.9) | |
Length of stay {Median (IQR)} | 9 days {3-27} | 9 days {4-24} | 0.441 |
Total charges {Median (IQR)} | $98,433 (29,509-306,504) | $85,121 (34,485-296,303) | 0.015 |
Female (%) | 4082 (45.4) | 776 (51.2) | <0.001 |
Payer (%) | |||
Medicaid | 5474 (60.9) | 1056 (69.6) | |
Medicare | 26 (0.3) | 0 (0) | |
Private insurance | 3131 (34.8) | 395 (26) | |
Self-pay | 115 (1.3) | 5 (0.3) | |
No charge | 11 (0.1) | 56 (3.7) | |
Other | 225 (2.5) | 3 (0.2) | |
Teaching status of the hospital (%) | |||
Metropolitan teaching | 8061 (89.7) | 1383 (91.2) | 0.133 |
Metropolitan non-teaching | 806 (9) | 119 (7.8) | |
Non-metropolitan hospital | 118 (1.3) | 14 (0.9) | |
Location of the hospital (%) | |||
Large metropolitan areas with at least one million residents | 5840 (65) | 1131 (74.6) | <0.001 |
Small metropolitan areas with less than one million residents | 3028 (33.7) | 371 (24.5) | |
Micropolitan areas | 107 (1.2) | 14 (0.9) | |
Not metropolitan or micropolitan (non-urban residual) | 11 (0.1) | 0 (0) | |
Median household incomes (%) | |||
$1-$43,999 | 2733 (30.4) | 501 (33.1) | 0.920 |
$44,000-$55,999 | 2397 (26.7) | 413 (27.3) | |
$56,000-$73,999 | 2405 (26.8) | 386 (25.5) | |
>$74,000 | 1375 (15.3) | 212 (14) |
CHD was present in 52.9% of the patients with PH, while BPD was found in 17.6% of the patients of which 71.6% were infants. Other common comorbidities included respiratory infections (29.2%) and sepsis (6%). Among PH patients, 7.6% required extracorporeal membrane oxygenation (ECMO) during the initial visit, 17.7% required intubation, while 14.9% required mechanical ventilation using a tracheostomy tube (Table 2).
Table 2
Multivariate logistic regression analysis to predict 30-day readmission
The comparison of various risk factors and comorbidities was done using logistic regression.
APR-DRG, All Patients Refined Diagnosis Related Group; ECMO, extracorporeal membrane oxygenation.
Risk Factors | Odds Ratio (95% CI) | p-Value |
Sepsis | 0.75 (0.64-0.89) | <0.001 |
Mechanical ventilation with intubation | 1.66 (1.4-1.96) | <0.001 |
Mechanical ventilation with tracheostomy tube | 1.35 (1.15-1.6) | <0.001 |
ECMO | 1.21 (0.97-1.52) | 0.09 |
Respiratory infection | 0.75 (0.67-0.85) | <0.001 |
Bronchopulmonary dysplasia | 1.11 (0.96-1.28) | 0.17 |
Congenital heart disease | 0.99 (0.88-1.11) | 0.87 |
Nitric oxide | 0.78 (0.55-1.11) | 0.17 |
Age | ||
Age < 1 | Reference | |
Age 1-7 years | 1.29 (1.12-1.48) | <0.001 |
Age 8-13 years | 1.44 (1.16-1.79) | <0.001 |
Age 14-18 years | 1.01 (0.8-1.26) | 0.96 |
Payer | ||
Medicaid | Reference | |
Medicare | 0 (0-) | 1.00 |
Private insurance | 0.7 (0.61-0.79) | <0.001 |
Self-pay | 0.26 (0.11-0.63) | <0.001 |
No charge | 0 (0-) | 1.00 |
Other | 1.27 (0.94-1.72) | 0.13 |
Teaching status of the hospital | ||
Metropolitan teaching | Reference | |
Metropolitan non-teaching | 0.88 (0.72-1.08) | 0.23 |
Location of the hospital | ||
Large metropolitan areas with at least one million residents | Reference | |
Small metropolitan areas with less than one million residents | 0.64 (0.57-0.73) | <0.001 |
Micropolitan areas | 0.72 (0.41-1.28) | 0.27 |
APR-DRG severity | ||
Minor loss of function (includes cases with no comorbidity or complications) | Reference | |
Moderate loss of function | 3.76 (1.49-9.47) | <0.001 |
Major loss of function | 5.82 (2.37-14.25) | <0.001 |
Extreme loss of function | 7.66 (3.13-18.76) | <0.001 |
Hospitals located in metropolitan areas with a population of less than one million were less likely to have 30-day readmissions for PH compared to those located in more populous areas {OR 0.62 (0.55-0.71), p < 0.001}. A comparison of teaching hospitals in large and small metropolitan areas showed that the former had higher rates of 30-day readmissions {OR 1.48 (1.37-1.68), p < 0.001}. However, it should be noted that compared to teaching hospitals from the small metro areas, those in the large metro areas were more likely to have patients with major or extreme loss of function as defined by categories 3 and 4, respectively, of APR-DRG {OR 1.28 (1.12-1.3), p < 0.001}, which may have resulted in the higher rate of 30-day readmissions.
Discussion
The pediatric PH 30-day readmission rate in this study was 14.4% (Table 1). This is very similar to the readmission rate of 14.7%, which was reported in a study of adults by Chatterjee et al. based on the NRD 2013 database. However, Awerbach’s 2018 study based on the PHIS database showed that readmission in pediatric PH was higher (at 26.3%) [11]. This may be due to the acuity of patients in the PHIS study [11,13]. The PHIS database includes more than 40 tertiary care children’s hospitals with most of them located in metropolitan areas with a population > one million, which in our study was associated with more patients having extreme loss of function APR-DRG severity compared to other levels of severity [14]. Since one-third of patients in our study were admitted to hospitals located in smaller metropolitan areas (population < one million), we sampled a more comprehensive population distribution as compared to the PHIS database, which may explain the lower readmission rate observed in this study (Table 1).
Similar to the PHIS study, a majority of PH cases in this study were associated with CHD, followed by BPD; similar results have been reported by other studies [15]. However, there has been an improvement in morbidity and mortality of these patients in the last few decades due to the advent of newer treatment modalities and early surgical repair of CHD [16]. This could also be a reason that while PH was present in many patients with CHD, it was not associated with increased readmission rates. In our study, both sepsis and respiratory infections were found to be associated with decreased readmissions in PH patients (Table 2). Although several pediatric studies have shown sepsis to be a risk factor in increased readmissions as well as an independent risk factor in increased morbidity and mortality in pediatric populations, our study did not show similar results [17-19]. A possible explanation could be due to transient PH in patients with sepsis; therefore, once the infection resolves during the initial admission, these patients were less likely to be readmitted than those without sepsis as a trigger [20]. Similarly, transient PH can occur in the setting of respiratory infections [21].
This study demonstrates that patients requiring invasive ventilation via either an endotracheal tube or tracheostomy tube had higher odds of readmission (Table 2) as compared to those who did not. Studies have shown mechanical ventilation as an independent risk factor for readmission [22,23]. The need for mechanical ventilation in patients with PH is associated with increased morbidity and mortality [24]. Similarly, the severity of illness based on APR-DRG shows that patients with extreme loss of function during the initial visit are more likely to be readmitted (Figure 1). Studies have shown that APR-DRG is an effective way to measure morbidity and mortality in ICUs and to correlate well with physiological scores for the same [25,26].
Figure 1
Outcome of the index visit based on the severity of illness
The severity of illness is based on All Patients Refined Diagnosis Related Group (APR-DRG) severity of illness scoring, and the difference between the two groups was found to be statistically significant (p < 0.001).
Similar to other studies, this study showed that patients with Medicaid as their primary insurance had a higher likelihood of readmission [27,28]. A study in the adult population has shown that about half of the Medicaid patients did not have an outpatient follow-up between their initial visit and unplanned 30-day readmission [29]. We also found that the average time for PH readmissions was 17 days (±7 days); therefore, outpatient visits should be scheduled in 10-24 days after discharge from inpatient admissions to prevent readmissions effectively.
Study limitations
This study has a few limitations. First, in an administrative database from which study data is derived, the accuracy is only as good as the codes entered. Therefore, like any other administrative database, there is a possibility that some disease processes may have been coded inaccurately. Second, there may be factors not recorded in a database that might contribute to readmissions; hence, a study of those factors could also help in predicting these readmissions in the future. Third, individual patient severity of illness scores is not possible in a database; thus, the calculation of severity of illness using the APR-DRG may have inherent inaccuracies.
Conclusions
This study demonstrates that while PH is associated with significant overall readmissions in pediatrics, the readmission rate is lower than that reported by the previous research. We found that patients with sepsis and respiratory infections during the initial visit were less likely to get readmitted within 30 days. Patients who required mechanical ventilation either via an endotracheal tube or tracheostomy tube were more likely to get readmitted. We found that the average time to readmission was 17 days, a finding that suggests that follow-up visits should be scheduled within two weeks of discharge, thereby potentially preventing some readmissions.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Copyright © 2021, Sehgal et al. This work is published under https://creativecommons.org/licenses/by/3.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Background and Objectives: Pulmonary hypertension (PH) leads to significant morbidity and mortality in pediatric patients and increases the readmission rates for hospitalizations. This study evaluates the risk factors and comorbidities associated with an increase in 30-day readmissions among pediatric PH patients.
Methods: National Readmission Database (NRD) 2017 was searched for patients less than 18 years of age who were diagnosed with PH based on the International Classification of Diseases, 10th Revision (ICD-10). Statistical Package for the Social Sciences (SPSS) software v25.0 (IBM Corp., Armonk, NY) was used for statistical analysis.
Results: Of 5.52 million pediatric encounters, 10,501 patients met the selection criteria. The 30-day readmission rate of 14.43% (p < 0.001) was higher than hospitalizations from other causes {Odds Ratio (OR) 4.02 (3.84-4.20), p < 0.001}. The comorbidities of sepsis {OR 0.75 (0.64-0.89), p < 0.02} and respiratory infections {OR 0.75 (0.67-0.85), p < 0.001} were observed to be associated with lower 30-day readmissions. Patients who required invasive mechanical ventilation via endotracheal tube {OR 1.66 (1.4-1.96), p < 0.001} or tracheostomy tube {OR 1.35 (1.15-1.6), p < 0.001} had increased unplanned readmissions. Patients with higher severity of illness based on All Patients Refined Diagnosis Related Groups (APR-DRG) were more likely to get readmitted {OR 7.66 (3.13-18.76), p < 0.001}.
Conclusion: PH was associated with increased readmission rates compared to the other pediatric diagnoses, but the readmission rate in this study was lower than one previous pediatric study. Invasive mechanical ventilation, Medicaid insurance, higher severity of illness, and female gender were associated with a higher likelihood of readmission within 30 days.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer