Hospital readmission not only increases health care costs, but also has a negative impact on the wellbeing of families.1 Consequently, the readmission rate is considered a key indicator of health care quality.2,3 Readmission is defined as repeated admission within 30 days of discharge, and is considered preventable when it occurs within 15 days of discharge and is related to the reason that led to the initial admission.4,5

In recent years, tools have been developed to identify risk factors for hospital readmission, such as the HOSPITAL score.7 This instrument includes, among other variables, parameters such as hemoglobin and sodium levels at discharge, the type of procedures performed, and the previous admissions of the patient as predictors of readmission. Although it has been validated in the pediatric population, the HOSPITAL score considers cancer as the sole indicator of chronicity, not taking into account other chronic conditions.

In the field of pediatrics, several studies show that children with medical complexity are at increased risk of preventable readmission, independently of their age or the severity of their condition.5,6 However, there is currently no prediction rule for pediatric readmission that includes complex chronic conditions (CCCs) as a risk factor.

The PedCom scale has recently been validated as a tool for identifying pediatric patients with CCCs, with a cut-off point of 6.5, a sensitivity of 98% and a specificity of 94%.8

The objective of the study was to analyze the clinical characteristics of non-oncological pediatric readmissions and to identify risk factors for readmission in the population of children with CCCs.

We designed a retrospective, analytical study of readmissions to the general pediatric inpatient ward of a tertiary care hospital with an annual admission volume of approximately 1500.

The study period spanned from January 2021 to June 2023. During this time, the hospital did not have a specific unit dedicated to the management of CCCs. We excluded readmissions from the neonatal, surgical, oncology, nephrology, and cardiology units from the analysis. Patient and readmission data were obtained by reviewing health records.

We defined readmission as repeated admission within 30 days of discharge and classified it according to the time from discharge as very early (< 48h), early (2–7 days), intermediate (8–15 days), and late (>16 days). Readmissions were also classified into four categories according to the clinical reason for the new admission4:

• 1

  Continuation or recurrence of the reason for initial admission.

• 2

  Acute decompensation or worsening of an underlying chronic condition related to the initial admission.

• 3

  Acute complication related to the problem causing the initial admission.

• 4

  Unrelated causes.

For example, in the case of a patient with epileptic encephalopathy and cerebral palsy previously admitted for respiratory dysfunction, the clinical reasons for readmission could include respiratory failure (category 1), status epilepticus (category 2), pneumonia (category 3), or acute gastroenteritis (category 4).

The study defined children with medical complexity/CCCs based on criteria adapted from Simon et al.9,10 as children meeting at least one of the following: significant chronic conditions in two or more body systems expected to last at least one year, or presence of one CCC in addition to continuous dependence on technology for at least 6 months or special health care needs. To date, the electronic health records system does not have a method in place to identify children with CCCs upon admission. We reviewed all the readmissions in the study period, and, if the involved patient met the Simon et al. criteria, they were considered to have a CCC.

We compared the variables included in the HOSPITAL score and PedCom8 scale in patients with CCCs and patients with acute conditions (Table 1).

The statistical analysis was performed with the software packages SPSS version 21.0 (SPSS Inc; Chicago, IL, USA) and R version 3.5.3. Quantitative variables were summarized as mean and standard deviation or median and interquartile range, depending on whether the data followed a normal distribution. Qualitative variables were expressed as percentages.

We conducted a comparative analysis using the Mann-Whitney U test and the χ2 test to assess differences in readmission rates between patients with and without CCCs. In the CCC group, we used the Kruskal–Wallis test to analyze differences according to the time of readmission (very early, early, intermediate, and late). We also used the Benjamini–Hochberg procedure for multiple comparisons.

To identify factors associated with the risk of readmission, we performed a multivariate linear regression analysis, assessing the association between the number of previous admissions and the time to readmission.

Statistical significance was defined as a P value of less than 0.05.

The study was approved by the Ethics Committee for Research with Medicines of the hospital (CEIM 457/20).

The analysis included 241 readmissions that occurred during the study period. The overall readmission rate was 3.3% in 2021, 2.5% in 2022, and 2.5% in 2023.

With regard to the reason for readmission, 14.9% of readmissions were due to the same cause as the previous admission, 16.2% to decompensation or worsening of the underlying chronic condition, and 7.5% to an acute complication of the condition that caused the previous admission.

The median length of stay upon readmission was 4 (3−7) days, and the median time to readmission was 13 (6–21) days. In respect of the time of readmission, 10% were very early (< 48h), 19.5% early (2–7 days) and 42.3% late (> 16 days).

Of the total readmissions, 41.5% were related to CCCs. Children with CCCs were older (2.1 [0.5–3.7] years vs. 0.3 [0.1−0.64] years; P<.01), predominantly male (76%), and readmitted more frequently due to decompensation of their underlying disease (28% vs 7.8%; P<.01) or for the same reason as the initial admission (26% vs 7.1%; P<.01) compared to patients with acute conditions.

The length of stay in the previous hospitalization was significantly longer in children with CCCs compared to children with acute conditions (six days [3–5] vs four days [2–6]; P<.01) (Table 2).

In addition, scores were significantly higher in readmitted children with CCCs compared to those without a CCC in both the HOSPITAL score (5 [3–7] vs 2 [1–4]; P<.01) and the PedCom scale (9 [5–14] vs 1 [0−2]; P<.01).

In the separate analysis of the CCC group, we found significant differences in the PedCom score according to the time of readmission. Early readmission was associated with higher scores compared to intermediate and late readmission: (14 [IQR, 9−14] vs 9 [IQR, 4−14] vs 6 [IQR, 4−10]; P<.01).

The reason for readmission also varied in relation to the time to readmission, although there the differences were not statistically significant. Very early readmissions (< 48h) were mostly associated with decompensation of the underlying condition (45.5% vs 9%) and unrelated causes (45.5%). Readmissions in patients with CCCs occurring 8 or more days after discharge were most frequently due to the same cause as the previous admission (43.5% vs 17.4%; P=.07). In the comparative analysis, there were no significant differences in age, sex, HOSPITAL score, number of annual visits, or number of annual admissions according to the time of readmission (Table 3).

Finally, the linear regression analysis showed an inverse association between the number of admissions in the past 12 months and the time to readmission (β coefficient=−0.65; t=−3.83; P<.01).

Nearly half of the patients readmitted in our study were children with medical complexity. For more than 10 years, studies published in the United States have shown a steady increase in the percentage of admissions involving these patients, who are also contributing to increasing hospital care costs. A study published in 2010 by Russell and Simon9 showed that children with medical complexity accounted for up to 40% of pediatric hospital charges, 42% of deaths, and 70%–90% of different forms of technology-assistance procedures. The results of our study show that the basic characteristics of readmissions are different in patients with CCCs compared to those without them. Specifically, they are readmitted more frequently for the same reason as their previous admission, which, according to the definition proposed by Goldfield et al.,4 suggests that these were potentially preventable readmissions. Previous studies in Argentina have demonstrated that respiratory dysfunction is the main cause of readmission in children with CCCs.6

Although the HOSPITAL score does not explicitly include medical complexity/CCCs as an item, our study found that these patients scored higher in this scale, suggesting an increased risk of readmission. In 2023, the HOSPITAL score was validated for the first time in a pediatric population with a sensitivity of 70.96% and a specificity of 78.29%, expanding its utility as a tool for predicting preventable readmissions in the pediatric population.11 However, the only item that assesses for the presence of a chronic condition is the item regarding discharge from an oncology service.

Our findings highlight the need to develop a specific tool for predicting readmissions in pediatric patients that considers medical complexity as a risk factor. This tool would enable the identification of patients at higher risk of readmission and the implementation of preventive strategies. Furthermore, in our study, longer lengths of stay were associated with a higher risk of readmission in children with CCCs, a trend that is already reflected in the HOSPITAL score for adults and children.

To date, there are no studies in the literature identifying risk factors for preventable readmission in children with CCCs based on the level of medical complexity. In Spain, the PedCom scale has been reviewed and validated for the purpose of identifying medically complex children.12 The findings of our study add a new dimension to the utility of this tool by demonstrating that scores on the PedCom scale are associated with the time of readmission among chronic patients. Specifically, patients with higher scores on the scale tend to be readmitted earlier.

In 2023, the PedCom scale was validated to assess medical complexity in patients with CCCs. A higher score indicates greater medical complexity on account of a greater need for chronic medication, special health care needs, or more severe functional limitations. This increased complexity translates into greater clinical vulnerability, which may explain why children with CCCs that score higher on the PedCom scale tend to be readmitted earlier.

In light of these results, we propose the following approach to prevent readmission: ensuring continuity of care after discharge through strategies such as hospital-at-home services, remote patient monitoring systems, or follow-up by primary care providers (with pre-discharge communication with the provider and early initiation of follow-up). Previous studies show that readmission rates can be reduced by carefully planning post-discharge follow-up and ensuring continuity of care.13,14 Our study also identified differences in the reasons for readmission among patients with CCCs. In this regard, the admission of patients with CCCs carries an increased risk of decompensation of the underlying condition (e.g., increased probability of seizures, respiratory failure) in the first week post discharge, followed by an increased risk of reinfection or relapse in the convalescence period that begins after that first week. New hospital-at-home models could contribute to ensuring continuity of care by detecting and treating complications early during the post-discharge convalescence period. This could prevent unnecessary readmissions, which would offer a significant clinical benefit and, following appropriate patient selection, achieve savings in health care costs while preventing functional decline.

On the other hand, our study shows that the time to readmission in patients with CCCs is inversely associated to the number of admissions in the past year. This implies that more stable chronically ill patients who experience fewer decompensations tend to be readmitted later. Therefore, another preventive strategy would be to anticipate recurrent decompensation in these patients. In this regard, specialized, multidisciplinary care programs focused on CCCs, as well as effective coordination with primary care, will be essential to enable an early response to signs of decompensation or clinical deterioration.15,16

This is the first study to propose specific measures aimed at preventing readmission of children with CCCs. However, it should be noted that the study has limitations intrinsic to its single-center retrospective design that could limit the generalizability of its results. The exclusion of readmissions to specialty care units (neonatology, surgery, oncology, nephrology, and cardiology) could be a source of selection bias; however, our study focused on children with CCCs admitted to the general pediatric ward to exclude readmissions scheduled by specialists managing medically complex children, which could not be considered preventable. At the time of the study, we considered the variables included in the current version of the PedCom scale (the preliminary version), so we did not exclude the number of hospital admissions. Future studies could analyze these readmissions considering the definition of CCC of the validated and expanded PedCom scale,12 which not only allows the identification of patients with CCCs, but also their classification according to the level of medical complexity. In addition, the expanded version takes into account socioeconomic variables as factors that could be relevant and were not considered in our study.

The analysis of hospital readmissions in children with CCCs showed that these patients have distinctive characteristics that could guide the implementation of preventive measures. Our findings suggest the need to develop and validate specific scores for patients with CCC and to design prevention strategies specific to this population. Unlike patients who are not medically complex, patients with CCCs who have longer lengths of stay and with higher PedCom scores are at increased risk of preventable readmission. This subgroup would particularly benefit from measures aimed at ensuring continuity of care, such as discharge planning, hospital-at-home, or coordinated follow-up with primary care.

This research did not receive any external funding.