Factors Associated With Return Visits at 7 Days After Hospital Discharge

This is a retrospective cohort study of patients discharged from an urban, academic, freestanding, tertiary-care pediatric hospital with 255 licensed inpatient beds, ∼15 000 admissions per year, and an annual ED census of 92 000 visits. The facility is staffed by hospitalists who provide coverage for select primary care pediatric practices and a teaching service with attending physicians, fellows, and resident physicians.

In this study, we included patient encounters during a 3-year period from October 2012 to September 2015. We also focused on encounters in the top 10 All Patient Refined Diagnosis-Related Groups (APR-DRGs) (version 32) inpatient hospitalizations. These were as follows: asthma; seizure; bronchiolitis and respiratory syncytial virus pneumonia; upper respiratory tract infection; nonbacterial gastroenteritis with nausea and vomiting; cellulitis and/or bacterial skin infection; digestive system diagnosis not elsewhere classified; pneumonia not elsewhere classified; diabetes; and respiratory signs, symptoms, and minor diagnoses. Patients >21 years of age at the initial visit were excluded.

Data were obtained from 2 sources: the Pediatric Health Information System (PHIS) and the hospital’s electronic medical record. The PHIS is a deidentified, administrative database maintained by the Children’s Hospital Association (Lenexa, KS), a consortium of 51 not-for-profit, tertiary-care children’s hospitals in the United States.

An index visit was defined as the initial admission encounter associated with a top 10 APR-DRG. For patients with multiple encounters during the study period, all subsequent hospitalizations within the top 10 APR-DRGs were also considered index visits.

The primary outcome was 7-day all-cause revisits after the index hospitalization. Revisits were defined as ED encounters or admissions after an index visit. Variables reflecting patient demographic, clinical, and hospital characteristics, which may be associated with unplanned hospital returns, were captured for analysis. For ease of comparative analysis, age was categorized into 7 groups based on previous studies: neonatal, ≤30 days; infant, >30 days to 1 year; toddler, >1 to 3 years; child, >3 to 9 years; preteenager, >9 to 12 years; teenager, >12 to 18 years; and young adult, >18 years.⁷ 

The collected patient-related clinical variables were heart rate at discharge¹⁰  (normal versus abnormal) and the Pediatric Medical Complexity Algorithm (PMCA) category for identifying chronic medical conditions. We speculated that heart rate at discharge may be associated with return visits on the basis of the pediatric Rothman Index, a novel metric that incorporates several clinical and laboratory parameters to predict readmission risk.¹¹ 

The captured hospital variables included length of stay (LOS) in days, surgical indicator (yes or no), ICU flag (yes or no), and weekday versus weekend discharge. A weekday was defined as any day between Monday at 7 am and Friday at 7 pm on the basis of previous studies as well as the unique availability of resources in our local environment.⁷ 

Continuous variables were reported as means with SDs, and categorical variables were presented as frequencies with percentages. Unadjusted and adjusted odds ratio (OR) estimates along with corresponding 95% confidence intervals (CIs) were reported on the basis of the bivariate and multivariable exploratory analyses, respectively, by using generalized estimating equations. Variables with P < .1 from the bivariate analyses were included in the initial regression models. A final model was developed to determine predictors for each of the outcomes, 7-day ED revisits and readmissions, by using the backward model selection approach and retaining only variables with P < .05 through the iterative process. All analyses were conducted by using SAS 9.3 (SAS Institute, Inc, Cary, NC).

The study was approved by the institutional review board.

There were 17 707 top 10 APR-DRG encounters for 13 642 patients identified during this study period. The encounter characteristics are summarized in Table 1. In the study population, the 7-day revisit rate was 4.2% (736 revisits). Bivariate analysis revealed that insurance status, being at the lower and upper ends of the age groups (ie, neonates, infants, teenagers, and adults), longer length of index hospital stay, and presence of a complex chronic condition were more likely to be associated with 7-day revisits (Table 2). Multivariable analysis revealed persistence of these associations.

The 7-day readmission rate was 2.3% (416 readmissions). Bivariate analysis revealed that race (especially African American versus white), being at the lower and upper ends of the age groups (ie, neonates, infants, teenagers, and adults), longer length of index hospital stay, and presence of a complex chronic condition were more likely to be associated with 7-day readmissions (Table 3). Multivariable analysis revealed similar findings in all categories except race.

In this study, we identified 3 factors that were associated with an increased risk of both 7-day revisits and 7-day readmissions for children discharged with common problems. These factors included age, LOS, and having a complex chronic medical condition. In addition, insurance status was associated with risk of revisit and race was associated with risk of readmission in the bivariate analysis.

Our study is consistent with a previous study, the results of which revealed younger age to be a risk factor for 7-day readmission.¹²  In addition, our 2.3% 7-day readmission rate was lower than that reported by Solutions for Patient Safety¹³  (4%), a collaborative of children’s hospitals dedicated to sharing outcomes and best practices. Interestingly, our results revealed that insurance status was associated with revisits. This may indicate challenges for appropriate transition to home care with certain insurance coverages. Moreover, our results revealed an association with race and early readmission, which may be a surrogate for readiness for discharge from the hospital for certain populations. This may also reflect the absence of necessary resources for home care.

In this study, we add to the current literature by examining pediatric revisits that did not result in readmission. ED revisits are not often included in the traditional definition of readmissions, but our findings support the importance of including this unique subset of pediatric patients in studies examining returns after hospital discharge. In addition, our focus was on early (7-day) return for a cohort of inpatients with the 10 most frequent discharge diagnoses. With regard to pediatric patient returns, we believe that 7 days may be a more meaningful time frame for identifying opportunities for intervention and prevention compared with the 30-day metric frequently used for adult patients.¹⁴  Our study was novel in that we combined variables from the PHIS, an administrative database, with patient-level clinical variables abstracted from the electronic medical records, such as heart rate and timing of discharge (weekend versus weekday), to assess associations for revisits.

The results of our study are consistent with previous data, revealing an association of 30-day readmissions in patients with complex chronic medical conditions. Using 7-day revisits may help identify missed opportunities for discharge preparation that may help reduce the need for return visits. Missed opportunities may stem from a lack of discharge readiness, a lack of caregiver understanding of illness management, a lack of access to follow-up care, or the persistence or progression of disease. In isolation, unlike in adult populations with chronic diseases, pediatric readmissions may be a flawed hospital performance measure.^15,16  However, the cumulative annual financial impact of these return visits can be significant for both hospitals and patient families. We found that younger patients were at a higher risk for revisits and readmissions. This population may benefit from further interventions at discharge, such as formal handoffs to primary care providers before discharge, assistance with medications, arranging follow-up transportation, dedicated education regarding postdischarge care, and early follow-up with the primary care provider.

This study had several limitations. It includes data from 1 pediatric institution. In addition, ED returns to other hospitals or to primary care offices were not evaluated. We also did not assess preventability or rationales for revisits.^4,17  Although a relationship between medications and risk of return visits has been previously described, we did not evaluate medication use at discharge.¹⁸  Lastly, we did not have information regarding adherence to evidence-based practices during transition from the ED to inpatient units, all of which have been identified as factors that may influence readmissions.^19–24 

In our study, we identified patient characteristics that may be associated with risk for early hospital return among pediatric patients discharged with the 10 most frequent APR-DRGs. This group can be targeted for intervention and enhanced postdischarge care coordination.