Unplanned Readmissions to Acute Care From a Pediatric Postacute Care Hospital: Incidence, Clinical Reasons, and Predictive Factors

Transcription

1 RESEARCH ARTICLE Unplanned Readmissions to Acute Care From a Pediatric Postacute Care Hospital: Incidence, Clinical Reasons, and Predictive Factors abstract OBJECTIVE: To identify the incidence, clinical reasons, and predictive factors for unplanned readmissions to acute care from a pediatric postacute care hospital. METHODS: A retrospective cohort analysis of all discharges between October 1, 2011, and September 30, 2013 (n = 298), in 1 pediatric postacute care hospital was conducted. Descriptive statistics were used to summarize the incidence and assess the clinical reasons for all readmissions to an acute care hospital. Logistic regression was used to identify predictive factors of any unplanned readmission to an acute care hospital. RESULTS: Thirty percent of all postacute care hospital discharges were unplanned readmissions to an acute care hospital. The primary clinical reasons for unplanned readmissions to acute care were respiratory decompensation (54%) and infection (20%). Requiring invasive mechanical ventilation, being <1 year of age, and having a postacute care length of stay <30 days were the 3 predictive factors. CONCLUSIONS: This is the first study to examine readmission to acute care from a postacute care hospital and to identify age, length of stay, and dependence on mechanical ventilation as predictive factors. Understanding which children are likely to require an unplanned readmission may allow providers to develop strategies to minimize this occurrence. Children who require extended hospital level of care may transfer to a pediatric postacute care hospital after an acute care hospitalization. Sometimes called pediatric long-term acute care hospitals, pediatric rehabilitation hospitals or pediatric specialty hospitals, the postacute care hospital provides medical care and rehabilitation services for children with a wide variety of medical diagnoses, therapeutic needs, and technology dependence. The average length of stay in a pediatric postacute care hospital has been reported to be 46 days, and the extended course of care for these children may include 1 or more readmissions to acute care before discharge home. 1 AUTHORS Jane E. O Brien, MD, 1 Helene M. Dumas, PT, MS, 1 Carol M. Nash, MS, 1 Rania Mekary, MSc, PhD 2 1 Research Center for Children with Special Health Care Needs, Franciscan Hospital for Children, Boston, Massachusetts; and 2 School of Pharmacy, Massachusetts College of Pharmacy and Health Sciences University, Boston, Massachusetts KEY WORDS pediatric hospital readmission, subacute care, long-term care, rehabilitation ABBREVIATIONS CI: confidence interval OR: odds ratio doi: /hpeds Address correspondence to Helene M. Dumas, PT, MS, Research Center, Franciscan Hospital for Children, 30 Warren St, Boston, MA hdumas@fhfc.org HOSPITAL PEDIATRICS (ISSN Numbers: Print, ; Online, ). Copyright 2015 by the American Academy of Pediatrics FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose. FUNDING: Support for this work was provided by Commonwealth of Massachusetts, Executive Office of Health and Human Services, Office of Medicaid (principal investigator: Dr O Brien). POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose. The rate of unplanned readmissions to an acute care hospital is an indicator used by payers and accrediting agencies to measure the quality of a hospital s performance. 2 It is unknown, however, what percentage of readmissions are children returning to acute care from postacute care hospitals. Reducing hospital readmissions is a significant consideration for all providers along the continuum of care and particularly for patient groups that are considered high risk or those 134 VOLUME 5 ISSUE 3

2 with high resource use, 3 such as children who require postacute care. In addition to being expensive, these unplanned events are disruptive to patient care and cause stress for a child and the child s family. 4 Hospitals that are designing integrative models of care and initiatives to improve patient safety and health outcomes, and decrease costs, need to focus on the entire continuum of care for children. Thus, the purpose of this study was to identify the incidence, clinical reasons, and predictive factors for unplanned readmissions to acute care from a pediatric postacute care hospital. METHODS Participants This is an institutional review board approved, retrospective cohort analysis of a convenience sample from 1 pediatric postacute care hospital in the northeastern United States. Electronic medical records of all patient discharges between October 1, 2011, and September 30, 2013, from the inpatient medical and rehabilitation units were analyzed (n = 298). Data were gathered and entered in to a projectspecific database. Data quality and reliability checks were conducted throughout the data collection phase of the study by random repeat data entry and comparison. One episode of care was excluded, as the child died while in the postacute care hospital, rendering a final sample size of 297. Procedures and Analysis Demographic and clinical information abstracted from the medical records for these 297 discharges were representative of 186 children and included the following: gender, race/ethnicity, referring acute care hospital, payer type, primary diagnosis, reason for admission to postacute care, second ary diagnoses and comorbidities, post acute care hospital admission number, admission and discharge day of the week and month of the year, discharge disposition (planned discharge to home, planned discharge to acute care, an unplanned readmission to acute care), and reason for readmission to an acute care hospital, as applicable. Using the primary diagnosis, secondary diagnoses, and reason for postacute care hospital admission, each case was assigned by the primary investigator to 1 of 5 previously described clinical program groups: noncomplex infants (diagnoses such as neonatal abstinence syndrome, prematurity with feeding issues), active rehabilitation (diagnoses such as acquired brain injury, spinal cord injury [no ventilator], cerebral palsy, and post orthopedic surgery), medically complex (chronic complex medical condition with 1 of the following: tracheostomy, gastrostomy, postsurgery [orthopedic, gastrointestinal], sepsis ruptured appendix, new onset or uncontrolled seizures, respiratory distress, dehydration, failure to thrive, malnutrition), and ventilator-dependent: noninvasive (eg, using bilevel positive airway pressure) and ventilatordependent: invasive (tracheostomy plus ventilation). 1 Standard descriptive statistics were used to summarize the demographic and clinical characteristics of the sample. Date of birth, date of admission, and date of discharge were used to calculate age at admission, age at discharge, and total length of stay in postacute care. Number of readmissions to acute care in <3 days, <15 days, and <30 days of admission to postacute care were calculated. Reasons for all planned and unplanned readmissions to acute care were sorted by clinical program group. To examine predictors of the independent variable, unplanned readmission to a pediatric acute care hospital, we estimated odds ratios (ORs) and 95% confidence intervals (CIs) by using the logistic regression model with the backward selection option. We first included the following covariates in a multivariate model: age at admission (years, continuous), length of stay at the postacute care hospital (days: <30, 30 [median = 29 days]), gender (male, female), ethnicity (white, nonwhite), admission and discharge day of the week and month of the year (continuous), number of admissions to postacute care (continuous), referral source (Boston hospitals, non-boston hospitals), payer type (public, private), and postacute care hospital clinical program group 1 (noncomplex infants, active rehabilitation, medically complex, ventilator-dependent: noninvasive, ventilator-dependent: invasive). Because of the level of clinical severity, the ventilator-dependent invasive group was considered as the reference category. The same multivariate model was re-run in a sensitivity analysis including length of stay in 6 categories (days: 0 7, 8 14, 15 30, 31 60, , >100) and age at admission in 5 categories (years: [infant]), >1 3, >3 7.5, >7.5 15, >15 years). The final multivariate model presented includes only the significant predictors retained by the backward selection option. As this is a retrospective cohort study and OR might significantly overestimate associations between risk factors and common outcomes, 5 we 135

3 conducted a second sensitivity analysis by using Proc Genmod with the binomial distribution and the link log to estimate the relative risk and the 95% CI by the log-binomial regression model. A third sensitivity analysis was conducted on the 186 cases representing each individual child, by using only their first admission to the postacute care hospital during the study period. SAS version 9.3 (SAS, Inc, Cary, NC) was used for all analyses, and P <.05 was considered statistically significant. RESULTS Incidence and Clinical Reasons for Unplanned Readmissions to Acute Care Demographics of the study sample are presented in Table 1. For 207 (70%) cases, discharge was planned (147 discharged to home and 60 planned discharges/readmissions to acute care). There were 90 (30%) unplanned readmissions to an acute care hospital. There were 14 (16%) readmissions to acute care in <3 days, 41 (46%) in <15 days, and 59 (66%) in <30 days of admission to the postacute care hospital. In this study sample, all of the children who returned to the acute care hospital, planned or unplanned, returned to the acute care hospital from which they were initially admitted unless a change was requested by the parent/ guardian or was due to medical emergency and travel distance. Fifty-eight (52%) of those children readmitted to acute care (unplanned) and returning to postacute care, did so only once. The mean number of discharges per child from postacute care was 1.82 (median 1.0, range 1 10 admissions). For the total sample of unplanned readmissions (n = 90), most readmissions to acute care were for cases in the ventilator-dependent: invasive group (n = 49, 54%), followed by the cases in the medically complex group (n = 27, 30%). Clinical reasons for planned and unplanned readmissions to acute care are listed for each clinical program group in Table 2 and Table 3. Predictive Factors for Unplanned Readmissions Of the 12 variables included in the logistic regression model, only 3 dependent variables were retained and were shown to be statistically significant. These variables were age at admission, length of stay in postacute care, and clinical program group. 1 For every additional year in age, (older) children had 7% lower odds of having an unplanned readmission to a pediatric acute care hospital as compared with younger children (OR 0.93, 95% CI ). Children >15 years had 83% lower odds of being readmitted to TABLE 1 Demographics of Study Sample (n = 297) Mean age at admission, y 3.8 (median 1.05, range ) Mean age at discharge, y 4.0 (median 1.2, range ) Mean length of stay, d (median 29, range 1 460) Gender, boys, n (%) 160 (54) Ethnicity, white, non-hispanic, n (%) 104 (35) Clinical program groups, n (%) Noncomplex infants: 41 (14) Active rehabilitation: 21 (7) Medically complex: 122 (41) noninvasive: 24 (8) invasive: 89 (30) acute care (OR 0.17, 95% CI ) as compared with infants (0 1 year; reference category). Children who were still in the postacute care hospital for 30 days (n = 31) after admission had 67% lower odds of an unplanned readmission to acute care (OR 0.33, 95% CI ) as compared with those who were in the postacute hospital for <30 days (n = 59) when readmitted to acute care. When length of stay was categorized into 6 categories, the results were in the same direction, suggesting that children who had been in the postacute care hospital the longest (>60 days), had the lowest odds of being readmitted to acute care (OR 0.05, 95% CI ) as compared with children who were in the postacute care hospital for 7 14 days (OR 0.17, 95% CI ), 15 to 30 days (OR 0.17, 95% CI ), 31 to 60 days (OR 0.16, 95% CI ), or 61 to 100 days (OR 0.06, 95% CI ). Children whose clinical program group 1 was categorized as noncomplex infants (OR 0.03, 95% CI ), medically complex (OR 0.18, 95% CI ), or ventilator: noninvasive (OR 0.27, 95% CI ) had significantly lower odds of unplanned readmission to acute care as compared with children in the ventilatordependent: invasive (reference) group (Table 4). These results did not change in the sensitivity analysis when we included only first admissions (n = 186) to the postacute care hospital (Table 4). The only exception was for children whose clinical program group was ventilator-dependent: noninvasive, as their decreased odds of having an unplanned transfer to an acute hospital became nonsignificant when compared with children in 136 VOLUME 5 ISSUE 3

4 TABLE 2 Reasons for Planned Discharge From Postacute Care Hospital by Clinical Program Group (n = 207) Noncomplex infants, n = 39, 19% Active rehabilitation, n = 14, 7% Medically complex, n = 95, 46% noninvasive, n = 19, 9% invasive, n = 40, 19% Planned Home Discharge, n = 148, 72% Other, n = 26, 13% Planned Discharge/Readmission to Acute Care Hospital, n = 59, 28% G-tube Placement, n = 15, 7% Respiratory Evaluation, n = 8, 4% Cardiac, n = 6, 3% Orthopedic, n = 3, 2% Organ Transplant, n = 2, 1% the ventilator-dependent: invasive category, most likely owing to loss of power (OR 0.35, 95% CI ). The OR and 95% CI obtained with Proc Genmod were not materially different from the OR for both the main and the sensitivity analyses (Table 5). DISCUSSION Pediatric postacute care hospitals play an important role in the continuum of care by providing care for a medically complex and fragile group of children, with the largest diagnostic groups being children with medical complexity and those dependent on invasive mechanical ventilation. 1 The purpose of this study was to identify the incidence, clinical reasons, and predictive factors for unplanned readmission to acute care from a pediatric postacute care hospital. Within the single-site postacute care hospital where this study was conducted, hospital policy supports the following criteria for emergency transfer to acute care: medical decompensation, not responsive to treatment (eg, unstable blood pressure, neurologic decompensation, respiratory failure), airway instability requiring intensive care (eg, emergent intubation, progressively increasing ventilator support), increasing medical complexity requiring invasive monitoring and increased staffing needs beyond the standard of care provided in postacute care, and/or acute changes in status requiring diagnostic capabilities or interventions not available in postacute care (eg, MRI, computed tomography, Doppler study, or surgical procedure). It is unknown if these criteria are similar to that of other pediatric postacute care hospitals. Children with medical complexity and children with ventilator dependence accounted for 79% of the sample in this single-site study of postacute care. Although estimates indicate they account for <1% of the pediatric population, children with medical complexity and technology dependence use almost one-third of all pediatric health care expenditures and make multiple transitions across providers and health care settings. 1,3,6,7 For instance, in 2006, children with medical complexity accounted for 10% of all US pediatric acute care hospital admissions, TABLE 3 Reasons for Unplanned Readmissions to Acute Care by Clinical Program Group (n = 90) Noncomplex infants, n = 2, 2% Active rehabilitation, n = 7, 8% Medically complex, n = 27, 30% noninvasive, n = 5, 6% invasive, n = 49, 54% Respiratory, n = 49, 54% Infection, n = 18, 20% Other, n = 8, 9% Cardiac, n = 7, 8% Equipment Malfunction, n = 6, 7% Gastrointestinal, n = 2, 2%

5 TABLE 4 Predictors of Unplanned Readmissions to a Pediatric Acute Care Hospital From a Pediatric Postacute Care Hospital By Using a Logistic Regression Model Variables Retained in the Model All Participants, n = 297 Participants With Only 1 Admission, n = 186 a n Cases/Total N MV-OR (95% CI) b n Cases/Total N MV-OR (95% CI) b Age at admission 90/ ( ) 40/ ( ) Length of stay in postacute hospital, d <30 31/ (Referent) 16/ (Referent) 30 59/ ( ) 24/ ( ) Clinical program group Ventilator: invasive 49/ (Referent) 17/ (Referent) Noncomplex infants 2/ ( ) 2/ ( ) Active rehabilitation 7/ ( ) 4/ ( ) Medically complex 27/ ( ) 13/ ( ) Ventilator: noninvasive 5/ ( ) 4/ ( ) a A sensitivity analysis was conducted in which only children with 1 admission were kept in the analysis. b Multivariate (MV)-OR is adjusted for age at admission, length of stay, and clinical program group; ORs were estimated from a logistic regression model by using Proc Logistic in SAS. 26% of pediatric hospital days, 40% of pediatric hospital charges, up to 92% of technology assistance procedures, and 435 inpatient deaths. 7 It has become imperative for health care providers to study and understand the needs of this population as models of care are developed. Data are available for unplanned readmissions to an acute care hospital for children and, in general, the incidence of unplanned readmissions has been reported as 8.4% at 15 days 8 and 6.5% at 30 days. 4,9 More specifically, the 30-day readmission rate for children who are typically developing and without chronic illness is 3.1%. 3 For children with a single complex chronic condition and/or technology assistance, the rate of readmission to an acute care hospital has been reported to be up to 25% within 30 days of a previous hospitalization. 3,6 For children newly dependent on invasive mechanical ventilation and discharged from the hospital, the 12-month incidence of nonelective readmission has been reported at 40%. 10 This is the first study examining readmission to acute care from a postacute care hospital. A recent study identifying children with chronic complex conditions reported a 19% incidence of readmission to the referring acute care hospital within 30 days of a subacute hospital discharge. 11 Thus, our overall unplanned readmission incidence TABLE 5 Predictors of Unplanned Readmissions to a Pediatric Acute Care Hospital From a Pediatric Postacute Care Hospital By Using a Log-Binomial Regression Model Variables Retained in the Model All Participants, n = 297 Participants With Only 1 Admission, n = 186 a n Cases/Total N MV-RR (95% CI) b n Cases/Total N MV-RR (95% CI) b Age at admission 90/ ( ) 40/ ( ) Length of stay in postacute hospital, d <30 d 31/ (Referent) 16/ (Referent) 30 d 59/ ( ) 24/ ( ) Clinical program group Ventilator: invasive 49/ (Referent) 17/ (Referent) Noncomplex infants 2/ ( ) 2/ ( ) Active rehabilitation 7/ ( ) 4/ ( ) Medically complex 27/ ( ) 13/ ( ) Ventilator: noninvasive 5/ ( ) 4/ ( ) a A sensitivity analysis was conducted in which only children with 1 admission were kept in the analysis. b Multivariate relative risks (MV-RR) were adjusted for age at admission, length of stay, and clinical program group; RRs were estimated from a log-binomial regression model by using Proc Genmod in SAS. 138 VOLUME 5 ISSUE 3

6 of 30% appears somewhat consistent with past studies for both acute and subacute care. Documenting the incidence of readmission to an acute care hospital for children who require postacute (or subacute) hospitalization is the first step in determining the magnitude of this issue and the efforts needed toward prevention. Consistent with previous research, 12 infections accounted for 20% of the reasons for return to acute care in this study. Respiratory decompensation, however, was the primary clinical reason for unplanned readmission to acute care in this sample. Given the large number of children with dependence on invasive and noninvasive ventilator support and considered medically complex (children with multisystem involvement), these were not surprising results. Respiratory decom pensation in this complex group of children has been noted to result from aspiration pneumonia, tracheitis, or a systemic illness. Knowing these are the clinical reasons for return to acute care, increased diligence with prevention and surveillance may help reduce the need for a return to acute care for treatment. In this study, clinical program group (eg, ventilator-dependent: invasive) was a predictor for returning to acute care after admission to a postacute care hospital. The clinical program group is reflective of the child s level of acuity, with dependence on invasive mechanical ventilation representing the highest level of severity and encompasses the impact of any additional diagnoses or clinical conditions. 1 For children discharged from subacute care and readmitted to acute care within 30 days, ventilator dependence was one of many factors associated with readmission. 11 Kun et al, 10 however, found no statistically significant associations between demographic and clinical data with nonelective readmissions to acute care from home for children dependent on mechanical ventilation. Gay et al 8 reported that infants (<1 year) was the age group with the greatest number of admissions (21%) within 15 days of discharge over a 1-year study period of readmissions to a pediatric acute care hospital. In this study, children who were younger also were more likely to be readmitted to acute care. Infants in postacute care, particularly those who are ventilator dependent, include a fragile group with chronic lung disease secondary to extreme prematurity and/or congenital anomalies and/or complex congenital heart disease. 1 In general, as these infants survive and grow, their conditions stabilize and therefore it is not surprising that younger children are more likely to be less stable in the postacute hospital setting. Readmission within 30 days of acute care discharge is a benchmark quality indicator for both adult and pediatric hospitals in efforts to improve patient outcomes, safety, and satisfaction. 2,9 The Centers for Medicare and Medicaid Services 2 encourages measuring readmission within 30 days rather than over a longer period, as the need for readmission may be affected by factors outside the hospital s control, such as illnesses, patients behavior, or care provided to patients after discharge, rather than the discharging hospital. With a median length of stay of 29 days and by using this common 30-day benchmark, we chose the cutoff of 30 days as a dependent variable to depict its association with readmission to acute care from postacute care. Being in the postacute care hospital for <30 days at the time of return to an acute care hospital was a predictive factor in unplanned readmission to acute care in this study. In a recent study examining readmission to acute care within 30 days of discharge from a subacute care hospital for children with complex chronic conditions, readmission was associated with the number of home medications, underlying chronic respiratory illness, pulse oximetry, home apnea monitoring, tracheostomy and/or ventilator dependence, length of stay, and number of follow-up appointments. The incidence of readmission increased to 29% for children discharged with >8 medications. 11 Differing from this current study in that the children were readmitted to acute care after discharge from subacute care, 2 of the predictor variables (length of stay and ventilator dependence) were the same, however, and considered together, these studies provide stronger evidence of the impact of these factors. The generalizability of this study is limited by the small sample size and inclusion of only 1 postacute care hospital. It is, however, the first report of readmissions to an acute care hospital from a pediatric postacute care hospital and could be easily replicated at other sites. Future work should include study of additional demographic, clinical, and institutional care factors that may contribute to an unplanned acute care readmission; evaluation of patient acuity at time of transfer to and from postacute care; identification of which readmissions are more likely preventable; and an examination of postacute care hospital discharges to home with return to an acute care hospital. In addition, future areas for study may 139

7 include examination of the criteria for discharge to postacute care, transition communication between providers, and interfacility transport protocols. CONCLUSIONS We found that 30% of all postacute care hospital discharges were unplanned readmissions to an acute care hospital, with the primary clinical reasons being respiratory decompensation (54%) and infection (20%). Requiring invasive mechanical ventilation, being <1 year of age at admission, and being in the postacute care hospital <30 days at time of readmission to acute care were the 3 predictive factors of unplanned readmission to an acute care hospital. Postacute care hospitals are an important part of the continuum of care and understanding which children are more likely to require a readmission to acute care may allow providers to develop strategies to minimize this costly and disruptive occurrence. REFERENCES 1. O Brien JE, Dumas HM. Hospital length of stay, discharge disposition, and reimbursement by clinical program group in pediatric post-acute rehabilitation. J Pediatr Rehabil Med. 2013;6(1): Centers for Medicare and Medicaid Services (CMS). Available at: Accessed February 18, Berry JG, Agrawal R, Kuo DZ, et al. Characteristics of hospitalizations for patients who use a structured clinical care program for children with medical complexity. J Pediatr. 2011;159(2): Berry JG, Ziniel SI, Freeman L, et al. Hospital readmission and parent perceptions of their child s hospital discharge. Int J Qual Health Care. 2013;25(5): Diaz-Quijano FA. A simple method for estimating relative risk using logistic regression. BMC Med Res Methodol. 2012; 12: Cohen E, Berry JG, Camacho X, Anderson G, Wodchis W, Guttmann A. Patterns and costs of health care use of children with medical complexity. Pediatrics. 2012; 130(6). Available at: cgi/content/full/130/6/e Simon TD, Berry J, Feudtner C, et al. Children with complex chronic conditions in inpatient hospital settings in the United States. Pediatrics. 2010;126(4): Gay JC, Hain PD, Grantham JA, Saville BR. Epidemiology of 15-day readmissions to a children s hospital. Pediatrics. 2011;127(6). Available at: full/127/6/e Berry JG, Toomey SL, Zaslavsky AM, et al. Pediatric readmission prevalence and vari - ability across hospitals [published correction appears in JAMA. 2013;309(10):986]. JAMA. 2013;309(4): Kun SS, Edwards JD, Ward SL, Keens TG. Hospital readmissions for newly discharged pediatric home mechanical ventilation patients. Pediatr Pulmonol. 2012;47(4): Jurgens V, Spaeder MC, Pavuluri P, Waldman Z. Hospital readmission in children with complex chronic conditions discharged from subacute care. Hosp Pediatr. 2014;4(3): Hain PD, Gay JC, Berutti TW, Whitney GM, Wang W, Saville BR. Preventability of early readmissions at a children s hospital. Pediatrics. 2013;131(1). Available at: www. pediatrics.org/cgi/content/full/131/1/e VOLUME 5 ISSUE 3