Preventing Hospitalizations in Children With Medical Complexity: A Systematic Review
BACKGROUND AND OBJECTIVES: Children with medical complexity (CMC) account for disproportionately high hospital use, and it is unknown if hospitalizations may be prevented. Our objective was to summarize evidence from (1) studies characterizing potentially preventable hospitalizations in CMC and (2) interventions aiming to reduce such hospitalizations.
METHODS: Our data sources include Medline, Cochrane Central Register of Controlled Trials, Web of Science, and Cumulative Index to Nursing and Allied Health Literature databases from their originations, and hand search of article bibliographies. Observational studies (n = 13) characterized potentially preventable hospitalizations, and experimental studies (n = 4) evaluated the efficacy of interventions to reduce them. Data were extracted on patient and family characteristics, medical complexity and preventable hospitalization indicators, hospitalization rates, costs, and days. Results of interventions were summarized by their effect on changes in hospital use.
RESULTS: Preventable hospitalizations were measured in 3 ways: ambulatory care sensitive conditions, readmissions, or investigator-defined criteria. Postsurgical patients, those with neurologic disorders, and those with medical devices had higher preventable hospitalization rates, as did those with public insurance and nonwhite race/ethnicity. Passive smoke exposure, nonadherence to medications, and lack of follow-up after discharge were additional risks. Hospitalizations for ambulatory care sensitive conditions were less common in more complex patients. Patients receiving home visits, care coordination, chronic care-management, and continuity across settings had fewer preventable hospitalizations.
Conclusions: There were a limited number of published studies. Measures for CMC and preventable hospitalizations were heterogeneous. Risk of bias was moderate due primarily to limited controlled experimental designs. Reductions in hospital use among CMC might be possible. Strategies should target primary drivers of preventable hospitalizations.
- ACSC —
- ambulatory care sensitive condition
- AOR —
- adjusted odds ratio
- CCC —
- complex chronic condition
- CI —
- confidence interval
- CMC —
- children with medical complexity
- CRG —
- clinical risk group
- CSHCN —
- Children With Special Health Care Needs
- ED —
- emergency department
- ICD-9 —
- International Classification of Diseases, Ninth Revision
- MAC —
- Maryland Access to Care
- MeSH —
- Medical Subject Headings
- MHI —
- Medical Home Index
- PACC —
- Pediatric Alliance for Coordinated Care
- PCP —
- primary care provider
- RN —
- registered nurse
Improvements in pediatric care have led to a substantial increase in the number of children surviving previously fatal complex conditions.1–4 The subsequent disability, vulnerability, and dependence on technology of these children can lead to significant family and health system consequences requiring intensive care coordination to achieve optimal outcomes.4,5 Nationally representative hospital samples have demonstrated that children with medical complexity (CMC) can be a small percentage of the population yet account for disproportionately high hospital utilization, including admissions, hospital days, and hospital charges.6–8 Though CMC represent ∼1% of the population, they may account for up to one-third of child health expenditures, with up to 80% of their cost due to inpatient care.8–10 At present, no consensus exists regarding how to define or reduce potentially avoidable hospitalizations in CMC. Preliminary studies contend that utilization might be reduced through certain care coordination activities,4,11–13 and some have used the patient-centered primary care medical home as a framework.14,15
Models commonly used to identify potentially preventable hospitalizations (eg, ambulatory care sensitive conditions [ACSCs]) may not apply to this particular population. For example, although gastroenteritis and dehydration may represent preventable hospitalizations in the general population, this may not be an appropriate designation in the child with a multiorgan system chronic illness, inability to tolerate enteral feedings, or a metabolic disorder. To understand and ultimately decrease hospitalizations in the complex population, drivers of preventable hospitalizations need to be identified. Interventions based on potentially modifiable drivers might lead to lower hospital utilization.
Because to our knowledge no review of preventable hospitalizations in CMC has been conducted, this systematic review aims to examine (1) studies characterizing potentially preventable hospitalizations in CMC, and (2) interventions aiming to reduce potentially preventable hospitalizations in these children. Such a review is critical to developing evidence-based strategies most likely to succeed in reducing unnecessary hospitalizations in this high-utilizing group of patients.
Data Sources and Article Selection
We searched for peer-reviewed English-language articles using Medline, Cochrane Central Register of Controlled Trials, Web of Science, and Cumulative Index to Nursing and Allied Health Literature databases from their originations, initially through May 30, 2013. The search strategy was developed with a biomedical librarian and carried out by using the AND operator to link combinations of key words and Medical Subject Headings (MeSH) terms from 3 groups of concepts: (1) children, (2) preventable hospitalization, and (3) chronic or complex illness. Additional articles were included after hand-searching the bibliographies of included articles. The comprehensive search strategy and database results are included (Supplemental Information 1). We updated the literature search to identify any new studies published through June 23, 2014.
Two reviewers (Dr Coller and Ms Saenz) used a structured screening protocol (Supplemental Information 2), refined after piloting on 5 representative articles, to independently screen titles, then abstracts, and finally full text articles. We identified studies focusing on pediatric patients with medical complexity that either characterized potentially preventable hospitalizations or tested interventions to reduce them. Using the conceptual frameworks for CMCs described by van der Lee and Cohen,4,16 we only included articles in which data from CMC were available to be analyzed. A third reviewer (Dr Nelson) resolved inclusion/exclusion disagreement between the primary reviewers.
Several specific exclusion criteria were defined before article screening. Studies exclusively looking at mental health diagnoses were excluded due to being outside the scope of the study aim. Because the organization and financing of the US health care system are sufficiently distinct from other industrialized countries, we only included studies occurring in the United States. We also excluded studies examining vaccine effectiveness or cost-effectiveness, unless they specifically examined activities to increase vaccination rates in CMC. The following study designs were excluded: case studies, letters to the editor, notes, clinical overviews, guidelines, and reviews or meta-analyses. Bibliographies of reviews were hand-searched to identify additional articles.
Data Extraction and Synthesis
A structured data collection tool was used to extract relevant data on study design, methods, study indicators for medical complexity and preventable hospitalization, outcomes, and findings. Data pertaining to hospitalization rates, hospital days, and/or hospital costs according to the study definition for preventable hospitalization were sought. Principal summary measures included odds ratios and differences in means or proportions. Results of studies were organized according to common themes of study focus. It was also noted if the study intervention, exposures, or findings were related to medical home principles as defined by the American Academy of Pediatrics. Finally, modifiable drivers of preventable hospitalization across studies were summarized.
Risk of bias was determined by 2 independent reviewers (Drs Coller and Sklansky) using a structured data collection tool. Observational study bias was assessed by using the Agency for Healthcare Research and Quality RTI Item Bank to Assess Risk of Bias and Confounding,17 and experimental study bias was assessed by using the Downs and Black checklist for assessing methodological quality of randomized and nonrandomized health care interventions.18 The Downs and Black checklist power assessment was modified from a 0 to 5 to a 0 to 1 scale, where the item was scored “1” if a power calculation or sample size calculation was present and “0” if there was no power/sample size calculation or an explanation of the appropriateness of the number of subjects. The wide variability in study designs, populations, and outcomes precluded pooling of data for meta-analysis.
The initial search yielded 484 titles after removing duplicates (Fig 1). After title screening, 237 titles remained for abstract screening. Of these abstracts, 195 did not meet our predefined article inclusion criteria, leaving 42 articles for full-text review. After full-text review, an additional 32 articles were excluded, leaving 10 articles for data extraction and bibliography review. Hand-search of the bibliographies from final articles produced an additional 7 articles, resulting in a total of 17 articles included in this systematic review (Tables 1, 2, and 3).
CMC were defined heterogeneously, broadly divided into either categorical diagnosis-based schema or noncategorical consequences-based schema. Categorical systems for identifying CMCs through groups of International Classification of Diseases, Ninth Revision (ICD-9) diagnostic codes representing particularly complex patient populations were used in 12 studies. Five studies used complex chronic conditions (CCCs) and 1 used clinical risk groups (CRGs); both of which are well-described approaches.19–21 Six other studies identified patients by using single diagnoses or specific groups of ICD-9 codes created by the investigators for the purposes of their study. Several studies defined CMCs by using a noncategorical, consequences-based approach, eg functional limitations on the Children With Special Health Care Needs (CSHCN) screener (n = 1), combinations of characteristics such as intensity of subspecialty use, organ system involvement, and past utilization (n = 2), or technology assistance (n = 5). Noncategorical approaches may capture a more diverse or inclusive cohort of children.4,21,22
Potentially preventable hospitalizations were defined in 1 of 3 ways: ACSCs (n = 5), readmissions (n = 8), or investigator-defined criteria such as after chart review or observing changes in hospitalization rates after an intervention (n = 7).
Each of these definitions has important considerations when interpreting findings. Results from several studies suggest that ACSCs may not be ideal markers of potentially preventable hospitalization in this population. First, ACSC hospitalizations may be significantly less common in CMC. Hospitalizations for ACSCs were less common for more complex patients in 2 separate national samples.23,24 Second, when ACSCs do occur among CMC, it is not clear that ambulatory care can prevent hospitalization. In the Maryland Access to Care (MAC) Medicaid experiment, investigators were only able to classify 38% of ACSC hospitalizations as avoidable.25 Armour et al26 found that among 81 patients with spina bifida and a hospitalization for a urinary tract infection (an ACSC), 73 had an ambulatory claim in the 7 days before hospitalization, with 47% of them receiving a diagnosis of urinary tract infections at that time. The study raises the question of how truly “ambulatory care sensitive” urinary tract infections were in this population when >90% of the patients received ambulatory care in the week before hospitalization.
Readmissions, assumed to frequently represent failures in care at 1 or more level, have important considerations as an indicator of preventable hospitalizations largely because there is no valid standard upon which a readmission can be labeled “preventable.”27–29 Hain et al28 devised a detailed 5-point Likert scale to rate preventability, which was independently applied by 4 physicians to chart reviews of 15-day readmission cases. They found that 20% of 15-day readmissions at their institution were potentially preventable, though overall agreement was difficult to achieve (and was specifically difficult in cases that were not clearly unpreventable). Finally, defining preventable hospitalizations from investigator-designed criteria are usually developed without confirming reliability and validity, making interpretation of results challenging. These definitions are also unique and therefore difficult to compare across studies.
Clinical Antecedents to Potentially Preventable Hospitalizations
Medical Technology, Device Complications
Technology assistance and complications are characteristics frequently associated with potentially preventable hospitalizations. In a nationally representative hospital discharge sample analyzing 365-day rehospitalizations, Berry et al23 found that only 5% of patients with no rehospitalizations had technology assistance compared with 53% who had ≥4 rehospitalizations (P < .001). Within this highest-utilizing group (ie, ≥4 rehospitalizations), technology complications were present in 8.7% of all admissions. Kun et al30 found that tracheostomy decannulation or obstruction and tracheitis were in 12% and 17% of readmissions in the year after discharge with home mechanical ventilation, respectively. Device complications were present in 8% of a randomly selected sample of 15-day readmissions of children with chronic illnesses from another center.31 Technology malfunctions were noted in 9% of admissions among patients participating in complex care programs affiliated with 4 children’s hospitals.32
Though challenging to interpret degrees of preventability with device complications, a sizable proportion may be preventable through improvements in surgical, medical, or home care. A prospective study of 248 patients, in which preventability of unscheduled ICU admissions was designated by investigators after chart review,33 revealed that 19% of admissions in the technology assisted chronic illness group were potentially preventable. Hain et al28 found that central venous catheter infections or ventricular shunt malfunctions were present in 8.5% of all 15-day readmissions and 43% of those rated potentially preventable.
Nonadherence to Recommended Hospital Follow-up
Follow-up after discharge is a frequently studied readmission predictor in adult studies; however, we found only 1 study addressing this question in our population.34 Despite no difference in having follow-up appointments at the time of discharge, lack of outpatient hematology follow-up and disease severity were significantly associated with readmission among 100 patients with sickle cell disease (R2 = 0.41, P < .001). Hematology follow-up occurred in 54% of the nonreadmitted group versus only 13% of the readmitted group (P < .001).
Several specific diagnoses appear to be more commonly associated with potentially preventable hospitalizations. In the 15-day readmissions study by Hain et al,28 readmissions after surgical hospitalizations were potentially preventable 39% of the time (significantly higher than after medical discharges [16%, P = .002]).28 National samples by Berry et al23 and Feudter et al35 identified neuromuscular and neurologic disorders as the most common CCCs associated with 365-day readmissions, and they were the second most common 15-day readmissions in the study by Gay et al.31 Among studies including ACSCs, asthma was the most common ACSC23,24,36; however, it is not known from these studies what fraction of the patients with asthma ACSC hospitalizations were medically complex.
Primary Care Experience and Continuity Across the Health System Landscape
Medical Home Characteristics, Organizational Capacity
Five studies15,25,37–39 assessed various principles from the American Academy of Pediatrics (AAP) medical home model, in particular coordination, accessibility, and family-centeredness. In a broad sample of 43 primary care practices with varying levels of success achieving the medical home principles measured by Medical Home Index (MHI) score, fewer overall hospitalizations were observed in practices with higher MHI scores (specifically higher levels of organizational capacity, data management, chronic condition management, and care coordination).37
The analysis of the MAC Medicaid Program by Gadmonski et al,25 which included 24/7 access to a medical home, revealed reduced avoidable hospitalizations after enrollment that approached statistical significance (adjusted odds ratio [AOR], 0.93; 95% confidence interval [CI]: 0.86–1.01). Similarly, a secondary data analysis of a national sample from the Medical Expenditure Panel Survey39 revealed that lower parent reports of family-centered care and realized access were associated with more hospitalizations among those with private insurance (IRR, 3.87; 95% CI: 1.23–12.13 and 3.45; 95% CI: 1.30–9.19, respectively). These associations were not present in the publicly insured group, and parent report of timeliness of care was not associated with hospitalizations in either group.
Palfrey et al15 reported the prepost evaluation results from the Pediatric Alliance for Coordinated Care (PACC) multisite pilot intervention in Massachusetts, and Gordon et al38 reported prepost findings from the Special Needs Program at the Children’s Hospital of Wisconsin. Both studied multifaceted, multidisciplinary interventions focused on coordination and partnership with particularly complex patients identified through eligibility criteria including combinations of subspecialty use, organ system involvement, technology assistance, developmental delays, past utilization, and future uncertainty or risk.
Key PACC activities included nurse practitioner involvement and home visits, consultation from a local parent of a CSHCN, modifications of office routines, individualized patient health plans, regularly scheduled continuing medical education, expedited referrals and communication with specialists/hospitals. Key Special Needs Program activities included unified care plans, coordination, and communication across inpatient and outpatient specialty and primary care. Additional activities included participation at health care visits, school, and with payers, home visits, outreach to primary care providers (PCPs) and community resources, registered nurse (RN) case management, psychosocial support, and physician availability 24/7. Both evaluations, though uncontrolled, revealed significant reductions in hospital utilization with relatively modest programmatic cost requirements.
Health Services Systems and Policies
Organization and Payer Structure
Investigators in upstate New York conducted an evaluation after a regional insurance company expanded funding for ambulatory care coordination and wraparound services (nursing, social work, psychology, occupational therapy/physical therapy, speech therapy, and special education) for children with chronic diseases.12 Over 10 years, hospitalizations for children with heterogeneous complex disorders reduced from 2796 to 1622 (R2 = 0.82, P < .001) at their hospital, without any simultaneous change observed in common acute hospitalizations (bronchiolitis, fracture, and appendicitis). The 30-day readmission rates were lower than a comparison academic hospital group (12.7% vs 15%); and overall hospitalization rates for chronic conditions were lower using appendectomy rates as a proxy denominator for the overall hospitalization rates (P < .01).
Another quasi-experiment, MAC,25 which included PCP assignment and managed care gatekeeping for emergency department (ED), inpatient and specialty care, improved reimbursement for primary care, and 24/7 access for patients in the Medicaid program, evaluated changes in hospitalization rates, ACSC hospitalizations, and an investigator-defined indicator of “avoidable hospitalizations.” The latter group was defined by prespecified combinations of ICD-9 codes with ambulatory and/or pharmacy claims after consensus from a modified-Delphi process with local experts. MAC enrollment was associated with a lower probability of avoidable (AOR, 0.89; 95% CI: 0.83–0.97), but not ACSC hospitalization (AOR, 0.96; 95% CI: 0.92–1.01) among children who used ambulatory care. After including ambulatory visit types in the final model (subspecialty, primary care, preventive, and ED), there was a nonstatistically significant reduction in avoidable hospitalizations with enrollment (AOR, 0.93; 95% CI: 0.86–1.01). Preventive visits were significantly associated with lower avoidable and ACSC hospitalizations (AOR = 0.70 and 0.83, respectively).
Beyond Hospital and Clinic Boundaries
Home Visitation and Community Resources
Both interventions by Gordon et al38 and Palfrey et al15 (described earlier) included home visitation and connection to community resources as key components. Neither was controlled nor designed to identify the specific program activities accounting for outcomes, so it remains unclear if these activities, specifically, were responsible for reduced hospitalizations.
Environmental Exposures and Triggers
In a study by Dosa et al,33 noncompliance, inappropriate supervision, and passive smoke exposure preceded 18% of all ICU admissions designated as potentially preventable and 8% of those for the children with technology assistance. When these investigators categorized the events leading to an admission, potentially preventable family and environmental factors accounted for 43% of all events leading to an admission.
Demographics: Increased Vulnerability, Less Modifiability
Public insurance,23,24,35 age,23,25,31,35,36 and nonwhite race/ethnicity23–25,35 were consistently associated with higher potentially preventable hospitalizations when included in analyses. Despite being largely unmodifiable, recognizing the increased vulnerability these characteristics introduce is important.
Based on the results from all included studies, a summary of key drivers of potentially preventable hospitalizations are presented in Table 4.
Strength of the Evidence
Most of the studies were rated as having at least medium risk of bias due to design flaws (Tables 5 and 6). Four studies used quasi-experimental designs, none of which were randomized. Three studies were uncontrolled prepost designs, and 1 was a natural experiment with a nonequivalent comparison group. Among the observational studies, though many attempted to account for potential confounders such as demographics and severity of illness, most were missing important variables: family income and employment, family and household structure, caregiver education, caregiver or patient health literacy, self-efficacy, mental health, adherence, community characteristics and resources, as well as distance to care, among others. Because several of the studies were small pilot or single-center studies, generalizability of findings to different health care settings or patient populations is also limited.
Evidence from this review suggests that hospitalizations can be reduced in CMC. Because these children account for as much as one-third of child health expenditures,8,10 and because hospital care may account for 80% of spending in this group,9 these findings are important. Progress in this field has the potential to lead to significant cost savings, higher care quality, and improved patient experience in a particularly vulnerable cohort. These findings lay the foundation for additional research by identifying candidate drivers, or root causes, of potentially preventable hospitalizations upon which interventions can be designed and tested.
Both home and immediate postdischarge environments appear to be high-yield focal points given their relationship to detrimental factors including adherence problems, inadequate technology use and maintenance, family stress, passive smoke and other environmental exposures, and insufficient follow-up or information transfer. Research with chronically ill adults suggests that increased patient activation (ie, the skills, knowledge, confidence and motivation to maintain one’s health and health care) is associated with less ED and hospital use as well as fewer readmissions.40–42 If these findings translate to caregivers of CMC, testing interventions designed to increase caregiver activation and improve care at home may be effective. Additionally, telehealth interventions with nurse chronic illness management have reduced ED use and preventable hospitalizations in elderly and Veterans Affairs patients,43,44 and such approaches may be germane to CMC receiving care at home.
High-quality hospital-to-home transitions practices, such as the Care Transitions Intervention developed for the geriatric population, might also be adaptable to families of CMC. Controlled studies of this intervention have revealed that combining postdischarge home visitation, self-care, and health coaching activities for patients and caregivers has led to improvements in caregiver activation and self-management with reductions in cost and readmissions in the months after discharge.45–48 Promising work that builds off the community health worker model49 and nurse case management programs50 have also led to reduced utilization after discharge. Refining and testing optimal home visit logistics, scope, and content as well as postdischarge activities are important next steps with the CMC population.
Several additional strategies should be sought to reduce CMC hospital use. Building on the findings of Cooley et al,37 Gordon et al,38 and Palfrey et al,15 effective models for written and verbal communication spanning across the inpatient-outpatient-community continuum will likely lead to reductions in hospital use. A landmark review by Kripalani et al51 highlighted the prevalence of deficits in communication transfer between hospital-based and primary care physicians, noting negative effects on quality of care in 25% of follow-up visits and possibly increased readmissions. Efforts to eliminate complications from medical devices and technologies, and efforts to anticipate, identify, and mitigate postsurgical complications also appear promising.
Practice-level redesign with an emphasis on chronic care management, organizational capacity, population-based management, and team-based care may also be particularly successful. Involving parents as stakeholders in these processes is necessary, and involving experts in redesign or chronic care management may be prudent for practices lacking these skills. Achieving success as a patient- and family-centered medical home may have potential for reducing hospitalizations. Finally, advocating or collaborating with payers to redesign payment structures for this population may be an effective approach at the health system level.
This review also underscores the need to better define preventable hospitalizations. Increasing numbers of studies demonstrate the challenges of interpreting readmission rates,27,52,53 including poor reliability in identifying preventable readmissions.28 To date, separation of preventable readmissions from all readmissions cannot be accomplished by using administrative data alone. ACSC hospitalizations may be as difficult to interpret in this population. A study reviewing physician and parent perspectives on the preventability of ACSCs in a sample of consecutively enrolled patients to a children’s hospital revealed that only 13% to 46% of hospitalizations for ACSCs were thought to be avoidable,54 and the authors did not look specifically at a medically complex cohort. Findings from our work suggest that ACSC hospitalizations may be less common for CMC than non-CMC.
A number of limitations should be considered. In addition to a relatively small number of published studies on this topic, the consistency and quality of this body of evidence is the major limitation of the review. The variability in patient populations, indicators of preventable hospitalization, and outcome measures limits synthesis of findings. The experimental designs are uncontrolled pilot studies or quasi-experimental nonequivalent comparisons. As such, drawing conclusions about causality from any of the interventions or exposures to changes in hospital use is challenging. We only included published, peer-reviewed publications, and it is possible that unpublished interventions exist. We were unable to formally assess risk of bias across studies such as through construction of a funnel plot, due to the limited number of included studies and heterogeneity of outcomes. Publication bias may have prevented intervention evaluations with negative findings from entering peer-reviewed journals. Finally, because of CMC definition heterogeneity, it is possible that ultimately preventing hospitalizations may require different approaches depending on the CMC population captured by the definition.
Nevertheless, while awaiting higher quality studies to be completed in the future, providers and health systems will continue to try to reduce hospital utilization in this population; and as such, developing strategies from the existing evidence is pragmatic. Given the relatively small number of published studies to date, intervention development might be aided by augmenting these findings with expert panel approaches such as the modified-Delphi Rand/UCLA Appropriateness Method.55
The call to reduce health care costs combined with the reality that expenditures and hospital use are disproportionately high among CMC means that reducing hospital utilization in this population is an important target for every health system. Research progress is challenged by relatively small patient numbers and heterogeneous diagnoses within single centers. Innovation and testing of new models that are informed by this study can be accomplished today, however, through opportunities offered by primary care research networks, quality improvement collaboratives, and partnership among the increasing number of complex care programs.9,32 This systematic review of the literature has identified several modifiable drivers of preventable hospitalization in CMC. Efforts targeting these factors specifically may lead to reductions in hospital use in this medically complex population.
- Accepted September 10, 2014.
- Address correspondence to Ryan J. Coller, MD, MPH, Department of Pediatrics, University of Wisconsin, Madison, 600 Highland Ave, Madison, WI 53792. E-mail:
Dr Coller conceptualized and designed the study, conducted primary data analysis, and drafted the initial manuscript; Dr Nelson assisted with data collection and analysis (resolved disagreements between primary reviewers) and reviewed and revised the manuscript; Dr Sklansky assisted with risk of bias data collection and analysis and reviewed the manuscript; Ms Saenz coordinated and supervised data collection, assisted with data analysis, and critically reviewed the manuscript; Drs Klitzner and Lerner assisted with project conceptualization and reviewed and revised the manuscript; Dr Chung contributed significantly to conceptualization, methodological supervision, technical oversight, and critically reviewed and edited earlier drafts; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This study was supported by grant R40MC25677 Maternal and Child Health Research Program, Maternal and Child Health Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Burke RT,
- Alverson B
- Wise PH
- Newacheck PW,
- Rising JP,
- Kim SE
- Cohen E,
- Kuo DZ,
- Agrawal R,
- et al
- Simon TD,
- Berry J,
- Feudtner C,
- et al
- Berry JG,
- Agrawal RK,
- Cohen E,
- Kuo DZ
- Viswanathan M,
- Berkman ND,
- Dryden DM,
- Hartling L
- Downs SH,
- Black N
- Feudtner C,
- Christakis DA,
- Connell FA
- Feudtner C,
- Christakis DA,
- Zimmerman FJ,
- Muldoon JH,
- Neff JM,
- Koepsell TD
- Alverson BK,
- O’Callaghan J
- ↵Gay JC, Hain PD, Grantham JA, Saville BR. Epidemiology of 15-day readmissions to a children’s hospital. Pediatrics. 2011;127(6). Available at: www.pediatrics.org/cgi/content/full/127/6/e1505
- Dosa NP,
- Boeing NM,
- Ms N,
- Kanter RK
- Feudtner C,
- Levin JE,
- Srivastava R,
- et al
- Todd J,
- Armon C,
- Griggs A,
- Poole S,
- Berman S
- Cooley WC,
- McAllister JW,
- Sherrieb K,
- Kuhlthau K
- Hibbard JH,
- Greene J
- Kind AJ,
- Jensen L,
- Barczi S,
- et al
- Kripalani S,
- LeFevre F,
- Phillips CO,
- Williams MV,
- Basaviah P,
- Baker DW
- ↵Feudtner C, Pati S, Goodman DM, et al. State-level child health system performance and the likelihood of readmission to children's hospitals. J Pediatr. 2010;157(1):98–102
- Flores G,
- Abreu M,
- Chaisson CE,
- Sun D
- Fitch K,
- Bernstein SJ,
- Aguilar MD,
- et al
- Copyright © 2014 by the American Academy of Pediatrics