Health care professionals generally recognize that a percentage of hospital readmissions are avoidable and undesirable as a result [1].  Avoidable readmissions (ARs) are undesirable because they can be costly, may reduce patient safety, and because they can create the perception of a lower quality of care (QC) level provided by a medical institution.  Avoidable readmissions have become important as well because policymakers, in response to provisions of the Affordable Care Act (ACA), are implementing incentive programs to reward or penalize hospitals on the basis of, among other criteria, readmission rates [2-6]. Moreover, ARs can be difficult to judge and predict, and they vary significantly according to hospital setting [7].    Avoidable readmissions in a pediatric hospital setting (hereafter: APRs) are unique in that they tend to be comparatively low in occurrence, and this in a setting where readmissions due to complex chronic conditions (CCCs) are common [8].

The purposes of this article are to (1) provide a brief summary of findings about APRs based on an informal review of the small amount of pediatric readmissions literature; (2) offer a commentary which takes a stance on the appropriateness of ACA penalty/reward incentives pending for children’s hospitals; and, (3) suggest a plan of action to help clarify the relationship between pediatric readmissions and care quality.

2. Informal Review of Pediatric Readmissions Literature:

Article Study Types

Nineteen articles published between 2011 and 2014 were reviewed informally as a basis for this commentary.  Thirteen were retrospective cohort analyses (RCs) [7-19], one an observational cohort [20], one a quantitative/qualitative intervention study [21], and four were reviews [22-25].  All articles bear directly upon pediatric readmissions in children’s hospitals, but do not include pediatric intensive care units (PICUs).  Four of the retrospective cohort analyses focused on overall characteristics of pediatric readmissions such as definitions of avoidable readmissions, risk factors, and statistical analyses of occurrence variations [7-9,15].  Four articles were related to coordinated care for pediatric patients with complex chronic conditions (CCCs) [10-12,16].  Five articles considered readmission rates relative to specific medical conditions [13,17-20].  Two focused on readmission preventability and reduction [8,22].  Two articles investigated readmission predictability [12,14].  One specifically evaluated the connection between APRs and perceptions of CQ [25].

Key Findings

• Based on the 19 articles analyzed in this literature review, APRs are:

• sually hard to define and identify (~40%); easier in cases of chronic illness [7,17]

• Variable in frequency across hospital settings [7,10,17,25] with limited predictability [20] Multifactorally determined according to:

• Medical condition [10-13,16-20]

• Care coordination characteristics9 [9,11,18]

• Non-medical variables [13,22]

• Best countered when multi-leveled, integrated interventions are applied [9,11,24], but, may increase in occurrence in high-quality care settings [8]. 

Overall, dependable conclusions about APRs are not easily achieved.  This is partially because such studies are relatively few.  The majority of completed studies indicate that APRs are multifactorally determined, and that they vary in frequency according to hospital setting and attendant medical and non-medical conditions.  But these indicated APR trends are conclusions based on a wide and inconsistent range of events that counted as “readmissions.”  Furthermore, while this review reveals possible correlations between, for example, certain medical and non-medical conditions increasing readmission rates, the review results do not clearly manifest criteria for judging which of these readmissions were, in fact, avoidable.

Time frames in which to count readmissions are not at all standardized.  In the 13 RCs in this review, time frames ranged from 15 days [8,15,20], to 30 days [7,11,12,16,19,20], to any 365-day period in a five-year frame [9].  Explicit rationales for time frame lengths were absent. Almost 70% of the RC studies (9 of 13 in this review), nonetheless, had APR time frames of either 15 days or 30 days.  Longer time frames occurred in studies involving chronic conditions [9], the presence of medical technology (ventilator) [17], and a multi-center predictability study [14]. At the moment, it is unclear how readmissions time-frame definitions affect study results.  

What counts or should count as an APR as well as APR rate variations and predictability are not easily established primarily because APRs are determined by a dynamic of medical factors (i.e., patient conditions) [9-13,16-20,22], hospital care coordination characteristics [9,11,18], and non-medical variables [13,22].  Only one study in this review investigated whether consensus defining an APR could be reached by care providers.  Consensus was reached ~40% of the time, but this included a higher percentage of consensuses when CMC or CCC conditions existed [17]. 

Frequency rates for APRs vary, and they tend to thwart prediction.  Based on data from 72 acute care pediatric hospitals, the 30-day unadjusted readmission rate for all hospitalized children was 6.5% with indications that this percentage varied significantly, especially because of chronic medical conditions, sometimes because of hospital setting (e.g., the hospital’s general readmission rate), and sometimes due to non-medical conditions [7].   Data about predicting APRs is inconclusive; one study by Feudtner et al. at a pediatric hospital addressed the idea of predicting readmissions and concluded that accurate prediction of readmission rates is possible when involving CCCs [14].  According to the authors, the information most important to forecasting readmission rates is the most recent hospital admissions data and the history of hospitalizations during the preceding 12 months.

Many studies corroborated that chronic medical conditions typically increase readmissions [7,11,13,20,22-24].  In a study covering 38 children’s hospitals in the U.S. focusing on CCC patients, a 16.7% readmissions rate was found [14].  This is ~10% percent higher than the pediatric mean estimated at 6.5% for children without CCC/CMC [7].  Two studies investigating chronic and non-chronic patients found higher readmission rates in the former [15,18].  It is still unclear, however, on the basis of the studies in this review, exactly how to judge which of the readmissions attendant with chronic conditions are avoidable. 

The same ambiguity holds true for other medical factors as well as non-medical factors.  For example, the number of discharge medications in CCC patients is associated with increased odds of readmission.  Patients with a cutoff of 8 medications were most associated with readmission.  Patients with >8 medications had a readmission rate of 29%, or 10% higher than the 19% rate found as the mean for patients with at least one CCC [16]. Patients needing mechanical devices are more apt to readmit.  A study of pediatric patients on home mechanical ventilators (HMVs) showed a 12-month non-elective readmission rate of 40% [17]. Differences in health care coverage and community health service availability were key to asthma patient readmission rates [13].

Many studies recognized the importance of care coordination for reducing pediatric readmissions [9,11,18,22].  Two studies with a CMC patient population focused on ideas of integrated care systems.  One of these showed the benefits of what authors called “Dedicated Places”-care coordination center points for involved caregivers [10].  Another concluded that what authors called “integrated clinics” could provide complex care in community-based settings with less direct tertiary care involvement [21].   One article focusing on pediatric patients with CCCs found that effective interventions should identify an individual or team to 1) design the inpatient-to-outpatient transition, and 2) offer telephone and home visitation support to families [16].  Family care is considered a major part of such coordinated care [17].  Confirmed that full implementation of the Asthma Care Process Model (CPM-CAC3) was associated with a significant (though delayed) decrease in readmissions [13]. One article found that among pediatric patients with non-complex conditions, higher nurse-to-patient ratios reduced the occurrence of avoidable readmissions [20]. It is worth noting, however, that a study at UCLA’s Mattel Children’s Hospital found that primary care physician follow-up may be related to higher, not lower, readmission rates [12].  This may be due to more focused care leading to greater recognition of medical issues needing hospital attention.

4. Avoidable Readmissions as Measures of Quality of Care

Quality care assessment and reductions in costs of care have been key motives for carrying out studies shedding light on ways to better identify avoidable readmissions and for ways of reducing their frequency.  For example, Berry et al. (2013) named quality of care perceptions as the contextual focal point for their study on variations in preventable readmissions rates in U.S [7].   On the other hand, the AHA (2011) report maintains that there may not be very good reason to do so [2], and the results of this review tend to corroborate the AHA report, especially in the case of pediatric hospitals.  First, readmission rates in general may not be clear measures of quality of care until what counts as an avoidable readmission can be better defined. The results of this review suggest that has yet to occur.  Second, there is some indication that hospitals with proven high-quality care see increased rates of readmission precisely because of the increased attention from caregivers.  One might expect this in a “high-quality” hospital setting.  Third, readmission rates in pediatric settings are so low that they cannot function as a dependable quality-of-care criterion.  Hain et al. conclude that random variation could account for rate differences when rates of readmissions are so low (e.g., typically 2%) [8].  (NOTE:  A review by Srivastava and Keren [25] that is not included as an article in this review, analyzes the Berry et al. results and agrees that variation in rates may be due to defects in quality.  They suggest, however, that a competing explanation involves factors particular to children and their readmissions to hospitals.  They maintain that a model should be devised that includes factors beyond the hospital itself such as parents’ health status or influences of parent-children relationships which can change over time.  In the meantime, the work of Berry et al. is a good starting point for more research.  Coller et al. also concluded that readmission rates are not a clear quality measure [12]). 

5. Commentary: Implications to the Key Findings

Government plans to implement penalty/reward incentive programs aimed at children’s hospitals seem poorly founded on many grounds.  First, if there is little consensus as to the very definition of an APR, it seems that application of rewards and penalties cannot be more than arbitrary.  Second, the assumption that APRs are related significantly to actual QC-and not just perceptions thereof-is not clearly defensible.  In a recent commentary, Alverson and O’Callaghan [26] conclude that there is only a tenuous connection at best between QC and APRs.  A major reason for this is the recognition that many factors that affect APR rates are beyond the control of the hospitals.  Gildemeister [5] is right to point out that care must be taken to make sure that risk adjustment performance measures avoid masking true instances of QC failure.  He goes slightly further, however, to suggest that looking at non-medical risk factors may get in the way of understanding instances of poor quality care.

This suggests a fundamental conflict of interest existing between the ACA and government reward/penalty programs on the one hand and children’s hospitals on the other.  Hospitals seem intent on finding more conclusive data about the effects of factors beyond hospital control on APRs; the government seems willing to proceed with implementing reward/penalty programs without such findings.  Holding such a stance too staunchly is cynical; it may also be unfair because if it underestimates the importance of factors out of a hospital’s control relative to APRs, then hospitals will bear an unfair burden.  Moreover, if such a policy were to continue over time, it is possible that investigations into extra-hospital factors may diminish because of bureaucratic inertia.  Unfairness to children’s hospitals would be “set in stone,” as it were.

More studies into all facets of pediatric readmissions are advisable.  Nearly all the articles in this review explicitly mentioned the need for further research into whatever particular facet of pediatric readmissions upon which they focused.  Whether on reduction in avoidable readmissions; on what factors cause avoidable readmissions; on what strategies are best for reducing the instances of such readmissions; or, on the relation to quality of care, the tendency of findings was, while not chaotic, toward complexity rather than resolution at this time.

Also, CMC and CCC patients are consistently shown to have the highest rates of readmission.  Consequently, analyzing current coordinated care strategies and designing and testing new ones specifically for these populations should perhaps be among the first orders of business, with one caveat:  there are indications that some proven elements of successful care coordination (e.g., making certain that discharged patients have timely primary care follow-up) have been shown, counter-intuitively, to be correlated with higher rates of readmission rather than lower rates.  The logic is thus:  with greater care coordination (an indicator of better care quality), there is likely to be greater recognition of treatment needs and more recommendations to address them, sometimes through readmission (a purported indicator of lesser care quality). 

 
Third, and perhaps of most immediate importance, policymakers should be actively dissuaded from implementing further-or urged to repeal current-incentive/disincentive structures built around the assumption that readmission rates are good indicators of care quality.  This is most evident, and most potentially unfair, where readmission rates are low in the first place (e.g., in pediatric settings).  Burstin suggests [5] two approaches to avoid unfairness: (1) pay providers not according to the rate of APRs but on improved health of patients over time; and, (2) group hospitals according to type before performing comparisons of APR rates.  In any event a focus on the connection of APRs and CQ seems misguided, and the preferred focus should be on better understanding the complex of factors creating higher (and lower) rates of readmission.  Unfair penalties to children’s hospitals could turn into a demoralizing disincentive.  Incentives that help to improve knowledge about APRs would be neither unfair nor demoralizing.

1. Vest JR, Gamm LD, Oxford BA, Gonzalez MI, Slawson KM et al. (2013) Determinants of Preventable Readmissions in the United States: A Systematic Review. Implement Sci 5: 1-28.

2. Examining the Drivers of Readmissions and Reducing Unnecessary Readmissions for Better Patient Care.(2011) American Hospital Association (AHA) 1-11.

3. Mittler JN, Hora OJL, Harvey JB, Press MJ, Volpp KG, et al. (2013) Turning Readmission Reduction Policies into Results: Some Lessons from a Multistate Initiative to Reduce Readmissions. Popul Health Manag 16: 255-260.

4. Sills MR (2016) Social Factors Impact Children’s Hospitals Readmissions (Interview). MedicalResearch.com.

5. Benson L (2016) Study: Health Care Providers Penalized for Patient Hardships. MPRNews.

6. News -Med (2016) Hospitals Serving Children May Face Financial Penalties for Readmissions, Researchers Find. NewsMedical Life Sciences & Medicine.

7. Berry JG, Toomey SL, Zaslavsky AM, (2013) Pediatric Readmission Prevalence and Variability across Hospitals. JAMA 309: 372-380.

8. Hain PD, Gay JC, Berutti TW, Whitney GM, Wang W, (2013) Preventability of Early Readmissions at a Children’s Hospital. Pediatr 181: e171-e181.

9. Berry JG, Hall DE, Kuo DZ, (2011) Hospital Utilization and Characteristics of Patients Experiencing Recurrent Readmissions within Children’s Hospitals. JAMA 305: 682-690.

10. Cady RG, Kelly AM, Finkelstein SM, Looman WS, Garwick  AW (2013) Attributes of Advanced Practice Registered Nurse Care Coordination for Children with Medical Complexity. J Ped Health Care 4: 1-8.

11. Cohen E, Lacombe-Duncan A, Spalding K (2012) Integrated Complex Care Coordination for Children with Medical Complexity: A Mixed-Methods Evaluation of Tertiary Care- Community Collaboration. BMC Health Serv Res 12: 1-11.

12. Coller RJ, Klitzner TS, Lerner CF, Chung PJ (2013) Predictors of 30-day Readmission and Association with Primary Care Follow-Up Plans. J Pediatr 163: 1027-1033.

13. Fassl BA, Nkoy FL, Stone BL, (2012) The Joint Commissions Children’s Asthma Care Quality Measures and an Asthma Readmissions. Pediatr 130: 482-491.

14. Feudtner C, Levin JE, Srivastava R (2009) How Well Can Hospital Readmission Be Predicted in a Cohort of Hospitalized Children? A Retrospective Multi-Center  Study.Pediatr 123: 286-293.

15. Gay JC, Hain PD, Grantham JA, Saville BR (2011) Epidemiology of 15-day Readmissions to a Children’s Hospital. Pediatr 127: e1505-e1512.

16. Jurgens V, Spaeder MC, Pavuluri P, Waldman  Z (2014) Hospital Readmission in Children with Complex Chronic Conditions Discharged from Subacute Care. Hosp Pediatr 4: 153-158.

17. Kun SS, Edwards JD, Ward SL, Keens TG (2012) Hospital Readmissions for Newly Discharged Pediatric Home Mechanical Ventilation Patients. Pediatr Pulmonol 47: 409-414.

18. Liu SYL, Pearlman DN (2009) Hospital Readmissions for Childhood Asthma: The Role of Individual and Neighborhood Factors. Public Health Rep 124: 65-78.

19. Neuman MI, Hall M, Gay JC (2014) Readmissions among Children Previously Hospitalized with Pneumonia. Pediatr 134: 100-109.

20. Tubbs-Cooley HL, Cimiotti JP, Silber JH, Sloane DM, Aiken LH (2012) An Observational Study of Nurse Staffing Ratios and Hospital Readmission among Children Admitted for Common Conditions. BMJ Qual Saf 22: 735-742.

21. Cohen E, Berry JG, Camacho X, Anderson G, Wodchis W, et al. (2012) Patterns and Costs of Health Care Use of Children with Medical Complexity. Pediatr 130: e1463-1470.

22. Auger KA, Kenyon CC, Feudtner C, Davis MM (2014) Pediatric Hospital Discharge Interventions to Reduce Subsequent Utilization: A Systematic Review. J Hosp Med 9: 251-260.

23. Chen AY, Schrager SM, Mangione-Smith R (2012) Quality Measures for Primary Care of Complex Pediatric Patients. Pediatr 129: 433-445.

24. Miller M (2013) Roles for Children’s Hospitals in Pediatric Collaborative Improvement Networks. Pediatr 131: S-215-S-218.

25. Srivistava R, Keren R (2013) Pediatric Readmissions as a Hospital Quality Measure. JAMA 309:396-398.

26. Alverson BKO, Callaghan J (2013) Hospital Readmission: Quality Indicator or Statistical Inevitability? Pediatr 132: 569.