November 2015, VOLUME136 /ISSUE 5

Interpreting Variability in the Health Care Utilization of Children With Medical Complexity

  1. Joanna Thomson, MD, MPHa,b and
  2. Samir S. Shah, MD, MSCEa,b,c
  1. aDivisions of Hospital Medicine and
  2. cInfectious Diseases, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio; and
  3. bDepartment of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio
  • Abbreviations:
    complex chronic condition
    children with medical complexity
  • The prevalence of children with medical complexity (CMC) has increased in recent decades.1,2 This increase is attributable, in part, to advances in medical care that improved survival of children with chronic, congenital, or critical illnesses. These children account for a disproportionate amount of inpatient utilization. Among 6.9 million pediatric discharges in 2006, just 10.1% of pediatric admissions included children with complex chronic conditions (CCC)3; however, these admissions accounted for 26.1% of pediatric hospital days and 40.6% of pediatric hospital charges.3 Few data, however, examine variability of utilization across health care systems.

    In this issue of Pediatrics, Ralston et al,4 by using an all payer claims database, examine variation in health care utilization in a population-based cohort of CMC across 4 children’s hospitals in New England. This cohort was defined based on revised CCC discharge diagnosis codes3 further refined to minimize bias by excluding children with disease processes necessitating care at 1 specific hospital (eg, bone marrow transplantation). Claims data provided information on utilization across the continuum of care. Health care encounters, imaging studies, and diagnostic testing were compared among hospitals. The entire utilization of each patient was attributed to the hospital in which the patient spent most inpatient days or the hospital of the provider hospital service area5 in which the patient had the most number of outpatient visits.

    Ralston and colleagues4 found high variation in rates of encounters, imaging, and diagnostics among the 4 hospitals: twofold variation in inpatient and intensive care days, threefold variation in head MRI, and fivefold variation in electrocardiography. Key differences in patient characteristics across the hospitals were adjusted for, notably median household income by zip code, percent Medicaid, and CCC diagnostic categorizations. This study highlights the substantial utilization by CMC; even children receiving care at the lowest utilizing hospital experienced over 400 inpatient days, 94 emergency department visits, and 917 office visits per 100 person-years.

    The study should be interpreted in the context of several limitations. The population of CMC is inherently variable. Utilization was not examined by condition- or illness-specific processes and did not account for illness severity or location of testing (inpatient versus outpatient). Therefore, it is not possible to understand which aspects of utilization represent effective and necessary care. Additionally, socioeconomic status is a potentially important driver of health care utilization, especially given the impact of CMC on families’ finances.68 Adjusting for only Medicaid incompletely accounts for socioeconomic status, as children in this study may receive Medicaid because of their degree of medical complexity rather than their family’s economic circumstances.9 Additionally, although the use of household income classified by zip code is a valid ecological measure, zip code socioeconomic heterogeneity may limit its utility.10 Census tract measures of poverty or income may better account for socioeconomic differences when patient-level measures are not available.

    So, how can we use the results of this study to advance the field? Unwarranted variation, defined as variation due to differences in health system performance, is the target of standardization and improvement in health care.10 Unwarranted variability in utilization has been demonstrated for many pediatric medical and surgical conditions.5,1114 It is likely that unwarranted variability is exaggerated in the care of CMC. With limited evidence and lack of consensus to guide practice, clinicians make medical decisions based on anecdotal personal experiences, engrained practice cultures, parental preferences, and their biases to intervene in the context of uncertainty.10,15,16 Subsequent studies should help us interpret unwarranted variability within condition- and disease-specific context to further our understanding of best practices.

    Although utilization as a measure of system performance provides a powerful examination of health care costs, complementary work is needed to determine care practices of value. Highlighting variation or canonizing low users should not be our primary goal. Disease-specific outcomes, when they are evidence-based and measurable, are more important than utilization patterns. For example, the Cystic Fibrosis Foundation tracks BMI and forced expiratory volume as indicators of health in patients with cystic fibrosis, with lesser importance placed on utilization.17 Unfortunately, given the heterogeneity of conditions that lead to medical complexity and the lack of evidence for many conditions, disease-specific outcomes might not always be possible to measure.

    There is promise in the evolution of outpatient and inpatient complex care clinical programs. Comprehensive and coordinated care programs for CMC have been associated with improved utilization outcomes; including decreased emergency department visits, decreased hospital days, and decreased costs.1821 Further study of outcomes reflecting quality of care (eg, care coordination) and/or patient- or family-centered priorities (eg, quality of life) may provide complementary data to utilization patterns.22

    In conclusion, variation in the utilization patterns for CMC likely indicates inefficiency and overuse. Further understanding the drivers of variation and identifying best practices will provide targets of focus as we strive to improve the health system caring for this important population.


      • Accepted August 4, 2015.
    • Address correspondence to Samir S. Shah, MD, 3333 Burnet Ave, ML 9016, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229. E-mail: samir.shah{at}
    • Opinions expressed in these commentaries are those of the author and not necessarily those of the American Academy of Pediatrics or its Committees.

    • FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

    • FUNDING: No external funding.

    • POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

    • COMPANION PAPER: A companion to this article can be found on pages 860, and online at


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.
    16. 16.
    17. 17.
    18. 18.
    19. 19.
    20. 20.
    21. 21.
    22. 22.