Outpatient Visits Before Ambulatory Care–Sensitive Hospitalization of Children Receiving Medicaid

Accepted Manuscript Title: Outpatient Visits before Ambulatory Care Sensitive Hospitalization of Children Using Medicaid Author: Neal A. deJong, Troy Richardson, Nicole Chandler, Michael J. Steiner, Matt Hall, Jay Berry PII: DOI: Reference:

S1876-2859(17)30495-3 https://doi.org/doi:10.1016/j.acap.2017.09.015 ACAP 1101

To appear in:

Academic Pediatrics

Received date: Revised date: Accepted date:

28-3-2017 8-9-2017 15-9-2017

Please cite this article as: Neal A. deJong, Troy Richardson, Nicole Chandler, Michael J. Steiner, Matt Hall, Jay Berry, Outpatient Visits before Ambulatory Care Sensitive Hospitalization of Children Using Medicaid, Academic Pediatrics (2017), https://doi.org/doi:10.1016/j.acap.2017.09.015. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Outpatient visits before ambulatory care sensitive hospitalization of children using Medicaid

Neal A deJong, MD, MPH,a Troy Richardson, PhD,b Nicole Chandler, MD,a Michael J. Steiner, MD, MPH,a Matt Hall, PhD,b Jay Berry, MD, MPHc Author affiliations: a. UNC Chapel Hill, General Pediatrics and Adolescent Medicine, 231 MacNider Hall, CB#7225, 301B S. Columbia Street, Chapel Hill, NC 27599 b. Children’s Hospital Association, 16011 College Blvd., Suite 250, Lenexa, KS 66219 c. Boston Children’s Hospital, 300 Longwood Ave., Boston, MA 02115 Address correspondence to: Neal deJong, MD, MPH; [email protected], phone 919-9662504, fax 919-966-3852 Running Title: Visits before ambulatory care sensitive hospitalization Key Words: Ambulatory Care Sensitive Conditions; non-Ambulatory Care Sensitive Conditions Word Counts: Abstract 250, Article 3,286 Funding Sources: Dr. deJong’s time was supported by a Health Resources and Service Administration (HRSA) National Research Service Award Grant T32 HP14001. HRSA played no role in the study. The Children’s Hospital Association purchased the rights to the data used in this study and allowed its use for the study at no cost. Potential Conflicts of Interest: None of the authors has any conflicts of interest to disclose. Abbreviations: ACSC (Ambulatory Care Sensitive Condition); OR (odds ratio); CI (confidence interval); CCC (complex chronic condition); ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification)

Abstract Objectives Hospitalizations for ambulatory care sensitive conditions (ACSC) are measured to indicate healthcare system quality, with the premise that fewer hospitalizations would occur with better preceding outpatient care. Our objective was to identify outpatient care received in the 7 days

Page 1 of 23

2

preceding acute pediatric hospitalizations, and to compare receipt of outpatient care by hospitalization type (ACSC vs. non-ACSC). Methods This was a retrospective, observational study using a 10-state database of Medicaid claims to identify outpatient visits within 7 days before acute, unplanned hospitalization for children ages 0-17 years. We used logistic regression to assess the relationship between hospitalization type and occurrence of a preceding outpatient clinic visit, controlling for patient age, race/ethnicity, type of Medicaid, and complex chronic conditions. Results Of 254,902 hospitalizations, 28.6% had a preceding outpatient visit. Thirty-five percent of hospitalizations were for ACSC. A greater percentage of ACSC vs. non-ACSC hospitalizations had a preceding outpatient visit (31.1% vs. 27.3%, p<0.001). In multivariable analysis, characteristics associated with a preceding outpatient visit were age <1 vs. 13-17 years [aOR 2.4 (95% CI 2.3-2.5)], ≥2 vs. 0 complex chronic conditions [aOR 1.9 (95% CI 1.8, 2.0)], Medicaid managed care vs. fee-for-service [aOR 1.2 (95% CI 1.2-1.3)], and ACSC vs. non-ACSC hospitalization [aOR 1.2 (95% CI 1.1-1.2)].

Conclusions Although receipt of outpatient care was modestly higher in children hospitalized with an ACSC, most hospitalized children did not receive preceding outpatient care. Further investigation is needed to assess why such a large proportion of children do not receive outpatient care before acute, unplanned hospitalization, especially for ACSC.

Page 2 of 23

3

What’s New: This study highlights potential missed opportunities to avert ACSC hospitalizations. The majority of children using Medicaid and hospitalized for an ambulatory care sensitive condition do not receive outpatient or emergency department care during the week before admission.

Authorship Statement: Conceptualized and designed the study: deJong, Richardson, Chandler, Steiner, Hall, Berry Coordinated and supervised data acquisition, cleaning, and analysis: Richardson, Hall, Berry Analysis and interpretation of data: deJong, Richardson, Chandler, Steiner, Hall, Berry Drafted the manuscript: deJong, Chandler, Berry Revised the manuscript: deJong, Richardson, Chandler, Steiner, Hall, Berry Approved the final manuscript: deJong, Richardson, Chandler, Steiner, Hall, Berry

Introduction Efforts to contain the use and cost of pediatric health services have focused on potentially preventable hospitalizations. These include hospitalizations for ambulatory care sensitive conditions (ACSC), which are “conditions for which good outpatient care can potentially prevent the need for hospitalization, or for which early intervention can prevent complications or more

Page 3 of 23

4

severe disease.”1 ACSC such as asthma exacerbations and dehydration from gastroenteritis are among the most common reasons children are hospitalized.2 Estimated costs for these hospitalizations are $4 billion annually.3 ACSC hospitalizations were conceptualized for quality measurement based on the premise that some might be avoided through timely and effective preceding outpatient care. However, little is known about how often children receive outpatient care before an ACSC hospitalization and how that frequency compares with hospitalizations for other reasons not considered ambulatory care sensitive. This information is important to understand, especially because ACSC hospitalizations are perceived to reflect health system performance leading up to admission. Two prior studies assessing primary care use before children’s hospitalizations – one in the 1980s in the northeastern U.S.4 and the other in the 1990s in Spain5 – found that about 50% of children had clinic visits in the days leading up to hospitalization. It is possible that advances in primary care for children in recent decades (e.g., patient-centered medical home) have led to greater use of outpatient care mitigating the risk of hospitalization for ACSC.6 Therefore, we assessed outpatient care received in the 7 days preceding unplanned, acute care hospitalizations in children, and compared receipt of outpatient care by whether the hospitalization was for an ACSC or not using a nationally representative database of children enrolled in Medicaid. Methods Study Design and Population This is a retrospective study of children ages 0-17 years using Medicaid from 10 programs in the Truven MarketScan© (Ann Arbor, MI) Database. A mix of statewide Medicaid programs and managed care organizations contribute claims data to this database, which is designed to be representative of the national, Medicaid-insured population. States are not

Page 4 of 23

5

identifiable in the database, and no other substitutes for the source of individual observations are available. Eligible children were hospitalized between January 31, 2009 and December 1, 2012 and continuously enrolled in Medicaid for at least one month before admission. For children admitted to the hospital more than once in the study period, we assessed their initial hospitalization only. We excluded hospitalizations for conditions typically not associated with lead-in time during which a child might be evaluated and treated appropriately in the outpatient setting. These hospitalizations included birth or pregnancy (Major Diagnostic Categories 14 and 157), planned procedures (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes for pediatric planned procedures endorsed by the National Quality Forum8), and malignancy (ICD-9-CM codes from prior studies9,10). Injury-related hospitalizations (Major Diagnostic Categories 21, 22, and 24) were included in the cohort because some reasons for injury-related hospitalization (e.g., child abuse) have potential to be detected and addressed through outpatient care. Because other injury-related hospitalizations (e.g., trauma from a motor vehicle accident) may not have this potential, we conducted a sensitivity analysis that excluded injury-related hospitalizations. The UNC Biomedical IRB exempted this study from IRB approval (Study #15-2818).

Outpatient Visits We identified outpatient primary care and subspecialty clinic visits using paid outpatient claims, within 7 days before each acute-care hospitalization. We also assessed emergency department (ED) visits that resulted in discharge to home within 7 days before hospitalization. ED visits that occurred the same day as hospitalization were not included because claims for

Page 5 of 23

6

those visits are subsumed into the hospitalization claims. We also did not include outpatient visits that occurred on the same day as a hospital admission because we felt that these might not allow sufficient lead-in time to address the health problem in an outpatient clinic. We assessed the reason for the outpatient visits using ICD-9-CM codes. To estimate the proportion of preceding outpatient visits that were likely unrelated to reasons for hospitalization, we identified diagnosis codes for health supervision visits and vaccinations that occurred prior to ACSC and non-ACSC hospitalizations.

Hospitalizations for Ambulatory Care Sensitive Conditions Based on prior studies of ACSC in children,1,11 ACSC hospitalizations included those with an ICD-9-CM principal diagnosis for the following conditions: acute otitis media or other upper respiratory tract infection, asthma, cellulitis, dental conditions, gastroenteritis or dehydration, pneumonia, seizures, or urinary tract infection. We did not include certain ACSC (e.g. tuberculosis) because of their rarity as a reason for hospitalization among U.S. children. The ACSC and corresponding ICD-9 codes used in this study are provided in Table 2.

Patient Characteristics Demographic characteristics included age (0-1, 2-5, 6-12, 13-18 years), gender, race/ethnicity (white, black, Hispanic, other), type of Medicaid (fee-for-service or managed care), and reason for Medicaid enrollment (disability or other). Clinical characteristics included type and number of complex chronic conditions (CCC). A CCC is “any medical condition that

Page 6 of 23

7

can be reasonably expected to last at least 12 months (unless death intervenes) and to involve either several different organ systems or 1 organ system severely enough to require specialty pediatric care and probably some period of hospitalization in a tertiary care center.”9 We identified CCC with ICD-9 codes developed by Feudtner et al.10 CCCs are associated with use of health services including outpatient and inpatient care.12,13

Statistical Analysis We conducted bivariable and multivariable analyses to assess the relationship between type of hospitalization (ACSC vs non-ACSC) and the receipt of preceding outpatient care within 7 days of admission. In bivariable analysis, a chi-square test was used to assess this relationship. We also used chi-square tests (with categorical data) and Wilcoxon rank-sum tests (with continuous data) to compare clinical and demographic characteristics between patients with vs. without a preceding outpatient visit and between patients hospitalized with vs. without an ACSC. In multivariable analysis, logistic regression was used to assess the relationship between type of hospitalization (ACSC vs. non-ACSC) and receipt of outpatient care in the 7 days before hospitalization, adjusting for patient characteristics. In the regression analysis, ACSC hospitalization was treated as a factor that, by definition, might be associated with preceding outpatient care. We chose this retrospective temporal orientation because ACSC hospitalizations are indicators of past health system performance, and because the type of hospitalization (ACSC vs. non-ACSC) is set from the nature of the acute illness, not the potential health services received in advance. Interaction terms between type of hospitalization and number of CCC were included to explore potential differences in how the presence of CCC related to receipt of

Page 7 of 23

8

outpatient care before ACSC vs. non-ACSC hospitalizations. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). P-values < 0.01 were statistically significant.

Results Study Population There were 1,077,713 eligible index hospitalizations in the study period. Of these, 822,811 (76.3%) were excluded for the following reasons: birth hospitalizations (72.4%); planned procedures (2.2%); hospitalizations for malignancy (0.9%); and lack of enrollment in Medicaid 7 days before admission (0.8%). After exclusions, 254,902 unplanned, acute-care hospitalizations were assessed. Hospitalized children 0-1 years old comprised the largest age category (31.5%), with the remainder (68.5%) divided evenly among ages 2-5, 6-12, and 13-17 years (Table 1). Fifty-five percent were male and 49.2% were non-Hispanic white. Most children were enrolled in a managed care plan (59.0%) and were eligible for Medicaid because of low family income (91.1%). Regarding complex chronic conditions (CCC), 16.8% had one and 6.5% had at least two. The most common CCCs were neurologic (6.4%), cardiovascular (6.0%), and metabolic (5.6%; Table 1).

Outpatient Care Prior to Hospitalization Twenty-nine percent of children had any outpatient visit in the 7 days before hospitalization; 19.9% were seen in primary care and 14.3% in specialty care (Table 1). The median number of days from the outpatient visit to hospital admission was 2 (IQR 1-4). Twenty-

Page 8 of 23

9

one percent of children had an ED visit resulting in discharge to home in the 7 days before hospitalization. Of the outpatient primary care and subspecialty clinic visits prior to ACSC hospitalization, 20,288 (5.2%) and 41,889 (4.2%) were for health supervision and administration of vaccinations, respectively. Fever (8.1% of visits before ACSC hospitalizations vs. 5.3% of visits before non-ACSC hospitalizations), upper respiratory tract infection (3.4% vs. 2.0%), cough (2.8% vs. 1.8%), and vomiting (2.4% vs. 1.6%) were acute illnesses commonly identified in outpatient visits before both types of hospitalizations, though all seemed to occur with relatively greater frequency before ACSC- than non-ACSC hospitalizations. Mental health diagnoses such as ADHD (1.9%) and depressive disorder (1.4%) occurred frequently in outpatient visits before non-ACSC hospitalizations. In bivariable analysis, the percentage of children receiving outpatient care before hospitalization varied significantly by all assessed demographic and clinical characteristics (Table 1). For example, children age 13-17 years had the lowest and infants <1 year the highest rates of preceding outpatient care, respectively (17.8% vs. 39.7%, p<0.001). Non-Hispanic black children had the lowest rates and Hispanic children the highest rates of preceding outpatient care (24.7 vs. 33.1%, respectively, p<0.001). Regarding number of complex chronic conditions, children with 0 CCC had the lowest and ≥2 CCC the highest rates of preceding outpatient care (26.4% vs. 45.5%, respectively, p<0.001). Regarding type of CCC, children with hematologic/immunologic or metabolic conditions had the lowest rates of outpatient care (33.3% and 33.9%, respectively), while children with gastrointestinal conditions or technology assistance had the highest rates (52.4% and 52.3%, respectively, p<0.001).

Page 9 of 23

10

Hospitalizations for Ambulatory Care Sensitive Conditions About one-third of hospitalizations were for ACSC (34.5%; Table 2). The most common conditions among 87,943 ACSC hospitalizations were asthma (24.3%), pneumonia (21.3%), and cellulitis (11.7%). The most common conditions among 166,959 non-ACSC hospitalizations were depression and psychoses (17.9%), bronchiolitis (14.3%), and appendicitis (4.6%). ACSC hospitalizations were shorter than non-ACSC hospitalizations [median 2 days, interquartile range (IQR) (1, 3) vs. 3 days (IQR 2, 5) for non-ACSC, p<0.001] and slightly less expensive [median $4,661 (IQR $2,647, $7,103) vs. $5,209 (IQR $3,115, $9,656) for non-ACSC; p<0.001]. Demographic and clinical characteristics varied significantly among children hospitalized for ACSC vs. non-ACSC (Table 3). For example, a greater percentage of ACSC vs. non-ACSC hospitalizations were for children ages 2-5 years (36.5% vs. 15.4%, p<0.001), for children with non-Hispanic black race/ethnicity (32.0% vs. 26.3%, p<0.001), and for children without a CCC (79.9% vs. 75.1%, p<0.001).

Relationship between Type of Hospitalization and Preceding Outpatient Care In bivariable analysis, children hospitalized for an ACSC vs. non-ACSC had a slightly higher percentage of a preceding outpatient visit (i.e, primary or specialty) (31.1% vs. 27.3%, p<0.001), primary care (22.4% vs. 18.6%, p<0.001), specialty care (15.0% vs. 13.9%, p<0.001), and care in an ED (22.2% vs. 20.8%, p<0.001; Table 2). In multivariable analysis, hospitalization for ACSC vs. non-ACSC was associated with slightly higher odds of having received preceding outpatient care [OR 1.2 (95% CI 1.1-1.2), p<0.001; Table 4]. Other characteristics associated with higher odds of outpatient care before hospitalization were age <1 year vs. 13-17 years [OR 2.4 (95% CI 2.3-2.5)], ≥2 vs. 0 CCC [OR 1.9 (95% CI 1.8-2.0)], 1 vs. 0

Page 10 of 23

11

CCC [OR 1.2 (95% CI 1.1-1.2)], eligibility for Medicaid because of disability vs. other reason [OR 1.6 (95% CI 1.5-1.7)], enrollment in Medicaid managed care vs. fee-for-service [OR 1.2 (95% CI 1.2-1.3)], and female gender [OR 1.1 (95% CI 1.1-1.1)]. Non-Hispanic black vs. NonHispanic white race/ethnicity was associated with lower odds of having received preceding outpatient care [OR 0.7 (95% CI 0.7-0.7)]. When we assessed interactions between type of hospitalization and number of CCC with the receipt of preceding outpatient care, children without a CCC hospitalized for an ACSC were more likely to have preceding outpatient care [OR 1.4 (95% CI 1.3-1.4); Table 4], but children with one or more CCCs were not [1 CCC: OR 1.1 (95% CI 1.0-1.1);, ≥2 CCCs: OR 1.1 (95% CI 1.0-1.2)]. Similar multivariable findings were observed when we removed injury-related hospitalizations in a sensitivity analysis.

Discussion This study’s findings suggest that less than a third of children with Medicaid receive outpatient care in the week before an ACSC hospitalization. The emergency department was a common site in which children received preceding care, though overall a higher percentage of children received preceding care in a primary care or specialty clinic. Children’s age, race and ethnicity, type of Medicaid plan, and presence of CCC had a larger effect on receipt of outpatient care prior to hospitalization than whether hospitalization was for an ACSC. Hospitals, outpatient clinicians, quality improvement specialists, and policymakers may find this information useful as they strive to understand the meaning and implications of ACSC hospitalization as a measure of healthcare system quality for children.

Page 11 of 23

12

Rates of preceding care have not been described recently for children with private insurance, who tend to have better access to a medical home and specialty care as well as fewer appointment barriers regarding transportation and work leave.15-17 The rate in the present study is lower than the 50% rate reported decades ago for U.S. children hospitalized for abdominal pain, meningitis, or asthma4 and for children in Spain hospitalized with an ACSC.5 Those studies, like ours, assessed preceding outpatient care for any reason. One study restricted measurement of outpatient care to visits for the same health problem leading to admission, and reported a lower rate of preceding care (25%).18 A recent study of Medicaid-enrolled children with type 1 diabetes hospitalized for diabetic ketoacidosis (DKA) found that 56% had an outpatient endocrinology visit in the 6 months before admission, a lower rate compared with those not hospitalized for DKA.19 Despite variation in use of pre-hospital outpatient care across these studies and ours, all find a substantial proportion of children not receiving outpatient care before hospitalization. Why is such a large proportion of children not receiving outpatient care before ACSC hospitalization? The concept of ACSC hospitalizations was developed and is applied under the premise that these are more potentially avoidable through high quality outpatient care than are other types of hospitalizations. Despite this premise, there is currently no accepted benchmark proportion of ACSC hospitalizations before which children should have received outpatient treatment. However, some might argue that if pediatric ACSC hospitalizations are truly preventable through high-quality outpatient care, then the proportion of children receiving outpatient care prior to hospitalization should likely be much higher than the 30% identified in this study. Alternative explanations are possible for the low proportion of preceding outpatient visits. For example, limiting our study cohort to hospitalized children could have selected ACSC

Page 12 of 23

13

patients with high illness severity; we were unable to measure hospitalizations avoided through outpatient care in children with lower severity ACSC. High severity could preclude the opportunity to receive outpatient treatment to the same degree for both ACSC and non-ACSC patients. If this were found true in subsequent studies, then ACSC hospitalizations in children might not be as sensitive to ambulatory care as currently perceived. Several patient characteristics assessed in this study had a stronger association with receipt of outpatient care than whether hospitalization occurred for an ACSC. Prior literature shows that these characteristics are associated with higher use of outpatient services in general, not just before hospitalization. For example, younger children are more likely to receive physician care for upper respiratory infections,20 and those under 1 year have frequent preventive visits. White and Hispanic children are more likely than children of other races and ethnicities to have outpatient visits;21 white children in particular have higher rates of subspecialty care referral and use.22,23 Medicaid managed care plans often successfully incentivize use of ambulatory care while diverting use of more expensive emergency department or inpatient care.18 Children with CCC are high users of all health services, including outpatient visits.24 Further investigation is necessary to assess how baseline levels of outpatient care use, including dimensions such as continuity of care, influence how families seek help during illnesses that could lead to hospitalization. ACSC hospitalizations in very young children and in those with medical complexity may merit additional consideration from those occurring in older and generally healthy children.25 In our study children ages 5 years old and younger experienced two-thirds of ACSC hospitalizations, even though they comprised only half of the cohort. ACSC may present with relatively higher severity in younger children, and therefore resulting hospitalizations may not be

Page 13 of 23

14

as preventable as those occurring in older children. For children with CCC, ACSC comprised under 30% of hospitalizations, but some of their most common reasons for hospitalization were ACSC, including seizures, asthma, and pneumonia (data not shown). Children with medical complexity often have close connections to health providers, and in our study receipt of preceding outpatient care for children with at least 1 CCC did not differ before ACSC and nonACSC hospitalizations. Prior efforts to assess the validity and meaning of ACSC have not specifically addressed potential differences in ACSC among children with vs. without medical complexity.26 For example, seizures are often included in ACSC studies,3,11,27 but treatment of them in the outpatient setting for children with complex neurological conditions (e.g., cerebral palsy, congenital brain malformation) has not been rigorously assessed. The study has several limitations. It cannot assess the effectiveness of outpatient care to prevent hospitalizations in children with ACSC; subsequent investigation with a different study design would be necessary. The data source for the study is limited to a single payer type, Medicaid, and the findings therefore may not generalize to children using other types of insurance. Using claims data, we relied on the accuracy of diagnosis codes to identify ACSC. Although errors in coding might exist, prior studies lead us to believe they did not have a substantial impact on our study’s findings.28 Moreover, the proportion of hospitalizations we identified as ACSC is similar to that reported in prior studies (33% and 37%)3,11 using nationally representative data. It is likely that some outpatient visits we identified were unrelated to the problem that led to hospitalization. However, review of the diagnosis codes associated with outpatient visits in our study demonstrated that only small proportions were for health supervision and vaccination, and that acute illness diagnoses seemed to predominate more before ACSC- than non-ACSC hospitalizations. We did not attempt to assess the relatedness of

Page 14 of 23

15

outpatient visits to subsequent hospitalizations because we felt that doing so without more clinically rich information (e.g., abstracted from chart review) might be misleading. In general, outpatient visits include substantially fewer diagnosis codes, and with less specificity, than do hospitalizations. We did not assess other sources of pre-hospital care, including retail clinics, home nursing, or care offered by telephone. There may be contextual factors we could not ascertain, such as differences among states, strength of relationship with a usual source of care, and health literacy, which might affect children’s use of outpatient and hospital care across the Medicaid programs in the study. Despite these limitations, our findings have important implications for Medicaid programs, policymakers, and clinical care delivery systems trying to reduce hospitalizations among children. ACSC hospitalization rates have declined for children in recent decades,29 but it is not known if further reductions may be possible through timely outpatient care. Further exploration is needed of the timing and trajectory of ACSC illnesses prior to, and in the absence of, hospitalization. Future studies should assess the following: 1) occurrence and content of outpatient visits for individual ACSC; 2) risk factors for subsequent hospitalization among those who receive outpatient care; 3) how the use of emergency department care affects the use of outpatient services and the occurrence of ACSC hospitalizations; 4) families’ perceptions of the severity of ACSC illnesses and their decisions about when and where to seek care. If further investigation demonstrates that many ACSC hospitalizations for children are not prevented by appropriate ambulatory care, then re-evaluation and revision of the definition and policy application of pediatric ACSC hospitalizations should occur. In the meantime, health systems and state Medicaid programs should continue working to optimize access to and use of outpatient care as children develop acute, potentially serious health problems.

Page 15 of 23

16

References 1.

AHRQ Quality Indicators - Guide to Prevention Quality Indicators: Hospital Admission for Ambulatory Care Sensitive Conditions. Rockville, MD: Agency for Healthcare Research and Quality;2001.

2.

Yu H, Wier LM, Elixhauser A. Hospital Stays for Children, 2009: Statistical Brief #118. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville, MD 2006.

3.

Lu S, Kuo DZ. Hospital charges of potentially preventable pediatric hospitalizations. Acad Pediatr. 2012;12(5):436-444.

4.

Perrin JM, Greenspan P, Bloom SR, et al. Primary care involvement among hospitalized children. Arch Pediatr Adolesc Med. 1996;150(5):479-486.

5.

Casanova C, Colomer C, Starfield B. Pediatric hospitalization due to ambulatory caresensitive conditions in Valencia (Spain). Int J Qual Health Care. 1996;8(1):51-59.

6.

Coller RJ, Klitzner TS, Saenz AA, Lerner CF, Nelson BB, Chung PJ. The Medical Home and Hospital Readmissions. Pediatrics. 2015;136(6):e1550-1560.

7.

Draft ICD-10-CM/PCS MS-DRGv29.0 Definitions Manual. 2012; https://www.cms.gov/ICD10Manual/version29-fullcode-cms/P0001.html. Accessed June 1, 2016.

8.

Schuster M BJ, Jang J, et al. Pediatric All-Condition Readmission Measure. Evidence Report on Measure No. 0129. Prepared by Harvard University Center of Excellence for Pediatric Quality Measurement. AHRQ Publication No. 14(16)-P008-3-EF. Rockville, MD: Agency for Healthcare Research and Quality; March 2015.

Page 16 of 23

17

9.

Feudtner C, Christakis DA, Connell FA. Pediatric Deaths Attributable to Complex Chronic Conditions: A Population-Based Study of Washington State, 1980–1997. Pediatrics. 2000;106(Supplement 1):205-209.

10.

Feudtner C, Feinstein JA, Zhong W, Hall M, Dai D. Pediatric complex chronic conditions classification system version 2: updated for ICD-10 and complex medical technology dependence and transplantation. BMC Pediatr. 2014;14:199.

11.

Parker JD, Schoendorf KC. Variation in hospital discharges for ambulatory care-sensitive conditions among children. Pediatrics. 2000;106(4 Suppl):942-948.

12.

Berry JG, Hall M, Cohen E, O'Neill M, Feudtner C. Ways to Identify Children with Medical Complexity and the Importance of Why. J Pediatr. 2015;167(2):229-237.

13.

Berry JG, Hall M, Neff J, et al. Children with medical complexity and Medicaid: spending and cost savings. Health Aff (Millwood). 2014;33(12):2199-2206.

14.

Measure Information Form: Value-Based Payment Modifier Program. Centers for Medicare and Medicaid Services;September 2015.

15.

Bisgaier J, Rhodes KV. Auditing access to specialty care for children with public insurance. N Engl J Med. 2011;364(24):2324-2333.

16.

Zickafoose JS, Gebremariam A, Clark SJ, Davis MM. Medical home disparities between children with public and private insurance. Acad Pediatr. 2011;11(4):305-310.

17.

Chung PJ, Garfield CF, Elliott MN, Carey C, Eriksson C, Schuster MA. Need for and use of family leave among parents of children with special health care needs. Pediatrics. 2007;119(5):e1047-1055.

18.

Gadomski A, Jenkins P, Nichols M. Impact of a Medicaid primary care provider and preventive care on pediatric hospitalization. Pediatrics. 1998;101(3):E1.

Page 17 of 23

18

19.

Crossen SS, Wilson DM, Saynina O, Sanders LM. Outpatient Care Preceding Hospitalization for Diabetic Ketoacidosis. Pediatrics. 2016;137(6).

20.

Saunders NR, Tennis O, Jacobson S, Gans M, Dick PT. Parents' responses to symptoms of respiratory tract infection in their children. CMAJ. 2003;168(1):25-30.

21.

Berdahl TA, Friedman BS, McCormick MC, Simpson L. Annual report on health care for children and youth in the United States: trends in racial/ethnic, income, and insurance disparities over time, 2002-2009. Acad Pediatr. 2013;13(3):191-203.

22.

Kuhlthau K, Nyman RM, Ferris TG, Beal AC, Perrin JM. Correlates of use of specialty care. Pediatrics. 2004;113(3 Pt 1):e249-255.

23.

Shenkman E, Tian L, Nackashi J, Schatz D. Managed care organization characteristics and outpatient specialty care use among children with chronic illness. Pediatrics. 2005;115(6):1547-1554.

24.

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):e1463-1470.

25.

Nelson BB, Coller RJ, Saenz AA, et al. How Avoidable Are Hospitalizations for Children with Medical Complexity? Understanding Parent Perspectives. Acad Pediatr. 2016.

26.

Measures of Pediatric Health Care Quality Based on Hospital Administrative Data: The Pediatric Quality Indicators. Agency for Healthcare Research and Quality;2006.

27.

Flores G, Abreu M, Chaisson CE, Sun D. Keeping children out of hospitals: parents' and physicians' perspectives on how pediatric hospitalizations for ambulatory care-sensitive conditions can be avoided. Pediatrics. 2003;112(5):1021-1030.

Page 18 of 23

19

28.

Gidengil C, Mangione-Smith R, Bailey LC, et al. Using Medicaid and CHIP claims data to support pediatric quality measurement: lessons from 3 centers of excellence in measure development. Acad Pediatr. 2014;14(5 Suppl):S76-81.

29.

Friedman B, Berdahl T, Simpson LA, et al. Annual report on health care for children and youth in the United States: focus on trends in hospital use and quality. Acad Pediatr. 2011;11(4):263-279.

Page 19 of 23

20 Table 1. Characteristics of the study population and bivariable analysis of the receipt of preceding outpatient care1 Patient Characteristics n=254,902 Age 0-1 year 2-5 years 6-12 years 13-17 years Gender Male Female Race/ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Other Type of Medicaid Managed care Fee-for-service Medicaid Eligibility Income or other Disability Complex Chronic Condition # 0 1 2 or more Complex Chronic Condition Type Neurologic Cardiovascular Metabolic Other congenital Technology assistance Hematologic/immunologic Gastrointestinal Respiratory Prematurity/neonatal Renal/urologic Preceding Care Any outpatient visit Primary care Specialty care Emergency department

Study Cohort n (Col %)

Preceding Outpatient Care Yes, Row %2

80,244 (31.5) 57,871 (22.7) 57,621 (22.6) 59,166 (23.2)

39.7 29.4 23.3 17.8

139,114 (54.6) 115,787 (45.4)

27.8 29.5

125,504 (49.2) 72,100 (28.3) 20,018 (7.9) 37,280 (14.6)

29.7 24.7 33.1 30.1

150,416 (59.0) 104,486 (41.0)

29.7 27.0

232,230 (91.1) 22,672 (8.9)

27.9 35.1

195,599 (76.7) 42,737 (16.8) 16,566 (6.5)

26.4 31.8 45.5

16,272 (6.4) 15,392 (6.0) 14,148 (5.6) 9,964 (3.9) 8,484 (3.3) 8,316 (3.3) 7,329 (2.9) 5,196 (2.0) 4,056 (1.6) 3,339 (1.3)

37.9 36.6 33.9 42.4 52.3 33.3 52.4 50.5 50.0 41.4

72,843 (28.6) 50,725 (19.9) 36,463 (14.3) 54,224 (21.3)

-

Abbreviations: IQR = interquartile range 1 Indicates ≥1 outpatient visit to primary or specialty care within 7 days preceding hospital admission. 2 All comparisons of characteristics by receipt of preceding outpatient care (i.e., Yes vs. No) are statistically significant with p<0.001 by chi-square tests.

Page 20 of 23

21 Table 2. Health services by type of index hospitalization

Health Services Most Common Conditions, MS-DRG3 (Col %) Condition 1 Condition 2 Condition 3 Hospital Resource Use, median (IQR) Length of stay, days Spending, $ Preceding Care, n (Col %) Any outpatient visit Primary care Specialty care Emergency department

Ambulatory Care Sensitive Condition1 n=87,943

Non-Ambulatory Care Sensitive Condition2 n=166,959

Asthma (24.3) Pneumonia (21.3) Cellulitis (11.7)

Depression (17.9) Bronchiolitis (14.3) Appendicitis (4.6)

2 (1,3) $4,661 (2,647, 7,103)

3 (2,5) $5,209 (3,115, 9,656)

27,307 (31.1) 19,711 (22.4) 13,219 (15.0) 19,503 (22.2)

45,536 (27.3) 31,014 (18.6) 23,244 (13.9) 34,721 (20.8)

1

Ambulatory Care Sensitive Conditions (ACSC; and ICD-9-CM codes) used in this study were asthma (493), pneumonia (481, 483, 485, 486, 482.2, 482.3), acute otitis media or other upper respiratory infection (381, 382, 460-463, 465, 472-474, 490, 034.0, 079.9, 466.0), gastroenteritis and dehydration (008.6, 008.8, 276.5, 558.9), cellulitis (289.3, 680-684, 686), seizures (345, 780.3), urinary tract infection (689, 690), and dental conditions (157-159).

2

All comparisons of characteristics by type of hospitalization (i.e., ACSC vs. non-ACSC) are statistically significant with p<0.001 by chi-square tests. 3

Medicare Severity-Diagnosis Related Group

Page 21 of 23

22 Table 3. Patient characteristics by type of index hospitalization

Patient Characteristics

Age 0-1 year 2-5 years 6-12 years 13-17 years Gender Male Female Race/ethnicity Non-Hispanic White Non-Hispanic Black Hispanic Other Type of Medicaid Managed care Fee-for-service Medicaid Eligibility Income or other Disability Complex Chronic Condition # 0 1 2 or more

Ambulatory Care Sensitive Condition 87,943 n (Column %)

Non-Ambulatory Care Sensitive Condition1 166,959 n (Column %)

27,803 (31.6) 32,110 (36.5) 20,607 (23.4) 7,423 (8.4)

52,441 (31.4) 25,761 (15.4) 37,014 (22.2) 51,743 (31.0)

46,830 (53.3) 41,113 (46.7)

92,284 (55.3) 74,674 (44.7)

38,762 (44.1) 28,176 (32.0) 7,119 (8.1) 13,886 (15.8)

86,742 (52.0) 43,924 (26.3) 12,899 (7.7) 23,394 (14.0)

54,026 (61.4) 33,917 (38.6)

96,390 (57.7) 70,569 (42.3)

81,366 (92.5) 6,577 (7.5)

150,864 (90.4) 16,095 (9.6)

70,250 (79.9) 12,932 (14.7) 4,761 (5.4)

125,349 (75.1) 29,805 (17.8) 11,805 (7.1)

1

All comparisons of characteristics by the type of hospitalization (i.e., ACSC vs. non-ACSC) are statistically significant with p<0.001 by chi-square tests.

Page 22 of 23

23 Table 4. Multivariable Analysis of the Receipt of Outpatient Care Before Hospitalization Characteristics Type of Hospitalization Ambulatory Care Sensitive Condition Referent = Non-ACSC Age 0-1 year 2-5 years 6-12 years Referent = 13-17 years Gender Female Referent = Male Race/ethnicity Non-Hispanic black Hispanic Other Referent = Non-Hispanic white Medicaid type Managed Care Referent = Fee-for-service Medicaid eligibility Disability Referent = Income or other Number of Complex Chronic Condition 1 2+ Referent = 0 CCC Interaction of ACSC hospitalization with CCC1 ACSC x Number of CCC 0 1 2+

Preceding outpatient care Odds ratio (99% CI) 1.2 (1.1, 1.2)*

2.4 (2.3, 2.5)* 1.5 (1.5, 1.6)* 1.2 (1.1, 1.2)*

1.1 (1.1, 1.1)*

0.7 (0.7, 0.7)* 1.0 (1.0, 1.1) 0.8 (0.7, 0.8)*

1.2 (1.2, 1.3)*

1.6 (1.5, 1.7)*

1.2 (1.1, 1.2)* 1.9 (1.8, 2.0)*

1.4 (1.3, 1.4)* 1.1 (1.0, 1.1) 1.1 (1.0, 1.2)

Abbreviations: ACSC (ambulatory care sensitive condition), CCC (complex chronic condition) 1 Odds ratios estimated by inputting values for ACSC and the specified value of each model parameter into the logistic regression equation, maintaining other model parameters at the mean. The referent group for the interaction is Non-ACSC x the specified value of each model parameter. * Statistically significant, all p<0.001.

Page 23 of 23