Reliance on Acute Care Settings for Health Care Utilization: A Comparison of Adolescents with Younger Children

Abstract

Objective:

Because a goal of the Affordable Care Act was to increase preventive care and reduce high cost care, the objective of this study was to evaluate current healthcare utilization and reliance on acute care settings among Medicaid-enrolled children.

Methods:

This was a retrospective cohort study of the 2015 Truven Marketscan Medicaid claims database among children 0-21 years old with at least 11 months of continuous enrollment. We calculated adjusted probabilities of healthcare utilization (any healthcare use and ≥ 1 health maintenance visit) and high acute care reliance (ratio of emergency department or urgent care visits to all health care visits > 0.33) by age and compared utilization between adolescents and younger children using multivariable logistic regression.

Results:

Of the 5,182,540 Medicaid-enrolled children, 18.9% had no healthcare visits and 47.3% had ≥1 health maintenance visit in 2015. Both healthcare use and health maintenance visits decreased with increasing age (p<0.001). Compared to younger children (0-10 years old), adolescents were more likely to have no interaction with the healthcare system (aOR 2.20; 95% CI 2.19, 2.21) and less likely to have health maintenance visits (aOR 0.40; 0.39, 0.40). High acute care reliance was associated with increasing age, with adolescents having greater odds of high acute care reliance (aOR 1.08 [1.08, 1.09]).

Conclusions:

Medicaid-enrolled adolescents have low rates of healthcare utilization and have high reliance on acute care settings. Further investigation into adolescent-specific barriers to health maintenance care and drivers for acute care is warranted.

INTRODUCTION

Adolescence is a critical period of opportunity for initiating and developing positive health behaviors that can have long lasting effects through adulthood. Although adolescents are a generally healthy age-group, this age group experiences high rates of mental health disorders and suicide, sexually transmitted infections, sexual and physical violence, substance use, and other conditions associated with risky behaviors.¹ Thus, continued engagement with primary care providers throughout adolescence is critical to establish healthy behaviors through the delivery of preventive health services and anticipatory guidance. Yet, adolescents are the least likely pediatric age group to access primary care services^{2, 3} and may over rely on acute care settings, such as the emergency department (ED) or urgent care (UC) for medical care.

The American Academy of Pediatrics recommends annual preventive care visits for all patients aged 11 to 21 years, including adolescent-specific guidelines for confidential screening and counseling for health risk behaviors such as sexual activity, tobacco and substance use.⁴ Unfortunately, there are many barriers to health care access for this age group, including lack of insurance, cost, transportation difficulties, limited primary care provider time, facilities without weekend and evening hours, adolescents’ perception of invincibility, and lack of clinician comfort with adolescent specific health issues.^5-7 Furthermore, privacy and confidentiality are of particular concern to adolescents, as they often forego needed care when they fear a possible breach in confidentiality.⁸

Underuse of primary care services, especially among adolescents, may result in overdependence on acute care settings (ED or UC) both for crisis intervention and routine care, especially for those at highest risk for poor health outcomes.⁹ Pre- Affordable Care Act (ACA) and Medicaid expansion data demonstrated that fewer than 15% of Medicaid-enrolled adolescents participated in yearly routine health maintenance exams,¹⁰ yet adolescents made more than 20 million ED visits annually.¹¹ Additionally, data from the 1990s reported that more than 1.5 million adolescents received their only medical care through EDs.^{12, 13} Prior to the ACA, older adolescents had the lowest rate of insurance coverage of any age group,¹⁴ and publicly insured children frequently relied on the ED for care.¹⁵ The impact of the ACA and Medicaid expansion on healthcare use patterns among adolescents remains understudied. Adams and colleagues recently demonstrated modest increases in adolescent preventive care visits post ACA implementation.¹⁶ Despite recent improvements in insurance coverage and access to primary care since ACA enactment, ED visits continue to rise and are a common source of care for many children in the US, often for non-acute complaints.¹⁷ Because data regarding adolescent health care utilization may be outdated, the objectives of this study were to describe current outpatient healthcare utilization among Medicaid-enrolled youth, reliance on acute care settings (EDs and UCs), and explore factors associated with high acute care reliance among adolescents. We hypothesized that adolescents would continue to have lower likelihood of healthcare utilization overall, have lower rates of routine health maintenance visits, and have higher acute care reliance.

METHODS

Study Design and Data Source:

We performed a retrospective cohort study using the 2015 Truven Marketscan Medicaid multistate claims database (IBM Watson Health, Armonk, New York). In 2015, the Marketscan database contained all fee-for-service (FFS) and capitated claims from 11 de-identified states, as well as enrollee information including year of birth, sex, race/ethnicity, and months of enrollment. This study was deemed exempt by the Children’s National Medical Center Institutional Review Board.

Study Population:

All children, ages 0 through 21 years with at least 11 months of continuous coverage in Medicaid from January 1, 2015 through December 31, 2015, were included in the dataset.^18-20 We excluded patients with complex chronic conditions using the 3M Health Information Systems (Salt Lake City, UT) software program Clinical Risk Groups (CRG).²¹ CRG is a validated measure^21-23 that uses diagnostic and procedure codes to categorize children by non-chronic and chronic conditions and to stratify them by clinical severity into 9 hierarchical groups (CRG 1-9). Patients with CRG categories of 7-9, as well as patients in CRG 6 with at least one serious chronic condition likely to result in progressive deterioration (e.g. dominant chronic), were designated as medically complex and excluded from our analysis. We excluded children with medical complexity from the analyses because this population would likely be high utilizers of healthcare. We also excluded any dental visits from our analyses.

Outcome Measures and Covariables:

The primary outcomes were healthcare utilization and high acute care reliance. With respect to healthcare utilization, we investigated two variables: (1) the proportion of children with any healthcare visits and (2) those with at least one health maintenance visit during the 1-year period. Health maintenance visits were identified using Healthcare Effectiveness Data and Information Set measurement specifications.²⁴ An outpatient healthcare encounter with any of the ICD-9-CM or ICD-10-CM codes for healthcare check or supervision for infant or child, general medical examination, or Current Procedural Terminology codes for preventive visits was considered a health maintenance visit. We adapted the definition by Kroner and colleagues¹⁵ to calculate acute care reliance as the percentage of all health care visits that occurred in the ED or UC settings. High acute care reliance was defined as the ratio of ED and UC visits to all healthcare visits >0.33, as per Kroner and colleagues¹⁵ definition. Thus, high acute care reliance was defined as making at least one-third of all healthcare visits to acute care settings. Our secondary objective was to identify factors associated with high acute care reliance among adolescents.

We adjusted for the following socioeconomic and patient characteristics in our multivariable models: age, gender, race/ethnicity, pay arrangement (e.g. FFS or capitated), and CRG category. Age was categorized by year from 0-21 years, and then subsequently dichotomized as younger children (0-10 years) and adolescents (11-21 years) for comparisons. Adolescents ages were also subcategorized into: early adolescence (11-14 years), middle adolescence (15-17 years), and late adolescence (18-21 years).⁹ Race/ethnicity was categorized as: white, non-Hispanic (NH); black, NH; Hispanic; and other. CRG was categorized as no chronic conditions (CRG 1-2) and chronic conditions (CRG 3-6a).

Data Analysis:

We used standard descriptive statistics to summarize the patient population. We calculated adjusted probabilities of healthcare utilization (e.g., any healthcare use and at least one health maintenance visit) and high acute care reliance by age over the 1-year study period. We performed multivariable logistic regression to identify factors associated with high acute care reliance within the adolescent subpopulation (ages 11-21 years). An alpha of 0.05 was used to determine statistical significance, and we reported adjusted probabilities and adjusted odds ratios with 95% confidence intervals. We used SAS Software version 9.4 (SAS Institute Inc., Cary, NC, USA) to perform all analyses.

RESULTS

In 2015, there were 7,294,581 Medicaid enrollees between the ages of 0 and 21 years in the Marketscan dataset, of which 5,182,540 met inclusion criteria (Figure 1). The median age was 9 years (IQR 5, 14), approximately half were female (49.4%) and of black, NH race/ethnicity (47.0%). The majority of enrollees had capitated payment (66.6%) and one-third had a chronic condition (Table 1).

Figure 1. Flow Diagram of Study Population.

Table 1.

Demographics of Study Population

Demographic		N= 5,182,540
Median Age, years (IQR)		9 (5,14)
Age Categories, %	0-10 years	56.9%
	11-21 years	43.1%
	11-14 years	21.3%
	15-17 years	25.3%
	18-21 years	31.8%
Male, %		50.6%
Race/Ethnicity, %	White, NH	47.0%
	Black, NH	33.2%
	Hispanic	8.4%
	Other	11.4%
Pay Arrangement, %	Fee-for-service	33.4%
	Capitated	66.6%
Chronic condition†, %		33.2%

Healthcare Utilization

Children experienced a median of 4 (IQR 2, 6) healthcare visits over the 1-year study period. Of the 1.9 million children that had at least one acute care visit, 77% sought care in EDs, 13.8% sought care in UC, and 9.2% had both ED and UC visits. There were 981,736 children (18.9% of the cohort) who had no healthcare visits during the study period. After adjustment for gender, race/ethnicity, pay arrangement, and presence of chronic condition, overall healthcare use declined (p<0.001) as age increased (Figure 2). Compared to younger children, a larger proportion of adolescents made no healthcare visits over the 1-year study period (24.9% of 11-21 year olds compared to 14.4% of 0-10 year olds [aOR 2.20; 95% CI 2.19, 2.21]). Furthermore, older adolescents had higher rates of no healthcare use when compared to younger adolescents (11-14 years: 21.3%; 15-17 years: 25.3% [aOR 1.23, 95% CI 1.23, 1.24]; 18-21 years: 31.8% [OR 1.66, 95% CI 1.65, 1.68]).

Figure 2. Healthcare Utilization and Acute Care Reliance by Age.

*adjusted for gender, race/ethnicity, pay arrangement, and presence of chronic condition

Among the entire study population, 47.3% had a health maintenance visit during the study period. After adjustment, the proportion of youth with a health maintenance visit during the one-year study period declined with increasing age (p<0.001) (Figure 2). When comparing adolescents to younger children, 34.7% of 11-21 year olds had a health maintenance visit compared to 56.8% of 0-10 year olds (aOR 0.40; 95% CI 0.39, 0.40). Compared to younger adolescents (11-14 year olds), older adolescents were less likely to have a health maintenance visit over the study period (11-14 years: 43%; 15-17 years: 35% [aOR 0.70, 95% CI 0.69, 0.70]; 18-21 years: 17% [aOR 0.27, 95% CI 0.27, 0.28]).

Acute Care Reliance

Among the study population with at least one healthcare visit during the study period (81.1%), 26.0% had high acute care reliance. Figure 2 provides a graph of adjusted probabilities of high acute care reliance by age. In adjusted analysis, adolescents had increased odds of high acute care reliance compared to younger children (aOR 1.08; 95% CI 1.08, 1.09).

Factors Associated with High Acute Care Reliance Among Adolescents

Compared to younger adolescents 11-14 years of age, older adolescents had increased odds of high acute care reliance (15-17 year olds: aOR 1.26 [1.25, 1.27]; and 18-21 year olds: aOR 2.21 [2.19, 2.23]). Furthermore, males had higher acute care reliance compared to females (aOR 1.10; 1.09, 1.11). Compared to white NH adolescents, black NH adolescents had higher odds of high acute care reliance (aOR 1.27; 1.26, 1.28), while Hispanic (aOR 0.66; 0.65, 0.67) and adolescents of other racial/ethnic groups (aOR 0.89; 0.88, 0.90) had lower odds of high acute care reliance. Additionally, adolescents with capitated pay arrangements had higher odds of high acute care reliance compared to those with FFS (aOR 1.42; 1.41, 1.43). Finally, healthy adolescents had higher odds of high acute care reliance compared to those with chronic conditions (aOR 1.24; 1.24, 1.26). (Table 2)

Table 2.

Factors Associated with High Acute Care Reliance Among Adolescents

Factors		High Acute Care Reliance aOR (95% CI)
Adolescent Age Categories	11-14 years	Reference
	15-17 years	1.30 (1.25, 1.27)
	18-21 years	2.21 (2.19, 2.23)
Male		1.10 (1.09, 1.11)
Race/Ethnicity	White, NH	Reference
	Black, NH	1.27 (1.26, 1.28)
	Hispanic	0.66 (0.65, 0.67)
	Other	0.89 (0.88, 0.90)
Capitated Pay Arrangement		1.42 (1.41, 1.43)
No Chronic Condition †		1.25 (1.24, 1.26)

DISCUSSION

These results support our hypothesis that Medicaid-enrolled adolescents have lower likelihood of healthcare utilization, make fewer health maintenance visits, and have higher acute care reliance than younger children. Almost one in four Medicaid-enrolled adolescents in 2015 did not interact with the healthcare system, and more than 60% of adolescents did not have a health maintenance visit over the one-year study period. However, of those who did interact with the health care system, more than 25% of all adolescents had high acute care reliance. Thus, despite Medicaid expansion, our data demonstrate that adolescents continue to have low rates of primary care use and high rates of acute care reliance when accessing healthcare.

Despite recent Medicaid expansion under the ACA, there has been an increase in ED usage by Medicaid patients^25-27 and the majority of ED care is for non-urgent complaints.¹⁷ As the number of UCs continues to rise, UCs have become a site of frequent health care access among youth,²⁸ especially among adolescents.²⁹ Our results further substantiate that adolescents in particular are more likely to forego primary care and access acute care settings for health care services. These data are consistent with prior findings that adolescents access primary care less often than any other age group.³⁰ This can result in fragmented care, as adolescents are accessing acute care for brief, problem-oriented concerns, rather than seeking comprehensive primary care.³⁰

Access to preventive care is particularly important for adolescents to help modify risky behavior and promote healthy habits.³⁰ However, many adolescents face significant barriers to health care access. Thus, adolescents may be more likely to rely on acute care settings given the convenience of evening and weekend hours, lack of necessity to make appointments, cost, and confidentiality concerns. We observed that adolescents, and specifically older adolescents, NH black, male, and relatively healthy adolescents, were more likely to rely on the acute care setting for health care access which is similar to the racial/ethnic disparities in healthcare receipt identified in prior studies.³¹

Given our findings that adolescents are more likely to rely on acute care settings for health care access, this setting may be uniquely positioned to serve as an important safety net for this vulnerable population.³² As such, the acute care setting can represent an additional setting for provision of preventive services to supplement care offered in primary care settings, with subsequent linkage back to the medical home. In fact, numerous EDs have demonstrated feasibility and efficacy of screening and interventions to target many behavioral health issues, including alcohol and substance use,^{33, 34} intimate partner violence,^{35, 36} and mental health.^{37, 38} Furthermore, in addition to the acute care setting, other non-traditional settings, such as school health clinics, for example, should be considered as additional sites for the provision of preventive health services to reach and intervene with this high risk population.

This study should be viewed in light of some potential limitations. Our findings from Medicaid-enrolled beneficiaries may not be generalizable to the privately insured or uninsured population. Additionally, eleven de-identified states contributed data for this study, and therefore, these results may not be nationally generalizable due to potential geographic differences in health care utilization. Furthermore, because the contributing states are de-identified, we do not know whether these states were involved in Medicaid expansion. Finally, as with all claims-related data, the reliability of the data depends on coding accuracy.

In conclusion, our study demonstrates that up to one in four adolescents have no interaction with the health care system. When adolescents do access health care, they often rely on acute care settings. This fragmented care may result in lack of receipt of preventive care services. Thus, future work should seek to understand adolescent-specific barriers to primary care utilization and drivers for high acute care reliance. Given adolescent reliance on acute care settings, integration of preventive health services into the acute care setting with linkage back to the medical home is critical to prevent poor health outcomes in this particularly vulnerable age group.

Abbreviations:

(ED)

Emergency department

(UC)

Urgent Care

(ACA)

Affordable Care Act

(FFS)

Fee-For-Service

(CRG)

Clinical Risk Group

(NH)

Non-Hispanic