Most Children With Medical Complexity Do Not Receive Care in Well-Functioning Health Care Systems

Abstract

OBJECTIVES:

To describe the access of children with medical complexity (CMC) to well-functioning health care systems. To examine the relationships between medical complexity status and this outcome and its component indicators.

PATIENTS AND METHODS:

Secondary analysis of children in the National Survey of Children’s Health combined 2016–2017 data set who received care in well-functioning health systems. Secondary outcomes included this measure’s component indicators. The χ² analyses were used to examine associations between child and family characteristics and the primary outcome. Adjusted multivariable logistic regression was used to examine relationships between medical complexity status and primary and secondary outcomes. Using these regression models, we examined the interaction between medical complexity status and household income.

RESULTS:

CMC accounted for 1.6% of the weighted sample (n = 1.2 million children). Few CMC (7.6%) received care in a well-functioning health care system. CMC were significantly less likely than children with special health care needs (CSHCN) (odds ratio, 0.3) of meeting criteria for this primary outcome. Attainment rates for secondary outcomes (families feeling like partners in care; receives care within a medical home; received needed health care) were significantly lower among CMC than CSHCN. Family income was significantly associated with likelihood of meeting criteria for primary and secondary outcomes; however, the relationships between medical complexity status and our outcomes did not differ by income level.

CONCLUSIONS:

CMC are less likely than other CSHCN to report receiving care in well-functioning health care systems at all income levels. Further efforts are necessary to better adapt current health care systems to meet the unique needs of CMC.

Children with medical complexity (CMC) are medically fragile, with multiple, severe chronic health conditions resulting in significant health service needs, major functional limitations, and high health care use.¹ Although a small segment of the pediatric population, CMC account for up to 1-third of all pediatric health care expenditures and are at high risk for poor health outcomes.^2–4 Furthermore, family caregivers of CMC report significant hardships and dissatisfaction with their children’s medical services, because many must navigate fragmented health care systems without adequate assistance.^5,6

In response, national pediatric organizations have promoted efforts aiming to improve the quality of CMC health care.⁷ However, because of the difficulty of identifying CMC from larger deidentified data sets, the proportion of CMC actually receiving care in well-functioning health systems is unclear. Among CMC, the relationship between socioeconomic status and receiving care within well-functioning systems is underexamined as well. Household family income is of particular concern given its negative relationship with health service inequites in other pediatric populations.⁸ For example, among children with special health care needs (CSHCN), households with higher family income are more likely to report experiencing shared decision-making and receive adequate care coordination.^9,10 It is unknown if the relationship between household income extends to key CMC health service outcomes.

To address these gaps in the literature, we used a new method to identify CMC from the National Survey of Children’s Health (NSCH) (see accompanying Brief Report¹¹) to (1) describe the proportion of CMC receiving care in well-functioning health care systems (and its individual component quality indicators) using measures contained in the NSCH combined 2016–2017 data set, (2) examine how the quality of care provided by current pediatric health systems differs for children of varying medical complexity, and (3) investigate the interaction between medical complexity status and household family income. We hypothesized that a low overall proportion of CMC would report receiving care in a well-functioning health care system or meet criteria for this measure’s component indicators. We also hypothesized that any negative associations between medical complexity status and receiving care in a well-functioning system (and its component indicators) might be reduced at higher family income levels.

Methods

Data Source

This is a secondary analysis of the combined 2016–17 NSCH data set. Funded by the Maternal and Child Health Bureau of the Health Resources and Services Administration, the NSCH is conducted annually by the US Census Bureau to produce national and state-level data on the physical and emotional health of all American children.¹² Specifically, it provides estimates for the Title V Maternal and Child Health Services Block Grant Program’s National Outcome and Priority Measures.¹³ Details about survey administration, sampling, and response rates are described elsewhere.¹⁴ We combined the 2016 and 2017 surveys to increase the sample size of CMC. This study received exempt status from the university’s institutional review board.

Study Population

CMC were identified by using the methodology described in our accompanying Brief Report. This approach uses NSCH survey items to identify children who experience diagnoses and health consequences in all 4 domains of the widely accepted framework by Cohen et al¹ of pediatric medical complexity (Fig 1). All children in the data set were categorized by medical complexity and special health care needs status: (1) children with no special health care needs (Non-SHCN), (2) CSHCN who were not medically complex, and (3) CMC.¹⁵

FIGURE 1.

Identification of CMC. CMC were those who identified as CSHCN according to the CSHCN screener¹⁴ and met inclusion criteria in all 4 domains of the medical complexity framework by Cohen et al.¹

Outcomes

Our primary outcome was the Title V Maternal and Child Health Services Block Grant Program’s National Outcome Measure 17.2 – Receiving care within a well-functioning health care system, which represent the Maternal and Child Health Bureau’s target goals for the quality of health systems caring for CSHCN.¹³ The National Outcome Measures are a performance measure framework reflecting the Title V program’s desired health outcomes for the country and have been incorporated into Healthy People 2020 and Bright Futures national guidelines.^16,17 Children meeting criteria for all 5 (6 if ages 12–17 years) of this measure’s component indicators were classified as having received care in a well-functioning health system. Component indicators (our secondary outcomes) included the following: (1) family feels like a partner in their child’s care, (2) receives care within a medical home, (3) preventive medical and dental visit in the past year, (4) adequate insurance coverage, and (5) no unmet needs and barriers to services. For children ages 12 to 17 years, whether a child receives preparation for transition to adult health care was the sixth component indicator. Each component indicator is a composite of multiple survey items with dichotomous or ordinal response options.

Child and Family Characteristics

We examined child and family characteristics in relation to our outcomes. Child demographics included age, sex, race and ethnicity, and insurance type. Family and household characteristics included family income (operationalized as a percentage of federal poverty level), parental employment status and educational attainment, primary household language, household family structure, and geographic census region.

Analysis

We summarized child and family characteristics by medical complexity status using descriptive statistics and χ² tests of association. Because our primary outcome is a quality metric for CSHCN, we compared CSHCN and CMC who received care in a well-functioning health care system with those who did not using χ² tests for categorical variables. To identify independent variables associated with the primary outcome, we also performed univariate logistic regression analysis for each covariate of interest. Variables significant in χ² testing (P < .05) or univariate logistic analyses were entered into a multivariable logistic regression model. To examine the relationship between medical complexity status and secondary outcomes, we compared attainment rates for all component quality indicators, for CSHCN and CMC, through additional χ² testing. We then performed additional multivariable logistic regression analyses for each secondary outcome adjusted for variables significant in previous χ² testing and univariate analyses. To examine whether the relationship between medical complexity status and primary and secondary outcomes varied at different household income levels, we examined the interaction between medical complexity status and household income using adjusted Wald test and seemingly unrelated estimation test, postestimation hypothesis testing that accounts for survey design.¹⁸

All analyses used person-level weights provided by the NSCH to generate national estimates.¹⁴ Because missing values for household family income were multiply imputed by using sequential regression methods by NSCH administrators, analyses involving the income variable were run across all data set imputations and combined to generate more conservative estimates.¹³ We tested for multicollinearity between all variables included in regression analyses by calculating variables’ variance inflation factor and tolerance. Variables with variance inflation factor >2.5 and/or tolerance <0.1 were not included in multivariable models.¹⁹ Statistical analyses were performed with Stata version 15.1 (StataCorp, College Station, TX).

Results

Among 73 387 211 children ages 0 to 17 living in the Unites States in 2016 to 17%, 1.6% were identified as medically complex, representing 1 156 126 children. CSHCN and Non-SHCN accounted for 17.2% (12 606 103 children) and 81.2% (59 624 982 children) of the pediatric population, respectively. Higher proportions of CMC were older (ages 12–17 years), Black and Hispanic, male, had public insurance, and lived in single-parent households with lower family income, parental employment, and parental educational attainment than Non-SHCN and CSHCN (all P < .001) (Table 1).

TABLE 1.

Child and Family Characteristics Overall and by Medical Complexity Status (N = 73 387 211)

Total sample (n = 73 387 211)		Medical Complexity Status
		Non-SHCN, %	CSHCN, %	CMC, %
Characteristic	Overall, %	59 624 982 (81.2)	12 606 103 (17.2)	1 156 126 (1.6)
Agea
0–5	32.3	35.6	18.1	16.0
6–11	33.9	32.9	38.3	35.9
12–17	33.8	31.5	43.6	48.1
Sexa
Male	51.1	49.5	57.2	66.9
Female	48.9	50.5	42.8	33.1
Racea
White	51.4	51.5	51.7	45.9
Black	13.1	12.0	17.6	19.3
Hispanic	24.7	25.3	21.8	27.1
Other	10.8	11.2	8.9	7.6
Insurance typea
Public	31.3	29.3	38.4	52.7
Private	57.1	59.4	49.8	22.0
Mix public and private	4.6	3.9	6.7	20.1
Unknown type	0.8	0.8	0.9	0.5
Uninsured	6.2	6.6	4.1	4.3
Family incomea
≥400% FPL	30.2	30.9	27.9	15.7
200% to 399% FPL	26.8	27.2	25.7	21.2
100% to 199% FPL	21.8	21.7	22.3	26.2
<100% FPL	21.2	20.2	24.1	36.9
Parental employmenta
≥1 parent employed	96.3	96.4	96.0	90.0
Unemployed	3.7	3.6	4.0	10.0
Highest household education levela
College or higher	50.4	51.3	47.6	35.3
Some college	23.8	23.2	25.6	31.3
High school or equivalent	20.9	20.6	21.7	27.4
Less than high school	4.9	4.9	5.1	6.1
Household family structurea
Two parents, married	67.1	69.6	56.9	49.7
Two parents, not married	8.9	8.8	9.2	12.7
Single parent	16.3	14.6	23.4	24.4
Other family type	7.7	6.9	10.6	13.2
Primary household languagea
English	85.6	84.3	91.2	88.9
Non-English	14.4	15.7	8.8	11.1
Regiona
North East	16.1	16.2	15.6	16.7
Midwest	21.2	20.9	22.7	22.4
South	38.5	37.8	40.8	45.2
West	24.2	25.1	20.9	15.7

All data weighted for survey to generate national estimates. %, column percent. FPL, federal poverty level.

^aP < .001 in χ² test for association with medical complexity status.

Overall, a low proportion of CSHCN and CMC in the United States (15.0%) received care in a well-functioning health care system in 2016–2017 (Table 2). A significantly smaller percentage of CMC (7.6%) met criteria for the primary outcome than other CSHCN (15.7%) (P < .001). Examining our primary outcome’s component quality indicators, the minority of CMC met criteria for receiving care within a medical home (27.8%) and services necessary for transition to adult health services (8.1%). Considerable segments of CMC also responded negatively about their families feeling like partners in care (17.7%), having adequate insurance coverage (37.4%), and receiving needed health care (25.5%). In χ² analysis, significantly smaller proportions of CMC than CSHCN had families who felt like partner’s in their child’s care, received care within a medical home, received needed health care, and received services necessary for transition to adult health care (all P < .05). A significantly higher percentage of CMC had preventive medical and dental visits in the previous year than other CSHCN. Overall, 50.6% of CMC ages 0 to 11 years met ≤3 quality indicators of a well-functioning health care system. Among CMC ages 12 to 17 years, 47.3% met criteria for ≤3 component indicators.

TABLE 2.

Rates of Receiving Care in a Well-Functioning Health Care System and Component Quality Indicators Among CSHCN and CMC (Weighted n = 13 761 570)

Quality Indicator	All CSHCN, %	CSHCN (No CMC), %	CMC only, %	Pa
Received care within a well-functioning health care system	15.0	15.7	7.6	<.001
Family feels like partners in child’s care	92.3	93.1	83.3	<.001
Child receives care within a medical home	42.8	44.1	27.8	<.001
Child had a preventive medical and dental visit	76.6	75.7	86.8	<.001
Child had adequate insurance coverage	64.0	64.1	62.6	.66
Child received (or was not frustrated in receiving) needed health care	88.0	89.2	74.5	<.001
Child received services necessary for transition to adult health careb	13.2	13.7	8.1	.03

All data were weighted to generate national estimates. %, column percent.

^aχ² analysis.

^bOnly children ages 12 to 17 y included.

In adjusted multivariable regression, CMC status remained negatively associated with our primary outcome (Table 3). Compared with other CSHCN, CMC had significantly lower odds of receiving care in a well-functioning system (odds ratio [OR], 0.3; 95% confidence interval [CI], 0.2–0.5).

TABLE 3.

Factors Associated With Receiving Care in a Well-Functioning Health Care System Among CSHCN and CMC (Weighted n = 13 761 570)

Characteristic	Received Care in a Well-Functioning Health Care System, %	Pa	Unadjusted Analysis, OR (95% CI)	Adjusted Analysisb OR, (95% CI)
Medical complexity status		<.001
CSHCN	15.7		Ref	Ref
CMC	7.6		0.4 (0.3–0.7)	0.3 (0.2–0.5)
Age		<.001
12–17	4.8		Ref	Ref
6–11	23.8		6.3 (5.0–7.9)	7.4 (5.8–9.4)
0–5	21.6		5.5 (4.2–7.3)	6.1 (4.6–8.1)
Sex		.102
Male	15.8		Ref	—
Female	13.9		0.9 (0.7–1.7)	—
Race		.074
White	16.1		Ref	Ref
Black	11.6		0.7 (0.5–0.9)	1.0 (0.7–1.4)
Hispanic	14.6		0.9 (0.7–1.2)	1.1 (0.8–1.5)
Other	16.8		1.1 (0.8–1.5)	1.0 (0.7–1.5)
Insurance		.48
Private	16.6		Ref	—
Public	14.6		0.9 (0.7–1.1)	—
Both public and private	14.3		0.8 (0.6–1.2)	—
Unknown type	16.2		1.0 (0.3–2.9)	—
Family income		<.001
≥400% FPL	40.4		Ref	Ref
200%–399% FPL	26.9		0.8 (0.6–1.0)	0.7 (0.5–0.9)
100%–199% FPL	17.0		0.7 (0.5–0.9)	0.8 (0.6–1.1)
<100% FPL	15.7		0.6 (0.4–0.8)	0.6 (0.4–0.9)
Parental employment		<.001
≥1 parent employed	19.7		Ref	Ref
Unemployed	12.6		1.1 (0.7–1.7)	1.5 (0.8–2.5)
Highest household education level		.012
College or higher	17.4		Ref	Ref
Some college	14.1		0.8 (0.6–0.9)	1.0 (0.8–1.3)
High school or less	13.1		0.7 (0.6–0.9)	1.1 (0.8–1.5)
Household family structure		.025
Two parents, married	16.7		Ref	Ref
Two parents, not married	11.9		0.7 (0.4–1.0)	0.6 (0.4–0.9)
Single parent	12.4		0.7 (0.6–0.9)	1.0 (0.7–1.3)
Other family type	15.8		0.9 (0.7–1.3)	1.1 (0.6–2.3)
Primary household language		.434
English	15.3		Ref	—
Non-English	12.7		0.8 (0.5–1.4)	—
Region		.148
North East	15.9		Ref	—
Midwest	15.1		0.9 (0.7–1.2)	—
South	13.5		0.8 (0.7–1.1)	—
West	17.1		1.1 (0.8–1.5)	—

All data weighted for survey to generate national estimates. %, row percent. FPL, federal poverty level; Ref, reference; —, not applicable.

^aχ² analysis.

^bMultivariable logistic regression model among CSHCN and CMC adjusted for all factors in the final model. Uninsured status omitted as it predicts perfect failure to meet outcome.

Child and family characteristics significantly associated with receiving care in a well-functioning system in χ² testing (P < .05) and/or univariate analysis included child age and race, parental employment status and educational attainment, family income, and family structure (Table 3). However, only child age, family income, and family structure remained independently associated in adjusted analysis. Children ages 12 to 17 years had significantly lower odds of meeting primary outcome criteria than other age groups. Living in households with 2 nonmarried parents was associated with 40% lower odds of receiving care within well-functioning systems than children living with married parents.

CMC status remained independently associated with several component quality indicators (secondary outcomes) (Table 4). CMC had significantly lower odds than other CSHCN of having their family feeling like partners in care (OR, 0.3; 95% CI, 0.2–0.5), receiving care within a medical home (OR, 0.4; 95% CI, 0.2–0.7), and receiving needed health care (OR, 0.3; 95% CI, 0.2–0.4). Conversely, CMC had significantly higher odds of receiving preventive medical and dental care in the previous year than other CSHCN.

TABLE 4.

Adjusted ORs of Receiving Care in a Well-Functioning System and Meeting Component Quality Indicators Among CSHCN and CMC (Weighted n = 13 761 570)

Quality indicator	CMC Versus CSHCN
	aOR	95% CI
Family feels like partners in child’s care	0.3	0.2—0.5
Child receives care within a medical home	0.4	0.2—0.7
Child had a preventive medical and dental visit	1.9	1.3—2.9
Child had adequate insurance coverage	0.8	0.6–1.1
Child received (or was not frustrated in receiving) needed health care	0.3	0.2—0.4
Child received services necessary for transition to adult health carea	0.7	0.5–1.1
Received care within a well-functioning health care system	0.3	0.2–0.5

All data weighted to generate national estimates. Multivariable logistic regression models among CSHCN and CMC adjusted for age, race, family income, parental employment, parental education, and family structure. aOR, adjusted odds ratio.

^aOnly children ages 12 to 17 years old included.

Household Family Income and Interaction With Medical Complexity Status

Family income was significantly associated with receiving care in a well-functioning health care system in both χ² testing and adjusted regression analyses (Table 3). Compared with CSHCN and CMC living in households with family income ≥400% FPL, CSHCN and CMC living in poorer households had 20% to 40% lower odds of achieving this outcome. This relationship persisted when examining the primary outcome’s component quality indicators individually (all P < .001 in χ² testing), except for receiving services needed for transition to adult health care. In adjusted analyses, CSHCN and CMC living in poorer households had 30% to 50% lower odds of having their families feel like partners in care, 30% lower odds of receiving care in a medical home, and 50% lower odds of receiving needed health care than children living in households at ≥400% FPL. The interaction between medical complexity status and family income was not significant in adjusted analysis of our primary outcome (P = .13) nor in adjusted analyses for any of our secondary outcomes (all P > .1).

Discussion

In this analysis of the NSCH combined 2016–2017 data set, we observed that a minority of US children receive care in a well-functioning health care system. Focusing on medical complexity status, we demonstrate CMC are less likely than other CSHCN to receive care in a well-functioning health care system and report meeting this measure’s component indicators. Although higher household family income was associated with significantly higher odds of meeting criteria for our primary outcome and several of its component indicators, we found no significant interaction between family income and medical complexity status; the negative relationship between medical complexity status and receiving care in a well-functioning system did not significantly differ at varying family income levels.

With these findings, we build on existing literature demonstrating that medical complexity status is associated with difficulty accessing high-quality health care.^2,3,10 This is concerning because CMC prevalence grows with advancements in medical care and technology.²⁰ In examining the component indicators of a well-functioning health care system, we saw the primary elements driving the disparity between CMC and other CSHCN were families feeling like partners in their child’s care, receiving care within a medical home, and receiving needed health services with minimal difficulty. Although CMC were more likely to report having preventive medical and dental visits in the previous year, we speculate this reflects the increased frequency in which CMC interact with health care systems in general, which includes more-frequent outpatient follow-ups and sick appointments with their primary care providers to receive preventive services.²¹

Multiple, concurrent factors likely explain why CMC families were less likely to report feeling like partners in their child’s care. Infrequent exposure to high clinical complexity, outside of residency training, can result in pediatric providers feeling uncomfortable and ill-equipped to care for CMC’s multiple medical conditions and medical technology requirements.²² Consequently, CMC families report difficulty establishing long-term relationships with health care provider(s) who oversee their child’s care and are able to collaboratively discuss questions, concerns, and plans of care.²³ Without a “go-to/in-charge” provider, family caregivers of CMC are responsible for balancing fragmented care plans of their child’s multiple subspecialists on their own.⁵ The unpredictable trajectories of CMC’s medical conditions also contribute. More than other children, CMC frequently require acute care services (ie, sick appointments, hospitalizations) in which short-term medical decisions are prioritized and made with unfamiliar providers.²⁴ Additionally, CMC’s inherently uncertain future can be challenging for both families and providers to reconcile, hindering discussions related to long-term planning and goals of care.²⁵ Pediatric physicians, especially, receive little formalized communication training and report discomfort with difficult conversations.^26–28 Consequently, instead of finding support, family caregivers of CMC report feeling judged, underestimated, and treated as nonequals in decision-making by their child’s providers.^29,30

Despite the emergence of the patient-centered medical home, and its focus on care coordination, as the central strategy to address CMC health service needs, a number of barriers prevent its widespread implementation.³¹ Many pediatric practices are financially compelled to structure themselves to focus on episodic, time-limited issues, making it challenging to adopt a chronic care model, which would better serve CMC.^24,32 The services and activities required to provide comprehensive care for CMC are difficult and time-intensive yet poorly compensated (even nonbillable).²² Pediatric providers’ high-patient volumes make it even harder to devote the necessary time.⁷ Although the number of pediatric “complex care centers” is growing, most are situated in urban academic institutions and not equally distributed geographically.^29,33 Therefore, for preventive and acute sick care, many CMC rely on community practices not officially connected to pediatric hospitals and CMC’s subspecialists.^34,35 However, such practices often have insufficient resources to employ the appropriate staff or create the infrastructure necessary for CMC care coordination, resulting in overworked support personnel lacking tools to connect families to community resources and keep multiple health care providers on the same page.^36,37 Furthermore, pediatricians report lacking the skills to lead multidisciplinary teams and effectively collaborate with nonphysician colleagues.³⁸ Consequently, CMC families often report being responsible for arranging complex care plans and advocating for their child without adequate assistance.³⁸

We also found children living in poorer households were significantly less likely to receive care in a well-functioning health care system or meet criteria for this outcome’s component indicators, which is consistent with existing literature demonstrating the relationship between family income and health care quality.^8,9,38 However, we did not observe a significant interaction between household income and medical complexity status for our primary or secondary outcomes, indicating the negative association between medical complexity status and receiving high-quality care persisted across all income levels. This is similar to the finding of Boudreau et al⁴⁰ that the beneficial association between care coordination and unmet specialty care needs was consistent for CSHCN at all income levels.

Taken together, we believe our study’s findings are further evidence that medical complexity is associated with unique challenges in navigating pediatric health care systems (eg, arranging home nursing) that many CMC families struggle with, regardless of their financial resources.⁴¹ Health care providers should not expect all CMC families to take the lead in crafting and arranging their child’s complex care plans.²³ Efforts to advance system- and provider-level reforms that promote true collaboration between families and CMC providers must be maintained.³⁰

In regards to our finding that older children (ages 12–17 years) had significantly lower odds of receiving care in a well-functioning system than younger children, we believe this is largely driven by our finding that a small overall proportion of CSHCN receive services necessary for transitioning to adult health care (Table 2). This is consistent with previous studies in which researchers examine receipt of transition care among CSHCN and reflects another aspect of CSHCN and CMC health care that warrants further attention.^10,42

This study has several important limitations. The NSCH is cross-sectional, so we cannot infer causality between medical complexity status and receiving care within well-functioning health care systems. Because the NSCH was not designed to focus on medical complexity, the method used to identify CMC among survey respondents may not map to all aspects of medical complexity. Our primary and secondary outcome measures are solely dependent on parent report. Therefore, in addition to issues of recall and sampling bias, measures do not account for perspectives of health care providers (or children themselves). The low overall proportion of children meeting our primary outcome likely reflects the composite nature of this measure. A simpler measure might have resulted in a higher proportion of children reporting having receive care in a well-functioning health system. We also recognize that some degree of tautology exists between our CMC inclusion criteria and chosen outcomes. Using a method to identify CMC based in part on their increased health service needs means that these children have more opportunities to not have these needs met.

Conclusions

The vast majority of CMC do not receive care in a well-functioning health care system. Medical complexity status is independently associated with lower likelihood of experiencing key aspects of high-quality health care. These challenges exist even for CMC families with the highest household incomes. Further efforts are necessary to create a system of care that meets the unique needs of CMC.