Abstract
BACKGROUND:
Children from socially disadvantaged families experience worse hospital outcomes compared with other children. We sought to identify modifiable barriers to care to target for intervention.
METHODS:
We conducted a prospective cohort study of hospitalized children over 15 months. Caregivers completed a survey within 3 days of admission and 2 to 8 weeks after discharge to assess 10 reported barriers to care related to their interactions within the health care system (eg, not feeling like they have sufficient skills to navigate the system and experiencing marginalization). Associations between barriers and outcomes (30-day readmissions and length of stay) were assessed by using multivariable regression. Barriers associated with worse outcomes were then tested for associations with a cumulative social disadvantage score based on 5 family sociodemographic characteristics (eg, low income).
RESULTS:
Of eligible families, 61% (n = 3651) completed the admission survey; of those, 48% (n = 1734) completed follow-up. Nine of 10 barriers were associated with at least 1 worse hospital outcome. Of those, 4 were also positively associated with cumulative social disadvantage: perceiving the system as a barrier (adjusted β = 1.66; 95% confidence interval [CI] 1.02 to 2.30), skill barriers (β = 3.82; 95% CI 3.22 to 4.43), cultural distance (β = 1.75; 95% CI 1.36 to 2.15), and marginalization (β = .71; 95% CI 0.30 to 1.11). Low income had the most consistently strong association with reported barriers.
CONCLUSIONS:
System barriers, skill barriers, cultural distance, and marginalization were significantly associated with both worse hospital outcomes and social disadvantage, suggesting these are promising targets for intervention to decrease disparities for hospitalized children.
Children from families with low income, of minority race and/or ethnicity, and with limited English proficiency (LEP) are at increased risk for worse outcomes after hospitalization, with longer length of stay (LOS), greater risk of readmission, and worse health-related quality of life (HRQL) compared with other children.1–11 However, the mechanisms by which these disparities occur are incompletely understood. Many factors likely contribute, occurring both within and outside the health care system and before and after hospital arrival. These include barriers to access and delayed presentation due to financial concerns, competing priorities, and/or previous negative experiences12–16; less family-centered communication, lower trust, and less partnership between families and providers17–20; increased rates of adverse events21–23; and provider bias.24,25
All of these factors must be addressed to achieve equity, but the tools, expertise, and time required for each will vary. For example, although efforts to address bias are essential, provider behavior is slow to change.26 It is therefore appealing to identify potentially modifiable attributes that occur within the health care setting to address in the near term. For example, disparate outcomes due to difficulty in navigating the health care system could be addressed through enhanced care coordination or patient navigation.
Thus, our objective was to identify potentially modifiable health care barriers experienced by families of hospitalized children to isolate promising targets for future intervention. We approached this in 3 phases: (1) identifying barriers to care associated with worse hospital outcomes, (2) identifying which of those barriers were also associated with greater social disadvantage, and (3) identifying the relative contribution of individual social disadvantage factors to the association with barriers to care.
Methods
Study Population
We prospectively studied a cohort of children admitted to a freestanding, quaternary-care children’s hospital between October 1, 2014, and December 31, 2015. The electronic medical record was used to screen children for eligibility. Children and caregivers were eligible if they were admitted to a medical or surgical unit, had not previously participated during the past 2 months, were not in strict reverse isolation secondary to immunocompromise, were not admitted for suspected child abuse, and spoke English or Spanish. Other languages were excluded because of a lack of validated survey measures. We excluded children admitted for scheduled chemotherapy or to the psychiatry, rehabilitation, or bone marrow transplant units because hospital outcomes such as LOS and HRQL are distinct for these patients and could significantly skew results. Eligible caregivers were approached and enrolled by research assistants within 72 hours of admission and asked to complete both an admission survey at enrollment and a follow-up survey 2 to 8 weeks postdischarge. Caregivers could self-administer the Web-based surveys or complete them via telephone with a bilingual research assistant.
Survey Administration
Both surveys are routinely administered to evaluate changes in HRQL resulting from hospitalization; we added previously validated measures to these ongoing surveys for this study. Most measures were caregiver reported, although children ≥8 years old were also invited to complete HRQL measures. Caregivers who completed the admission survey were e-mailed the follow-up survey link or called, depending on preference, 2 to 4 weeks after discharge. Seventy-two percent of completions occurred between 2 and 6 weeks after discharge. Caregivers of patients readmitted within 30 days were ineligible for the follow-up survey.
Survey Measures
Survey measures included sociodemographic disadvantage markers (eg, low income), caregiver experiences, abilities and attitudes that may affect health care system navigation (hereafter referred to as “barriers to care”), and HRQL during and after hospitalization. The previously validated barriers-to-care measures (Table 1) were chosen because each relates to the interaction between the caregiver and the health care system, has a known relationship with health or health care outcomes, has a potential relationship with sociodemographic characteristics, and could be modifiable with the right intervention (as opposed to measures such as race and/or ethnicity). Barriers-to-care measures were scored so that higher scores reflected greater barriers.
TABLE 1.
Description of Data Elements Used in the Study, Including Social Disadvantage Factors, Caregiver-Reported Barriers-To-Care Measures, and Hospital Outcomes
| Measure | Construct and Description | Data Source |
|---|---|---|
| Social disadvantage factors | ||
| US census question followed by preferred language for care60 | Interpreter need: 2-item, sequential approach with excellent sensitivity and specificity for identifying individuals who are likely to benefit from interpretation during medical care | Hospital administrative data and admission survey |
| Caregiver report of highest educational attainment | Low education: defined as caregiver with a high school degree or less | Admission survey |
| Caregiver report of annual family income | Low income: annual household income <$30 000 | Admission survey |
| Primary payer for current hospital stay | Public insurance: Medicaid and Medicare; no insurance and charity care were also classified with public insurance | Hospital administrative data |
| US census questions regarding race and ethnicity | Minority race and/or ethnicity: caregiver report of child’s race and/or ethnicity as anything other than non-Hispanic white | Hospital administrative data |
| Barriers-to-care measures | ||
| Barriers to Care Questionnaire: System48 | System as a barrier: 1 item measuring the caregiver’s perception of how well the health care system has worked for their child in the past 3 mo (score range 0–100) | Admission survey |
| Barriers to Care Questionnaire: Skills48 | Skills barriers: 8 items measuring the perceived mismatch between the caregiver’s skills and what the health care system requires of them; items include the degree to which caregivers have enough information about how the system works, the degree to which they understand doctor’s orders, their ability to get referrals to specialists, and their ability to get enough help with paperwork (score range 0–100) | Admission survey |
| Cultural Distance Scale47 | Cultural distance: 4 items measuring how similar or different the caregiver feels they are compared with the child’s doctors, with questions about how they speak, how they reason about problems, communication styles, and general values in life (score range 0–100) | Admission survey |
| Barriers to Care Questionnaire: Marginalization48 | Marginalization barriers: 11 items measuring feelings of marginalization and disrespect within the health care system; items include not being listened to, impatient or intimidating doctors, not being treated with respect, and being judged on your appearance, ancestry, or accent (score range 0–100) | Admission survey |
| Health Care System Distrust scale61 | System distrust: 5 items measuring trust in the health care system; items capture the degree to which the caregiver agrees, for example, that the system lies to make money, covers up its mistakes, and discriminates on the basis of race and ethnicity (score range 5–25) | Admission survey |
| Barriers to Care Questionnaire: Expectations48 | Expectation barriers: 7 items measuring ways in which the health care system might fail to meet expectations for acceptable care; items include lack of coordination between the child’s doctors and other parts of the health care system, medical errors, and doctors treating symptoms without finding the cause of an illness (score range 0–100) | Admission survey |
| Barriers to Care Questionnaire: Knowledge and Beliefs48 | Knowledge barriers: 4 items measuring perceived discordance between the caregiver’s and providers’ beliefs and recommendations for treatment and health; items include disagreeing with doctors’ orders and doctors and nurses having different ideas about health than the caregiver does (score range 0–100) | Admission survey |
| Perceived Efficacy in Patient-Physician Interactions62,63 | Health care self-efficacy: 10-item tool measuring the caregiver’s confidence in interacting with health care professionals and organizations; items include confidence in knowing what questions to ask, getting a doctor to do something about their main health concern for their child, and understanding what the doctor said (score range 10–50) | Admission survey |
| Wake Forest Physician Trust Scale64 | Distrust of providers: 10-item tool measuring trust in the specific physician(s) providing care to the child; items include the degree to which the caregiver believes that the doctors only care about what is best for the child, that the doctors are thorough and careful, and that they will do whatever it takes to get the child needed care (score range 10–50) | Follow-up survey |
| Partnership-building subscale of Parents' Perceptions of Physicians' Communicative Behavior65 | Lack of partnership: 3 items assessing provider-caregiver partnership; items include the degree to which the doctor encouraged the caregiver to express worries and concerns and the degree to which they asked for the caregiver’s opinions on what to do for the child’s medical condition (score 3–18) | Follow-up survey |
| Hospital outcomes | ||
| PedsQL Generic Core or Infant Scales27,66 | HRQL: 23–45 items (depending on child age) assessing physical and psychosocial functioning collected at admission and follow-up. The change score is calculated as the difference between the follow-up and admission survey scores (each scaled 0–100). For analyses, caregiver report was used for patients <13 y old. For teenagers, self-report was used when available; otherwise, caregiver report was used. | Admission and follow-up surveys |
| Hospital LOS | Measured as a continuous variable to the one-tenth of a day | Hospital administrative data |
| Unplanned 30-d readmissions32 | 30-d unplanned readmissions: readmissions to inpatient or observation status within 30 d of discharge after excluding readmissions with procedure codes that are likely to be related to a planned procedure | Hospital administrative data |
For HRQL, changes in physical functioning after hospital discharge were assessed by using the Pediatric Quality of Life Inventory 4.0 Generic Core or Infant Scales (PedsQL) physical functioning subscale.27,28 We used the physical functioning subscale because hospital care most strongly influences a child’s physical (rather than psychosocial) functioning.29,30 For analyses, caregiver-proxy report was used for patients <13 years old. For teenagers, self-report was used when available; otherwise, caregiver-proxy report was used.
Additional Data
Hospital administrative records were used to collect data on child age, sex, insurance type, caregiver-reported child race and/or ethnicity, preferred language for medical care, medical complexity, LOS (measured to the one-tenth of a day), and unplanned 30-day readmissions. Medical complexity was determined by using the Pediatric Medical Complexity Algorithm, which classifies children as having no chronic conditions, noncomplex chronic conditions, or complex chronic conditions.31 Inpatient readmissions were classified as unplanned by using the methods developed by Berry et al32 (Table 1).
The hospital’s institutional review board approved this study.
Analytic Approach
Analysis occurred in 3 phases: First, we identified caregiver-reported barriers to care associated with hospital outcomes (Fig 1). To identify potential targets for intervention to improve outcomes for children with social risk, in phase 2, we identified those barriers associated with both worse outcomes and a cumulative social disadvantage score. To better tailor future interventions, in phase 3, for those barriers identified as potential targets in phase 2, we identified the relative contribution of each social disadvantage factor included in the social disadvantage score.
FIGURE 1.
Schematic of the 3 phases of the analytic approach.
Phase 1: Identifying Barriers to Care Associated With Hospital Outcomes
For phase 1, the 10 survey-based barriers-to-care scores were the main predictors. Outcomes were modeled on the basis of 10-unit differences in the barriers-to-care scores.
Hospital outcomes were HRQL change scores, LOS, and unplanned 30-day readmissions. LOS was winsorized such that the 99th percentile was assigned to the most extreme values to prevent skewing of the results.
We modeled HRQL change scores as a continuous outcome using linear regression and unplanned readmissions as a dichotomous outcome using logistic regression. LOS was modeled by using a generalized linear model with a log-link and γ distribution. Models accounted for clustering on patient and controlled for patient age and medical complexity.
Phase 2: Identifying Barriers to Care Associated With Social Disadvantage Score
The social disadvantage score was the main predictor in this analysis. To create this score, 1 point was assigned for each of the following: public or no insurance, minority race and/or ethnicity, low income, low education, and interpreter need (score range 0–5; Table 1). We chose this approach on the basis of previous studies demonstrating clear evidence for an additive rather than threshold relationship for a variety of risks and outcomes in children to reflect the way in which risk factors tend to aggregate together to produce a disproportionate impact on a subset of individuals.33–38 Approximately 24% of participants declined to report income or did not know; therefore, we used multiple imputation, with 20 rounds of imputation, for missing income information before calculating the risk score.39,40
Nine of the 10 barriers-to-care scores (Table 1) that were identified in phase 1 as being significantly associated with ≥1 worse hospital outcome served as the outcomes for these analyses; the majority were associated with LOS.
We modeled each barriers-to-care score using linear regression, with robust SEs accounting for clustering on patient and controlling for patient age and medical complexity. We imputed missing income data using multiple imputation with chained equations. We conducted analyses on each multiply imputed data set and pooled the results to generate final results.41 We modeled the social disadvantage score as a continuous variable to assess for a linear trend in its relationship to each of the 9 barriers-to-care scores.
Phase 3: Identifying the Relative Contribution of Social Disadvantage Factors to Overall Associations
All possible combinations of the 5 individual factors constituting the social disadvantage score served as the main predictors.
We used the 5 barriers-to-care scores demonstrating positive linear relationships with the social disadvantage score in phase 2 as the outcomes in these analyses.
To identify the relative contribution of each social disadvantage factor, we fitted a linear regression model for all barriers-to-care scores, including all social disadvantage factors, as dichotomous predictors. We controlled for age and medical complexity, used robust SEs accounting for clustering within patient, and addressed missing data with multiple imputation. We then assessed the predicted marginal effects on that barriers-to-care score from each combination of social disadvantage factors. In this way, we used our models to predict how much of a problem a caregiver would report each barrier to care to be for every combination of individual social risk factors.
Results
During the 15-month study, there were 5992 families eligible for the admission survey. Of those, 1083 (18.1%) declined, 503 (8.4%) had no parent or guardian available, 455 (7.6%) were unable to be approached within 72 hours, 300 (5.0%) completed the main survey but did not complete measures for this study, and 3651 (60.9%) completed at least 1 of the barriers-to-care measures. Of those who completed the admission survey, 1734 (47.5%) also completed the follow-up survey. Less than 3% of admission surveys (84 of 3651) and <2% of follow-up surveys (19 of 1734) were completed by caregivers who were repeat study participants.
Of survey participants, almost half were publicly insured and approximately half were non-Hispanic white; ∼10% were likely to need interpretation (Table 2). Approximately 40% had a bachelor’s degree or more, and >40% reported an annual household income >$50 000. Before imputation, 35% of respondents had no social disadvantage factors, 28% had 1, 14% had 2, 11% had 3, 8% had 4, and 4% had 5.
TABLE 2.
Characteristics of Patients and Caregivers Who Completed Admission Survey Questions and Follow-up Surveys
| Admission Survey (N = 3651) | Admission and Follow-up Surveys (N = 1734) | |
|---|---|---|
| Insurance type, n (%) | ||
| Commercial | 1971 (54.1) | 948 (54.7) |
| Public | 1674 (45.9) | 786 (45.3) |
| Male sex, n (%) | 1899 (52.1) | 885 (51.0) |
| Race and/or ethnicity, n (%) | ||
| Non-Hispanic white | 1922 (52.7) | 935 (53.9) |
| African American or Black | 156 (4.3) | 54 (3.1) |
| Hispanic | 711 (19.5) | 360 (20.8) |
| Asian American | 220 (6.0) | 100 (5.8) |
| Other, multiracial, or unknown | 636 (17.5) | 285 (16.4) |
| Language for care, n (%) | ||
| English | 3321 (91.1) | 1538 (88.7) |
| Spanish | 301 (8.3) | 184 (10.6) |
| Other | 23 (0.6) | 12 (0.7) |
| Likely to need interpretation, n (%)a | 280 (7.9) | 165 (10.1) |
| PMCA | ||
| Nonchronic | 1227 (33.6) | 567 (32.7) |
| Noncomplex chronic | 1042 (28.6) | 480 (27.7) |
| Complex chronic | 1381 (37.8) | 687 (39.6) |
| Income, $ | ||
| <15 000 | 282 (7.7) | 111 (6.4) |
| 15 000–29 000 | 462 (12.7) | 220 (12.7) |
| 30 000–50 000 | 402 (11.0) | 187 (10.8) |
| >50 000 | 1627 (44.5) | 789 (45.5) |
| Declined or missing | 878 (24.0) | 427 (24.6) |
| Parent highest education, n (%) | ||
| Less than high school | 333 (9.2) | 180 (10.5) |
| High school graduate | 599 (16.6) | 277 (16.1) |
| Some college | 1262 (34.9) | 544 (31.7) |
| Bachelor’s degree or more | 1425 (39.4) | 716 (41.7) |
| Patient age, y, mean (SD) | 6.9 (5.5) | 6.9 (5.5) |
| LOS, d, mean (SD) | 3.1 (5.6) | 3.1 (4.4) |
| PedsQL change score, mean (SD)b | ||
| Total | — | 22.0 (24.1) |
| Physical | — | 28.9 (33.6) |
| Psychosocial | — | 15.1 (21.7) |
| 30-d unplanned readmission, n (%) | 387 (10.6) | 126 (7.3) |
| Cumulative social disadvantage scorec | ||
| 0 | 958 (34.8) | 478 (37.0) |
| 1 | 768 (27.9) | 360 (27.8) |
| 2 | 395 (14.4) | 160 (12.4) |
| 3 | 295 (10.7) | 129 (10.0) |
| 4 | 225 (8.2) | 99 (7.7) |
| 5 | 110 (4) | 67 (5.2) |
PMCA, Pediatric Medical Complexity Algorithm; —, not applicable.
This combines self-reported degree of English proficiency with preferred language for communicating on the basis of survey responses (Table 1).
Only shown for follow-up because the calculation of the change score requires admission and follow-up results.
Shown only for those with complete data (before multiple imputation): N = 2751 for the admission survey and N = 1293 for the follow-up survey.
Phase 1: Caregiver-Reported Barriers to Care Associated With Hospital Outcomes
Nine of 10 barriers-to-care scores were significantly associated with at least 1 hospital outcome (Table 3). Skill barriers and distrust of providers were significantly associated with less improvement in HRQL, although these differences were below the minimal clinically important difference of 4.5 (scale 0–100).27 Three barriers-to-care scores were associated with increased odds of unplanned readmission: the system as a barrier (odds ratio [OR] 1.05; 95% confidence interval [CI] 1.01 to 1.10), expectation barriers (OR 1.08; 95% CI 1.02 to 1.15), and knowledge barriers (OR 1.07; 95% CI 1.10 to 1.14). Eight barriers-to-care scores were associated with increased LOS, with increases ranging from 3% to 12% for each 10-point increase in the reported barrier; this amounts to an additional 2 to 9 inpatient hours beyond a typical LOS. Lack of partnership was not associated with any hospital outcomes.
TABLE 3.
Adjusted Associations Between 10-Unit Change in Caregiver-Reported Barriers and Hospital Outcomes
| Barrier to Care | Hospital Outcomes | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| HRQLa | LOSb | Readmissionc | |||||||
| n | β Coefficient | 95% CI | n | GLM Coefficient | 95% CI | n | OR | 95% CI | |
| System barrier | 1304 | −.34 | −1.02 to 0.35 | 3390 | 1.03d | 1.02 to 1.05 | 3390 | 1.05d | 1.01 to 1.10 |
| Skill barrier | 1257 | −1.74d | −2.67 to −0.81 | 3273 | 1.04d | 1.02 to 1.06 | 3273 | 1.01 | 0.96 to 1.06 |
| Cultural distance | 1318 | −1.85 | −2.05 to 0.35 | 3395 | 1.04d | 1.01 to 1.06 | 3395 | 1.01 | 0.94 to 1.08 |
| Marginalization barriers | 1259 | −.18 | −1.27 to 0.91 | 3225 | 1.06d | 1.04 to 1.09 | 3225 | 1.03 | 0.96 to 1.10 |
| System distrust | 1280 | −.06 | −0.97 to 0.84 | 3292 | 1.11d | 1.01 to 1.22 | 3292 | 1.34 | 0.98 to 1.82 |
| Expectation barriers | 1267 | .30 | −0.69 to 1.30 | 3276 | 1.07d | 1.05 to 1.10 | 3276 | 1.08d | 1.02 to 1.15 |
| Knowledge barriers | 1300 | .46 | −0.60 to 1.51 | 3362 | 1.07d | 1.04 to 1.09 | 3362 | 1.07d | 1.01 to 1.14 |
| Health care self-efficacy | 1247 | .62 | −0.88 to 2.12 | 3216 | 1.12d | 1.04 to 1.20 | 3216 | 1.11 | 0.88 to 1.39 |
| Distrust of providers | 1245 | −1.64d | −2.97 to −0.32 | 1685 | 1.01 | 0.92 to 1.11 | 1685 | 1.12 | 0.81 to 1.54 |
| Lack of partnership | 1262 | −.65 | −1.74 to 0.45 | 1713 | 0.99 | 0.80 to 1.21 | 1713 | 1.14 | 0.53 to 2.45 |
Adjusted for age and medical complexity and clustered on patient. GLM, generalized linear model.
HRQL physical functioning change score (range 0–100) modeled by using linear regression. Caregiver-proxy report was used for patients <13 y old. For patients ≥13 y old, self-report was used when available; otherwise, caregiver-proxy report was used.
LOS winsorized at the 99th percentile and modeled by using a GLM with log-link and γ distribution.
OR for 30-d unplanned readmission modeled by using logistic regression.
Statistically significant result: P < .05.
Phase 2: Identifying Barriers to Care Associated With Social Disadvantage
Of the 9 barriers-to-care scores associated with worse hospital outcomes, 5 were significantly associated with cumulative social disadvantage (Table 4). Four of these had a positive relationship, indicating that greater social disadvantage was associated with greater barriers to care. These included the system as a barrier (adjusted β 1.66; 95% CI 1.02 to 2.30), skill barriers (β 3.82; 95% CI 3.22 to 4.43), cultural distance (β 1.75; 95% CI 1.36 to 2.15), and marginalization barriers (β .71; 95% CI 0.30 to 1.11). One barrier to care, system distrust, had an inverse relationship with social disadvantage, meaning that greater social disadvantage was associated with less distrust of the health care system (β −.33; 95% CI −0.43 to −0.23).
TABLE 4.
Adjusted Association Between Social Disadvantage Scores and Caregiver-Reported Barriers to Care
| Barriers to Care | n | Social Disadvantage Score, β | 95% CI | P |
|---|---|---|---|---|
| System barrier | 3219 | 1.66 | 1.02 to 2.30 | <.001 |
| Skill barriers | 3143 | 3.82 | 3.22 to 4.43 | <.001 |
| Cultural distance | 3221 | 1.75 | 1.36 to 2.15 | <.001 |
| Marginalization barriers | 3103 | .71 | 0.30 to 1.11 | .001 |
| System distrust | 3137 | −.33 | −0.43 to −0.23 | <.001 |
| Expectation barriers | 3121 | .02 | −0.43 to 0.47 | .91 |
| Knowledge barriers | 3196 | .23 | −0.16 to 0.63 | .24 |
| Health care self-efficacy | 3052 | −.12 | −0.25 to 0.01 | .06 |
| Distrust of providers | 1489 | −.15 | −0.3 3 to 0.02 | .09 |
Adjusted for medical complexity and child age and clustered on patient.
Phase 3: Identifying the Relative Contribution of Social Disadvantage Factors to Overall Associations
Across all 5 barriers to care included in phase 3, the various social disadvantage factors we considered clearly had a cumulative effect on the degree of barrier reported, with low income being associated with the most pronounced increases in barriers, as seen in Supplemental Fig 2. Interpreter need, in contrast, was associated with more modest increases or decreases in reported barriers depending on which barrier was being considered. For example, both alone and combined with other factors, interpreter need was associated with increased skill barriers and cultural distance but decreased marginalization barriers and system distrust.
Discussion
In this study of 3651 caregivers of hospitalized children, we found that caregiver-reported barriers to care related to the system, skills, cultural distance, and marginalization were associated with worse hospital outcomes, namely longer LOS and/or increased 30-day readmissions. These barriers were more likely to occur among families with greater social disadvantage. We also found that low income was the strongest predictor of reporting greater barriers to care across categories; interpreter need was associated with less markedly increased, or in some cases decreased, reported barriers to care.
This study adds to existing literature on barriers to care for disadvantaged children by exploring which barriers may contribute to disparate outcomes in the hospital setting. Previous studies have documented the wide range of health care barriers faced by children from low-income, minority, and LEP families, including worse access to primary and specialty care,16,42,43 lower family centeredness and worse communication once in the health care system,19,20 worse comprehension of diagnoses and instructions,44–46 and lower trust in providers.18,47 Each could plausibly contribute to the worse outcomes observed across studies for low-income and minority children.1–10 Our results highlighted 4 specific caregiver-reported barriers as promising for intervention: the system as a barrier, skill barriers, marginalization barriers, and cultural distance.
The system as a barrier was a single, global item reflecting how responsive the health care system has been to the child’s needs.48 We found that caregivers who reported the system as a barrier had children with increased LOS and higher odds of unplanned readmission. This single item could help identify those who would benefit from additional support. Potential interventions include patient navigation,49,50 enhanced care coordination,51–53 or connecting patients to primary care medical homes.54
Skill barriers measured the caregiver’s assessment of their own skills for navigating the health care system and included items related to communication, comprehension, and having enough health care knowledge.48 Skill barriers are associated with caregiver-reported quality of primary care and child HRQL.48 In our study, caregivers reporting skill barriers had children with increased LOS, and these barriers were strongly associated with social disadvantage. Although caregiver skills are not readily modifiable, interventions to decrease the skills required of caregivers would render their current skills adequate for navigating the system. Interventions could include patient navigation, low-literacy educational materials, improved interpreter use, and simplified, streamlined referral processes, among others.
Cultural distance reflected the degree of similarity or difference that caregivers perceived between themselves and their children’s providers.47,55 It framed culture in terms of shared values, communication, and reasoning rather than asking about shared heritage. Cultural distance is associated with trust, quality of care, satisfaction, and intention to adhere along with degree of reported pain in controlled laboratory studies, suggesting a powerful role in influencing experience.47,56,57 We found that caregivers reporting greater cultural distance had children with increased LOS, and degree of cultural distance was strongly associated with social disadvantage. Interventions to bridge cultural distance could include hiring providers and staff to reflect the patient populations, cultural humility and communication training, and culturally concordant patient navigators or patient advocates.
Marginalization barriers reflected caregivers feeling disrespected or otherwise marginalized within the health care system.48 Among children with asthma, marginalization barriers are associated with fewer primary care or subspecialty visits, worse quality of primary care, and worse HRQL.48,58 We found that caregivers reporting marginalization barriers had children with increased LOS. Potential interventions addressing marginalization barriers include equity, diversity, and inclusion training for providers and staff; hiring to promote a diverse workforce; redesigning hospital spaces to make them welcoming and easy to navigate; and patient advocates to help caregivers speak up when they have not been treated with respect.
Low income was the strongest predictor of greater reported barriers to care, alone and across combinations, for each barrier we examined. Tailoring interventions to these families would be expected to lead to greater improvements in experienced barriers and child hospital outcomes.
Interpreter need had the weakest association with reported barriers to care; in some cases, it predicted modest increases, and in other cases, it predicted decreased barriers. Although this may reflect the high-quality interpreter services at the study hospital, it more likely reflects differential survey-response patterns59 and perhaps low expectations. Because most LEP caregivers are immigrants with previous experience of another health care system, they may have lower expectations of the US system and therefore report fewer subjective difficulties. This may also help explain the otherwise surprising finding that greater social disadvantage was associated with lower system distrust in our population, primarily driven by those with interpreter need. This is important to remember when planning intervention evaluations because it may not always be helpful to compare reported barriers between LEP and English-proficient populations.
This study has several limitations. It was conducted at a single, quaternary-care institution, so results may not be generalizable. Participation in hospital-wide survey studies tends to select less disadvantaged families, so these results likely do not reflect the full range of disadvantaged families within our institution, and we were unable to compare respondents to nonrespondents because of institutional review board limitations.29 We allowed repeat participation, which although infrequent could have influenced our results. In addition, the enrollment survey was administered while caregivers were physically in the hospital, which could have led to positive response bias. The social disadvantage score, although useful for capturing the cumulative effects of social disadvantage, is deficit based and does not capture caregiver and family strengths that may mitigate those effects. Finally, we had missing income data for nearly one-quarter of our survey respondents. We used multiple imputation to retain those families in the analysis and account for the resulting uncertainty in our estimates, but additional uncertainty may remain.
Conclusions
In this study, we identified 4 potentially modifiable barriers to care as promising targets for intervention: the system as a barrier, skill barriers, cultural distance, and marginalization barriers. These barriers were associated with both greater social disadvantage and worse hospital outcomes. Low income had the strongest association with reported barriers to care, whereas interpreter need had the weakest, which is consistent with previously described survey-response patterns. Interventions to help the health care system better meet patients and families where they are by bridging cultural differences, decreasing the skills required of families, creating a more inclusive environment, and assisting with system navigation have potential to improve experiences and outcomes for hospitalized children and their families.
Footnotes
The funders had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the article; or decision to submit the article for publication.
Dr Lion conceptualized and designed the study, performed the analyses, and drafted the manuscript; Dr Zhou provided critical input on the study design and analysis and reviewed the manuscript; Drs Ebel and Penfold assisted with the study design, participated in analysis planning and interpretation, and reviewed and revised the manuscript; Dr Mangione-Smith participated in the study design, contributed to analytic planning and interpretation of the results, and critically revised the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grant K23 HD078507; principal investigator: Dr Lion). In-kind support was provided by the Seattle Children’s Hospital Outcomes Assessment Program. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
References
- 1.Lion KC, Wright DR, Desai AD, Mangione-Smith R. Costs of care for hospitalized children associated with preferred language and insurance type. Hosp Pediatr. 2017;7(2):70–78 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lopez MA, Cruz AT, Kowalkowski MA, Raphael JL. Factors associated with high resource utilization in pediatric skin and soft tissue infection hospitalizations. Hosp Pediatr. 2013;3(4):348–354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pati S, Lorch SA, Lee GE, Sheffler-Collins S, Shah SS. Health insurance and length of stay for children hospitalized with community-acquired pneumonia. J Hosp Med. 2012;7(4):304–310 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wallander JL, Fradkin C, Chien AT, et al. Racial/ethnic disparities in health-related quality of life and health in children are largely mediated by family contextual differences. Acad Pediatr. 2012;12(6):532–538 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fedele DA, Molzon ES, Eddington AR, Hullmann SE, Mullins LL, Gillaspy SG. Perceived barriers to care in a pediatric medical home: the moderating role of caregiver minority status. Clin Pediatr (Phila). 2014;53(4):351–355 [DOI] [PubMed] [Google Scholar]
- 6.Quittner AL, Schechter MS, Rasouliyan L, Haselkorn T, Pasta DJ, Wagener JS. Impact of socioeconomic status, race, and ethnicity on quality of life in patients with cystic fibrosis in the United States. Chest. 2010;137(3):642–650 [DOI] [PubMed] [Google Scholar]
- 7.von Rueden U, Gosch A, Rajmil L, Bisegger C, Ravens-Sieberer U. Socioeconomic determinants of health related quality of life in childhood and adolescence: results from a European study. J Epidemiol Community Health. 2006;60(2):130–135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Beck AF, Simmons JM, Huang B, Kahn RS. Geomedicine: area-based socioeconomic measures for assessing risk of hospital reutilization among children admitted for asthma. Am J Public Health. 2012;102(12):2308–2314 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berry JG, Hall DE, Kuo DZ, et al. Hospital utilization and characteristics of patients experiencing recurrent readmissions within children’s hospitals. JAMA. 2011;305(7):682–690 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Beck AF, Huang B, Simmons JM, et al. Role of financial and social hardships in asthma racial disparities. Pediatrics. 2014;133(3):431–439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Anderson BR, Fieldston ES, Newburger JW, Bacha EA, Glied SA. Disparities in outcomes and resource use after hospitalization for cardiac surgery by neighborhood income. Pediatrics. 2018;141(3):e20172432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sobo EJ, Seid M, Reyes Gelhard L. Parent-identified barriers to pediatric health care: a process-oriented model. Health Serv Res. 2006;41(1):148–172 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Seid M, Sobo EJ, Gelhard LR, Varni JW. Parents’ reports of barriers to care for children with special health care needs: development and validation of the barriers to care questionnaire. Ambul Pediatr. 2004;4(4):323–331 [DOI] [PubMed] [Google Scholar]
- 14.Karaca-Mandic P, Choi-Yoo SJ, Lee J, Scal P. Family out-of-pocket health care burden and children’s unmet needs or delayed health care. Acad Pediatr. 2014;14(1):101–108 [DOI] [PubMed] [Google Scholar]
- 15.Ngui EM, Flores G. Satisfaction with care and ease of using health care services among parents of children with special health care needs: the roles of race/ethnicity, insurance, language, and adequacy of family-centered care. Pediatrics. 2006;117(4):1184–1196 [DOI] [PubMed] [Google Scholar]
- 16.Flores G, Tomany-Korman SC. Racial and ethnic disparities in medical and dental health, access to care, and use of services in US children. Pediatrics. 2008;121(2). Available at: www.pediatrics.org/cgi/content/full/121/2/e286 [DOI] [PubMed] [Google Scholar]
- 17.Horn IB, Mitchell SJ, Joseph JG, Wissow LS. African-American parents’ perceptions of partnership with their child’s primary care provider. J Pediatr. 2011;159(2):262–267 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Horn IB, Mitchell SJ, Wang J, Joseph JG, Wissow LS. African-American parents’ trust in their child’s primary care provider. Acad Pediatr. 2012;12(5):399–404 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ashton CM, Haidet P, Paterniti DA, et al. Racial and ethnic disparities in the use of health services: bias, preferences, or poor communication? J Gen Intern Med. 2003;18(2):146–152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Street RL, Jr, Gordon H, Haidet P. Physicians’ communication and perceptions of patients: is it how they look, how they talk, or is it just the doctor? Soc Sci Med. 2007;65(3):586–598 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lion KC, Rafton SA, Shafii J, et al. Association between language, serious adverse events, and length of stay among hospitalized children. Hosp Pediatr. 2013;3(3):219–225 [DOI] [PubMed] [Google Scholar]
- 22.Divi C, Koss RG, Schmaltz SP, Loeb JM. Language proficiency and adverse events in US hospitals: a pilot study. Int J Qual Health Care. 2007;19(2):60–67 [DOI] [PubMed] [Google Scholar]
- 23.Bartlett G, Blais R, Tamblyn R, Clermont RJ, MacGibbon B. Impact of patient communication problems on the risk of preventable adverse events in acute care settings. CMAJ. 2008;178(12):1555–1562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Santry HP, Wren SM. The role of unconscious bias in surgical safety and outcomes. Surg Clin North Am. 2012;92(1):137–151 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Institute of Medicine. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: The National Academies Press; 2003 [PubMed] [Google Scholar]
- 26.Grimshaw JM, Shirran L, Thomas R, et al. Changing provider behavior: an overview of systematic reviews of interventions. Med Care. 2001;39(8,suppl 2):II2–II45 [PubMed] [Google Scholar]
- 27.Varni JW, Seid M, Kurtin PS. PedsQL 4.0: reliability and validity of the Pediatric Quality of Life Inventory version 4.0 generic core scales in healthy and patient populations. Med Care. 2001;39(8):800–812 [DOI] [PubMed] [Google Scholar]
- 28.Varni JW, Burwinkle TM, Seid M, Skarr D. The PedsQL 4.0 as a pediatric population health measure: feasibility, reliability, and validity. Ambul Pediatr. 2003;3(6):329–341 [DOI] [PubMed] [Google Scholar]
- 29.Desai AD, Zhou C, Stanford S, Haaland W, Varni JW, Mangione-Smith RM. Validity and responsiveness of the pediatric quality of life inventory (PedsQL) 4.0 generic core scales in the pediatric inpatient setting. JAMA Pediatr. 2014;168(12):1114–1121 [DOI] [PubMed] [Google Scholar]
- 30.Kourtidou S, Slee AE, Bruce ME, Wren H, Mangione-Smith RM, Portman MA. Kawasaki disease substantially impacts health-related quality of life. J Pediatr. 2018;193:155–163.e5 [DOI] [PubMed] [Google Scholar]
- 31.Simon TD, Haaland W, Hawley K, Lambka K, Mangione-Smith R. Development and validation of the pediatric medical complexity algorithm (PMCA) version 3.0. Acad Pediatr. 2018;18(5):577–580 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Berry JG, Toomey SL, Zaslavsky AM, et al. Pediatric readmission prevalence and variability across hospitals [published correction appears in JAMA. 2013;309(10):986]. JAMA. 2013;309(4):372–380 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Larson K, Russ SA, Crall JJ, Halfon N. Influence of multiple social risks on children’s health. Pediatrics. 2008;121(2):337–344 [DOI] [PubMed] [Google Scholar]
- 34.Bauman LJ, Silver EJ, Stein RE. Cumulative social disadvantage and child health. Pediatrics. 2006;117(4):1321–1328 [DOI] [PubMed] [Google Scholar]
- 35.Appleyard K, Egeland B, van Dulmen MH, Sroufe LA. When more is not better: the role of cumulative risk in child behavior outcomes. J Child Psychol Psychiatry. 2005;46(3):235–245 [DOI] [PubMed] [Google Scholar]
- 36.Atzaba-Poria N, Pike A, Deater-Deckard K. Do risk factors for problem behaviour act in a cumulative manner? An examination of ethnic minority and majority children through an ecological perspective. J Child Psychol Psychiatry. 2004;45(4):707–718 [DOI] [PubMed] [Google Scholar]
- 37.Sameroff AJ, Seifer R, Barocas R, Zax M, Greenspan S. Intelligence quotient scores of 4-year-old children: social-environmental risk factors. Pediatrics. 1987;79(3):343–350 [PubMed] [Google Scholar]
- 38.Sameroff AJ, Seifer R, Baldwin A, Baldwin C. Stability of intelligence from preschool to adolescence: the influence of social and family risk factors. Child Dev. 1993;64(1):80–97 [DOI] [PubMed] [Google Scholar]
- 39.Molenberghs G, Kenward MG. Missing Data in Clinical Studies. Chichester, United Kingdom: John Wiley & Sons, Ltd.; 2007 [Google Scholar]
- 40.Carpenter J, Kenward MG. Multiple Imputation and Its Applications. Chichester, United Kingdom: Wiley; 2013 [Google Scholar]
- 41.White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med. 2011;30(4):377–399 [DOI] [PubMed] [Google Scholar]
- 42.Stevens GD, Seid M, Mistry R, Halfon N. Disparities in primary care for vulnerable children: the influence of multiple risk factors. Health Serv Res. 2006;41(2):507–531 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Seid M, Stevens GD, Varni JW. Parents’ perceptions of pediatric primary care quality: effects of race/ethnicity, language, and access. Health Serv Res. 2003;38(4):1009–1031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Wang NE, Gisondi MA, Golzari M, van der Vlugt TM, Tuuli M. Socioeconomic disparities are negatively associated with pediatric emergency department aftercare compliance. Acad Emerg Med. 2003;10(11):1278–1284 [DOI] [PubMed] [Google Scholar]
- 45.Harris LM, Dreyer BP, Mendelsohn AL, et al. Liquid medication dosing errors by Hispanic parents: role of health literacy and English proficiency. Acad Pediatr. 2017;17(4):403–410 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Glick AF, Farkas JS, Nicholson J, et al. Parental management of discharge instructions: a systematic review. Pediatrics. 2017;140(2):e20164165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Saha S, Sanders DS, Korthuis PT, et al. The role of cultural distance between patient and provider in explaining racial/ethnic disparities in HIV care. Patient Educ Couns. 2011;85(3):e278–e284 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Seid M, Opipari-Arrigan L, Gelhard LR, Varni JW, Driscoll K. Barriers to care questionnaire: reliability, validity, and responsiveness to change among parents of children with asthma. Acad Pediatr. 2009;9(2):106–113 [DOI] [PubMed] [Google Scholar]
- 49.Dohan D, Schrag D. Using navigators to improve care of underserved patients: current practices and approaches. Cancer. 2005;104(4):848–855 [DOI] [PubMed] [Google Scholar]
- 50.Chin MH, Clarke AR, Nocon RS, et al. A roadmap and best practices for organizations to reduce racial and ethnic disparities in health care. J Gen Intern Med. 2012;27(8):992–1000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Weiss MA, Marchese S, Zhang L. Effective care management for children with special health care needs in the era of value-based payment. Clin Pediatr (Phila). 2019;58(9):949–956 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Hamline MY, Speier RL, Vu PD, et al. Hospital-to-home interventions, use, and satisfaction: a meta-analysis. Pediatrics. 2018;142(5):e20180442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Janevic MR, Stoll S, Wilkin M, et al. Pediatric asthma care coordination in underserved communities: a quasiexperimental study. Am J Public Health. 2016;106(11):2012–2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Garg A, Jack B, Zuckerman B. Addressing the social determinants of health within the patient-centered medical home: lessons from pediatrics. JAMA. 2013;309(19):2001–2002 [DOI] [PubMed] [Google Scholar]
- 55.Beach MC, Saha S, Korthuis PT, et al. Patient-provider communication differs for black compared to white HIV-infected patients. AIDS Behav. 2011;15(4):805–811 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Losin EAR, Anderson SR, Wager TD. Feelings of clinician-patient similarity and trust influence pain: evidence from simulated clinical interactions. J Pain. 2017;18(7):787–799 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Street RL, Jr, O’Malley KJ, Cooper LA, Haidet P. Understanding concordance in patient-physician relationships: personal and ethnic dimensions of shared identity. Ann Fam Med. 2008;6(3):198–205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Sadreameli SC, Riekert KA, Matsui EC, Rand CS, Eakin MN. Family caregiver marginalization is associated with decreased primary and subspecialty asthma care in head start children. Acad Pediatr. 2018;18(8):905–911 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Weech-Maldonado R, Elliott MN, Oluwole A, Schiller KC, Hays RD. Survey response style and differential use of CAHPS rating scales by Hispanics. Med Care. 2008;46(9):963–968 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.Karliner LS, Napoles-Springer AM, Schillinger D, Bibbins-Domingo K, Pérez-Stable EJ. Identification of limited English proficient patients in clinical care. J Gen Intern Med. 2008;23(10):1555–1560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Shea JA, Micco E, Dean LT, McMurphy S, Schwartz JS, Armstrong K. Development of a revised Health Care System Distrust scale. J Gen Intern Med. 2008;23(6):727–732 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Tarini BA, Christakis DA, Lozano P. Toward family-centered inpatient medical care: the role of parents as participants in medical decisions. J Pediatr. 2007;151(6):690–695, 695.e1 [DOI] [PubMed] [Google Scholar]
- 63.Maly RC, Frank JC, Marshall GN, DiMatteo MR, Reuben DB. Perceived efficacy in patient-physician interactions (PEPPI): validation of an instrument in older persons. J Am Geriatr Soc. 1998;46(7):889–894 [DOI] [PubMed] [Google Scholar]
- 64.Hall MA, Zheng B, Dugan E, et al. Measuring patients’ trust in their primary care providers. Med Care Res Rev. 2002;59(3):293–318 [DOI] [PubMed] [Google Scholar]
- 65.Street RL., Jr Physicians’ communication and parents’ evaluations of pediatric consultations. Med Care. 1991;29(11):1146–1152 [DOI] [PubMed] [Google Scholar]
- 66.Varni JW, Burwinkle TM, Seid M. The PedsQL as a pediatric patient-reported outcome: reliability and validity of the PedsQL Measurement Model in 25,000 children. Expert Rev Pharmacoecon Outcomes Res. 2005;5(6):705–719 [DOI] [PubMed] [Google Scholar]