Impact of concurrent hospice care on primary care visits among children in southern rural Appalachia

Abstract

Introduction:

The purpose of the study was to test the effect of receiving pediatric concurrent hospice care on primary care visits.

Methods:

This retrospective study was limited to pediatric decedents younger than 21 years with a hospice service claim from 2011–2013. Our outcome of interest concerned whether concurrent hospice care impacted primary care visits.

Results:

Of the 460 pediatric decedents in southern rural Appalachia, 42% continued to visit their primary care provider during hospice enrollment while 51% received concurrent hospice care. Concurrent hospice care was significantly related to pediatric primary care visits (β = 2.31 p <0.001).

Discussion:

Study findings revealed that receipt of concurrent hospice care impacted primary care. Children in concurrent care were twice as likely to continue to receive care from their primary care provider. This finding is consistent with our hypothesis; however, the magnitude of the finding was unexpected given their residence in medically underserved areas.

Introduction

Children in southern rural Appalachia are some of the sickest in our nation, even at end of life (Gitterman et al., 2016). Appalachia is a US region that extends from Georgia to New York across 13 states (Appalachian Regional Commission (ARC), 2022). Southern Appalachia in particular includes Alabama, Virginia, Georgia, West Virginia, Tennessee, Kentucky, Mississippi, South Carolina, and North Carolina (ARC, 2022). Lack of access to healthcare services starts at birth with limited prenatal care resources and continues to end of life with restricted access to pediatric hospice care (Mixer et al., 2014). As one of the most medically underserved areas (MUA) in the nation, the region has provider shortages and limited healthcare infrastructure with significant care fragmentation (Fornehed et al., 2022). MUAs are defined as areas within the US where there are not sufficient primary care providers to meet the needs of that community. These communities typically have high poverty and infant mortality rates (HRSA, 2022b).

What is known about primary care at end of life has focused mostly on adult patients (Ankuda et al., 2017; Ding et al., 2022; Gao et al., 2020; Howard et al., 2017; Kim & Tarn, 2016; Kinder et al., 2022; Liu et al., 2019; Mitchell et al., 2022). Although this literature suggests that maintaining relations with a primary provider during hospice or palliative care enrollment might provide a critical link to continuous care at end of life (Kim & Tarn, 2016; Liu et al., 2019), few studies have examined the issue among pediatric patients. In a 2005 study conducted in Massachusetts, Goldstein and Rimer (2013) interviewed bereaved parents and discovered that primary care involvement was very limited at end of life for children. Among pediatric Medicaid beneficiaries in California with a life-limiting condition during 2007 to 2010, Lindley and Nageswaran (2017) reported that adolescents between 15–20 years with a usual source of primary care were more likely to use hospice care.

What is unknown about primary care at end of life for children is still considerable, especially given the implementation of pediatric concurrent hospice care (Affordable Care Act (ACA), section 2302) that enables Medicaid beneficiaries under 21 years to receive medical care for their terminal illness while in hospice care. Although it is not common in adult hospice, pediatric patients in hospice have historically maintained a relationship with their primary care providers (Fornehed et al., 2022). Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) services are an example of preventative care that is still available to children in standard hospice (Medicaid.gov, 2022). Concurrent hospice care expanded the opportunity for engagement with primary care as a provider of treatments, medications, and therapies for children at end of life. No studies could be identified that examined the impact of concurrent care on primary care, especially in medically underserved areas such as rural southern Appalachia.

Understanding the relationship between concurrent hospice care and primary care visits is clinically relevant. Increasing awareness of the federally-mandated concurrent hospice care is important for nurses providing primary care. Additionally, advance practice nurses may need further education in pediatric end-of-life care such as End of Life Nursing Education Consortium (ELNEC) in order to be able to provide quality coordinated interdisciplinary pediatric hospice care. Therefore, the purpose of the study was to test the effect of receiving concurrent hospice care on primary care visits for children in southern rural Appalachia.

Methods

This study was a retrospective, quasi-experimental study using administrative data. The sample was limited to pediatric decedents who were younger than 21 years with a hospice service claim between January 1, 2011 to December 31, 2013. Children were included if they resided in the southern region of Appalachia: Alabama, Virginia, Georgia, West Virginia, Tennessee, Kentucky, Mississippi, South Carolina, and North Carolina (ARC, 2021). Rural residence using home zip codes was also an inclusion criterion (Health Resources & Services Administration (HRSA), 2022a). Children with no date of birth, date of death, or service level Medicaid claims were excluded. The study was reviewed and approved by the University of Tennessee, Knoxville Institutional Review Board.

National Medicaid data files covered from 2011 to 2013. Data on community characteristics were added from the 2010 American Community Survey of the U.S. Census Bureau. Data about hospice providers was added from the CMS Hospice Provider of Services files and Hospice Utilization and Payment files.

The main outcome variable of primary care was whether or not children visited a primary care provider during hospice care. The Medicaid files were used to identify primary care visits. The independent variable was concurrent versus standard hospice care. Concurrent care was defined as simultaneously obtaining non-hospice medications on the same day as hospice care based on Medicaid claims activity dates (Mor et al., 2016, 2019). Standard care was defined as receiving no non-hospice medications during hospice enrollment. Children’s demographic characteristics comprised age, gender, race (including Caucasian, Black, or Other), complex chronic condition, comorbidities, mental/behavioral health conditions, and technology dependence. Hospice characteristics were organizational size measured in the number of employees (less than 50 or more than 50 employees), ownership (for-profit or non-profit/government), years of operation, and pediatric program. Community characteristics included educational level (no high school, high school, or more than high school) and median household income (< $50,000/yr or ≥ $50,000/yr).

Data Analysis

Descriptive statistics were calculated, and the analysis was conducted using instrumental variable analysis (IVA) (Brookhart, 2010). The instrumental variable was the number of medical providers seen upon hospice admission (i.e., medical team size). To estimate the relationships between concurrent care and primary care visits, a two-stage Least Square Method (2SLS) estimator was used (Angrist & Pischke, 2009). The first step of the 2SLS equation assessed the validity of using medical team size as the instrument. The second step equation estimated the effect of enrollment in concurrent hospice care vs. standard hospice care on visiting primary care providers.

Results

Table 1 presents the characteristics of children in the sample. The final cohort in the study included 460 pediatric decedents in southern rural Appalachia. Over forty percent of the sample (42.4%) continued to visit their primary care provider during hospice enrollment, while 50.9% of the children received concurrent hospice care. The average age of the children was 8 years. The most common race was White (52.2%) and least common was Black (15.2%). Eighty-six percent of children had complex chronic conditions and were technology dependent (55.2%). Comorbidities were present in 70.8% children with mental and behavioral health problems in half of the sample (54.5%). Among the hospices that provided care for the children, nearly half were large hospices (48%), with one-third (33%) having a dedicated pediatric hospice program. The area where these children resided was characterized by 36.5% having at least some high school education, while 83.26% resided in communities with an income range of <$50K/year.

Table 1.

Sample Characteristics of Southern Rural Appalachian Children (N = 460)

Variables	No.	% or Mean (std)
Dependent Variable
Primary care visits	195	42.39%
Independent Variable
Concurrent hospice care	234	50.87%
Covariates
Demographic Characteristics
Age (years)	460	8.27 (7.14)
Gender - female	225	48.91%
Race
White	240	52.17%
Black	70	15.22%
Other	150	32.61%
Complex chronic condition	396	86.09%
Comorbidities	326	70.87%
Mental/behavioral health conditions	251	54.57%
Technology dependence	254	55.22%
Hospice Characteristics
Hospice Size - >50 employees	221	48.04%
Hospice ownership - non-profit/government	194	42.17%
Hospice operation (years)	460	22.63 (7.52)
Hospice pediatric program	152	33.04%
Community Characteristics
High school education	168	36.50%
No high school education	96	20.95%
More than high school education	199	42.55%
Household income <$50K	383	83.26%

Note: STD, standard deviation

The correlation results between the instrumental variable and the exposure are displayed in Table 2. The first-stage equation found that the instrumental variable was highly correlated with the independent variable. Medical team size was significantly related to concurrent hospice care (β = 0.05, p< 0.001). The test of the instrumental variable strength, which was run simultaneously with the first-stage equation, suggested that medical team size was a strong and valid instrument with a statistically significant F-statistic > 10 and partial R² < 0.05.

Table 2.

Assessment of Instrumental Variable – First Stage Equation (N=460)

Variables	Concurrent Hospice Care Use β (95%CI)
Instrument
Medical Team Size	0.05 (0.064, 0.076)***
Covariates
Demographic Characteristics
Age	0.01 (0.001, 0.003)***
Female	−0.01 (−0.024, 0.006)
Caucasian	0.01 (−0.019, 0.023)
Black	−0.01 (−0.029, 0.011)
Health Characteristics
Complex Chronic Condition	−0.05 (−0.065, −0.034)***
Comorbidities	0.02 (−0.016, 0.051)
Mental/Behavioral Conditions	0.01 (−0.018, 0.020)
Technology Dependence	0.01 (−0.033, 0.044)
Hospice Characteristics
Size >50 employees	−0.01 (−0.037, 0.008)
Ownership Non-profit/Government	0.05 (0.026, 0.069)***
Years of Operation	0.01 (−0.001, 0.002)
Pediatric Program	0.01 (−0.011, 0.034)
Community Characteristics
No High School Education	−0.01 (−0.004, 0.001)
High School Education	−0.01 (−0.002, 0.001)
Household Income ≤ $50,000/yr	0.06 (0.039, 0.083)***
F-statistic	>10
Partial R2	0.05

^*p< 0.05,

^**p <0.01,

^***p< 0.001

Note: CI, Confidence Intervals

Note: All analyses controlled for study years

Table 3 shows the results of the instrumental variable regression analysis estimating the effect of concurrent hospice care on primary care visits. From this regression analysis, concurrent hospice care was significantly related to primary care visits (β = 2.31, p <0.001). Significant covariates related to primary care visits included younger age children (β = −0.03, p < 0.05) and Caucasian children (β = 0.30, p < 0.05), non-profit/government hospice services (β = −0.42, p < 0.05), and hospice services with pediatric programs (β = 0.40, p <0.05). No other covariates were associated with primary care visits.

Table 3.

Instrumental Variable Regression Analysis (N=460)

Variables	Primary Care Visits β (SE)
Independent Variable
Concurrent Hospice Care	2.31 (0.18)***
Demographic Characteristics
Age	−0.03 (0.01)**
Gender	0.03 (0.11)
Caucasian	0.30 (0.11)**
Black	0.15 (0.16)
Health Characteristics
Complex Chronic Condition	−0.29 (0.22)
Multiple Complex Chronic Conditions	0.33 (0.21)
Mental/Behavioral Conditions	0.15 (0.15)
Technology Dependence	−0.14 (0.16)
Hospice Characteristics
>50 employees	0.09 (0.14)
Non-profit/Government	−0.42 (0.15)**
Years of Operation	−0.01 (0.01)
Pediatric Program	0.40 (0.14)**
Community Characteristics
No High School Education	0.02 (0.01)
High School Education	−0.02 (0.01)
≤ $50,000/yr Household Income	0.18 (0.22)

Note:

^*p< 0.05,

^**p <0.01,

^***p< 0.001

Note: SE, Standard Error

Note: All analyses controlled for study years

Discussion

As one of the first studies to examine children in hospice care in southern rural Appalachia, the children in this region presented with a unique profile. Decedents were relatively young and suffered with severe medical complexities, even at end of life. They had a higher rate of multiple comorbidities, mental/behavioral health problems, and technology dependence such as ventilation compared to the national average for children in hospice care (Keim-Malpass et al., 2021). The children in southern rural Appalachia also resided in communities with significant poverty and low education.

The study findings revealed that receipt of concurrent hospice care impacted primary care. Children in concurrent care were twice as likely to continue to receive care from their primary care provider than their peers in standard care. Although this finding is consistent with our hypothesis, the magnitude of the finding was unexpected given that Appalachian children in the study resided in medically underserved areas (HRSA, 2022b). One possible explanation for the finding is that many primary care providers in this region often must perform obstetric services and deliver babies. With the foundation of the relationship at birth, children and their families will continue to receive care from these primary care providers well into their adolescent years. Additionally, it is common for primary care clinicians who have grown up in these rural communities to come back where their extended families live and enjoy a broad scope of practice (MacQueen et al., 2018; Nasim et al., 2021; Rosenblatt, 2004; Walker et al., 2021). These health care professionals may even have family in the community they return to, making them a trusted part of the family (Abernathy & Byerley, 2019; Johnson et al., 2018; MacQueen et al., 2018). More research is needed to reveal the determinants of provider migration of providers into or out of the southern rural Appalachian communities.

Another interesting finding was that as children aged, they were less likely to visit their primary care clinicians in concurrent care versus standard hospice care. This finding was expected and consistent with data on primary care utilized about adolescents in the general populations (Rand & Goldstein, 2018). Many rural health systems provide core primary health care services needed by younger children even at end of life, thus relieving the family from costly travel to the hospital, decreasing the burdens of losing time from work, and eliminating long wait times at the hospital (Habeger & Venable, 2018; Rosenblatt, 2004).

This study had several limitations. First, the results may not be generalized to pediatric non-Medicaid populations. There may be different findings among children and adolescents with private or commercial insurance. However, Medicaid does insure most children who have life-limiting illness in the United States (Keim-Malpass et al., 2021). Second, these data may also not be generalized to children outside the region of southern rural Appalachia. Nasim et al. (2021) revealed a practice difference in urban areas of Appalachia when compared to rural areas. Rural community family practice providers would have a greater tendency than their urban counterparts to continue with a broader or even an increase in scope of practice to care for the rural community. Finally, this study is based on data from 2011 to 2013. Rural Appalachian family and provider views of care of patient at end of life may have changed since that time. However, evidence suggests that even in the year 2020, there is very limited to no access to pediatric hospice care in this region of the country (Mixer et al., 2014).

In summary, a rigorous, head-to-head evaluation of two health care delivery systems to improve care consistency for children at end of life was conducted. Pediatric concurrent hospice care was positively related to continued visits with primary care. These findings have implications for pediatric nurse practitioners caring for this population.