Microsoft Word - Document5
A Mixed Methods Analysis of the Place-Related Risk and Protective Factors for
Hospital Utilization among Children with Complex Chronic Conditions
Shannon M. Hudson, PhD, RN, Alumnus CCRN 1
Gayenell S. Magwood, PhD, RN, Alumnus CCRN 2
Marilyn A. Laken, PhD, RN, FAAN 3
Martina Mueller, PhD 4
Susan D. Newman, PhD, RN, CRRN 5
1Assistant Professor, College of Nursing, Medical University of South Carolina,
hudsonsm@musc.edu
2 Associate Professor, College of Nursing, Medical University of South Carolina,
magwoodg@musc.edu
3 Professor, College of Nursing, Medical University of South Carolina,
lakenm@musc.edu
4 Associate Professor, College of Nursing, Medical University of South Carolina,
muellerm@musc.edu
5 Associate Professor, College of Nursing, Medical University of South Carolina,
newmansu@musc.edu
Abstract
Purpose: Children with complex chronic conditions (CCC) have high health care needs
and utilization. The purpose of this study was to identify place-related risk and protective
factors associated with hospital utilization among children with CCC.
Methods: A mixed methods study was conducted to generate a risk profile of place-
related factors associated with hospital utilization. The quantitative phase of the study
consisted of retrospective review of 216 urban hospital medical records of infants and
young children with CCC. The qualitative phase of the study included interviews with
parents/caregivers and focus groups with health care providers (HCPs) of children with
CCC. Results from multiple regression modeling and directed content analysis were
merged using a side-by-side table organized by ecological level.
Findings: Place-related risk and protective factors on multiple ecological levels were
identified. Key place-related factors associated with hospital utilization were more
complex conditions, positive relationships with HCPs, more parent/family resources, and
having fewer place-related resources.
Conclusions: The results of this study suggest patterns of health care utilization among
infants and young children with CCC are influenced by place-related factors. Parent
relationships with HCPs and comfort with care, hospital resources, and parent resources
determine location of care. Results of this study also imply place-related disparities in
access to care, especially among children in smaller metropolitan areas and rural-
dwelling children.
Keywords: rural population; rural health services; infant; child, preschool; chronic disease
A Mixed Methods Analysis of the Place-Related Risk and Protective Factors for
Hospital Utilization among Children with Complex Chronic Conditions
Children with complex chronic conditions (CCC) are a vulnerable population of
children who, with their families, face increased emotional, psychological, physical and
financial difficulties. Children with CCC form a subpopulation of children with special
health care needs (McPherson et al., 1998), and are described as having a health
condition expected to last at least 12 months that substantially affects one or more body
systems. These conditions lead to the need for pediatric specialty care and hospitalization
(Feudtner, Christakis, & Connell, 2000), resulting in high health care utilization rates and
associated health care costs. For example, children with CCC have longer pediatric
emergency department (ED) length of stay and higher hospital and pediatric intensive
care unit (PICU) admission rates (O'Mahony et al., 2013). Health care utilization rates
among children with CCC significantly increased between 1997 and 2006, and in 2006
hospitalizations among children with CCC accounted for 26% of pediatric inpatient days
and 41% of pediatric hospital charges (Simon et al., 2010). Moreover, issues pertaining to
location such as geographic proximity to health care, access to care, and availability of
resources may influence health care utilization patterns among children with CCC
(Pollack et al., 2004; Yamamoto et al., 1995).
Identifying place-related factors associated with health care utilization is critical
since rural-dwelling children are more likely to have unmet health care needs, at least one
ED visit within a year (DeVoe, Krois, & Stenger, 2009), higher asthma hospitalization
rates (Knudson, Casey, Burlew, & Davidson, 2009), higher use of medical settings for
dental needs (Martin, Vyavaharkar, Veschusio, & Kirby, 2012), and limited access to
pediatric subspecialty providers (Mayer, 2008). Rural areas also see higher infant
mortality rates and higher rates of obesity in children (Wright et al., 2008). Therefore,
this study was designed to address the research question: What multi-level place-related
factors put children with CCC at risk for or protect them against hospital admissions or
ED visits?
Theoretical Framework
The risk and protective factors model (Hawkins, Catalano, & Miller, 1992) and
social ecological theory (SET) (Bronfenbrenner, 1994) served as the theoretical
frameworks for this study. According to the risk and protective factors model, certain
aspects of an individual or the environment lead to an increased likelihood of a particular
outcome (risk factors) while other aspects are associated with a lesser likelihood of the
outcome (protective factors) (Hawkins et al., 1992; Nagy & Fawcett, 2011). Risk factors
have varying degrees of modifiability and some, such as race/ethnicity, are impossible to
change. According to this model, protective factors mediate or moderate the effect of the
risk factor on the outcome (Hawkins et al., 1992). In fact, among persons with similar
risk factors, those with protective factors have improved outcomes. The domains of
factors in this framework are: individual/peer, school, family, and community.
The risk and protective factors (RPF) model has been widely used to address drug
use and other risk behaviors in adolescents and has guided the development of a
community program for addressing adolescent health behavior (Hawkins et al., 2008).
The model has also been used as the theoretical framework in studies investigating
hospital readmission behaviors (Bachrach, Schwarz, & Bachrach, 2003; Bulkow,
Singleton, Karron, Harrison, & the Alaska RSV Study Group, 2002; Garcia-Aymerich et
al., 2000; Garcia-Aymerich et al., 2003; Garcia-Aymerich et al., 2004; Garcia-Aymerich
et al., 2001; Guevara, Young, & Mueller, 2001). However, no study was identified in
which the RPF model was used to investigate hospital admissions and ED visits in
children with CCC. In this study, aspects pertaining to place, such as rural residence and
proximity to care, were assessed as risk or protective factors for hospital admissions or
ED visits in infants and young children with CCC.
Social ecological theory was used in conjunction with the RPF model during the
qualitative analysis in this study. According to SET, drivers of behavior exist on multiple
levels: intrapersonal, interpersonal, family, community, society, and political. Additional
tenets of SET posit that behavioral influences on each level interact and successful
behavior change strategies address influences on multiple levels (Sallis, Owen, & Fisher,
2008). Interview and focus group questions were formulated using the key concepts of
the RPF model and SET, and during data analysis, initial coding categories were created
using the guiding frameworks. In addition, particular attention was paid during data
analysis to quotes and phrases that indicated differences or disparities pertaining to place.
Methods
Design
A convergent parallel mixed methods study was conducted using simultaneous,
independent quantitative and qualitative methods. The quantitative phase of the study was
conducted using a retrospective cohort design with medical record review. During the
qualitative phase of the study, key informant interviews with parents/primary caregivers
(subsequently referred to as “parents”) and focus groups with health care providers
(HCPs) were conducted using qualitative descriptive design and analyzed using directed
content analysis. While the purpose of the larger study was to investigate the risk and
protective factors associated with hospital admissions and ED visits in children with
CCC, the focus of this sub-study was to identify place-related factors associated with
hospital utilization in infants and young children with CCC. Findings from the larger
study are presented elsewhere; this sub-study was conducted with a focus on the place-
related variables and the place-related qualitative data. The primary objectives were to 1)
investigate the role of urban versus rural residence on hospital utilization, 2) explore the
perceptions of parents and HCPs regarding place-related factors associated with hospital
utilization in children with CCC, and 3) organize issues pertaining to location into an
ecological risk and protective factors model.
Setting. The study was conducted at a 453-bed level III trauma regional hospital
that serves a 15-county area of South Carolina (McLeod Health, 2012) with more than
875,000 residents (United States Census Bureau [USCB], 2013). The hospital is the main
provider of acute care pediatric services in the area, with a pediatric general care floor, a
pediatric intensive care unit, and some pediatric specialty providers, such as pediatric
intensivists and a pediatric cardiologist. There is no freestanding pediatric hospital in the
region. This setting was ideal for assessing the possible effects of location on hospital
admissions and ED visits because several community hospitals in surrounding rural
counties transfer higher-needs patients to the urban hospital.
The US Office of Management and Budget’s (OMB’s) definition of rurality was
adopted for this study. According to this definition, a metropolitan area has at least
50,000 residents, a micropolitan area has an urban center of at least 10,000 but not more
than 50,000 residents, and a non-metropolitan area has fewer than 10,000 residents in a
central area (USCB, n.d.). Micropolitan and non-metropolitan statistical areas are
considered rural, while metropolitan statistical areas are considered urban. This study was
conducted in an urban county primarily surrounded by rural counties.
Sample. Medical records of children with CCC born between January 1, 2008 and
November 25, 2011 were reviewed. Records for children having an index hospitalization
with a diagnosis on the list of complex chronic conditions outlined by Feudtner and
colleagues (2000) along with at least one additional hospital admission or ED visit within
365 days were selected for further analysis. Additional details can be found in Hudson,
Mueller et al. (2014a).
Any parent who was at least 18 years of age and was the caregiver of a child born
between January 1, 2008 and November 25, 2011 with a diagnosis that met the definition
of complex chronic condition (Feudtner et al., 2000) and at least one additional
hospitalization or ED visit was recruited to participate in key informant interviews. HCPs
working locally in pediatric acute care, pediatric primary care, or emergency care were
recruited to participate in focus groups. Participating HCPs included physicians, nurses,
ancillary care providers, and other disciplines; each HCP had worked for at least 1 year
caring for children with CCC, and directly cared for children with CCC in their current
role.
Data Collection. Data collection processes are presented in greater detail in the
reports of the quantitative (Hudson, Mueller, et al., 2014a) and qualitative (Hudson,
Newman, et al., 2014b) phases of the larger study. A hand search of the medical records
was performed to identify a cohort of children with CCC born between January 1, 2008
and November 25, 2011 (n = 216). The demographic and clinical data collected included:
age, sex, race/ethnicity, type of insurance, county of residence, zip code of residence,
technological dependence, primary diagnosis for each admission, and number and type of
secondary diagnoses for each admission. Measures of hospital admissions and ED visits
collected were: index (or initial) hospitalization admission date, index hospitalization
discharge date, number of additional hospital admissions and/or ED visits, and dates for
each subsequent hospitalization or ED visit.
Qualitative data collection was carried out using key informant interviews and focus
groups. Questions specific to location such as perceptions of the roles of transportation,
distance to health care providers, and community resources were included in the
interview and focus group interview guides. The findings from interviews with parents
and focus groups with HCPs were compared, as were findings from interviews with rural-
dwelling and urban-dwelling parents. Institutional Review Board (IRB) approval was
obtained from the study site and the principal investigator’s (PI’s) institution
(Pro00016037) prior to data collection.
Data Analysis. Data accuracy was ensured by verifying data entry with the source
document, and by checking for missing data and outliers. Descriptive statistics were
calculated to assess the characteristics of rural-dwelling and urban-dwelling children with
CCC. Linear regression analyses were conducted to examine whether associations were
present between urban versus rural residence and measures of hospital utilization.
Outcome variables were: length of stay (LOS) during the index hospitalization, number
of hospital admissions and/or ED visits (occurrences) following the index hospitalization,
number of days between the index hospitalization and the first subsequent hospital
admission or ED visit, and mean number of days between all subsequent hospital
admissions and/or ED visits. After bivariate analyses between urban versus rural
residence and each outcome variable, multiple regression analyses were conducted to
determine whether the addition of urban versus rural residence influenced the relationship
with predictors and outcome measures. Multiple regression analyses were conducted with
the set of demographic and clinical variables with and without the residence variable as
possible predictors of the outcome variables to evaluate differences. Stepwise linear
regression modeling with backward deletion was performed with the level of significance
for variable entry set at p < .10, and the level of significance for variable removal set at p
> .15 to reduce the probability of a type II error.
Qualitative data were analyzed using directed content analysis (Hsieh & Shannon,
2005) guided by an ecological risk and protective factors framework. To begin the
analysis, initial coding categories were chosen using the ecological risk and protective
factors framework that emerged during the qualitative phase of the larger study (Hudson,
Newman, et al., 2014b). Quotes that seemed to fit into a category were highlighted on the
interview or focus group transcript. Transcripts were reviewed again, and highlighted
segments were coded into one or more categories. Phrases associated with place,
location, or residence were sought. Lincoln and Guba’s (1985) framework for developing
trustworthiness was adopted to enhance the quality of this portion of the study. The PI
kept a field journal and recorded personal influences to promote reflexivity. Subject
checking was conducted to verify findings.
Results
Demographic and clinical characteristics of rural- versus urban-dwelling children
from the medical record review (n = 216) are presented in Table 1. Urban-dwelling and
rural-dwelling groups differed in hospital utilization. Rural-dwelling children had a
significantly longer average index hospitalization LOS than urban-dwelling children but
statistically significantly fewer occurrences following the index hospitalization.
Characteristics of hospital utilization by residence are outlined in Table 2.
Table 1
Medical Record Sample Demographic and Clinical Data by Urban versus Rural Status
Variable Urban
N = 144
Proportion (n) or Mean +
SD (Range)
Rural
N = 70
Proportion (n) or
Mean + SD (Range)
p-value
Age
Age (in months) 31.5 + 13.1 (7.0- 52.0) 31.9 + 13.8 (6.0- 52.0) .72
Gender
Male 48.6 (70) 52.9 (37) .56
Female 51.4 (74) 47.1 (33) .56
Race/ethnicity
White 31.9 (46) 42.9 (30) .12
Black 63.9 (92) 50.0 (35) .05
Asian/Pacific Islander 0.7 (1) 0 .49
Native American/Alaskan
Native
0 2.9 (2) .04
Hispanic or Spanish
origin
2.1 (3) 2.9 (2) .73
Other 1.4 (2) 1.4 (1) .98
Procedures
None 79.2 (114) 77.1 (54) .74
Tracheostomy 2.8 (4) 2.9 (2) .97
Gastrostomy 14.6 (21) 15.7 (11) .83
Permanent indwelling
catheter
2.1 (3) 1.4 (1) .74
Ventriculoperitoneal shunt 5.6 (8) 7.1 (5) .65
Index admission primary diagnosis category
Neuromuscular 6.3 (9) 5.7 (4) .88
Cardiovascular 1.4 (2) 2.9 (2) .46
Respiratory 11.1 (16) 15.7 (11) .34
Renal 0 1.4 (1) .15
Hematological/
Immunological
2.1 (3) 1.4 (1) .74
Metabolic 0 4.3 (3) .01
Perinatal/birth 72.0 (105) 67.1 (47) .38
Other 6.3 (9) 1.4 (1) .12
Index hospitalization
number of secondary
diagnoses
6.4 + 5.4 (0- 29)
Median = 5.0
7.5 + 5.9 (0- 21)
Median = 7.0
.27
Per patient mean number of
secondary diagnoses
(subsequent events)
1.0 + 1.5 (0- 13.5)
Median = 0.6
1.2 + 1.4 (0- 8)
Median = 1.0
.16
Insurance
Public 88.2 (127) 91.4 (64) .47
Private 15.3 (22) 14.3 (10) .85
Self-pay 2.8 (4) 1.4 (1) .54
Number of types of insurance
One type (or self-pay) 93.8 (135) 92.9 (65) .80
Two types 6.3 (9) 7.1 (5) .80
Family status
(n = 103) (n = 49)
Single-parent 46.6 (48) 46.9 (23) .95
Two-parent 48.5 (50) 49.0 (24) .95
Other relative 2.9 (3) 4.1 (2) .73
Other non-relative 1.9 (2) 0 .32
Number of siblings (n = 95)
1.3 + 1.4 (0- 9)
(n = 47)
1.6 + 1.9 (0- 7)
.74
The parents of 11 children expressed interest in participating, and the parent(s) of 8
children (n = 13) were enrolled in the study; 27 HCPs provided informed consent, and 24
participated in focus groups. The demographic profiles of parents and HCPs are
presented in Hudson, Newman, et al. (2014b). The majority of parents resided in an urban
county (84.6%, 11/13); however, 36.3% (4/11) of these parents described living in more
sparsely populated areas of the county. Characteristics of hospital utilization by
residence are outlined in Table 2.
Table 2
Hospital Admissions and ED Visits by Urban versus Rural Status
Variable Urban
N = 144
Proportion (n), or
Mean + SD (Range)
Rural
N = 70
Proportion (n), or
Mean + SD (Range)
p-value
Number of occurrences
following index
hospitalization
3.1 + 2.8 (1-16)
Median = 2.0
2.2 + 1.9 (1-8)
Median = 1.5
.02
Number of days between
index hospitalization
discharge and first subsequent
occurrence
N = 144
115.2 + 100.0 (1-357)
Median = 75.0
N = 70
106.9 + 98.7 (1-356)
Median = 73.0
.48
Per patient mean number of N = 90 N = 31 .68
days between subsequent
occurrences
108.1 + 81.5 (.5-
330.0)
Median = 94.0
117.6 + 87.1 (5.0-
308.0)
Median = 88.2
Index hospitalization LOS in
days
22.4 + 31.3 (0-117)
Median = 4.0
37.2 + 43.5 (1-194)
Median = 14.0
.003
Merged findings by ecological level
Qualitative and quantitative findings were merged using side-by-side analysis for
convergence and divergence. Merged findings are presented as follows by ecological
level.
Individual child. Parents and HCPs perceived that the complexity and severity of
the child’s condition influenced the location of care. Parents also perceived that the
complexity of their child’s condition mandated they bring their child to a particular ED or
hospital. A father described one of the reasons he chose to bring his daughter to a
preferred urban hospital: “I don’t think at [hospital name], I don’t think they have a
special floor for pediatrics and taking care of a baby that has [a CCC].” Results from
multiple regression analyses supported these beliefs. Reliance on a technological device
(such as a tracheostomy or gastrostomy tube) was included in the prediction models for
the number of occurrences that followed the index admission (R2 = .28, F(9, 203) =
8.56, P < .001) (Table 3), the number of days between the index admission and the first
subsequent occurrence (R2 = .10, F(5, 207) = 4.47, P = .001) (Table 4), and the mean
number of days between occurrences that followed the index admission (R2 = .16, F(5,
122) = 4.58, P = .001) (Table 5).
Table 3
Coefficients for the Final Model for Predictors of the Number of Occurrences Following
the Index Hospitalization
Predictor B Std. error t p Bivariate
r
Partial r
Constant 3.556 0.687 5.173 .022
Rural versus
urban
-0.932 0.329 -2.834 .005 -.158 -.195
Other
race/ethnicity
-1.169 0.707 -1.655 .100 -.088 -.115
Neurological
diagnosis
1.358 0.659 2.061 .041 .152 .143
Cardiovascular
diagnosis
2.824 1.130 2.498 .013 .134 .173
Other
diagnosis
-1.119 0.731 -1.530 .128 -.060 -.107
Age in months 0.057 0.012 4.913 <.001 .340 .326
No
technological
device
-1.598 0.380 -4.201 <.001 -.269 -.283
Private
insurance
-0.918 0.431 -2.131 .034 -.114 -.148
Self-pay 1.646 1.015 1.621 .107 .098 .113
Table 4
Coefficients for the Final Model for Predictors of the Number of Days between the Index
Hospitalization and the First Subsequent Occurrence
Predictor B Std. error t P Bivariate
r
Partial
r
Constant 93.327 24.932 3.743 <.001
Rural versus urban -6.422 14.020 -0.458 .647 -.037 -.032
Hematology/Immunology
diagnosis
-104.215 48.202 -2.162 .032 -.122 -.149
Mean number of
secondary diagnoses
during subsequent
occurrences
-51.325 30.338 -1.692 .092 -.152 -.117
No technological device 55.753 16.435 3.392 .001 .238 .229
Self-pay -66.854 43.674 -1.531 .127 -.062 -.106
Table 5
Coefficients for the Final Model for Predictors of the Mean Number of Days between
Subsequent Occurrences
Predictor B Std. error t p Bivariate
r
Partial r
Constant 91.780 73.582 1.247 .215
Rural versus
urban
45.871 27.586 1.663 .099 .091 .149
Mean number of
secondary
diagnoses
during
subsequent
occurrences
-135.387 68.912 -1.965 .052 -.203 -.175
Age in months 2.443 1.004 2.443 .016 .179 .215
No
technological
device
54.642 30.493 1.792 .076 .182 .160
Public insurance -91.586 40.045 -2.287 .024 -.225 -.203
Thus, children with CCC severe enough to require a technological device had more
hospital admissions or ED visits (occurrences) and fewer days between occurrences.
Number of secondary diagnoses was statistically significant in the prediction models for
the number of days between occurrences following the index admission and for the index
admission LOS (R2 = .73, F (5, 207) = 113.56, P < .001) (Table 6). These results further
supported parents’ and HCPs’ beliefs by indicating children with a higher number of
secondary diagnoses during hospitalizations had fewer days between occurrences and a
longer index hospitalization LOS than those with fewer secondary diagnoses.
Parent-child relationship. Parents and HCPs frequently mentioned the influence of
parents’ preferences on location of care. According to HCPs, this preference can place
children at risk for greater utilization because parents spend time driving from distant
locations during which the child’s condition deteriorates. But parents believed spending
the extra time driving to a preferred hospital protected children because urban centers
have needed equipment and personnel with expertise.
Table 6
Coefficients for the Final Model for Predictors of the Index Hospitalization Length of
Stay
Predictor B Std. error t p Bivariate
r
Partial r
Constant -31.760 5.844 -5.435 <.001
Rural versus
urban
6.137 2.653 2.313 .022 .169 .159
No
technological
device
5.754 3.138 1.834 .068 -.108 .126
Neurological
diagnosis
9.661 5.420 1.783 .076 -.153 .123
Number of
index
hospitalization
secondary
diagnoses
5.625 0.248 22.663 <.001 .842 .844
Age in months 0.247 0.094 2.632 .009 .056 .180
One mother described how she decides whether to travel to a more distant urban center:
When I’m in the car with her, I’m, like, ‘Can I make it? Do I need to go to [the
rural ED]?’ I think if it came to the point where I thought, okay, this is enough and
I was by [the rural hospital] she would go to [the rural hospital]. So I just feel, I
guess, comfortable enough to drive the extra 20 minutes and get to where I know
she’ll be seen immediately and get taken care of.
Many times, parents’ decisions are influenced by their knowledge and experience
with taking their child with CCC to particular hospitals. One mother said that because her
child’s condition is severe she is taken “immediately back [to be seen], no matter what.
Soon as I walk in [to preferred hospital’s name], we are back…we don’t even sign in. We
go automatically back.” Because she does not want her child to wait to be seen, the
mother always brings her child to this ED.
Parents’ access to care and ability to choose a particular hospital is influenced by
their available resources. Those who have transportation describe weighing the severity
of the child’s condition against the time needed to drive to the preferred hospital, whereas
the decisions of those without readily available transportation are determined by the
proximity of care. Parents who lived in close proximity to an ED said they took their
children to the ED when they didn’t have the resources, such as a vehicle or bus fare, to
take them to see their primary care provider (PCP) across town. Similarly, parents stated
lacking money or gas influenced their decisions and access to care. A rural-dwelling
mother discussed the problems she faced with her husband working out-of-town and her
lack of a telephone: “I get worried sometimes because…what if she has a seizure and I
can’t get up with my husband. If she has a seizure, do I just let her lay there? Is she gonna
die or snap to?”
Parents’ level of comfort with managing the child’s condition was another influence
on location of health care. Parents and HCPs described situations during which parents
tried to manage the child’s condition at home before bringing the child to be seen. HCPs
believed rural-dwelling parents attempted to manage the child’s condition longer than
urban-dwelling parents because of the challenges associated with having a child
hospitalized far from home. According to one HCP, rural-dwelling parents “know that if
my kid really is sick enough to be in the hospital then I’m going to have a problem with
my other kids; I’m going to have a problem with my job.”
Multiple regression analyses of medical record data supported the position that lack
of resources is associated with increased hospital utilization. Public insurance and/or
self-pay status were statistically significantly associated with higher hospital utilization
(ie, more occurrences and fewer days between occurrences). Public insurance was
considered a proxy for lack of resources. Although Medicaid standards vary by state,
child eligibility is based on parents’ income (Centers for Medicare and Medicaid
Services, n.d.).
Family relationships and structure. Parents and HCPs perceived that the
proximity of extended family and friends and their availability to provide support
influenced location of health care. HCPs believed extended family members were
essential sources of transportation and influenced parents’ decisions regarding place of
care. In addition to providing transportation to parents, extended family members
watched siblings so that parents could take children with CCC to out-of-town specialists’
appointments. Parents who had available extended family described the importance of
this resource while parents who did not have friends or family near expressed frustration
with the lack of this resource.
Heath care provider relationships. Parents and HCPs frequently mentioned
factors pertaining to health care relationships. Both rural- and urban-dwelling parents
described choosing to take their child to a hospital farther from home because of their
personal preferences. One father described the reasons he preferred a particular hospital:
“It makes me feel more comfortable…it’s more of a home setting, the doctor’s more
friendly.” Relationships with HCPs were built on previous experiences and were also
influenced by the interactions the parents observed between the staff and the child. The
parents of a child with frequent hospitalizations said the following: “Since we’ve been up
here, like we got to know the nurses, the nurses got to know us, so its […] more
comfortable. You feel more comfortable, and I guess [the child] feels more comfortable
because of the familiar faces.” HCPs agreed that familiarity with a child was beneficial.
According to one HCP, a lack of familiarity “puts you even further behind [in the child’s
care] because you know this child has something, and you just don’t know what it is.”
Beyond being familiar with individual children, staff’s expertise with caring for children
with CCC in general was a perceived benefit. One HCP described the effect of rural
staff’s familiarity with children with CCC:
…sometimes the comfort level of the physician and the staff, the more complex a
child is, they’re very uncomfortable or they may not do everything to begin with
that they need to do, and that child continues to deteriorate in the emergency room
before they make a phone call [to a more urban hospital].
Parents described a lack of satisfaction with some rural hospitals because they
believed staff “couldn’t handle” or “didn’t understand” the child’s condition. Parents
preferred staff members who not only “knew how to handle [the child’s condition] and
knew all the medicines to give her” but also interacted with the child through talk and
play.
Parents and HCPs believed communication was a factor that influenced health care
utilization. According to HCPs, effective communication between providers at smaller
community hospitals and those at centers capable of providing a higher level of care was
essential to optimal care delivery for children with highly complex conditions. Through
effective communication, HCPs in rural areas could discuss conditions with HCPs in
centers providing higher levels of care who also may be more familiar with the conditions
as well as with individual children with CCC. Parents also valued effective
communication with PCPs and were willing to travel an hour or more to find a PCP who
“is really listening and valuing what I have to say and taking that into consideration to
make the right judgment call for my kids.”
Health care system environment. Parents’ familiarity with the processes at a
particular hospital influenced utilization. Parents described knowing, for instance, that
the wait time at an ED would be longer and deciding to visit another hospital’s ED
instead. However, in the case of an emergency, parents abandoned their preferences for a
particular location and instead took the child to the nearest hospital. Parents perceived,
and HCPs agreed, that some locations lacked equipment or expertise to adequately treat a
child with CCC. One mother relayed her perception of this lack of expertise:
Basically, I went to the ER with her one morning and he [the physician] told me,
‘Oh well, you shouldn’t put her through more than what she has to go through,
she’s already sick enough, don’t keep running up here every time you feel like
something’s wrong.’ So that was the last time I went to [hospital] because
something actually was wrong. Because we came here [preferred hospital] and she
got admitted.
Parents believed some hospitals lacked trained staff able to care for children with
CCC. The lack of local specialists and other health care services needed by children with
CCC meant that all parents, including those who reported urban residence, faced access
to care issues and were required to travel an hour or more for care. Parents also believed
the stress of traveling out of town to see specialists would be relieved by the availability
of local clinics or home visits. The lack of health personnel in rural locations also
included out-of-hospital services. According to one mother, in her community emergency
medical service (EMS) providers are not located at a central station so ambulances could
take an hour to reach the child. When the child had a seizure, the parents “had to go to the
EMS station” to receive timely treatment.
Community environment. Parents and HCPs perceived that a lack of community
resources and organizations influenced rural health care utilization. One HCP argued that
an absence of resources in rural locations could substantially affect care:
The more rural (further out they are), the less resources are available to go out
there because there are some places that only have physical therapy that cover this
area. So you could be in a spot and have a really chronic child and have no
resources available to you because you live there.
HCPs believed identifying a PCP in each community who was comfortable with
caring for children with CCC would benefit these children. Parents did not discuss a lack
of health care organizations in rural areas, and instead mentioned the lack of community
groups. One mother knew of a parents’ support group in a more metropolitan area, but
not of a local group. She attested to location-associated limitations when she said, “I can’t
go three hours, two and a half hours away every time I feel like I need to have a support
group.” Rural versus urban residence was statistically significant in the prediction models
for index admission LOS, number of occurrences, and mean number of days between
occurrences that followed the index admission. Rural residence was associated with
fewer occurrences and more days between occurrences. This finding seems to refute
parents’ and HCPs’ beliefs; however, rural residence was associated with a longer index
hospitalization LOS.
Discussion
The findings led to an ecological model of place-related risk and protective factors
for hospital admissions and ED visits among children with CCC (Figure 1). Hospital
utilization differed between rural- and urban-dwelling children with CCC.
Increased complexity, or severity of the child’s condition, was the only biological
risk factor associated with hospital utilization identified in this study. Despite
participants’ perceptions that increased complexity led to the need for care at urban
centers, rural residence was associated with fewer hospital admissions and ED visits and
more days between hospitalizations and visits. This finding may result from data being
collected solely at one urban hospital.
All parent participants reported using other urban or rural hospitals, thus findings from
the medical record review most likely underestimate true utilization rates. While rural
residence was in some cases associated with lower rates of utilization, it was also
associated with a longer index hospitalization LOS. This finding may suggest rural-
dwelling children are sicker on transfer to an urban center, or may suggest more time is
needed for rural-dwelling parents to prepare the home environment for children with
CCC prior to hospital discharge. Park and colleagues found that rural-dwelling children
with liver transplant did not have significantly poorer health outcomes, but tended to be
sicker at the time of transplantation than urban-dwelling children (Park et al., 2011).
The lower rates of hospitalizations or ED visits and higher number of days between
hospitalizations or visits may also indicate disparities in access to care among rural-
dwelling children. Access disparities have been discovered between low-income rural-
and urban- dwelling children (DeVoe et al., 2009). According to Laditka and colleagues
(2009), rural-dwelling children are more likely to be hospitalized for ambulatory care
sensitive conditions, findings which suggest disparities in access to primary care.
Similarly, Roy, McGinty, Hayes, and Zhang (2010) report rural-dwelling children with
chronic illness have higher rates of hospitalization than urban-dwelling children that may
be related to environmental, social, or access to care issues.
Many of the parents in this study resided in an urban county, but nearly all of these
parents discussed place-specific issues such as utilization of rural hospital services,
choice of one urban hospital over another, or travel to a more metropolitan area for
specialty services. Although this study was conducted at a hospital in an urban county,
many pediatric specialty providers are located in the largest metropolitan areas of South
Carolina, approximately 80 to 130 miles away. The lack of local specialty providers
corresponds with Mayer’s (2008) findings that rural and smaller metropolitan areas had
poorer geographic access to pediatric specialty providers. The finding that place-related
issues were prevalent among both urban- and rural-dwelling participants suggests access-
to-care issues are faced not only by the most rural-dwelling children but also by children
with complex chronic conditions in smaller metropolitan areas.
When considered in conjunction with the Rural Nursing theory (Long & Weinert,
1989), findings from this study have particular relevance for nurses in rural-dwelling
areas. For instance, according to Rural Nursing theory, rural-dwelling people equate work
with health and are often viewed as postponing health care until the illness has advanced
to the point of requiring hospitalization. Findings from this study support this statement
and may suggest that nurses and primary care providers be more easily accessible to
rural-dwelling children with CCC. However, Rural Nursing Theory also posits that rural
nurses and other rural HCPs may need to have extended time with communities to gain
acceptance and trust, which can affect access if no established providers are available.
This study suggests that parents and caregivers of children with CCC seek providers with
whom they have an established, trusting relationship. When viewed through the lens of
Rural Nursing theory, this need for a trusting relationship may be more critical for rural-
dwelling children with CCC and their families.
Limitations
Many of the parents who reported living in an urban county also reported using
small community hospital services. This finding suggests that classifying rural versus
urban residence by county may not accurately reflect rural versus urban utilization.
Another limitation results from data having been collected only at one urban hospital.
Nearly all of the interview participants described using other hospitals or EDs indicating
an underestimation of hospital utilization in the quantitative portion of the study. This
limitation may also affect the relationships between rural versus urban status and hospital
utilization found in this study. Another limitation was that qualitative findings indicated
additional quantitative variables should have been collected. For instance, no quantitative
variables on the health care provider relationships level were available to compare with
qualitative findings.
Conclusion
The results from this study reveal differences in hospital utilization between rural-
and urban-dwelling children with CCC. The findings provide evidence of access issues
and extend the problem beyond rural counties to urban counties with rural communities
and smaller metropolitan areas. Further study is needed to develop a more comprehensive
understanding of patterns of hospital utilization among rural-dwelling children with CCC
to include admissions and ED visits at rural and urban hospitals, primary care utilization,
and outpatient services utilization. Knowledge gained from further study can support and
build upon the risk and protective factors model developed in this study and interventions
can be designed to minimize risk factors and strengthen protective factors for hospital
admissions and ED visits among children with CCC.
Acknowledgements
This study was funded in part by a grant from the Sigma Theta Tau Gamma Omicron at-
Large Chapter. This study was supported by the South Carolina Clinical & Translational
Research (SCTR) Institute, with an academic home at the Medical University of South
Carolina
References
Bachrach, Virginia R. Galton, Schwarz, Eleanor, & Bachrach, Lela Rose. (2003).
Breastfeeding and the risk of hospitalization for respiratory disease in infancy: A
meta-analysis. Archives of Pediatric and Adolescent Medicine, 157(3), 237-243.
http://dx.doi.org/10.1001/archpedi
.157.3.237
Bronfenbrenner, U. (1994). Ecological models of human development. International
Encyclopedia of Education (2nd ed., Vol. 3). Oxford: Elsevier.
Bulkow, Lisa R., Singleton, Rosalyn J., Karron, Ruth A., Harrison, Lee H., & the Alaska
RSV Study Group. (2002). Risk factors for severe respiratory syncytial virus
infection among Alaska Native children. Pediatrics, 109(2), 210-216.
http://dx.doi.org/10.1542/peds.
109.2.210
Centers for Medicare and Medicaid Services. (n.d.). InsureKidsNow.gov: connecting kids
to coverage. Retrieved from http://www.insurekidsnow.gov/chip/index.html
DeVoe, J. E., Krois, L., & Stenger, R. (2009). Do children in rural areas still have
different access to health care? Results from a statewide survey of Oregon's food
stamp population. Journal of Rural Health, 25(1), 1-7.
http://dx.doi.org/10.1111/j.1748-0361.2009.00192.x
Feudtner, C., Christakis, D. A., & Connell, F. A. (2000). Pediatric deaths attributable to
complex chronic conditions: A population-based study of Washington State, 1980-
1997. Pediatrics, 106(1), 205-209. http://dx.doi/org/10.1542/peds.106.1.S1.205
Garcia-Aymerich, J., Barreiro, E., Farrero, E., Marrades, R. M., Morera, J., & Anto, J. M.
(2000). Patients hospitalized for COPD have a high prevalence of modifiable risk
factors for exacerbation (EFRAM study). European Respiratory Journal, 16, 1037-
1042. http://dx.doi.org/10.1034/j.1399-3003.2000.16f03.x
Garcia-Aymerich, J., Farrero, E., Felez, M. A., Izquierdo, J., Marrades, R. M., & Anto, J.
M. (2003). Risk factors of readmission to hospital for a COPD exacerbation: A
prospective study. Thorax, 58, 100-105. http://dx.doi.org/10.1136/thorax.58.2.100
Garcia-Aymerich, J., Marrades, R. M., Monso, E., Barreiro, E., Farrero, E., & Anto, J. M.
(2004). Paradoxical results in the study of risk factors of chronic obstructive
pulmonary disease (COPD) re-admission. Respiratory Medicine, 98(9), 851-857.
http://dx.doi.org/10.1016/j.rmed.2004.02.019
Garcia-Aymerich, J., Monso, E., Marrades, R. M., Escarrabill, J., Felez, M. A., Sunyer,
J., . . . Efram Investigators. (2001). Risk factors for hospitalization for a chronic
obstructive pulmonary disease exacerbation. American Journal of Respiratory and
Critical Care Medicine, 164(6), 1002-1007.
http://dx.doi.org/10.1164/ajrccm.164.6.2006012
Guevara, J. P., Young, J. C. C., & Mueller, B. A. . (2001). Do protective factors reduce
the risk of hospitalization in infants of teenaged mothers? Archives of Pediatric and
Adolescent Medicine, 155, 66-72. http://dx.doi.org/10.1001/archpedi.155.1.66
Hawkins, J. D., Brown, E. C., Oesterle, S., Arthur, M. W., Abbott, R. D., & Catalano, R.
F. (2008). Early effects of Communities That Care on targeted risks and initiation of
delinquent behavior and substance use. Journal of Adolescent Health, 43, 15-22.
http://dx.doi.org/10.1016/j.jadohealth.2008.01.022
Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for
alcohol and other drug problems in adolescence and early adulthood: implications
for substance abuse prevention. American Psychological Bulletin, 112(1), 64-105.
http://dx.doi.org/10.1037/0033-2909.112.1.64
Hsieh, H., & Shannon, S. E. (2005). Three approaches to qualitative content analysis.
Qualitative Health Research, 15(9), 1277-1288.
http://dx.doi.org/10.1177/1049732305276687
Hudson, S.M., Mueller, M., Hester, W. H., Magwood, G.S., Newman, S.D., & Laken,
M.A. (2014a). At-risk characteristics for hospital admissions and ED visits. Journal
for Specialists in Pediatric Nursing, 19, 183-193.
http://dx.doi.org/10.1111/jspn/12068
Hudson, S.M., Newman, S.D., Hester, W.H., Magwood, G.S., Mueller, M., & Laken,
M.A. (2014b). Factors influencing hospital admissions and emergency department
visits among children with complex chronic conditions: a qualitative study of
parents' and providers' perspectives. Issues in Comprehensive Pediatric Nursing,
37(1), 61-80. http://dx.doi.org/10.3109/01460862.2013.855844
Knudson, A., Casey, M., Burlew, M., & Davidson, G. (2009). Disparities in pediatric
asthma hospitalizations. Journal of Public Health Management Practice, 15(3),
232-237. http://dx.doi.org/10.1097/01.PHH.0000349739.81243.ea
Laditka, J. N., Laditka, S. B., & Probst, J. C. (2009). Health care access in rural areas:
evidence that hospitalization for ambulatory care-sensitive conditions in the United
States may increase with the level of rurality. Health & Place, 15(3), 761-770.
http://dx.doi.org/10.1016/j.healthplace.2008.12.007
Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Newbury Park, CA: Sage
Publications, Inc.
Long, K. A. , & Weinert, C. (1989). Rural nursing: Developing the theory base. Scholarly
Inquiry for Nursing Practice, 3(2), 113-127.
Martin, A. B., Vyavaharkar, M., Veschusio, C., & Kirby, H. (2012). Rural-urban
differences in dental service utilization among an early childhood population
enrolled in South Carolina Medicaid. Maternal Child Health Journal, 16, 203-211.
http://dx.doi.org/10.1007/s10995-010-0725-1
Mayer, M. L. (2008). Disparities in geographic access to pediatric subspecialty care.
Maternal Child Health Journal, 12, 624-632. http://dx.doi.org/10.1007/s10995-007-
0275-3
McLeod Health. (2012). Physician recruitment FAQ's. Retrieved from
http://www.mcleodhealth.org/MRMC/physician_faq.cfm
McPherson, M., Arango, P., Fox, H., Lauver, C., McManus, M., Newacheck, P. W., . . .
Strickland, B. (1998). A new definition of children with special health care needs.
Pediatrics, 102(1), 137-139. http://dx.doi.org/10.1542/peds.102.1.137
Nagy, J., & Fawcett, S. B. (2011). Understanding risk and protective factors: Their use in
selecting potential targets and promising strategies for interventions. The
Community Toolbox. Retrieved from
http://ctb.ku.edu/en/tablecontents/sub_section_main_1156.aspx
O'Mahony, L., O'Mahony, D. S., Simon, T. D., Neff, J., Klein, E. J., & Quan, L. (2013).
Medical complexity and pediatric emergency department and inpatient utilization.
Pediatrics, 131(2), e559-565. http://dx.doi.org/10.1542/peds.2012-1455
Park, K. T., Nanda, P., Bensen, R., Strichartz, D., Esquivel, C., & Cox, K. (2011). Effects
of rural status on health outcomes in pediatric liver transplantation: a single center
analysis of 388 patients. Pediatric Transplantation, 15, 300-305.
http://dx.doi.org/10.1111/j.1399-3046.2010.01452.x
Pollack, H. A., Dombkowski, K. J., Zimmerman, J. B., Davis, M. M., Cowan, A. E.,
Wheeler, J. R., . . . Freed, G. L. (2004). Emergency department use among
Michigan children with special health care needs: an introductory study. Health
Services Research, 39(3), 665-692. http://dx.doi.org/10.1111/j.1475-
6773.2004.00250.x
Roy, S. R. , McGinty, E. E., Hayes, S. C. , & Zhang, L. (2010). Regional and racial
disparities in asthma hospitalizations in Mississippi. Journal of Allergy and Clinical
Immunology, 125(3), 636-642. http://dx.doi.org/10.1016/j.jaci.2009.11.046
Sallis, J. F., Owen, N., & Fisher, E. B. (2008). Ecological models of health behavior. In
K. Glanz, B. K. Rimer & K. Viswanath (Eds.), Health behavior and health
education (4th ed., pp. 465-485). San Francisco, CA: Jossey-Bass.
Simon, T. D., Berry, J. G., Feudtner, C. , Stone, B. L., Sheng, X., & Bratton, S. L. (2010).
Children with complex chronic conditions in hinpatient hospital settings in the
United States. Pediatrics, 126(4), 647-655. http://dx.doi.org/10.1542/peds.2009-
3266
United States Census Bureau. (2013, June 11, 2014). State & county quick facts.
Retrieved from http://quickfacts.census.gov/qfd/states/45000.html
United States Census Bureau. (n.d.). Metropolitan and micropolitan. Retrieved from
http://www.census.gov/population/metro/
Wright, B., Bellinger, J. , Brahmbhatt, M. Y., Brokks, K. , Gangadaram, M. K., Garner,
D. Q., . . . Wilkes, J. (2008). State rural plan for South Carolina. Columbia, SC:
South Carolina Rural Health Research Center.
Yamamoto, L. G., Zimmerman, K. R., Butts, R. J. , Anaya, C. , Lee, P. , Miller, N. C. , . .
. Leung, Y. (1995). Characteristics of frequent pediatric emergency department
users. Pediatric Emergency Care, 11(6), 340-346.
http://dx.doi.org/10.1097/00006565-199512000-00003