Article Text
Abstract
Objectives To identify the factors associated with paediatric emergency department (ED) visits and parental perceptions of the avoidability of their child's ED visit.
Design Cross-sectional study by performing secondary analysis of 2010–2011 Iowa Child and Family Household Health Survey data.
Setting State-wide representative population-based sample of families with at least one child in the state of Iowa in the USA.
Patients/participants Among the eligible households, 2386 families completed the survey, yielding a cooperation rate of 80%.
Exposure/intervention Presence of a medical home.
Main outcome measures Child visiting an ED in the past year; parents believing that ED care could have been provided in a primary-care setting.
Results Among children who needed medical care in the past year, 26% visited an ED. Younger children, non-Hispanic black children, non-Hispanic others, children whose parents were not married, children who were from food-insecure households and had poorer health status were more likely to visit an ED. Having a medical home was not associated with ED visits (OR=0.80, 95% CI 0.61 to 1.04), even after stratifying by the child's health status. About 69% of parents who took their child to an ED agreed that ED care could have been provided in a primary-care setting. Parents of children with public insurance, those who were not referred to the ED and those who could not get routine care appointments were more likely to report a primary-care preventable ED visit.
Conclusions The majority of parents believed that paediatric ED visits could be avoided if adequate primary-care alternatives were available. Expanding access to primary care could lead to a reduction in avoidable ED visits by children.
- primary care
- emergency care systems, admission avoidance
- paediatrics, paediatric emergency medicine
Statistics from Altmetric.com
Key messages
What is already known on this subject?
Avoidable emergency department (ED) visits constitute a significant public health burden with health outcomes and ethical and economic implications. The medical home model is proposed as one way to reduce them, but the evidence to support this association is not well established. Parents are the primary decision-makers for child's healthcare, and their perceptions of primary-care treatable ED visits have not been examined.
What might this study add?
Food insecurity is strongly associated with ED visits, whereas having a medical home is not. Among children who visited an ED, about 70% of parents said care could have been provided in a primary-care setting. Although medical home was not associated with parental perceptions of avoidability, inability to get routine care appointments, public insurance and referral from healthcare provider were associated.
Introduction
Avoidable emergency department (ED) visits have been implicated as a major contributing factor to overcrowding of the EDs and escalating healthcare costs in the USA,1 ,2 as well as internationally.3–5 Apart from the economic costs, avoidable use of EDs has been associated with poorer health outcomes, diversion of the emergency personnel from emergent cases and other ethical implications.1 ,2 In the USA, paediatric populations comprise about a quarter of all ED visits even though health insurance and access to healthcare is usually better for children.6 A major factor that differentiates paediatric ED visits is that the decision to seek ED care is made by parents, rather than the children themselves. This makes it important to understand the factors that impact parents’ decision and perceptions of what constitutes an ‘emergency’.
Several studies have tried to answer this question by surveying parents after the hospital triage determined a child's ED visit to be non-urgent.3 ,4 ,7–9 There is disagreement in the literature as to whether parental perception of severity of their child's illness is comparable with the hospital nurse triage status. Although some studies indicate that parental assessment is comparable with emergency nurse's triage scores,10 another study from Australia found that even though the triage found a condition to be non-urgent, the parents perceived the problem to be severe.3 Few studies have reported reasons like convenience and long wait to see a primary care provider (PCP) behind parents taking their child to ED for non-urgent care.8 However, no studies so far have looked at what parents perceive to be a primary-care preventable ED visit, regardless of the hospital-determined triage status.
The American Academy of Pediatrics introduced the concept of a medical home in 1967,11 which has evolved since then to indicate the presence of a regular source of healthcare that is comprehensive, coordinated, patient centred and culturally competent.12 Important attributes for primary care practices to be considered a medical home in the USA are adoption of health information technology, including electronic health records, open same-day scheduling and expanded hours and continuity of care by serving as a source of regular care over time.13 It is important to note that what comprises a ‘medical home’ varies considerably between the USA and other countries.13
The medical home model has been proposed as a way to reduce avoidable ED visits. However, the evidence to support this claim is questionable, as the literature is scarce and contradicting.14 ,15 One way to answer this question is to ask parents who brought their child to ED if they think care could have been provided by a PCP if one was available to them, as these can be defined as ‘primary-care treatable’ ED visits.16
The primary purpose of the current study is to examine parental perceptions of the avoidability of their child's ED visit, regardless of the hospital triage outcome, using a state-wide population-based sample of families with children. Secondary objective of the study is to examine the factors associated with paediatric ED visits, especially the association of medical home with ED visits, with the hypothesis that children with a medical home will be less likely to visit an ED, especially when its avoidable, than children without a medical home.
Methods
Data source
The study was conducted using data from the Iowa Child and Family Household Health Survey (IHHS) for the year 2010–2011. The IHHS is a population-based state-wide survey that used an address-based sampling approach to get a state-wide representative sample of households with one randomly selected child between the ages of 0 and 17 years. Hispanics and African–American families were oversampled to get stable estimates for subgroup comparisons.17 The dataset contained information on 2386 families with at least one child residing in the state of Iowa. The cooperation rates, defined as eligible households that agreed to participate in the survey in accordance with American Association for Public Opinion Research guidelines,18 was 80% (figure 1). Of the 2386 families who responded, 527 completed the survey online while 1859 completed the survey through telephone, with no significant differences in their characteristics.17
The primary outcomes of interest were child visiting an ED in the past year and if they did visit an ED, did the parents think that ED care could have been provided in a primary-care setting, like a doctor's office or clinic, if one had been available. The total numbers of ED visits made by the child in the past year were available, but as less than 10% made more than one ED visit, we used a dichotomous response variable.
Conceptual model
We propose that a child's use of an ED is associated with personal and familial socioeconomic–demographic factors. Andersen's behavioural model of health services’ usage19 served as the basis of our conceptual model. Explanatory factors were chosen a priori and then grouped into the three main domains as outlined by Andersen's model. Predisposing characteristics are usually non-changeable inherent traits that are associated with healthcare usage, such as age, race and education. Enabling factors are those that either facilitate or hinder healthcare usage, such as health insurance and medical home. Finally, need is either direct or indirect perceived measure of the requirement for healthcare usage. Each of the explanatory factors was derived from a single survey item, except food security and medical home, which were derived from multiple survey items as follows:
Food insecurity: the following three survey questions were used based on a validated screening tool:20 in the last 12 months, (1) food we bought did not last, and we did not have money to get more, (2) did any adult in the household cut his/her meal size or skip meals because there was not enough money for food, (3) were you ever hungry but did not eat because there was not enough money for food. If parents responded positively to any of the three questions, the household was categorised as being food insecure.
Medical home: a composite indicator of whether or not a child had a medical home was constructed using 17 survey questions that tapped into various conceptual components of a patient-centred medical home. The detailed methodology of indicator development can be found elsewhere.17 ,21
Statistical analysis
Bivariate analyses were conducted using logistic regression to examine crude ORs between explanatory factors and outcome variables. Bivariate analyses results were used to inform multivariable analyses by including those variables with p values ≤0.10 in bivariate analysis in the multivariable logistic regression models. If a pair of explanatory variables was correlated, one was chosen based on conceptual significance to avoid multicollinearity. Final parsimonious models, modelling the adjusted ORs of visiting an ED and parents perceiving the child's ED visit to be avoidable were developed using a manual backward elimination method (p≤0.10 was used to keep variables in the model) based on the likelihood ratio test. Ninety-five per cent CIs were used as precision estimates. All data were analysed using SAS V.9.3. Statistical significance was determined at p<0.05. The research protocol was approved by the University of Iowa Institutional Review Board.
Results
Data describing the sample distribution and results of bivariate analysis are presented in table 1. A majority of the parents surveyed were married (82%), identified their child's race/ethnicity as non-Hispanic white (88%), age >6 years (72%) and reported that the child had some form of health insurance (97%). This reflects the demographic characteristics of the state of Iowa. About 9% families reported household food insecurity, while 20% did not have a medical home for their children. Bivariate analysis revealed that many explanatory factors were significantly associated with the odds of having an ED visit while only child's age, type of health insurance, referral by healthcare provider and ability of getting routine care appointments were associated with parental perceptions of the avoidability of ED visit (table 1).
About 59% (n=1409) of the respondents reported that they needed some form of medical care for their child in the past year. Among these, 26% (n=371) of the respondents reported taking their child to an ED at least once in the past year. Among those who reported an ED visit for their child, the parents of 68.7% (n=250 of 364) of children reported that the care could have been provided at a primary-care setting if it was available (figure 1). About 36% of children with ED visits were referred by a healthcare provider to the ED, and more than half of these children's parents thought that ED visit could have been avoided if a PCP was available.
The multivariable logistic regression, modelling the odds of ED visit, revealed that predisposing characteristics, enabling factors and need were all significantly associated with ED visits (table 2). Children aged 1–5 years had 60% greater odds of visiting an ED than children aged 13–17 years. Non-Hispanic black children and non-Hispanic others had greater odds of having an ED visit than non-Hispanic white children. Children whose parents were not married had more than twice the odds of visiting an ED than children whose parents were married. Children from food-insecure households had 69% greater odds of visiting an ED than children from food-secure households. Finally, children with a poorer reported health status were at greater odds of visiting an ED than children with better health status. Medical home was not found to be associated with having an ED visit, even after it was stratified by child's health status.
Multivariable analysis revealed that only enabling factors were significantly associated with parents’ perceptions that care could have been provided in a primary-care setting. After adjusting for other variables in the model, parents of children with public insurance had 75% greater odds of reporting an avoidable visit compared with those with private insurance. Parents of children who were not referred to the ED by a healthcare provider had almost twice the odds of reporting an avoidable ED visit than parents of children who were referred. Finally, parents of children who did not get routine care appointments as soon as needed always also had greater odds of reporting an avoidable ED visit (table 3).
Discussion
The results support the hypothesis that predisposing characteristics, enabling factors and need are all significantly associated with a child visiting an ED. The study also provides evidence that in the USA, 70% of the parents believe that the paediatric ED visit could be avoided if primary-care alternatives were available. Although the presence of a medical home was not significantly associated with fewer ED visits overall, the ability to get routine care appointments as needed was significantly associated with parental perceptions of the avoidability of their child's ED visit.
There is some evidence indicating that the presence of a medical home is associated with fewer ED visits and better health outcomes.1 ,22 Our study results could not confirm this finding as our estimates of association between medical home and ED visits did not reach statistical significance in either bivariate or multivariable analyses, even though the OR estimates were in the same direction as the previous studies. David et al15 found that medical home was associated with fewer ED visits only among patients with chronic illnesses, but not among others. Our survey had a low response (20%) for the chronic disease question due to a skip pattern. So, we stratified the effect of medical home on ED visits by parents’ report of their child's health status. However, no association was found between medical home and ED visits within those health status categories either (results not shown).
Over the past two decades, food insecurity has been proposed as a predictor of health outcomes and health services usage.23 Studies indicate that food insecurity is associated with acute healthcare usage, including ED visits, in vulnerable populations.23 ,24 However, no studies have examined the relationship between food insecurity and paediatric ED visits using a population-based state-wide sample. Our study provides evidence that household food insecurity is significantly associated with increased ED visits by children in Iowa. Although other enabling factors, such as type of health insurance and family income, were significantly associated in bivariate analyses, the multivariable model revealed that food insecurity was most strongly associated with ED visits. It is possible that food insecurity is a proxy for several social determinants of health, and future studies should examine potential mediating factors between food insecurity and acute healthcare usage.
In our study, 69% of the parents whose child visited an ED reported that the care could have been provided in a primary-care setting, indicating the possible avoidable nature of the ED visit. This is much higher than a systematic review of literature, which shows that, on average, 37% of all ED visits were avoidable25 and another study that found 58% of paediatric ED visits to be non-urgent.26 However, the criteria to determine what constitutes an avoidable ED visit vary greatly across studies. Our study used parents’ assessment based on their recollection of a past event, which could differ significantly from hospital triage reports or the algorithm-based assessment used by most studies. Our study found that in the USA, parents of children with public insurance were much more likely to report an avoidable ED visit than parents of children with private insurance. This highlights the poor state of access to primary care for publicly insured children in the USA, and is consistent with literature that examined ED visits for ambulatory care sensitive conditions.26 Such variation in avoidable ED visits by type of health insurance would not be applicable to other countries with universal public insurance system. We also found that parents of children who were referred to the ED by a healthcare provider were less likely to report a primary-care preventable ED visit compared with the parents of children who were not referred to the ED. This is not surprising as these parents had already made contact with a non-ED provider, most likely the child's PCP, and the referral reaffirmed their decision to take the child to the ED. However, more than half of the parents who were referred to the ED by a healthcare provider reported that the ED visit could have been avoided if primary-care alternatives were available. This interesting finding could possibly be attributed to automated referral messages at the PCP office phone after hours, or could potentially indicate the presence of inappropriate referral patterns, as suggested by some studies.8 ,27 The variation in ED visits by PCP referral can also vary depending on certain countries’ policy that require a referral for all services received outside of the primary care physician. Finally, although medical home was not associated with parental perceptions of the avoidability of ED visit, parents who reported an inability to get routine care appointments were twice as likely to report a primary-care preventable ED visit. This finding highlights the importance of regular and reliable access to primary care in reducing avoidable ED visits by children, which is confirmed by other studies,28 and could be an effective way to reduce avoidable paediatric ED visits.29 ,30
Our study had several limitations. First, the assessment of avoidability of the ED visit was reported by parents, and could not be compared with the assessment by a healthcare professional at the ED. There is debateable evidence on the level of agreement between parental assessments of the severity of their child's illness and the assessment performed by the healthcare professional,16 and our study could not address the issue. Second, the sample size was relatively small to examine factors associated with parental perceptions of the avoidability of ED visit. Although we did identify several important factors, but may have missed others due to lack of enough statistical power. Furthermore, parents were asked if ED visit could have been avoided if primary care was available to them many days to months after the actual visit occurred, which could lead to recall bias. It is also possible that the parents believed that the situation was a true emergency at the time, but after an extensive work-up was followed by minimal intervention, they may have decided in retrospect that ED visit could have been avoided. Finally, these results may not be generalisable to other states or other countries that have a different demographic profile or healthcare system.
In conclusion, paediatric ED visits in the state of Iowa were not significantly associated with medical home, but were strongly associated with household food insecurity. Parents perceive that a vast majority of the ED visits can be avoided if they have access to primary care at the time of the ED visit. Expanding access to primary care could lead to a reduction in such avoidable ED visits, as evidenced by few initiatives.29 ,30 The Affordable Care Act passed in 2010 in the USA has several provisions to improve access to primary care and care coordination, and, if implemented successfully, has the potential to reduce avoidable ED visits.
References
Footnotes
Contributors The manuscript has seven authors and contributors; of whom, the first author was a PhD candidate at the time of initial submission. This manuscript is a product of AS dissertation work, which has been accomplished by constant participation, supervision and guidance by the dissertation committee members who are the other six authors. AS conceptualised the research question, study design, analysed the data, drafted the initial version of the manuscript and revised and approved the final manuscript as submitted. DJC conceptualised the study, refined the analysis approach, reviewed and revised the manuscript and approved the final manuscript as submitted. MPJ reviewed the data analysis, guided the analytical approach, reviewed and revised the manuscript and approved the final manuscript as submitted. RAK conceptualised the study, reviewed and revised the manuscript and approved the final manuscript as submitted. ETM conceptualised the study, critically reviewed and revised the manuscript and approved the final manuscript as submitted. CTB provided contextual translation of many emergency department-specific concepts, critically reviewed and revised the manuscript and approved the final manuscript as submitted. PCD conceptualised the study question, research approach, was responsible for survey instrument development, data collection, composite measures development, reviewing the analysis, reviewed and revised the manuscript and approved the final manuscript as submitted.
Competing interests None declared.
Ethics approval University of Iowa Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.