Introduction

At any given moment in the United States there are over 400,000 children in foster care, and 264,746 children entered the foster care system in 2014 (United States Department of Health and Human Services, Administration for Children and Families, Administration on Children, Youth, and Families, Children’s Bureau 2015). Children in foster care have greater medical, dental, developmental, behavioral, and psychosocial needs than children not in the foster care system, and due to the pervasiveness of these health issues are considered children with special health care needs (Council on Foster Care, Adoption, and Kinship Care, Committee on Adolescence, & Council on Early Childhood 2015). Studies show that up to 80% of children in foster care have at least one medical problem, up to 80% have a significant mental health need, almost 40% have a significant oral health issue, up to 60% under age 5 have a developmental issue, and more than 40% of school age children have an educational concern (Chernoff, Combs-Orme, Risley-Curtiss and Heisler 1994; Council on Foster Care, Adoption, and Kinship Care et al. 2015; Halfon et al. 1995; Jee et al. 2010; Leslie et al. 2005; Simms et al. 2000; Steele and Buchi 2008; Szilagyi 2012). These needs are often unmet at foster care entry due to fragmented medical care, poor access to preventative health services, and a chaotic or deprived home environment prior to foster care (Council on Foster Care, Adoption, and Kinship Care et al. 2015). Entry into the foster care system provides an opportunity to address the myriad of health and psychosocial needs of a child and ensure that unmet needs are addressed.

Health care coordination (HCC) is a process that facilitates the linkage of children and families with appropriate services and resources in an effort to achieve optimal health, and HCC has been shown to improve quality of care, particularly for children with special health care needs (Council on Children with Disabilities 2005). Advantages include improved integration of services provided by different specialties, easier access to multiple resources and specialties, and improved communication amongst providers (Council on Foster Care, Adoption, and Kinship Care et al. 2015; Jee et al. 2010; Leslie et al. 2005). These in turn decrease unnecessary health care utilization and cost through reduced hospital admissions, reduced length of hospital stay, reduced inpatient charges, and decreased utilization of emergency services (Council on Children with Disabilities 2005; Criscione et al. 1995; Walsh et al. 2002). In addition, families of children with special health care needs who feel they receive adequate HCC services report improved family-child outcomes compared to families who feel the HCC they receive is inadequate (Turchi et al. 2009). Families with adequate HCC report decreased unmet specialty care needs, decreased problems with referrals for specialty care, fewer missed school days due to illness, and fewer emergency department visits (Boudreau et al. 2014; Turchi et al. 2009).

Children in foster care are in particular need of HCC services, as this population accounts for a disproportionately high level of health care service utilization and cost, with studies estimating those costs being up to twelve times higher for a child in foster care than for a child not in the system (Allen and The Center for Health Care Strategies, Inc. 2008; Child Welfare League of America 2011; Rubin et al. 2004). Provision of HCC services presents a particular challenge in this high needs population due to frequently changing placements and child welfare workers, the need for consent for certain interventions and services, lack of complete historical medical information, and other barriers (Hillen and Gafson 2015; Simms et al. 2000; Szilagyi 2012). Further, families play a primary partner in a child’s HCC efforts and have important knowledge about their child’s conditions (American Academy of Pediatrics, Committee on Hospital Care 2003; Council on Children with Disabilities 2005). Children in foster care may lose the benefit of having a biological parent partnering in coordination efforts, and often lack a historian with detailed knowledge of health issues.

Legislators at the federal level have recognized the importance of HCC for children in foster care, and have tasked states to develop systems to ensure that health care needs are met. Both the 2008 Fostering Connections to Success and Increasing Adoptions Act (Pub L No. 110-351) and The Child and Family Services Improvement Innovation Act of 2011 (Pub L No. 112-34) emphasize the need for development of coordinated systems to care for this population. States have developed strategies to address this mandate and provide HCC for children in foster care, with various models emerging including medical home sites, foster care evaluation and intake sites, specialized primary care sites, and state administrative programs (Jee et al. 2015). In response to the federal mandate, Wisconsin’s Department of Health Services and Department of Children and Families partnered with Children’s Hospital of Wisconsin to develop a Medicaid Foster Care Medical Home program entitled “Care4Kids” in 2014. This HCC model of care delivery is active in 6 southeastern Wisconsin counties. Each child enrolled in the Care4Kids program has an assigned health care coordinator to ensure that health care needs are met in a timely fashion. The Care4Kids program development team reviewed available literature to determine how to most effectively utilize the limited HCC resources to be most impactful for the children served.

Studies looking at HCC for children in foster care have largely focused on strategies for coordination once the child has been established in foster care. Development of a “health passport” to capture important medical, educational, and social information and accompany the child throughout placement has been one strategy to improve coordination (Lindsay et al. 1993). Ensuring that children are cared for within a primary care medical home has also been shown to improve information sharing and coordination for these children (Homer et al. 2008; Medical Home Initiatives for Children with Special Needs Project Advisory Committee 2002). Limited data exists about how to stratify children at foster care entry based on unmet health care needs, so that HCC services can be targeted to the children who will benefit most. These HCC services could then assist children get those unmet needs fulfilled within the first few weeks in care, in addition to the standard health services required for all children at foster care entry. Therefore, we aimed to (1) utilize a Triage Tool at foster care entry to stratify children based on unmet health care needs and (2) determine if the Triage Tool is an accurate predictor of health care utilization as indicated by cost during a child’s first weeks in care.

Method

Participants

The Milwaukee Child Advocacy Center (MCAC) provides an Initial Foster Care Health Exam (IFCHE) within the first few days of a child entering foster care in our community. On January 1, 2014, medical providers at the MCAC began to utilize a Triage Tool (TT) to assign children a Triage Level based on their unmet health needs at the time of the IFCHE. This Triage Level was documented in the medical record and communicated to the Care4Kids HCC team.

Our study was a retrospective chart review of consecutive youth ages 0–18 years seen at the MCAC for an IFCHE from January through July of 2014. Inclusion into the study required that the child was enrolled in the Care4Kids Foster Care Medical Home program at the time of the IFCHE as well as the subsequent 60 days, as this allowed study staff access to their Medicaid records during the study time period. Children never enrolled in the Care4Kids program (due to placement outside of the Care4Kids six county region or biological parents opting their child out of the program), or enrolled for less than the 60 day time period were excluded. During this time period an IFCHE was performed on 620 children in foster care seen at the MCAC. Of those children, 445 were enrolled in the Care4Kids program at the time of the IFCHE and for the subsequent 60 days, and were thus eligible for inclusion in this study.

Procedure

The TT was developed by medical providers at the MCAC, drawing upon their collective experience providing medical care for children entering foster care. The tool was piloted, evaluated, modified, and finalized in the months prior to the launch of the Care4Kids program, and was implemented in January of 2014. The TT categorized children into 3 levels of unmet need, which included medical, psychiatric, immunization, or any other unmet needs (Table 1). Level 1 represented the highest level of unmet need, where HCC services were needed urgently (next day). Level 2 represented a moderate level of unmet need, where HCC services were needed within the week. Level 3 represented the baseline level of unmet need, where routine HCC services were needed. In addition to assigning each child a Triage Level, medical providers also documented a reason for assigning a child as Level 1 or Level 2. Providers were allowed to choose one or more reasons for assigning a given level.

Table 1 The 3 Triage Level categories assigned to a child at the time of the Initial Foster Care Health Exam, and the criteria for medical, psychiatric, immunization, or other unmet needs used to stratify a child into each Level
Full size table

It is important to note that medically complex children were stratified as Triage Level 2 in this study. Medical complexity was defined as a child being followed by three or more specialty providers and/or having three or more health problems. Though a medically complex child could have all health needs met at the time of foster care entry, it was felt that they were at higher risk for having a need emerge, so should be triaged higher than the general foster care population (Table 1).

Measures

The electronic medical record (EMR) was reviewed from the MCAC visit to determine demographic information as well as the Triage Level and reason for level assignment. Medicaid health care service utilization and health care coordination data was then reviewed for each child for the first 60 days following the IFCHE. Total Medicaid cost as well as specific costs for HCC services, urgent and emergency care services, and mental health services were calculated from the data set to determine utilization of services. The charges that were included in each of the above categories were determined by the billing codes, by a provider’s listed specialty, and by the place of service. A description of the codes used to determine a charge’s category is listed in Table 2. Total charges were determined by calculating the sum of the charges that occurred within 60 days of the IFCHE visit date. These costs served as a marker of service utilization for each child in the study population.

Table 2 Medicaid codes, facility, and provider types used to calculate costs for health care coordination, urgent care, emergency care, and mental health care utilization categories
Full size table

Data Analyses

Data analysis was performed using SPSS statistical software (Version 22.0; Chicago, IL). The Mann-Whitney/ Kruskal-Wallis test was used to compare the costs by level. Pearson and Spearman correlations as well as Scatterplots were used to examine the correlation between continuous variables. Log transformation was used for cost related variables. A p-value of <0.05 was considered significant.

This study was approved by the Institutional Review Board of Children’s Hospital of Wisconsin. Informed consent was not obtained as both the IFCHS and the use of the TT are considered a part of routine medical care, and would have occurred regardless of this study.

Results

Of the 620 children in foster care seen at the MCAC for an IFCHE between January 1 and July 31, 2014, 445 were enrolled in the Care4Kids program at the time and for the subsequent 60 days, and were thus eligible for inclusion in this study. There was no significant difference in demographic characteristic between the 620 children seen at the MCAC and the 445 enrolled in the study. The average age of the study population was 6.7 years, with the majority of the children between 0–6 years. There were slightly more female than male children, and most were African American followed by Caucasian. The majority of children were placed in a foster home, most commonly with a relative. Full demographic information for the 445 enrolled children can be found in Table 3.

Table 3 Demographic information for study population
Full size table

Of the 445 children enrolled in the study, 44 (9.9%) were assigned a Level 1, 127 (28.5%) were assigned a Level 2, and 274 (61.6%) were assigned a Level 3 using the Triage Tool (Fig. 1). Children assigned to Level 1 most often had unmet immunization needs (45.5%) followed by unmet medical needs (36.4%) and unmet psychiatric needs (20.5%). Children assigned Level 2 most often had unmet medical needs (52.0%) followed by unmet immunization needs (41.0%) and unmet psychiatric needs (16.5%) (Fig. 1). For children assigned Level 2, 9.5% were considered medically complex.

Fig. 1
figure 1

Percent of children assigned each Triage Level using the Triage Tool. For Triage Levels 1 and 2, the reason(s) that the Level was assigned (Medical, Psychiatric (Psych), Immunization (Imm), and/or Other). Note: More than one reason may be selected for a given Level, so percentages will add up to >100%

Full size image

Medicaid health care service and health care coordination cost during the first 60 days was examined based on Triage Level, and is summarized in Fig. 2. The average total cost of Medicaid services for children assigned Level 1 was $891, for Level 2 was $772, and for Level 3 was $835. These results were not statistically significant. Similarly, there was no statistically significant difference found in cost for emergent/urgent care services, health care coordination services, or mental health services when stratified by Triage Level (Fig. 2).

Fig. 2
figure 2

Total Medicaid costs, Emergent/Urgent (ER/UC) care costs, Health Care Coordination (HCC) costs, and Mental Health (MH) costs during first 60 days per Triage Level assigned. There are no statistically significant differences

Full size image

Costs for Triage Levels 1 and 2 were examined based upon the reason that the Level was assigned (medical, psych, immunization, or other). For Triage Level 1, costs were highest in the 60 days following the IFCHE when “psych” was chosen as a reason ($1660.75), followed by medical ($738.84), immunization ($625.37), and other ($544.98). The cost difference was not statistically significant based on reason level was assigned for Level 1. For Triage Level 2, costs were highest in the 60 days following the IFCHE when “other” was chosen as a reason ($906.05), followed by psych ($658.68), medical ($642.86), and immunization ($608.37). This cost difference was also not statistically significant based on reason level was assigned for Level 2. Each child with Triage Level 1 or 2 could have one or more than one reasons selected for that level.

Demographic factors were also examined to determine correlation to service costs during the first 60 days following the IFCHE. Medicaid costs based on age, gender, and race were not statistically significant. However, Medicaid costs based on placement type did prove statistically significant. Youth placed in a group home incurred significantly higher health care costs during the first 60 days ($1109.97) than youth placed in a relative foster home ($736.48) or non-relative foster home ($845.47) (Fig. 3). Youth placed in a group home also incurred significantly higher emergent/urgent care costs ($102.23) than a youth in another placement setting ($33.60). There was no statistically significant difference in costs for mental health or care coordination services. The percentage of youth residing in a group home assigned to Triage Levels 1, 2, and 3 was not significantly different than the percentage assigned those Levels in other placement settings, thus the level of unmet need at foster care entry was similar regardless of placement type.

Fig. 3
figure 3

Total health care costs incurred during the first 60 days following the Initial Foster Care Health Exam categorized by placement setting. Costs were significantly higher for youth placed in a group home vs. those placed in relative or non-relative foster care

Full size image

In addition to cost analysis, time to next appointment following the IFCHE was reviewed for each child to determine if Triage Level assign impacted how quickly a child would be seen in follow-up. There was no significant difference in time to next appointment based on a child’s Triage Level, and overall a child was seen 19.3 days following the IFCHE.

Discussion

Our study examining use of a Triage Tool at foster care entry to stratify children based on unmet health care needs has three main findings. First, the tool is clinically useful as a health needs triage mechanism, and provides a better understanding of the types of unmet needs a child has at foster care entry. Second, the Triage Tool is not an accurate predictor of health care utilization as indicated by costs within the first 60 days in care. Third, youth placed in a group home setting incur significantly higher costs than youth placed in a traditional foster home.

Results showed that just under 10% of children entering foster care had the highest level of unmet need, 28.5% had a moderate level, and the majority of children had the lowest level of unmet need. This distribution allows targeted focus of health care coordination resources to those most in need of services. Literature is lacking on utilization of a TT in this population, and we were able to demonstrate the feasibility of use in a clinical setting, and the clinical utility of using such a tool to improve communication of a child’s unmet needs between the medical provider and health care coordination staff.

Our data also helps define what needs are unmet at foster care entry and the costs incurred for a child during the first 60 days in care. Our findings confirm that this population’s high level of medical, dental, developmental, behavioral, and psychosocial needs translate to high health care utilization and costs. The average monthly cost of $376 per child was well above the non-foster care Medicaid average of $70 within our Medicaid health plan. This internal data is consistent with findings in other studies that a child in foster care is more likely to incur higher health care costs than other children receiving Medicaid (Council on Foster Care, Adoption, and Kinship Care et al. 2015; Kenney et al. 2009; Kuo et al. 2015; Takayama et al. 1994).

The Triage Tool was not an accurate predictor of health care utilization as indicated by cost. The TT developed and used in this project was meant to identify unmet needs at a single point in time, so that those needs could be addressed by the health care coordination team. This study showed that unmet needs do not translate into increased cost and utilization within the first several weeks of foster care entry, since unmet needs are not necessarily costly. For example, a child behind on immunizations would be assigned a Level 1 or 2 at foster care entry, but would be able to receive these immunizations at a relatively minimal cost. Conversely, a child could be getting all their health needs met, but incur relatively high expenses due to the nature of their diseases (for example multiple medications, supplies, and office visits may be costly in the management of a stable Type 1 Diabetic). Therefore, tools designed to reflect a child’s health status as opposed to their unmet health needs may more accurately predict utilization and cost over time (Nelson and Berwick 1989; Varni et al. 2005).

Independent of age, this study found that group home placement is correlated with increased health care utilization and cost vs. placement in a traditional foster home. In particular, emergent and urgent care costs were significantly higher for youth in group homes than for children in other placement settings. This finding suggests that unique strategies may be warranted to provide care coordination and service delivery to those in a group home setting, in particular to promote primary care service utilization to mitigate the need for more emergent and urgent care. One intervention that has been shown to be valuable in the general population to improve utilization of preventative services is mobile clinics that bring services to the patient instead of requiring individuals to go to clinics for appointments (Hill et al. 2012; Kahn et al. 2003). Another intervention that has been shown effective with youth is the utilization of social media and technology to increase health knowledge and improve health behaviors and outcomes (Korda and Itani 2013; Webb et al. 2010). These and other unique strategies should be applied to foster children residing in group homes, and the effectiveness, utilization, and health care costs associated with such interventions should be investigated.

This study has several limitations. First, it is a retrospective chart review so there may be error in reading and interpreting historical medical records. Second, this was a single-site study, performed at an urban Child Advocacy Center. Reproducing this work in other geographical locations as well as other clinic settings would further validate the findings. Third, Medicaid health service and coordination claims data was reviewed for only 60 days post-IFCHE. This is a relatively short time period and it is likely that not all unmet needs were addressed during this timeframe. However, the TT was developed to identify short-term unmet needs, so this timeframe was felt to be most appropriate for the purposes of this study. Fourth, our analysis focuses on cost as an indicator of utilization. A more in-depth analysis of specific utilization patterns would allow better understanding of services accessed, but was beyond the scope of this project. Another limitation, as discussed above, is that this tool reflects the level of unmet needs at one point in time, rather than on an ongoing basis. Utilization of this tool in combination with a health status measurement tool may give a more complete picture of a child’s overall health needs and status.