3Assessing the Impact of Care Coordination for CSHCN

Publication Details

Search Criteria

A search of the literature was conducted using the following databases and search terms. The PubMed database, the Social Sciences & Humanities Index, and the Social Services Abstracts databases were searched (1988 – February 2006; includes HealthSTAR since 2000). For these databases, a broad search was conducted using MeSH headings and terms:

  • “Disabled children” or “Chronic illness[Multi]” or “Chronic disease” or “Catastrophic illness” or “Special health care needs” or “Special needs” or “Activity limitations.”
  • In conjunction with: “infant” or “child” or “child, preschool” or “adolescent.”
  • In conjunction with “Care coordination” or “Case management” or “Medical home” or “Family” or “Parents.”

These search criteria were designed to reflect the scope of this technical review and to capture articles that were likely to most directly relate to the definitions of CSHCN and care coordination used in this review. (For an assessment of a more comprehensive definition of care coordination, see the review by McDonald and colleagues.25) Of the generated references, those considered for detailed review also had to be empirical evaluations of care coordination programs that involved more than one diagnostic group. These criteria were chosen because of the expressed focus on empirical evaluation. All evaluation studies regardless of study outcomes (including utilization, costs and health outcomes) were included in the detailed review. Articles concerned with interventions directed at only one diagnostic group (e.g. asthma) were not included in the review. The concern here was that such experiences may be too narrow and not reflect appropriately the needs of a diverse population of children with special health care needs, the central focus of this review. For example, asthma is of far higher prevalence, often of lower severity, and generally requires fewer services than many of the other conditions of concern in assessing care coordination for children with special health care needs. In addition, a forthcoming systematic review addresses comprehensively quality improvement strategies for asthma care, including those that improve care coordination. In contrast, there has not been a recent review of generic care coordination programs designed to address the needs of children with a variety of clinical disorders. The articles generated by the search criteria above were examined to identify those that were specifically evaluative in nature. These studies and their relevant references were reviewed by one of the authors (PHW).

Recently, West and colleagues described a suggested approach to rating the quality and strength of empirical studies.38 The authors presented a selected set of scales and checklists that clinicians, policymakers, and researchers can use to assess study quality and the strength of scientific evidence. We used these assessment methods to review the identified studies.39 (See Appendix A for review criteria).

Review Findings

Despite considerable efforts to describe and prescribe care coordination initiatives for CSHCN, there is a striking paucity of empirical evidence regarding its ultimate impact. For example, the 1999 AAP statement on care coordination cited only 2 evaluative studies in support of its recommended strategies; only one had been published in the prior 15 years. The 2005 AAP policy statement cited only three studies that included children with special health care needs.

Our review identified seven studies published in the past 15 years that evaluated the impact of care coordination interventions on health care utilization, costs, or health outcomes.28, 3842 The nature of the coordination programs, the study designs, the health status and social status of the children served, and the outcomes measured varied considerably. A summary of these programmatic and evaluative elements are presented in Table 1. A summary of the study quality characteristics of the reviewed studies are presented in Table 2.

Table 1. Characteristics of evaluations of care coordination interventions for children with special health care needs.

Table 1

Characteristics of evaluations of care coordination interventions for children with special health care needs.

Table 2. Summary of Quality Elements of Reviewed Studies.

Table 2

Summary of Quality Elements of Reviewed Studies.

Criscione and colleagues using a randomized control trial evaluated a coordinated care model on hospital admissions for a population of people (primarily adolescents and adults) with developmental disabilities.40 Individuals in the coordinated care group had shorter average lengths of stay and lower hospital charges than did the group receiving standard care, especially when charges were adjusted for case mix. The authors calculated substantial savings associated with this reduced use of hospitalization services.

A study by Fields et al. described the experience of a coordination model for 28 technology-dependent children living at home.28 The coordination of care was intensive and was led by care coordinators from a community-based home care agency. Although a longitudinal case-series design, the number of children studied was inadequate to ascertain the impact of care coordination on utilization, costs, or health outcomes. However, parents did report a high level of satisfaction with care coordination to the extent that it allowed their children to be cared for at home, an important consideration in the care of these seriously disabled children.

Liptak et al. reported a descriptive study of a hospital-based care coordination program in Rochester, New York.41 The authors compared hospitalization patterns among children with chronic illness admitted to the major tertiary pediatric hospital in Rochester with those of a group of other tertiary pediatric hospitals in other areas of the country and with national data. The study reported that hospital length-of-stay and associated costs were lower in Rochester than in the comparison groups. Approximately, half of the costs of the coordination program could not be recovered from standard charges to insurance plans. Rather, supplemental funds were obtained from a capitation agreement with local insurers to support the care coordination program. The authors calculate that the savings from reduced hospital charges far exceeded the costs associated with the coordination program. The authors do not account for the fact that hospitalization rates for children in Rochester appear to be significantly lower in general than other studied cities.42

Perhaps, the most instructive study regarding the impact of care coordination for CSHCN was performed by Palfrey and colleagues, reporting on the results of the Pediatric Alliance for Coordinated Care (PACC) model.43 Using a pre/post survey of parents, the authors assessed an intervention model which relied on nurse coordinators based in a consortium of primary care practices in the Boston area. The intervention model was a medical home strategy that relied on six components:

1.

The services of a designated pediatric nurse practitioner.

2.

Consultation from a local parent of a child with special health care needs.

3.

Modification of office routines.

4.

Implementation of an individualized health plan (IHP).

5.

Regularly scheduled continuing medical and nursing education.

6.

Expedited referrals and communication with specialists and hospital-based personnel.

A total of 150 children with a variety of major health problems were recruited in six pediatric practices. Although this group was derived from a total of 222 referred families, it was unclear what portion of the 72 families not enrolled in the study resulted from exclusion criteria or refusal to participate, an important distinction in assessing the potential for selection bias. Families were administered surveys prior to the intervention and two years after the intervention was initiated. Thirty-three of the original 150 families were not available for the follow-up survey, leaving 117 for analysis. The study found that parent satisfaction with pediatric primary care improved subsequent to the medical home intervention. Specifically, there was improvement in being able to speak with the same nurse by phone, ease in obtaining letters of medical necessity, receiving resources, obtaining acute medical care as well as specialty referral, receiving prescriptions, understanding the child's condition and setting goals for care, and improving the relationship with the child's physician. These improvements were noted to be greatest for children with the most severe conditions. However, overall satisfaction with their primary care source did not change as satisfaction levels were high prior to the intervention. Emergency room visits and school days missed did not change significantly over the study period. Parents reported a reduction in days they missed work and in their child's need for hospitalization. However, given the study pre/post design, the authors could not assess the extent to which these findings were due to the child growing older rather than program effects. The authors estimated that the care coordination program cost approximately $400 annually for each child enrolled. This figure did not include any estimate of savings or additional expenditures due to possible changes in utilization resulting from care coordination activities. Although the studied children were generally quite severely affected by their conditions, this study represents the first community practice-based trial of a medical home model for children with special health care needs.

Farmer and colleagues utilized the PACC medical home model developed by Palfrey et al. to improve care coordination among CSHCN in three university-affiliated primary care practices in the rural Midwest.44 A total of 51 children were studied via a pre/post parental survey; this represented approximately one-third of the children eligible for the study as the remaining two-thirds declined participation or could not be reached for the follow-up survey. Post intervention, parents reported reduced caregiver strain, parents' missed work days, their children's school absences, and ambulatory care visits. Hospitalizations did not decrease. Parental satisfaction with care coordination and access to mental health services improved but satisfaction with overall primary care fell slightly subsequent to the intervention. No cost data were provided. Two studies evaluated the impact of coordinative interventions on the mental health outcomes for CSHCN. A study by Chernoff et al.45 utilized organized linkages to “experienced mothers” and child life specialists to provide support and logistical expertise to families engaged in the program. Although not specifically directed at care coordination, the nature of the intervention included guidance and support for coordinating care elements. In direct response to calls for more rigorous evaluation of coordinative and mental health interventions for CSHCN, this study used a randomized, controlled trial design to assess the impact of a 15 month program. The measured outcomes were focused on four areas: adjustment and role skills; depression; anxiety, and self-perception. Overall, the intervention had modest results. In part, this may have been due to the relatively small number of children who entered the trial with highly abnormal scores on the administered tests. The intervention seemed to have its greatest positive effects in children initially found to be at highest risk for the examined psychosocial outcomes.

A similar study was conducted by Pless et al.46 which evaluated the impact of a nursing intervention program to improve the adjustment of children with chronic disorders. Children recruited from a variety of specialty clinics were randomized to a program utilizing nurses to address parental stress, parental competency and overall family functioning. Control families received standard nursing care from the specialty clinics. In an effort to restrict substantial nursing involvement to the program group, families to be entered into the program group were recruited from clinics without standard nursing involvement. However, this strategy yielded too few recruitments and was abandoned. The portion of the control and program groups derived from clinics with preexisting and ongoing nursing involvement was not provided. Outcomes were generally positive although somewhat mixed, particularly among different age groups. This complex set of findings may have been due in part to the varied utility of the measures used for different developmental ages and underlying disabling conditions.

Although not a trial of care coordination per se, a study by Antonelli and Antonelli provided important insight into the costs associated with care coordination for CSHCN in primary care practice.47 The authors documented in detail all care coordination activities performed in a primary care pediatric practice over a 95 day period. Assessed activities included telephone discussions, contacts with consultants, form processing for schools, camps, etc, meetings, written reports for SSI or other agencies, letters, chart reviews, and patient-focused research. Costs were calculated on the basis of time spent on the care coordination elements multiplied by the average salary of the office personnel performing the service. Over the study period, 774 encounters on 444 separate patients were logged. Half of the encounters were considered related to issues not traditionally considered medical, such as processing referrals, consulting with schools, and oversight for psychosocial problems. Based on national salary data, the annual cost of the care coordination activities in this care model for the studied CSHCN ranged between a 25th percentile of $22,809 and a 75th percentile of $33,048, which per child averaged $51 and $71, respectively. The authors concluded that these costs while considerable were not prohibitive for their practice setting. However, the study also underscored the relatively large costs associated with care coordination that are not directly reimbursable under current payment mechanisms. The cost figures in this study were substantially lower than those estimated by Palfrey et al. for the PACC model. However, the children entered into the PACC program were a more severely affected group than those in the Antonelli and Antonelli study, which included children with the kind of broad range of severities usually found in general pediatric practices.