 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Health Policy Research Program, Public Policy Center
b Departments of Preventive and Community Dentistry
d Pediatrics
c Child Health Specialty Clinics, University of Iowa, Iowa City, Iowa
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
DESIGN. Data from 2 sources were matched at the individual level: (1) the 2002 Iowa Consumer Assessment of Health Plans Study survey of Medicaid enrollees and (2) Iowa Medicaid administrative claims, encounter, and enrollment files.
PARTICIPANTS. The subjects were 1140 children aged 6 months to 12 years for whom both sources of data were available.
MAIN OUTCOME MEASURE(S). Outcomes measures included medical homeness, as developed by a scale of items in the Consumer Assessment of Health Plans Study survey, and outpatient costs, as determined from Medicaid administrative data.
RESULTS. From the regression models, we found that (1) for all Medicaid-enrolled children, outpatient costs were significantly higher for female children and children and youth with special health care needs, (2) for children and youth without special health care needs, costs were significantly higher for female children, those with a personal doctor or nurse, and those with more of a medical home, and (3) for children and youth with special health care needs, costs were significantly higher for those in a lower health state, for those in health maintenance organization 2, and for older children.
CONCLUSIONS. Although the degree of medical homeness was not related to outpatient costs for children and youth with special health care needs, medical homeness may affect inpatient costs more than outpatient costs for children and youth with special health care needs and should be investigated further.
Key Words: medical home • CAHPS • CYSHCN • outpatient costs
Abbreviations: AAP—American Academy of Pediatrics • CYSHCN—children and youth with special health care needs • CAHPS—Consumer Assessment of Health Plans Study • TANF—Temporary Assistance for Needy Families program • FPL—federal poverty level • CAHMI—Children and Adolescent Health Measurement Initiative • HMO—health maintenance organization
In a 1992 report, the American Academy of Pediatrics (AAP) formally proposed the concept of a medical home for children that goes beyond having a regular source of medical care.1 The medical home was further defined in 2002 by the AAP Medical Home Initiative for Children With Special Health Care Needs Project Advisory Committee with 7 dimensions of care identified as necessary to have a medical home (accessible, family centered, continuous, comprehensive, coordinated, compassionate, and culturally effective).2 Children and youth with special health care needs (CYSHCN) have been identified as a group that could particularly benefit from the coordinated care provided by a medical home. In the Maternal and Child Health Bureau’s "National Agenda for Children With Special Health Care Needs," having a medical home is identified as one of the 6 critical indicators of progress.3
In general, costs of health care for CYSHCN are 
3 times higher than for other children because of the higher need for and use of services.4 Although the presence of a medical home should provide improvements in access to and quality of care, there is less agreement about the impact on health care costs, especially costs for outpatient care. Without a medical home, costs could be lower because of unmet need for medical and behavioral or emotional care. Alternatively, the lack of a medical home could lead to higher costs, because care may be sought from more expensive emergency departments, walk-in clinics, or hospitals.2 Within a medical home, well-coordinated primary care could reduce overall medical costs by reducing unnecessary hospitalizations and preventing future needs. Or, conceivably, costs for outpatient care could increase by increasing the receipt of needed primary and preventive services.
Two different approaches have been used to operationalize the AAP definition of the medical home from existing data. Using data from the National Survey of Children With Special Health Care Needs, Strickland et al5 evaluated the impact of a medical home on access to care, defining medical home as that which provides (1) a usual place for care, (2) a personal doctor or nurse, (3) no difficulty obtaining needed referrals, (4) needed care coordination, and (5) family-centered care.
Bethell et al6 compared alternative methods for identifying a medical home from existing survey data sets, including using items from the Consumer Assessment of Health Plans Study (CAHPS) survey instrument. They used 34 items from the CAHPS–Children With Chronic Conditions survey that were deemed relevant to the AAP definition of the medical home to create a measure of medical home. After calculating mean scores from domain scores, they were transformed on a scale of 0 to 100. Although Bethell et al suggested that the concept of having a medical home was conceptually a continuum, they and Strickland et al both defined the medical home as a yes/no rate-based measure in their published analyses.
The purpose of this study was to (1) operationalize the AAP definition of the medical home as a continuum (medical homeness) using existing data from a CAHPS-based survey instrument and (2) evaluate if the degree of medical homeness impacted the cost of outpatient health care.
There are 2 related primary hypotheses for this study. First, as the level of medical homeness increased for children without special health care needs, the costs of outpatient medical care would increase. This should be a result of greater use of routine and preventive services. Second, as the level of medical homeness increased for CYSHCN, the costs of outpatient medical care would decrease. This should be a result of better care coordination and prevention.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The 2 sources of data used in this study were (1) the 2002 Survey of Iowa Medicaid Enrollees and (2) Iowa Medicaid administrative claims, encounter, and enrollment files. These are relatively unique sources of data in that utilization and cost information can be matched with survey responses at the individual level. The CAHPS 2.0 survey was the foundation of the Survey of Iowa Medicaid Enrollees with the addition of the Children and Adolescent Health Measurement Initiative (CAHMI) items to identify CYSHCN.7–9 A mixed-mode survey process was used to collect information from a stratified random sample of 2831 children who had been enrolled for at least 6 months in 4 different Medicaid managed care plans. The survey was fielded from October 2001 to February 2002, and 1226 parents/guardians responded for a 43% response rate (for more information about the Medicaid enrollee survey see ref 10). Of the 1226 responses, 1140 (93%) met CAHPS criteria for a valid survey (ie, at least 50% of applicable items were completed) and had complete data on the demographic and health-status items relevant to this analysis. Index development and analysis were conducted on this sample of 1140 Iowa Medicaid-enrolled children.
Medical Homeness Index Construction
A retrospective approach was used to create an index score from the survey data indicating the degree of medical homeness for each child. All items in the survey were first examined for face validity in measuring any of the 7 AAP domains of a medical home. The staff of the Iowa Medical Home Initiative was involved in the face-validity item review. Eleven questions were ultimately identified as measuring a particular aspect of one of the medical home domains on the basis of a balance of conceptual salience and empirical testing (Table 1). Four questions were added to the index for CYSHCN about care coordination, specialist referrals, and ability to receive behavioral/emotional health care on the basis of the salience to this population.
|
To combine the responses for questions with different response categories and create a single score for medical homeness, responses for each item were first linearly transformed to a 0–100 possible range for uniformity. The medical home score was then computed by summing the transformed (0–100) responses across the 11 items (15 for the CYSHCN scale) and dividing by the number of items completed, yielding a medical home index score with a possible range of 0 to 100. Because of skip-out patterns in the survey instrument, the number of items completed for the medical homeness index ranged from 1 to 11 or 1 to 15. The average number of completed items for all children on the 11-item scale was 8.2 (SD: 2.8). Cronbach's 
for the 11-item scale was .84 for all children, .79 for non-CYSHCN only, and .89 for CYSHCN. The 15-item index for CYSHCN had an value of .91.
There were a number of survey items that were considered for inclusion but dropped for conceptual, internal-reliability, or multicolinearity reasons: (1) "whether the child has a personal doctor or nurse" was deemed more appropriate to enter into the model as a separate measure because it was not measuring the experience within the medical home like the other items and was not highly correlated with other items in the scale; (2) "time since last preventive visit" was not highly correlated with other items in the scale; (3) "problems getting prescription medication" were judged to be more likely affected by plan and pharmacy than having a medical home; (4) "problem finding a personal doctor when joining Medicaid" may not be currently relevant, because the item does not indicate how long ago the person joined Medicaid; (5) "delays in care due to need for plan approval" were considered more of a health plan issue than a medical home issue; and (6) "got in to see specialist if needed" was not relevant for enough children and was not highly correlated with the other medical home–related items. Although there were no separate measures for culturally effective care related to language barriers or interpreters, doctor communication and staff helpfulness/courtesy do reflect some degree of cultural effectiveness.
Outpatient Health Care Utilization and Costs
Information was collected from Iowa Medicaid administrative data for all children participating in Medicaid managed care plans who had been enrolled for at least 6 months before the initial time of the survey in October 2001 (n = 1140). Administrative data were used to determine utilization and cost of outpatient health care procedures from October 2000 to October 2001. These data were subsequently matched to the survey information at the level of the individual child. Outpatient health care procedures were defined as all health care services billed by a health care provider on a Health Care Financing Administration (now called the Center for Medicare and Medicaid Services) 1500 form. Provider-submitted charges on the administrative data were used to measure the cost of the outpatient care.
Analysis Plan
Distributions of both the 11-item and the 15-item medical home indices were highly negatively skewed, with median scores of 91 and 88, respectively. Therefore, quartile measures are presented and nonparametric methods (
2, Mann-Whitney U, and Kruskal Wallis tests) were used in the bivariate significance tests. The primary dependent variable in the study, outpatient costs, was then modeled using multivariate ordinary least-squares regression. Independent variables included demographic characteristics, health plan, health status, having a personal doctor or nurse, and the medical homeness scale score. Because of multicolinearity and conceptual salience, only 1 measure of the child's health status was used in each model. The presence or absence of a special health care need was used as the health-status indicator in the cost model for "all children," whereas the excellent-to-poor health-status scale was used to indicate health status when modeling children with and without a special need separately. The data in the regressions were weighted back to all children in the TANF who were enrolled for at least 6 months.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
50% of these children scoring <80 on the 0–100 medical homeness scale. Also, there was a decreasing trend in mean medical homeness score for children with reported excellent health status to those with reported poor health status. The same decreasing trend direction, but slightly weaker, was seen for younger age groups to older age groups.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Other findings from the bivariate analyses support common understandings about the relationships of health status and existence of a special health care need on both outpatient costs and medical homeness scores. Specifically, outpatient costs increase as reported health status decreases, whereas medical homeness scores decrease as health status decreases. Similarly, having a special health care need is associated with higher outpatient cost and a slightly lower medical homeness score. Neither outpatient costs nor medical home scores were associated with the reported family income.
This study is important as one of the few that have been able to match outpatient utilization data with survey information at the individual level, allowing for the evaluation of relationships between CYSHCN, medical homeness, and outpatient costs. In this exploratory study, a reasonably comprehensive measure of medical homeness with adequate face validity and internal reliability was developed retrospectively from the commonly used CAHPS, Living With Illness Measures, and CYSHCN survey instruments that measure most constructs of the AAP definition of medical home. As the ability to link data increases and there is more access to data such as CAHPS through the National CAHPS Benchmarking Database, improved development of an operational definition of medical homeness from these data should be considered, including measures of continuous care.
Although medical home can be measured as a yes/no construct, a medical homeness scale was chosen for this study to evaluate the concept of medical home as a continuum. Domain scores could also be developed for each of the 7 AAP domains of a medical home before creating the final medical homeness score. However, as shown in Table 1, the questions from the CAHPS survey overlapped various medical home domains, making it difficult to separate domains and leading to the creation of a single measure in this exploratory approach. Additional evaluation of the contribution of each of the 7 domains to the value of having a medical home should be undertaken.
Of importance to policy makers is the relatively high rate of CYSHCN in the TANF-eligible Medicaid managed care population. Twenty-six percent of the children in Medicaid managed care plans were identified as having a special health care need, which is significantly higher than the 17% found in a statewide study of all children in Iowa using the same CAHMI CYSHCN screening instrument or the 12.6% found in the National Survey of Children With Special Health Care Needs.4,8 This has important implications for Medicaid programs, because the children enrolled through the TANF are generally considered to be healthier than the CYSHCN who are enrolled in the Social Security Income program. This high proportion of CYSHCN in Medicaid managed care indicates that care coordination and other services important for CYSHCN should be available to all children in Medicaid.
Medicaid programs should also be encouraged to add the CYSHCN screener and either the Living With Illness measures or additional modules from the CAHPS survey for Children With Chronic Conditions instrument to their regular consumer-assessment surveys. These modules can work particularly well for Medicaid programs using telephone data collection, in which the insertion of different modules for children identified as having a special need is more flexible and cost-effective.
Because this was a retrospective approach to developing a medical homeness scale, the study was limited to data that were already collected. A prospective approach would have the obvious advantage of developing the questions more directly to the AAP medical home domains; however, it would not be as generalizable to programs using CAHPS data. The CAHPS instrument in current use is the 3.0 version and, although not greatly different from the 2.0 version, has some differences that should be considered with the analysis of more recent data. All administrative data have the limitation of being collected for billing purposes and were not designed for research. Any services that were provided but not submitted for reimbursement or not accurately coded could affect the quality of utilization and cost data used in this study.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
We thank Adrienne Akers (Early Intervention Research Institute) for assistance with this study.
|
FOOTNOTES |
|---|
Address correspondence to Peter C. Damiano, DDS, MPH, Health Policy Research Program, University of Iowa, Public Policy Center, 227 S Quadrangle, Iowa City, IA 52242. E-mail: peter-damiano{at}uiowa.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  N. Murphy and B. Christian Disability in Children and Young Adults: The Unintended Consequences Arch Pediatr Adolesc Med, October 1, 2007; 161(10): 930 - 932. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
