Does Access to a Medical Home Differ According to Child and Family Characteristics, Including Special-Health-Care-Needs Status, Among Children in Alabama?
OBJECTIVE. Our goal was to examine relationships among access to a medical home, special-health-care-needs status, and child and family characteristics in one Southern state. We hypothesized that access to a medical home is influenced by several family and child sociodemographic characteristics, including special-health-care status.
METHODS. We used data from the 2003 National Survey of Children's Health. The study sample comprised all Alabama resident children. The main dependent variable was a medical home; the primary independent variable classified children according to children-with-special-health-care-needs status. We controlled for child age, gender, race, family structure, health status, insurance coverage, household education, and poverty. We first explored means or proportions for the study variables and then estimated multivariate logistic regression models.
RESULTS. Children with special health care needs were significantly more likely than children without special health care needs to have a personal doctor or nurse, to have a preventive health care visit in the previous 12 months, and to have good communication with their provider. Children with special health care needs were also more likely to experience problems accessing specialty care, equipment, or services. Being uninsured, living at or near the federal poverty level, in a household where no one completed high school, being black, having less than excellent or good health, and living in a nontraditional family structure were characteristics associated with being less likely to have a medical home. In general, children-with-special-health-care-needs status was not related to having a medical home, but dependency on prescription medicine was.
CONCLUSIONS. Assuring that all children, irrespective of family income, have access to and are enrolled in health insurance plans will move us closer to the national goal of having a medical home for all children, especially those with a special health care need, by 2010.
The term “medical home” entered into use almost 4 decades ago, described as a central place for information regarding a child's medical care, especially for children with special health care needs (CSHCN).1,2 The American Academy of Pediatrics (AAP) first defined the medical home as a concept and method of service delivery, rather than a specific location, in 1992.3 Later, the AAP expanded and operationalized the definition, stating “…medical care of infants, children, and adolescents ideally should be accessible, continuous, comprehensive, family centered, coordinated,4 compassionate, and culturally effective.”5 The latest clarification contains 7 components covering 37 specific topics.6 Although these recent policy statements and interpretations have added clarity to the definition, measurement issues persist.7–9
Defining and measuring CSHCN, including children with chronic conditions expected to last for an extended period of time, presents additional challenges. Varying definitions and measurements have resulted in estimates ranging from 2% to >30%.10–13 According to the Maternal and Child Health Bureau (MCHB): “Children who have special health care needs are those who have … a chronic physical, developmental, behavioral or emotional condition and who also require health and related services of a type and amount beyond that required by children generally.”14
An evolving national agenda for CSHCN has included a call to provide comprehensive community-based systems of care for CSHCN by 2000, later extended to 2010.15,16 The MCHB further articulated outcomes. Core Outcome 2 states: “CSHCN will receive coordinated, ongoing, comprehensive care within a medical home.”14
This study examines relationships among access to a medical home, CSHCN status, and child and family characteristics in 1 Southern state. Improved understanding of how access to a medical home varies by special-needs status and other factors may enhance understanding of disparities in accessibility and availability of a medical home. We hypothesized that access to a medical home is influenced by family and child sociodemographic characteristics, including special health care need status.
We used data for all Alabama children (N = 2167) from the 2003 National Survey of Children's Health (NSCH). Human subjects review was not required for this study. The survey design is described briefly in the article by Kogan and Newacheck17 in this issue; more in-depth information can be found elsewhere.18 Our main dependent variable was a medical home measure constructed by using 3 criteria: (1) the child must have a personal doctor or nurse (if not, no additional questions were asked); (2) children with a personal doctor or nurse must have at least 1 preventive medical care visit in last 12 months (if not, no medical home); and (3) children with a personal doctor or nurse and preventive care visit must also consistently get needed care. For this criterion, the respondent must answer 4 questions: (1) a personal doctor or nurse “usually or always” spends enough time and communicates well; (2) the child “usually or always” gets needed care and advice from a personal doctor or nurse; (3) “few or no problems” accessing needed specialist care, services, or equipment; and (4) a personal doctor or nurse “usually or always” follows up after the child receives specialist care, services, or equipment. These 4 questions pertain to family-centered, compassionate, culturally appropriate, accessible, comprehensive, and coordinated care. The questions and scoring algorithms for the Medical Home section of the NSCH were developed through the work of an advisory group convened by the MCHB and the AAP (D. Read, MPH, written communication, February 6, 2006).
Our primary independent variable, CSHCN status, was computed by using the CSHCN screener questions. The first part of each screener question asked whether a child: (1) needed/used medicine prescribed by a doctor (other than vitamins); (2) needed/used more services than is usual; (3) was limited/prevented in ability to do what most same-aged children can do; (4) needed/got special therapy; and/or (5) had an emotional, developmental, or behavioral problem that needed treatment. Those answering “yes” were asked whether the condition was because of a medical, behavioral, or other health condition, and if yes, whether that condition has/is expected to last 12 months or longer. We constructed 3 definitional domains to further describe CSHCN: dependency on prescription medicine (question 1), service use (questions 2, 4, and 5), and functional limitation (question 3). All 3 parts of any screener question were answered “yes” for the child to be categorized as a CSHCN or to qualify for a definitional domain.
Control variables used in the analysis included child's age, gender, race, family structure, health status, number of children in family <18 years of age, highest household education, health insurance, and poverty. Health insurance coverage over the past 12 months was categorized as consistent, inconsistent (insured, but not continuously over the past 12 months), or uninsured. Poverty was measured from <100% to >400% of the federal poverty level (FPL), with an additional category for children with missing poverty information.
For most variables, <1% of the observations were missing, but ∼10% of the observations were missing poverty data. The multivariate results derived from dropping these missing observations might be biased if data on poverty were not missing at random. We created a dummy variable for missing poverty and recoded the missing values on existing poverty categories to zero. Missing values on all other variables were not included in the calculation of summary statistics or in the multivariate analysis.
We first explored the means and proportions for our medical home index and its components. We also assessed the independence between the distribution of medical home and its components by CSHCN status. Because our data were survey weighted, we used a modified test statistic that has an F distribution to assess the independence between discrete random variables.19 Next, we estimated the proportion of children with medical home access for each of the independent variables. Confidence intervals (CIs), SEs, and unadjusted odds ratios (ORs) are reported.
We estimated multivariate logistic regression models where medical home access was our dependent variable and CSHCN was our main independent variable. We controlled for individual child and family demographic and socioeconomic characteristics. Because the survey uses a complex sample design with clustering of children within households and stratification of households within states, we used the sampling weights to produce estimates representative of Alabama's child population. SEs of the estimated parameters were corrected by using weights provided in the NSCH data set. We used Stata 9.020 for analysis and followed the statistical approach suggested in the NSCH methodology report to estimate the correct SEs of the parameters.18
To assess goodness of fit, we used the F-adjusted mean residual test to compare weighted observed and predicted probabilities.21 If the difference is high, then the F value obtained was high, and we rejected the null hypothesis of a close fit between observed and predicted probabilities. A small F value would preclude rejecting this null hypothesis.
Table 1 shows the proportion of children meeting each and all of the medical home needed care components. Approximately 48% of non-CSHCN Alabama children (n = 430086) had a medical home, as did 51% of those classified as CSHCN (n = 104336). Significantly fewer non-CSHCN (85%) than CSHCN (91%) had a personal doctor or nurse (P < .05), the baseline criterion for a medical home. Among CSHCN, ∼89% also had a preventive medical visit during the past 12 months, and 85% reported enough time and good communication with their personal doctor or nurse. These proportions were also significantly higher than for non-CSHCN children (75% and 79%, respectively). Notably, CSHCN were significantly more likely to experience greater problems accessing specialty care, services, or equipment (17% vs 6%), thus less likely to meet this criterion pertaining to accessible and comprehensive care.
More than 90% of both groups of children consistently received needed care and advice from a personal doctor or nurse. Only 60% to 65% of respondents for both groups reported that their personal doctor or nurse usually/always followed up after the child received the specialist care. This last component measures care coordination and has the lowest proportions for any of the medical home components.
For each independent variable, the proportions with medical home access are presented in Table 2. Among the nonexclusive definitional domains for CSHCN, 55% of the children with dependency on prescription medications had access and were 1.8 times more likely than non-CSHCN to have a medical home. Forty-four percent of the CSHCN who qualified in the service use domain and 37% in functional limitations had a medical home, but neither of these groups was significantly more likely than non-CSHCN to have a medical home. Characteristics associated with being less likely to have a medical home included black race, less than excellent or very good health, and living in a family not 2-parent biological or adoptive. The proportion of children having a medical home decreased as age and poverty increased and was lower in households in which the highest education was less than high school and for uninsured children.
Table 2 also presents ORs for the association of the independent variables with medical home. Because medical home access did not vary significantly by gender, it was dropped from subsequent analysis. Importantly, the dummy variable for missing poverty was significantly associated with medical home, suggesting that data on poverty was not missing randomly with respect to the distribution of the dependent (medical home) variable.
Results of multivariate logistic regression to identify factors associated with access to a medical home are given in Table 3. Contrary to conventional wisdom, CSHCN in Alabama were no more likely than non-CSHCN to have a medical home. The final model also indicated that children <12 years of age and those living in families with more children <18 years are more likely than older children and those in families with fewer children <18 years to have a medical home. In addition, being black, in less than excellent or good health, living in a household in which the highest education level was less than high school and/or the income was below 100% FPL, being in a family with other than 2 biological, step, or adoptive parents or a single mother, and being uninsured in the past 12 months were characteristics significantly associated with being less likely to have a medical home. Having inconsistent health insurance in the past 12 moths was associated with lower likelihood of medical home, but the result was not statistically significant. The F value for the mean residual test (0.80) resulted in a P value of .61, suggesting that we cannot reject the null hypothesis that the model has a good fit.
A separate model estimated the adjusted odds for access to medical home for the 3 CSHCN domains. Adjusted for the same variables in Table 3, dependence on prescriptions was associated with access to medical home (adjusted odds ratio [aOR]: 1.77; 95% CI: 1.15–2.71), whereas service use (aOR: 0.69; 95% CI: 0.42–1.14) and functional limitations (aOR: 0.75; 95% CI: 0.35–1.58) were associated with lower odds of having a medical home, although this result was not statistically significant.
Although the proportion of non-CSHCN children in Alabama (48%) and CSHCN (51%) with a medical home compares favorably with the national percentages of 47% and 44%, respectively, these statistics indicate considerable room for improvement in meeting AAP medical home criteria.6 Examination of the separate CSHCN definitional domains indicated that only prescription drug dependency was significantly related to having a medical home. Previous research showed that children who take prescription medicine may differ from CSHCN who qualify for other definitional domains.22 In our study, children dependent on prescription medicine were more likely to have a medical home, and although not statistically significant, children with service use or functional limitations were less likely to have a medical home controlling for other family or child factors.
Overall, being uninsured, living at or near the FPL, in a household where no one completed high school, being black, having less than excellent or good health, living in a nontraditional family structure, and being an adolescent were all associated with less likelihood of a medical home. Many of these factors are well-established risk indicators for poor health and developmental outcomes, and lacking the kind of medical care that is provided in a medical home can add to this risk. Conversely, having a medical home in which “accessible, continuous, comprehensive, family centered, coordinated”,4 “compassionate and culturally effective”5 care is delivered may ameliorate these risks.
Our results indicate that having special health care needs is only one, and not the strongest, factor associated with a child's likelihood of having a medical home. Children who visit the doctor more often—younger children, those in households with more younger children, and those dependent on prescription medications for some condition—are more likely to have a medical home. Having a medical home is clearly linked to access to care, and access to care is strongly influenced by continuity in insurance coverage.23 Improving insurance coverage may be critical to reaching the goal of a medical home for all children.24 Child health insurance rates in Alabama have improved, particularly because of the start of the State Children's Health Insurance Program in 1998; however, the state's child poverty rate at 24% remains higher than the national rate of 18%.25 Alabama's large black and growing Latino populations could benefit from the medical home model for health care delivery, which can meet cultural, familial, and medical needs in a comprehensive and coordinated manner.
Our findings also have policy and practice implications. Although access to care coordination did not differ significantly between children with and without special health care needs, the finding that only 60% to 65% of children received follow-up to needed care has considerable clinical and practice significance. Other national studies have found care coordination to be the least widely implemented medical home component.26 CSHCN programs need to develop strategies to better support medical homes for CSHCN through care coordination and community-based services. In addition to providing clinical services for CSHCN, Alabama's state CSHCN program, Children's Rehabilitation Service (CRS), provides care coordination. Care coordination is the key to the philosophical and practical implementation of care in a medical home.27,28
Recognizing that increased technology use is another strategy to communicate CSHCN needs more effectively to medical homes, CRS is creating an electronic mailbox system through which specialty clinic reports and updated service plans are sent to the primary care physician via a secure Web site. In addition, the medical home will do online referrals to CRS through the secure Web site. By expanding their clinical evaluation program, CRS will provide more support to local physicians to increase their coordination and communication efforts, thereby promoting access to medical homes. Although reimbursement and time constraints continue to challenge the implementation of the medical home and specifically care coordination, state efforts, such as those provided by CRS, will be important elements in moving the national CSHCN agenda forward.
Our results must be interpreted in light of several limitations. First, the measures of CSHCN status and medical home are self-reported and subject to the usual pitfalls of such data. For example, one of the components of the medical home index is whether the child received needed care, but respondents may differ on the “threshold” they used to report whether the child received care. Thus, parameter estimates may measure the effect of the independent variables on the “threshold.” Standard logistic regression models assume a constant threshold across observations, with that threshold typically normalized to zero. Second, our operational measure of the medical home, based on the work of Bethell et al,7 differs somewhat from that used for the core MCHB outcome measure9 and medical home definition used in the National Survey of Children With Special Health Care Needs and other surveys.7 These differences may affect estimates and policy implications.
Both program planners and CSHCN researchers can use these findings. Assuring that all children, irrespective of family income, have access to and are enrolled in health insurance plans will move us closer to the national goal of having a medical home for all children, especially those with a special health care need, by 2010.15 Researchers can use the results to design studies that better measure the concepts and components inherent in the definitions of medical home and CSHCN. Building care systems that include the components and support necessary to assure the availability of a medical home for all children requires such an interdisciplinary endeavor that responds to changing demographic and epidemiologic conditions. Integration of principles of population-based public health with pediatric practice may assist in identifying both needed health systems reform and required changes in delivery of pediatric services, especially for CSHCN.29
- Accepted September 15, 2006.
- Address correspondence to Beverly A. Mulvihill, PhD, University of Alabama, School of Public Health, Department of Maternal and Child Health, RPHB 320, 1530 Third Ave South, Birmingham, AL 35294-0022. E-mail:
The authors have indicated they have no financial interests relevant to this article to disclose.
- ↵American Academy of Pediatrics, Council on Pediatric Practice. Pediatric Records and a “Medical Home.” Standards of Child Care. Evanston, IL: American Academy of Pediatrics; 1967:77– 79
- ↵Sia C, Tonniges TF, Osterhus E, Taba S. History of the medical home concept. Pediatrics.2004;113(5 suppl) :1473– 1478
- ↵American Academy of Pediatrics, Task Force on Definition of the Medical Home. The medical home. Pediatrics.1992;90 :774
- ↵American Academy of Pediatrics, Committee on Children With Disabilities. Care coordination: integrating health and related systems of care for children with special health care needs. Pediatrics.1999;104 :978– 981
- ↵American Academy of Pediatrics, Committee on Pediatric Workforce. Culturally effective pediatric care: education and training issues. Pediatrics.1999;103 :167– 170
- ↵American Academy of Pediatrics, Medical Home Initiatives for Children With Special Needs Project Advisory Committee. The medical home. Pediatrics.2002;110 :184– 186
- ↵Bethell CD, Read D, Brockwood K; American Academy of Pediatrics. Using existing population-based data sets to measure the American Academy of Pediatrics definition of medical home for all children and children with special health care needs. Pediatrics.2004;113(5 suppl) :1529– 1537
- ↵Ireys HT, Katz S. The Demography of Disability and Chronic Illness Among Children: Mosby's Resource Guide to Children With Disabilities and Chronic Illness. St Louis, MO: Mosby; 1997:3– 12
- ↵McPherson M, Arango P, Fox H, et al. A new definition of children with special health care needs. Pediatrics.1998;102 :137– 140
- ↵US Department of Health and Human Services. Healthy People 2000: National Health Promotion and Disease Prevention Objectives. Washington, DC: US Department of Health and Human Services; 1991
- ↵US Department of Health and Human Services. Healthy People 2010. Conf ed. Washington, DC: US Department of Health and Human Services; 2000
- ↵Kogan MD, Newacheck PW. Introduction to the volume on articles from the National Survey of Children's Health. Pediatrics.2007;119(suppl 1) :S1– S3
- ↵Blumberg SJ, Olson L, Frankel M, Osborn L, Srinath KP, Giambo P. Design and operation of the National Survey of Children's Health. Vital Health Stat 1.2005;(43):1–124
- ↵Rao J, Scott AJ. On χ2 tests for multiway contingency tables with cell proportions estimated from survey data. Ann Stat.1984;12 :44– 60
- ↵Stata Corp. Stata Statistical Software [computer software]. College Station, TX: Stata Corp; 2006
- ↵Archer KJ, Lemesha S. Goodness-of-fit test for a logistic regression model fitted using survey data. Stata J.2006;6 :97– 105
- ↵Newacheck PW, McManus M, Fox HB, Hung YY, Halfon N. Access to health care for children with special health care needs. Pediatrics.2000;105 :760– 766
- ↵Kogan MD, Newacheck PW, Honberg L, Strickland B. Association between underinsurance and access to care among children with special health care needs in the United States. Pediatrics.2005;116 :1162– 1169
- ↵Annie E. Casey Foundation. Annie E. Casey Foundation Kids Count Data Book, 2005. Baltimore, MD: Annie E. Casey Foundation; 2005
- ↵Antonelli RC, Antonelli DM. Providing a medical home: the cost of care coordination services in a community-based, general pediatric practice. Pediatrics.2004;113(5 suppl) :1522– 1528
- ↵Palfrey JS, Sofis LA, Davidson EJ, Liu J, Freeman L, Ganz ML. The Pediatric Alliance for Coordinated Care: evaluation of a medical home model. Pediatrics.2004;113(5 suppl) :1507– 1516
- ↵Goldhagen J. Integrating pediatrics and public health. Pediatrics.2005;115(4 suppl) :1202– 1208
- Copyright © 2007 by the American Academy of Pediatrics