 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODOLOGY. I used a retrospective case-control design. Low birth weight and normal birth weight infants in the Florida Medicaid program (1995–1999) matched on demographic factors were compared on their relative odds of achievement of preventive care goals and Medicaid program retention using multivariate logistic regression (n = 7510). A subsample of infants with linked maternal claims was used to measure the association of maternal health promotion during the prenatal period.
RESULTS. Low birth weight infants had lower relative odds of achieving preventive care goals during the first year and of retaining Medicaid coverage after the first year compared with infants of normal birth weight. However, when maternal health promotion during the prenatal care period was held constant, birth weight was no longer related to achievement of preventive care goals and program retention. Maternal health promotion was positively associated with achievement of preventive goals and program retention for all infants regardless of birth weight.
CONCLUSIONS. Some Medicaid-enrolled low birth weight infants are at risk for poor health supervision and poor continuity of care through failure to retain coverage. The disproportionate odds of poor health promotion among mothers of low birth weight infants explain much of this deficit. States may want to prioritize preventive care supervision and program reenrollment for children of mothers with evidence of low health promotion. This recommendation is particularly important for infants of low birth weight.
Key Words: Medicaid • maternal health promotion • program retention • preventive care
Abbreviations: AFDC—Aid to Families With Dependent Children • OBRA—Omnibus Budget Reconciliation Act • PCCM—primary care case management • HMO—health maintenance organization • ICD-9—International Classification of Diseases, Ninth Revision • AAP—American Academy of Pediatrics • OR—odds ratio • CI—confidence interval
Of the 35% of infants who are eligible for Medicaid services each year, at least 8% are born at a low birth weight (<2500 g).1–3 It has been well established that children born at a low birth weight are at greater risk for poor health and developmental delay.4 Close management of their care and enhanced developmental services have been shown to ameliorate these sequelae.5 Successful treatment for low birth weight infants depends on continued developmental assessment (such as that performed in the context of preventive care visits) and on continuity of care (such as that delivered by a designated primary care provider). This article evaluates whether low birth weight infants enrolled in Medicaid achieve preventive care goals during their first year of life and whether they remain enrolled within the Medicaid program at the end of their first year, compared with Medicaid-enrolled infants of normal birth weight. This is the first article, to my knowledge, to examine these issues specifically within the Medicaid program.
The ability of state Medicaid programs to address the needs of low birth weight infants is a function of several factors. All infants born to Medicaid-eligible mothers qualify for Medicaid services for the duration of their first year through federal mandate. This policy increases the odds both of close supervision through enriched preventive care and of continuity of care through the first year. Ideally, after the first year, these children continue to receive care and supervision through the Medicaid program or through an alternate source of coverage such as private insurance. However, the Medicaid program has historically had difficulty retaining eligible enrollees.6 For example, a recent study suggests that 28% of children leave Medicaid within a year, half of these becoming uninsured although they are still eligible for Medicaid participation.7 In addition, by virtue of their eligibility for the program, Medicaid-enrolled low birth weight infants are coincidentally exposed to the complicating effects of poverty, which have been shown to increase the odds of poor outcomes for low birth weight infants.4
A large part of the ability of Medicaid to both care for and retain eligible low birth weight enrollees may also be confounded by a pattern of poor health promotion that can be traced to the prenatal period. As Alexander and Kotelchuck8 write in their review of the effectiveness of prenatal care, inadequate maternal use of prenatal care may be considered as a "proxy indicator" of an underlying low maternal propensity toward health promotion, A mother's low propensity toward health promotion can have 2 implications for her child. First, a mother with low health promotion may engage in poor health behavior during pregnancy (eg, tobacco use) that, in turn, is associated with low birth weight. Second, a mother's low propensity toward health promotion may be manifest in the postpartum period through lower odds of seeking preventive care for her child as well as failure to maintain her child's Medicaid enrollment. Several studies have documented that poor maternal health promotion in the prenatal period is associated with poor use of preventive services for the child, although none specifically within a Medicaid population.9–11 The association of poor health promotion in the prenatal period with poor retention of health coverage has not yet been discussed in the literature.
In this article, I evaluate the relative odds of achievement of preventive care goals and remaining enrolled in Medicaid among low birth weight infants compared with infants of normal birth weight. In addition, I assess the relationship of poor maternal health promotion in the prenatal period with the relative odds of preventive care achievement during the first year of life and remaining enrolled in Medicaid after the first year among low and normal birth weight infants.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
To measure maternal health promotion, I calculated the number of months before birth that the mother's first prenatal care visit occurred. If the first visit occurred at least 3 months before the infant's birth date, I evaluated that infant's mother as obtaining "early prenatal care," thus exhibiting greater propensity toward health promotion. If the infant's mother did not initiate prenatal care until 
3 months before the infant's birth date, I evaluated that infant's mother as receiving "late prenatal care," thus exhibiting a lower propensity toward health promotion. Although this definition is one frequently used for assessment of prenatal care, it would have been preferable to control for gestational age.8 Because low birth weight is associated with preterm labor, it is possible that late entry into prenatal care as measured could reflect a shortened overall pregnancy period for some of the sample. Without information on gestational age at delivery, I was not able to correct this error. So, although I find that mothers who initiated prenatal care in the last trimester were more likely to have low birth weight infants (57% vs 43%), I cannot determine to what extent this reflects poorer baseline health promotion among mothers of low birth weight infants or a shortened third trimester resulting from premature birth. However, those mothers with strong health promotion who happened to deliver sufficiently preterm to satisfy my definition of low maternal health promotion would bias my results toward the null hypothesis of no effect.
Data Source and Study Measures
I used linked enrollment and claims files from the Florida Medicaid program to identify date of birth, preventive care use (as recorded using ICD-9 codes v03–v06 and v20.2 and Current Procedural Terminology codes 90701–90748, 99381–99392, W9843, and W9881), program retention, first-year expenditures, and demographic characteristics (gender; black, Hispanic, white, and other race/ethnicity; AFDC or OBRA eligibility; 1995–1996, 1996–1997, or 1997–1998 birth year; and county of residence) for each infant. I also used the linked enrollment and claims files to categorize infants as normal or low birth weight using ICD-9 code 765 for disorders relating to short gestation and low birth weight. Excluding extremely premature infants as noted previously, these codes generally imply a birth weight between 1000 and 2500 g. Although low birth weight is often underreported using ICD-9 coding in claims data, the overall low birth weight rate for the sample of 7.2% in 1997–1998 is within range of the overall low birth weight rate of 9.0% estimated for the Florida Medicaid population for 1998 using birth certificate data.12 Error resulting from underreporting through ICD-9 coding should bias the results toward the null hypothesis because unidentified low birth weight infants will be counted as of normal birth weight.
I used linked enrollment, claims, and family case files to assess sibling participation in Medicaid, maternal age, whether the mother was enrolled in Medicaid before the index pregnancy, and the month of the mother's first prenatal care visit (as recorded using ICD-9 codes v22 and v23, and Current Procedural Terminology codes 99201–99205 and 99211–99215).
I defined appropriate preventive care as achievement of the 6-month American Academy of Pediatrics (AAP) preventive care goal (receipt of at least 4 preventive care visits) at some point during the first year.13 I used this generous definition to allow for delays resulting from variation in the preventive care schedule for low birth weight infants and because full compliance with AAP recommendations is rare among Medicaid populations.14 Forty-eight percent of the matched sample met this defined preventive care goal. Fifty percent of the subsample of infants with linked maternal claims met the goal.
I defined program retention as remaining within the Medicaid program for the 14th and 15th month after birth, allowing families 2 months to complete the reenrollment process. Eighty-six percent of the matched sample satisfied this definition of program retention. Eighty-five percent of the subsample of infants with linked maternal claims satisfied this definition, corresponding with a similar estimate of disenrollment among infants covered by public insurance in Tennessee.15 I was not able to determine who was ineligible for reenrollment at the end of the first year, so I assume all enrollees were eligible for the purposes of analysis. As a consequence, my calculated rates of program retention are underestimated. However, any bias resulting from this error should be in the direction of the null hypothesis given the established association between low birth weight status and poverty. In other words, I would expect infants who were no longer eligible for participation in Medicaid after the first year because of achieving higher income would be more likely to be of normal birth weight. One indication that the magnitude of this error is reasonably small is a recent estimate based on Current Population Survey data that 14% of eligible but uninsured children disenrolled annually from the Florida Medicaid program during the years 1998–2001.7
Analysis
I used logistic regression to compare infants of low and normal birth weight on the relative odds of achievement of preventive care goals in the first year and program retention after the first year, adjusting for gender, race/ethnicity, Medicaid eligibility category, sibling status, birth year, and county fixed effects.16 I also adjusted the retention model for total first-year expenditures by quartile because previous research has suggested a positive association between health care expenditures and retaining health coverage.5 Attributable primarily to expenses related to and directly after birth, the low birth weight infants in the sample had an average expenditure amount more than 5 times that of normal birth weight infants (see Table 1).
Among the subsample of infants with linked maternal claims, I used logistic regression to evaluate the degree to which maternal health promotion, as measured by late entry into prenatal care, is associated with variations in preventive care achievement and Medicaid retention among low and normal birth weight infants. First, I estimated 2 main-effects models to estimate the relative odds of preventive care achievement and retention as a function of low birth weight and late entry into prenatal care, adjusting for whether the mother was an adolescent and for the infant's gender, race/ethnicity, eligibility category, sibling status, birth year, and county of residence. I also adjusted the retention model for the log of the infant's first-year expenditures. I then tested a second-order interaction term of maternal health promotion by birth weight status in both models to evaluate how the association between birth weight status and the relative odds of preventive care achievement and program retention might vary with maternal health promotion.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The importance of maternal health promotion (as measured by early or late entry into prenatal care) is reflected among the subsample models estimating the achievement of preventive care goals and program retention. Greater maternal health promotion in the prenatal period was positively related to the relative odds that the infant would receive preventive services in the first year and that the infant would remain enrolled. These findings on the overall importance of maternal health promotion within this Medicaid population add to the existing literature on the correlates of maternal prenatal behavior.8–10
The positive effects of maternal health promotion on achievement of preventive care goals and retention of coverage were similar for both low and normal birth weight infants, although slightly stronger among low birth weight infants in the context of program retention. It is important to note, however, that the definition used to measure AAP preventive care goal achievement was generous to allow for possible delays in preventive care necessitated by poor health status associated with low birth weight. It may be that there are finer differences between low and normal birth weight infants using a more sensitive measure. Also, I was not able to distinguish if the preventive care was delivered within the context of a well-child or sick visit, whereas it may be that only preventive care delivered within the context of a well-child visit is a true correlate of maternal health promotion. It is possible that low birth weight infants had more sick-visit opportunities for preventive care given their higher likelihood of compromised health.
In this study, the most important difference between low and normal birth weight infants regarding achievement of preventive care goals and remaining enrolled in Medicaid was their baseline degree of maternal health promotion, as measured by when their mothers first sought prenatal care. The fact that health promotion was lower among mothers of low birth weight infants compared with mothers of normal birth weight infants is not surprising. Some of the infants of mothers satisfying the definition of poor maternal health promotion may have been preterm births because of an inability to correct for gestational age. Also as discussed, one of the correlates of low maternal health promotion may be low birth weight as a result of other negative health promotion behaviors such as tobacco use. However, this article provides evidence that these health promotion attitudes endure into the postpartum period and are associated with lower relative odds that low birth weight infants receive preventive services and benefit from continuity of care within the Medicaid program.
Given the increased need for developmental supervision and continuity of care among children born at low birth weight, these findings have important implications for the Medicaid program. States may want to prioritize preventive care supervision and program reenrollment for children of mothers with evidence of low propensity toward health promotion. This recommendation is particularly important for low birth weight infants, for whom the consequences of poor health promotion may be significant, costly, and enduring. One possible point of intervention may be just after the infant's birth when mothers report prenatal care use while applying for their child's birth certificate. At this juncture, the infant's birth weight status is known and the mother's degree of health promotion can be easily assessed. As Kogan et al (1998)9 have described, identification of at-risk infants at delivery could lead to interventions such as leveraging the involvement of managed care providers to closely follow these infants, targeted parent education regarding the importance of preventive care and continuity of care, and enhanced case management services. Such intervention would be relatively low cost and, if effective, could greatly improve health and developmental outcomes for Medicaid-enrolled low birth weight infants.
|
FOOTNOTES |
|---|
Address correspondence to Shanna Shulman, PhD, Mathematica Policy Research, Inc, 955 Massachusetts Ave, Suite 301, Cambridge, MA 02139. E-mail: sshulman{at}mathematica-mpr.com
The author has indicated she has no financial relationships relevant to this article to disclose.
Dr Shulman's current address is Mathematica Policy Research, Inc, 955 Massachusetts Ave, Suite 801, Cambridge, MA 02139.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
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. |
||
|
||
