| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Obstetrics and Gynecology, Brown University Medical School, The Center for Statistical Sciences, Brown University, Providence, Rhode Island; and Robert Wood Johnson Clinical Scholars Program and VA Center for Practice Management and Outcomes Research, Ann Arbor, Michigan.
Address reprint requests to: Maureen G. Phipps, MD, MPH, Women & Infants Hospital of Rhode Island, 101 Dudley Street, Providence, RI 02905; E-mail: mphipps{at}wihri.org.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: We combined the comprehensive 1996 and 1997 United States birth cohorts to compare postneonatal mortality rates among maternal age groups. With postneonatal death as our main outcome measure, we used multivariable logistic regression to model adjusted odds ratios.
RESULTS: The postneonatal mortality rate for infants born to mothers 15 years old and younger was substantially higher (3.2 per 1000) than that of infants born to mothers 23–29 years old (0.8 per 1000) and remained substantially higher after adjusting for maternal race or ethnicity. Even after adjusting for maternal race or ethnicity, prenatal care utilization, and marital status, infants born to early adolescent mothers had a three-fold higher risk (odds ratio 3.0, 95% confidence interval 2.5, 3.6) of postneonatal death compared with adult mothers.
CONCLUSION: Healthy infants born to early adolescent mothers are at increased risk of postneonatal death. Many of these deaths are potentially preventable; therefore, developing targeted postnatal support services specifically designed to address the needs of healthy infants born to adolescent mothers might have a positive effect on the lives of these children.
Infants born to mothers 15 years old and younger are at increased risk of death within the first year after birth compared with infants born to older mothers. The elevated risk is consistent across racial and ethnic groups and can be partly explained by early adolescent mothers’ higher rates of premature delivery and low birth weight infants. Although biologic immaturity might explain some of the elevated risk associated with young maternal age, poor social conditions surrounding pregnancy and childbirth also appear to significantly affect the risk of poor birth outcomes.1,2
Few studies have explored the association between maternal age and infant death among full-term, normal birth weight infants without congenital anomalies—in other words, healthy infants.3 Neonatal deaths, deaths within 28 days after birth, are often attributed to complications from birth, such as premature delivery, low birth weight, and other biologic causes such as congenital anomalies. However, postneonatal deaths, deaths more than 28 days after birth, are often associated with infection, congenital anomalies, sudden infant death syndrome, and intentional as well as unintentional injuries.3,4 We know that infants born to early adolescent mothers have higher neonatal mortality rates secondary to higher rates of preterm delivery and higher rates of delivering low birth weight infants. However, we do not know whether healthy infants born to early adolescent mothers are also at increased risk of postneonatal death compared with healthy infants born to adult mothers. If infants born to adolescent mothers are at increased risk of neonatal death due to poor social conditions, then postneonatal infants might also be at risk, especially if similar social conditions influence the adolescent mother and her infant.
We examined postneonatal mortality rates for infants identified as healthy at birth and compared these rates among five predetermined maternal age groups. We also estimated the risk of postneonatal death associated with maternal age, adjusting for race and several other factors. We specifically addressed the question, "Are normal birth weight, full-term infants born to early adolescent mothers at greater risk of postneonatal death compared with such infants born to adult mothers?"
|
METHODS |
|---|
|
TOP
ABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We included only singleton, first births to mothers 12 to 29 years old. Multiple births and higher-order births were not included in this analysis because these infants are at higher risk of morbidity and mortality compared with singleton births. Furthermore, there are relatively few higher-order births in the youngest age group. Our definition of healthy infants included infants with gestational age of 37 weeks or greater and birth weight of 2500 grams or greater. Infants who survived the neonatal period (first 28 days after birth) were included, except for those identified as having congenital anomalies on their birth certificate. We limited the racial stratification of our data set (n = 1,830,350) to three main subgroups that were sufficiently large to permit meaningful analyses (non-Hispanic white [n = 1,246,863], non-Hispanic black [n = 303,699], and Mexican Americans [n = 279,788]).
Demographic information included in the data set was extracted from birth certificate records. There is no information in this data set about the location of the infant after birth or who cared for the infant after discharge from the hospital. Maternal age was classified into previously established age categories that have been shown to predict risk.5 These age groups are 15 years old and younger (n = 48,294), 16–17 years old (n = 192,937), 18–19 years old (n = 326,300), 20–22 years old (n = 427,348), and 23–29 years old (n = 835,471).
Statistical analyses were performed using SAS 8.0 (SAS Inc., Cary, NC), S-PLUS (Insightful Corp., Seattle, WA), and STATA (Stata Corp., College Station, TX) statistical software. The main outcome measure, post-neonatal infant death, is defined as infant death between 28 and 365 days after birth (n = 2516). All postneonatal mortality rates were calculated per 1000 infants who survived the neonatal period. Logistic regression was used to model postneonatal mortality risk by maternal age group. Because postneonatal death is a rare event in both exposure groups, the adjusted odds ratios obtained from the logistic regression models adequately approximate the associated relative risks.
The 23- to 29-year-old age group was used as the reference group for all models. Covariates were chosen on the basis of the literature and biologic plausibility for influencing the main outcome.4,6,7 We evaluated the following covariates along with appropriate interaction terms in our models: adequacy of prenatal care utilization, maternal race or ethnicity, self-reported tobacco use, self-reported alcohol use, and marital status. Prenatal care utilization was categorized according to the R-GINDEX measure.8 We compared no prenatal care and inadequate prenatal care utilization to moderate and adequate prenatal care utilization. This measure is used as a proxy for health care access and utilization for the child.9 Marital status, a dichotomous variable in this analysis, was used as a measure of social support. Although this marker does not capture nonmarital relationships,10 it is the best proxy measure for social support using vital statistics data.
The father’s name is sometimes missing on the infant’s birth certificate. Previous research has suggested that missing information about the father on the birth certificate is associated with increased risk of infant death.11 Most (62%) of the birth certificates in our database did not contain information on the father of the infant for single mothers who were 15 years old or younger. We created a dichotomous variable to indicate when information about the father was missing and used it as a covariate in the regression analysis. To fully determine the effect of this variable, we developed a model specific to the subset of single mothers, because children born to married women automatically have the father’s name recorded on their birth certificate.
We examined the underlying cause of death as identified by the International Classification of Disease, Ninth Revision (ICD-9) codes listed on the death certificate. We compared the causes of death among individual categories. Although not all deaths in the following categories result from neglect or abuse, we grouped them in our analysis of deaths from possible neglect or abuse: inhalation and ingestion of food or other object causing obstruction of respiratory tract or suffocation (E911-E912), accidental mechanical suffocation (E913), other accidental causes and adverse effects (E800-E910, E914-E949), child battering and other maltreatment (E967), other homicide (E960-E966, E968-E969), and sudden infant death syndrome (798.0). We calculated postneonatal mortality rates separately for each of these causes of death by maternal age group (data not shown). We then combined these ICD-9 codes to create a single dichotomous variable for postneonatal death from causes resulting from possible neglect or abuse. Independent associations were evaluated using logistic regression with covariates in the models as outlined above.
Kaplan-Meier curves and box plots were used to examine the distribution of survival times among infants who did not survive through the first year. Specifically, we examined the data to determine whether death times differed among maternal age groups. We considered the possibility that if infants born to mothers of different ages are dying from different causes, they might die at different times.
This data set represents the entire, finite, population of births from the United States for the years 1996 and 1997. Therefore, we can exactly determine the proportion of infants who did not survive the first year. However, in a finite population, this proportion is not necessarily the probability that an infant will not survive the first year in another time. The proportion from the whole population in 1996 and 1997 serves as an estimate of the probability for other infants. As a result, confidence intervals are necessary when we estimate an infant’s risk of dying during the first year, but such confidence intervals are not used when we describe the patterns of infant deaths for the specific population in our data set.
|
RESULTS |
|---|
|
TOP
ABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
|
shows how often several traditional maternal risk factors were reported by maternal age group. For birth records with information reported, the early adolescent group had the highest percentage of unmarried mothers and the highest percentage of inadequate prenatal care utilization. However, early adolescent mothers did not have high rates of self-reported tobacco or alcohol use.
|
). When adjusted for maternal race or ethnicity, marital status, and prenatal care utilization, the early adolescent group still had a three-fold higher risk of postneonatal death compared with the adult reference group. The risk is somewhat elevated for 16–17 year olds, 18–19 year olds, and 20–22 year olds; however, the greatest risk is in the youngest age group of mothers.
|
). This result was determined after controlling for paternal acknowledgement on the birth certificate and prenatal care utilization. In the Mexican-American group, the risk of postneonatal death in the youngest age group of mothers was elevated; however, the risk was essentially the same for the 18- to 19-year-old and the 20-to 22-year-old groups compared with the 23- to 29-year-old reference group. In contrast, among both the non-Hispanic white and non-Hispanic black mothers, risk was elevated for all three teen-mother groups and modestly elevated in the 20–22-year-old group.
|
Multivariable analysis for the risk of postneonatal death from potential abuse reveals that the crude odds ratio for infants born to early adolescent mothers is five times greater than that for the 23–29-year-old reference group (Table 5). Even with adjustments for maternal risk factors, the risk remains almost four times as great for infants born to early adolescent mothers compared with the adult reference group. Again, the risk of postneonatal death from potential abuse in the older adolescent groups are also elevated compared with the 23–29-year-old reference group, suggesting a gradient of higher risk with younger maternal age.
|
where the denominator is the total number of births for the specific maternal age group), those infants are not dying at different times (denominator is the number of postneonatal deaths for the specific maternal age group). Hence, there does not appear to be a relationship between cause of death and timing of death during the postneonatal period. |
DISCUSSION |
|---|
|
TOP
ABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Young maternal age is associated with a higher risk of postneonatal death from possible abuse or neglect. In fact, young maternal age per se is a marker for neglect or abuse of the mother, as most states would define a pregnancy at 15 years old and younger a result of statutory rape.14 The underlying social condition that predisposes the early adolescent to get pregnant probably continues after the birth of her child. Early sexual activity has been associated with a history of childhood sexual abuse.15,16 Spouse or partner abuse before pregnancy is a strong risk factor for subsequent abuse.17 Spouse or partner abuse also increases the likelihood of child abuse.18 Our findings are consistent with previous studies reporting that children born to abusive families are at higher risk of postneonatal death.19
We could not assess the effect of paternal age on postneonatal infant death for infants born to early adolescent mothers because of the large amount of missing data about fathers. However, simply not identifying the father was an important risk factor. When we included not reporting the father as a covariate in our statistical model, the odds ratio for postneonatal death in infants born to early adolescent mothers decreased from three times to approximately two times more than the control group. Although some effort has been focused on identifying paternity for single mothers, this effort is often in response to state financial needs and the desire to maximize levels of child support.20 Investigation into the reason the mother does not want to identify the father or that the father does not want to be identified on a child’s birth certificate is necessary, not just for financial reasons, but because this information could help to explain this risk factor.
Since 1994, when the American Academy of Pediatrics introduced the "Back to Sleep" campaign to inform the public about the risks associated with prone sleeping positions, there has been a substantial decrease in the rate of death from sudden infant death syndrome.21 It is unclear whether this campaign has been more effective with older mothers compared with younger mothers. Our data suggest that this may be the case; infants are dying from sudden infant death syndrome at approximately the same ages for all mothers; however, there is a higher rate of death from sudden infant death syndrome in infants born to early adolescent mothers.
Although using vital statistics data gives us enormous analytic power, it leaves us without information on some variables that we would like to analyze, such as socioeconomic data. We were not able to use education as a proxy for socioeconomic status because adolescents 15 years old and younger are not expected to have graduated from high school; therefore, this marker would not yield meaningful information. Furthermore, socioeconomic status and young maternal age cannot be separated because maternal age less than 16 years old might be a marker for poverty. In this case, the variables young maternal age and poverty are classically confounded, and any amount of additional data on socioeconomic status would not affect our analysis. Simply controlling for age also indirectly controls for socioeconomic status because mothers 15 years old and younger also have low socioeconomic status. To speak meaningfully about the effect of socioeconomic status beyond that of age, for young women, a referent group of young women with high socioeconomic status must be conceptualized, which is probably not plausible because most young adolescent mothers live in poverty.22
Because this analysis was limited to healthy children, we likely underestimated the risk of postneonatal death for infants born to early adolescent mothers. Low birth weight and preterm infants are at higher risk of abuse compared with healthy infants,23 and early adolescent mothers have higher rates of these poor birth outcomes. Had we included these infants in our analysis, the findings would have been even more troubling.
There is reason to be concerned about the children born to early adolescent mothers. Based on this study as well as previous studies, we suggest that support for early adolescent mothers and their infants not be limited to prenatal and perinatal care. Most efforts aimed at preventing infant death focus on the care of premature and low birth weight infants. Unfortunately, a seemingly healthy infant born to an adolescent mother has a significantly lower chance of living to celebrate a first birthday than an infant born to an older mother. Because many of these deaths can be prevented, developing and evaluating targeted postnatal support services for healthy infants born to adolescent mothers could have a dramatic effect on the postneonatal mortality rate. Studies designed to uncover the specific risk factors that put children born to early adolescent mothers at greater risk of postneonatal death would help direct appropriate health policy and prevention programs.
|
Footnotes |
|---|
Received February 8, 2002. Received in revised form March 25, 2002. Accepted April 4, 2002.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Fraser AM, Brockert JE, Ward RH. Association of young maternal age with adverse reproductive outcomes. N Engl J Med 1995;332:1113–7.
3. Druschel CM, Hale CB. Postneonatal mortality among normal birth weight infants in Alabama, 1980 to 1983. Pediatrics 1987;80:869–72.
4. Seigel E, Scurletis TD, Abernathy JR, Surless KB. Postneonatal deaths in North Carolina, 1959–1963. N C Med J 1966;27:366–71.[Medline]
5. Phipps MG, Sowers MF. Defining early adolescent child-bearing. Am J Public Health 2002;92:125–8.
6. Semenciw RM, Morrison HI, Lindsay J, Silins J, Sherman GJ, Mao Y, et al. Risk factors for postneonatal mortality: Results from a record linkage study. Int J Epidemiol 1986; 15:369–72.
7. Mayerberg DZ, Carpenter RG, Myerberg CF, Britton CM, Bailey CW, Fink BE. Reducing postneonatal mortality in West Virginia: A statewide intervention program targeting risk identified at and after birth. Am J Public Health 1995;85:631–7.
8. Alexander GR, Kotelchuck M. Quantifying the adequacy of prenatal care: A comparison of indices. Public Health Rep 1996;111:408–19.[Medline]
9. Freed GL, Clark SJ, Pathman DE, Schectman R. Influences on the receipt of well-child visits in the first two years of life. Pediatrics 1999;103:864–9.
10. Bird ST, Chandra A, Bennett T, Harvey SM. Beyond marital status: Relationship type and duration and the risk of low birth weight. Fam Plann Perspect 2000;32:281–7.[Medline]
11. Gaudino JA Jr, Jenkins B, Rochat RW. No fathers’ names: A risk factor for infant mortality in the State of Georgia, USA. Soc Sci Med 1999;48:253–65.
12. Satin AJ, Leveno KJ, Sherman ML, Reedy NJ, Lowe TW, McIntire DD. Maternal youth and pregnancy outcomes: Middle school versus high school age groups compared with women beyond the teen years. Am J Obstet Gynecol 1994;171:184–7.[Medline]
13. Cooper LG, Leland NL, Alexander G. Effect of maternal age on birth outcomes among young adolescents. Soc Biol 1995;42:22–35.[Medline]
14. Donovan P. Can statutory rape laws be effective in preventing adolescent pregnancy?Fam Plann Perspect 1997; 29:30–4, 40.[Medline]
15. Fiscella K, Kitzman HJ, Cole RE, Sidora KJ, Olds D. Does child abuse predict adolescent pregnancy?Pediatrics 1998; 101:620–4.
16. Stock JL, Bell MA, Boyer DK, Connell FA. Adolescent pregnancy and sexual risk-taking among sexually abused girls. Fam Plann Perspect 1997;29:200–3, 227.[Medline]
17. Martin SL, Makie L, Kupper LL, Buescher PA, Morracco KE. Physical abuse of women before, during, and after pregnancy. JAMA 2001;285:1581–4.
18. Augustyn M, Zuckerman B. Impact of violence on children. In: Behrman RE, Kliegman R, Jenson HB, eds. Nelson textbook of pediatrics. 16th ed. Philadelphia, PA: W.B. Saunders;2000:109–10.
19. Roberts J, Lynch MA, Golding J. Postneonatal mortality in children from abusing families. BMJ 1980;281:102–4.
20. Dye JL, Presser HB. The state bonus to reward a decrease in ‘illegitimacy’: Flawed methods and questionable effects. Fam Plann Perspect 1999;31:142–7.[Medline]
21. Malloy MH, Freeman DH. Birth weight- and gestational age-specific sudden infant death syndrome mortality: United States, 1991 versus 1995. Pediatrics 2000;105: 1227–31.
22. Centers for Disease Control and Prevention. Poverty and infant mortality—United States 1988. MMWR Morb Mortal Wkly Rep 1995;44:923–7.
23. Johnson CF. Abuse and neglect of children. In: Behrman RE, Kliegman R, Jenson HB, eds. Nelson textbook of pediatrics. 16th ed. Philadelphia, PA: W.B. Saunders;2000:111–2.
This article has been cited by other articles:
|
|
 |
|
  C. Rosengard, L. Pollock, S. Weitzen, A. Meers, and M. G. Phipps Concepts of the Advantages and Disadvantages of Teenage Childbearing Among Pregnant Adolescents: A Qualitative Analysis Pediatrics, August 1, 2006; 118(2): 503 - 510. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Raatikainen, N. Heiskanen, P. K. Verkasalo, and S. Heinonen Good outcome of teenage pregnancies in high-quality maternity care Eur J Public Health, April 1, 2006; 16(2): 157 - 161. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rosengard, M. G. Phipps, N. E. Adler, and J. M. Ellen Psychosocial Correlates of Adolescent Males' Pregnancy Intention Pediatrics, September 1, 2005; 116(3): e414 - e419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kitson Young maternal age was associated with increased risk of postneonatal death in full term, healthy infants Evid. Based Nurs., April 1, 2003; 6(2): 57 - 57. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
