Short interpregnancy intervals are associated with a number of adverse outcomes for both mother and neonate, including increased risk of preterm birth, low birth weight, and preeclampsia,1–5 making prevention of short interpregnancy intervals a public health priority in the United States. Specifically, the 2020 Healthy People objectives call for a 10% reduction of pregnancies that occur within 18 months of a previous birth.6
There has been little systematic national surveillance of pregnancy interval length in the United States. Instead, most relevant studies focus on single states and key subpopulations.7,8 National-level baseline estimates in Healthy People 2020 relying on unpublished bivariate analyses from the National Survey of Family Growth suggest that approximately one third of interpregnancy intervals in the United States are below this 18-month threshold.6 However, to our knowledge, there has been no national-level investigation of the correlates of short interpregnancy intervals in a multivariate framework.
The potential association between pregnancy intentions and interpregnancy intervals is of particular interest. Strategies to reduce unintended pregnancy may indirectly effect the prevalence of short interpregnancy intervals; in particular, reducing mistimed pregnancies (those reported as occurring sooner than a woman desires) would increase interpregnancy interval length. Moreover, a 2001 study in Denmark found that unplanned pregnancy was associated with an increased risk of interpregnancy interval of 9 months or less.9 Given high rates of unintended pregnancy in the United States,10 similar investigation of the relationship between pregnancy intention and short interpregnancy intervals using national-level data is needed.
Our research objective was to identify the characteristics of women associated with short interpregnancy intervals to inform programs and policies aimed at reducing the occurrence of short interpregnancy intervals. We use nationally representative data to analyze the prevalence and correlates of short interpregnancy intervals, giving particular attention to the role of pregnancy intention.
MATERIALS AND METHODS
The Institutional Review Board of the second author's organization (Department of Health and Human Services identifier institutional review board 00002197) determined that the project was exempt from Institutional Review Board approval. The data from this study were drawn from the 2006–2010 National Survey of Family Growth, a periodic national probability survey of the noninstitutionalized population aged 15–44 years in the United States conducted by the National Center for Health Statistics. The survey used a multistage, stratified, clustered sampling frame to collect interviews continuously from June 2006 to June 2010. Methods of data collection and dissemination of the public use data set were reviewed by the Institutional Review Board at National Center for Health Statistics for protections of human subjects. Further methodologic details are available elsewhere.11,12 Face-to-face interviews were conducted with 12,279 women who answered detailed retrospective questions about their pregnancy experiences; the response rate was 78%. Data for each reported pregnancy were contained in a separate data file and linked to the primary respondent file.
The interpregnancy interval was calculated as the time elapsed between the conception date of any second or higher-order birth (hereafter labeled as index pregnancy) and the date of a previous birth. Pregnancies conceived within 18 months after a previous birth were classified as having a short interpregnancy interval, the convention of previous studies and the Healthy People 2020 objective.
This analysis excluded all pregnancies not ending in live birth (ie, miscarriage, abortion, and stillbirth) because of likely underreporting of these other pregnancy outcomes and the relevance of birth spacing for perinatal health. We also excluded all multiple births from the preceding and index pregnancies. To reduce issues of retrospective reporting bias and to focus on experiences during a period of current policy interest, we limited our sample to second or higher-order singleton births born in the 5 years preceding the interview (n=2,265); the preceding singleton live births used to calculate the interval could have occurred outside this 5-year window. We also excluded pregnancies with implausible interpregnancy interval lengths because of erroneous reporting (n=5) and pregnancies with missing covariates (n=7). Thus, the final sample size for analysis included 2,253 second and higher-order singleton births that occurred within 5 years of the interview date, were preceded by a singleton birth, and had complete data.
Independent variables selected for this analysis include the following measures of mothers' sociodemographic characteristics: race or ethnicity; union status at conception of index pregnancy; Medicaid-funded delivery of index pregnancy; and completed education at the time of the interview. Measures related to childbearing include number of births before the index pregnancy, age at conception of the index pregnancy, and age at initiation of childbearing, as well as pregnancy intention of the index pregnancy. Pregnancy intention was determined from a series of questions women were asked to assess their feelings right before they got pregnant. Based on their responses, each pregnancy was classified following conventional measurement approaches for this concept as intended (wanted and on time, later than wanted, or indifferent), mistimed (wanted but occurring sooner than desired), or unwanted.13
The present study used three analytical approaches. First, we used simple and multivariate logistic regression to estimate unadjusted and adjusted odds ratios for the relationship between short interpregnancy intervals and maternal demographic and childbearing characteristics. Multivariate models included all of the measures with a significant bivariate relationship to short interpregnancy intervals (α=0.05); union status, parity, and education were retained in the models for theoretical reasons. Because age at first birth and age at most recent conception are essentially a linear combination of the interpregnancy interval length for most pregnancies in the sample, we estimated models controlling for each of these factors separately. Model 1 controls for age at conception of the index pregnancy, and model 2 controls for age at first birth. The second analytical approach focused on pregnancies occurring from the short interpregnancy interval and used simple logistic regression to examine variation in pregnancy intention by maternal demographics among these pregnancies.
Finally, we estimated a hypothetical estimate of the share of short interpregnancy intervals if unintended pregnancies (mistimed and unwanted pregnancies) were averted. For this calculation, all pregnancies in women who reported they had not wanted to have any (more) children were removed from both the numerator and the denominator. For pregnancies that were reported by the mother as mistimed (occurring too soon), we recalculated the interpregnancy interval length by adding the number of months by which the pregnancy was reported to be mistimed to the actual interval. For example, if the observed interpregnancy interval was 12 months but the mother reported that the pregnancy occurred 9 months before desired, the new interpregnancy interval would be 21 months and would be shifted to a longer interpregnancy interval. Interpregnancy interval lengths for pregnancies reported as intended were unchanged. All analyses were weighted and used the svy command prefix in Stata 12.1 to adjust for the complex survey design of the National Survey of Family Growth.
Among the 2,253 pregnancies in our sample, the average interpregnancy interval was 34.0 months. One third (35%) were conceived within 18 months of a previous birth, meeting our criteria of a short interpregnancy interval (Table 1). The majority of pregnancies had an interpregnancy interval of 18 months or more, with 50% occurring at 18–59 months and 16% occurring at 60 months or more.
There is some evidence of associations between measures of pregnancy intention, childbearing history, and short interpregnancy intervals (Table 2). Pregnancies reported as mistimed or unwanted were significantly more likely to have short interpregnancy intervals compared with pregnancies reported as intended (unadjusted odds ratios 4.3 and 1.8, respectively). Short interpregnancy intervals were significantly inversely associated with age at conception of the pregnancy. In contrast, births to women initiating childbearing before age 30 years were significantly less likely to have shorter interpregnancy intervals than births to women aged 30 years and older at first birth.
There was limited variation in the share of short interpregnancy intervals by other core demographic measures (Table 2). Births to non-Hispanic black women were significantly more likely than those to Hispanic women to have short interpregnancy intervals, as were births with deliveries paid for by Medicaid. However, maternal union status at conception and maternal education were not associated with short interpregnancy intervals at the bivariate level.
Results from multivariate analyses predicting the likelihood of having a short interpregnancy interval are shown in Table 3. Although both models include different measures of age, the results are generally similar to those found in the bivariate results. Both models also indicate that union status, which was not significant at the bivariate level, is a significant predictor of interpregnancy interval length; the adjusted odds of having a short interpregnancy interval were higher among births to married compared with single women. Additionally, model 1 provides evidence that births were more likely to have short interpregnancy intervals in high-parity women or in women with a college degree, whereas model 2 suggests that interpregnancy intervals are more likely among Medicaid delivery births or among non-Hispanic black women.
Table 4 further explores the associations between pregnancy intentions and short interpregnancy interval by examining the proportion of all short interpregnancy interval pregnancies (n=791) that were intended. We find that 45% of these pregnancies were reported as intended by the mother and this varied significantly across all of the childbearing and demographic measures examined. Short interpregnancy interval pregnancies to more advantaged mothers were more likely to be intended; 59–70% were intended among births to those aged 30 years or older at first birth, college graduates, and those not using Medicaid to pay for delivery. Similarly, approximately half of the short interpregnancy interval pregnancies were intended among births to white women, those married at conception, and those aged 30–44 years at the most recent birth.
We calculated the extent to which preventing unintended pregnancies would reduce the share of all pregnancies with short interpregnancy intervals by assuming that all unwanted pregnancies in our sample were averted and that mistimed pregnancies were appropriately timed by the mother. Based on these counterfactual assumptions, we estimate that the prevention of unintended pregnancies would reduce the proportion of short interpregnancy intervals overall from 35% to 23%.
Using recent nationally representative data, we estimate that more than one out of three second or higher-order singleton births occur after a short interpregnancy interval. Nearly 7% are conceived within 6 months of a previous birth.
Age plays an important role in short interpregnancy intervals for two distinct groups of women. First, we identify short interpregnancy intervals as a correlated and troublesome outcome of second births to teenage mothers; two thirds of births in this age group had a short interpregnancy interval. Although teenage mothers comprise only a small number (6%) of all second and higher-order births, additional interventions are needed to address suboptimal birth spacing in this population. Second, women with a first birth at age 30 years or older are more likely to experience short interpregnancy intervals than those initiating childbearing earlier, suggesting that closer birth spacing is a response to later initiation of childbearing. This premise is supported by the finding that among pregnancies with short interpregnancy intervals in women initiating childbearing after age 30 years, nearly three out of four were intended pregnancies. For this group, short interpregnancy intervals appear to be a choice and not an unintended outcome.
With 55% of short interpregnancy interval pregnancies unintended, helping women achieve their desired pregnancy intentions is the "low-hanging fruit" for public health interventions to reduce the share of short interpregnancy interval pregnancies. Improvements in contraceptive use for women can further reduce rates of unintended pregnancy and, by extension, short interpregnancy intervals. Long-acting reversible contraceptives, such as intrauterine devices and implants, seem particularly well-suited to lengthening the interpregnancy interval.14,15 However, this approach will only go so far because we estimated that alleviating all unintended pregnancies among these second and higher-order births—an exceptionally lofty goal—would still leave 23% with short interpregnancy intervals.
Further supporting the idea that closely spaced births may be part of a strategy for family building was the finding that more than half of short interpregnancy interval births to more advantaged women were reported as intended. Increasing interpregnancy interval length among intended births is more challenging and likely requires health care providers to educate and counsel patients about the potential negative health consequences of short interpregnancy intervals. Further research is needed to reevaluate the evidence base for negative health consequences of short interpregnancy intervals among the substantial number of intended births to more advantaged women. Any suggestion of promoting longer pregnancy intervals for these women must weigh benefits against the potential health risks and decreased fecundity associated with increasing maternal age at birth.16,17
The choice of an 18-month cut-off to define a short interpregnancy interval in this analysis was based on the indicator used in Healthy People 2020. Although the literature suggests that interpregnancy intervals less than 18 months are associated with increased risk, it is important to note that even within this 18-month window, the level of risk likely decreases as interval length increases.1 Additionally, although the Healthy People 2020 objective is limited to reduction of short interpregnancy intervals, there is evidence that interpregnancy intervals more than 60 months are also detrimental to maternal and neonatal health.1 Considering that 16% of pregnancies in our sample had interpregnancy intervals more than 60 months, these data suggest that approximately half (51%) of interpregnancy intervals in the United States are outside generally recommended standards.
The National Center of Health Statistics, as part of the Healthy People 2020 objectives, tracks short interpregnancy intervals using the same National Survey of Family Growth data we analyzed here. However, National Center of Health Statistics reports as their baseline measure the share having a short interpregnancy interval during the 5 years preceding the interview as opposed to the share. Because each woman can only provide a single pregnancy to the numerator, their measurement approach is biased away from measuring short interpregnancy intervals; women with shorter intervals may have more than one pregnancy during the period and thus have relevant pregnancy experiences excluded. Our pregnancy-based measure conceptually parallels the stated objective to reduce the share of short interpregnancy interval pregnancies and is methodologically sound in incorporating all reported pregnancies ending in live birth. Future monitoring should use a pregnancy-based measure, as we have here.
Although linked vital records often are used to assess the causes and consequences of interpregnancy length, they are unable to provide pregnancy intention status, which, as demonstrated here, is a key determinant of interpregnancy length. Another source of data, the Pregnancy Risk Assessment Monitoring System, provides information on interpregnancy length and pregnancy intention, but not all states participate in data collection efforts. Therefore, this study drew strength from its ability to link interpregnancy intervals and maternal characteristics, including pregnancy intentions, from a recent nationally representative large sample of pregnancies.
Although the National Survey of Family Growth data are widely utilized and considered highly reliable and valid, there are always limitations to self-reported data. Relevant to this study, women may potentially misreport birth dates or gestational age of a child, resulting in a miscalculation of the length of the interpregnancy interval. External validation with medical records of this self-reported data was not possible; however, any misreporting is likely random and should not bias the observed relationships. The response rate of the National Survey of Family Growth was 78%, which may have resulted in underrepresentation of certain high-risk groups. Likewise, our exclusion of multiple births also may have affected our estimate of short interpregnancy interval pregnancies. Finally, there has been concern about bias in the retrospective reporting of pregnancy intentions if women adjust their reporting of births toward more intended farther from the actual time of conception.18 Limiting the analyses to a 5-year retrospective period minimizes this concern and follows methodologic approaches established in previous analyses of this measure from the National Survey of Family Growth.10,13
1. Conde-Agudelo A, Rosas-Bermudez A, Kafury-Goeta AC. Birth spacing and risk of adverse perinatal outcomes: a meta analysis. JAMA 2006;295:1809–23.
2. Conde-Agudelo A, Rosas-Bermudez A, Castaño F, Norton MH. Effects of birth spacing on maternal, perinatal, infant, and child health: a systematic review of causal mechanisms. Stud Fam Plan 2012;43:93–114.
3. Khoshnood B, Lee KS, Wall S, Hsieh HL, Mittendorf R. Short interpregnancy intervals and the risk of adverse birth outcomes among five racial/ethnic groups in the United States. Am J Epidemiol 1998;148:798–805.
4. Klerman LV, Cliver SP, Goldenberg RL. The impact of short interpregnancy intervals on pregnancy outcomes in a low-income population. Am J Public Health 1998;88:1182–5.
5. Zhu BP, Rolfs RT, Nangle BE, Horan JM. Effect of the interval between pregnancies on perinatal outcomes. N Engl J Med 1999;340:589–94.
8. Bloebaum L, Baksh L, McGarry J, Carapezza D, Hossain S, Streeter N. Short interpregnancy spacing in Utah. PRAMS PERSPECTIVES 2006. Available at: http://health.utah.gov/mihp/pdf/IPI.pdf
. Retrieved January 4, 2013.
9. Nabukera SK, Wingate MS, Kirby RS, Owen J, Swaminathan S, Alexander GR, et al.. Interpregnancy interval and subsequent perinatal outcomes among women delaying initiation of childbearing. J Obstet Gynaecol Res 2006;34:941–7.
10. Finer LB, Zolna MR. Unintended pregnancy in the United States: incidence and disparities, 2006. Contraception 2011;84:478–85.
11. Lepkowski JM, Mosher WD, Davis KE, Groves RM, van Hoewyk J, Willem J. National Survey of Family Growth, Cycle 6: sample design, weighting, imputation, and variance estimation. Vital Health Stat 2 2006;142:1–82.
12. Groves RM, Mosher WD, Lepkowski J, Kirgis NG. Planning and development of the continuous National Survey of Family Growth. Vital Health Stat 1 2009;48:1–64.
13. Mosher WD, Jones J, Abma JC. Intended and unintended births in the United States: 1982–2010. Natl Health Stat Rep 2012;55:1–28.
14. Increasing use of contraceptive implants and intrauterine devices to reduce unintended pregnancy. ACOG Committee Opinion No. 450. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009;114:1434–8.
15. Speidel JJ, Harper CC, Shields WC. The potential of long-acting reversible contraception to decrease unintended pregnancy. Contraception 2008;78:197–200.
16. Cleary-Goldman J, Malone FD, Vidaver J, Ball RH, Nyberg DA, Comstock CH, et al.. Impact of maternal age on obstetric outcome. Obstet Gynecol 2005;105:983–90.
17. Menken J, Trussell J, Larsen U. Age and infertility. Science 1986;233:1389–94.
18. Joyce T, Kaestner R, Korenman S. The stability of pregnancy intentions and pregnancy-related maternal behaviors. Matern Child Health J 2000;4:171–8.