the association of change in maternal marital status between births and adverse pregnancy outcomes...

The association of change in maternal maritalstatus between births and adverse pregnancyoutcomes in the second birth

Victoria L. Holta,b, Nancy L. Danoffa,c, Beth A.Muellera,d and Marcia W. Swansona

Departments of aEpidemiology and bHealth Services, School of PublicHealth and Community Medicine and cDepartment of Pediatrics,School of Medicine, University of Washington, and dDivision of PublicHealth Sciences, Fred Hutchinson Cancer Research Center, Seattle,Washington, USA

Summary. The relationships between change in marital status betweentwo consecutive births and adverse pregnancy outcomes at the secondbirth were investigated using linked Washington State 1980±93 whitesingleton birth certificates. Women who were married at the first birthhad lower low birthweight (LBW) and small-size-for-gestational-age(SGA) rates at that birth than single women, and women married at thesecond birth had lower LBW, SGA and preterm delivery rates at thatbirth, regardless of marital status at the first birth. Adjusted relative risks(RR) of LBW and SGA were significantly increased for initially marriedwomen who were single at the second birth compared with those whoremained married (RR = 1.4 and 1.3, respectively). Risks of LBW andSGA were significantly decreased among initially single women whomarried by the second birth, compared with those remaining single(RR = 0.7 for LBW and 0.8 for SGA). We conclude that the largelyunstudied subgroup of previously married women is at increased risk foradverse pregnancy outcomes. Public health policy and programmesdirected at high-risk mothers and infants should be aware of the specificphysical and emotional needs of this group of child-bearing women.

Paediatric and Perinatal Epidemiology 1997, 11, Suppl. 1, 31±40

Address for correspondence: Dr V. L. Holt, University of Washington, School of Public Healthand Community Medicine, MCH Program, Box 357230, Seattle, WA 98195±7230, USA.

31# 1997 Blackwell Science Ltd.

5 sup DISC

Introduction

The number and proportion of births to unmarried women have increased over

the past decade in the US.1±6 Over 1.2 million infants nationwide were born to

unmarried women in 1992, an 82% increase from the number reported in 1980.6

The greatest increase in the proportion of births to unmarried women has been

among white women, from 11% of births in 1980 to 22% in 1991, but there has

been a substantial increase among African±American women as well (from 55 to

68%).3,6 These trends are concerning because of the association of unmarried

status with increased incidence of low birthweight and infant mortality.7,8 In 1991,

8% of infants born to unmarried white US mothers and 15% of infants of

unmarried African±American mothers were of low birthweight, compared with 5

and 11% of the corresponding-race infants born to married mothers.7 The adverse

health, neurodevelopmental, and emotional consequences of low birthweight are

manifold, and the cost of neonatal intensive care alone for US low birthweight

infants is estimated at $2.9 billion yearly.9±11

Most studies examining marital status as a risk factor for poor pregnancy

outcomes such as infant mortality and low birthweight have focused on

differences between women of differing marital status at the time of a single

birth.8,12±17 Although all have shown an increased risk of poor outcomes for single

women relative to married women, the possibility that marital status is a marker

for associated unmeasured factors cannot be ruled out in these analyses. Only one

study has explored the relationship of marital status and risk of poor pregnancy

outcome in two consecutive livebirths. This study, based on first and second births

recorded in the Norwegian Medical Birth Registry, found that the risk of perinatal

mortality for infants of mothers single at both births was twice that for infants of

mothers married at both births, while infants of mothers who were single at the

first birth and married at the second birth were not at increased risk.18 Although

Skjaerven and Irgens did not report on the mortality risks associated with a

change in marital status from married to single, the findings of a British study of

single births suggest that women who are divorced at the time of childbirth may

have the highest adverse outcome risks.15

The purpose of the current study was to examine the effect of change in marital

status on risk of low birthweight, specifically its two components, preterm

delivery and small-size-for-gestational-age. We wished to determine whether

women who were married at their first birth and single at the next, or single at the

first birth and married at the next, were at increased or decreased risk for adverse

pregnancy outcomes when compared with reference groups of women with the

same marital status at the first birth who did not change marital status at the

second birth. By including women who were married at one birth but unmarried

at the next, we were able to examine a previously unstudied subset of women

potentially at high risk for poor pregnancy outcomes.

# 1997 Blackwell Science Ltd. Paediatric and Perinatal Epidemiology, 11, Suppl. 1, 31±40

32 Victoria L. Holt et al.5 sup DISC

Methods

We used a linked file of Washington State birth certificates to obtain a cohort of

white women with known marital status who had their first and second singleton

livebirths in Washington State USA, between the linked file's inception in 1980

and 1993, the last year for which linked data were available. Details of the linked

file's construction have been reported elsewhere.19 The study population included

all white women who were married at their first birth and single at the second

(n = 2849), single at the first birth and married at the second (n = 13 263), single at

both births (n = 12 976) or married at both births (n = 114 183). Excluded from

analysis were the 551 women with missing marital status at one or both births.

Women with a missing birthweight for one or both births (n = 360) also were

excluded from the cohort, yielding a final study size of 142 911. Women of non-

white races constituted less than 13% of total births in the linked file, and no non-

white race accounted for more than 6% of births. These small numbers,

particularly when further divided into four exposure categories, precluded the

meaningful inclusion of non-white women in the analyses.

Marital status at both births, categorised as `married' and `single', was

determined from a checkbox on the birth certificate. While procedures to gather

birth certificate information vary from hospital to hospital in Washington State,

this variable is ascertained usually from direct query of the parturient and is one of

the most completely reported items on the birth certificate. Three outcome

variables were defined: low birthweight (LBW), preterm delivery (PTD), and

small-for-gestational-age (SGA). Calculation of low birthweight (< 2500 g)

percentages included all study subjects, while calculation of preterm delivery

(delivery at less than 37 completed gestational weeks) and SGA percentages

excluded the 5679 women with missing or improbable (under 22 weeks or over 49

weeks) values for length of gestation at the first birth and 5513 with such values at

the second. Length of gestation was calculated using the clinical estimate of

gestational age present on the birth certificate for all births from 1987 to 1993. For

earlier births or those with missing clinical estimate of gestational age, gestational

length was determined by subtracting the date of last menstrual period from the

infant's birth date. There was no appreciable difference in percentage of infants

classified as SGA by method of calculation of gestational length. Small-for-

gestational-age, defined as birthweight under the 10th percentile at a given

gestational age, was determined using gender-specific birthweight-gestational

age distributions of non-Hispanic white infants reported by Williams et al.20 These

distributions were based on over two million California singleton births be-

tween 1970 and 1976, and were computed from date of birth and last men-

strual period according to the American Academy of Pediatrics convention. A

computer screening method was used to eliminate the error in gestational ages

resulting from misdated menstrual periods, and the calculated mean 10th


Marital status and pregnancy outcomes 33

5 sup DISC

percentile values agree very closely with two clinically based studies of gestational

age and birthweight.20

To examine the associations between the outcomes LBW, PTD and SGA and

changed marital status from one birth to the next, we used Mantel±Haenzel

estimates of relative risks (RRs) and test-based 95% confidence intervals.21 In these

analyses, women who were married at both births (MM) were the referent group

compared with women married at their first birth and single at the next (MS).

Women single at both births (SS) were the referent group compared with women

single at the first birth and married at the next (SM). Potential confounders con-

sidered in these analyses were maternal age at first birth (< 20 years, 20±29 5 30),

onset of prenatal care (1st trimester vs. later), interbirth interval (< 2 years vs. 2 or

more years), and the dichotomous variables: smoking at second birth, congenital

malformations at second birth, and medical complications in second pregnancy.

Smoking data were available for births from 1984 onwards. Adjustment for

confounders was made if they altered the RR associated with LBW, PTD, or SGA

by at least 10%.

Results

Women who were married at both births began child-bearing later than those who

were single at one or both births (mean ages at first birth: MM = 25.0 years,

MS = 21.1 years, SM = 20.3 years, SS = 19.7 years), and over half of the women

who were single at their first birth were teenage mothers (Table 1). Women who

were married at both births were least likely to smoke during either pregnancy, and

most likely to obtain first trimester prenatal care. Women who were married at the

first birth and became single by the second birth were three times as likely to smoke

as women who remained married and 17% less likely to receive early prenatal care

during that first, married, pregnancy. Women who were married at the second

birth did not change their smoking behaviour from the first pregnancy to the

second, while both groups of women who were single at the second birth were

more likely to smoke during the second pregnancy than during the first. The timing

of onset of prenatal care did not vary by birth among women who were married at

the first birth, but both groups of initially single women were more likely to get

early care during their second pregnancy than their first. Approximately one-third

of MM and SS women had an interbirth interval of less than 2 years, while women

who changed marital status between births were less likely to have a short interval.

The father's identity was recorded on both birth certificates for nearly all of the MM

women, and the majority of women in the other three groups did not state

paternity for one or both births. Almost all of the MM women had the same partner

for both births, as did 30% of the SM women and 8±11% of MS and SS women.

Both groups of women who were married at the first birth were less likely than

those who were single to have a low birthweight or preterm outcome at that birth


34 Victoria L. Holt et al.

5 sup DISC

(Table 2). Women who were married at the second birth were less likely to have a

low birthweight, preterm or SGA outcome at that birth than single women,

regardless of marital status at first birth. The percentage of women with adverse

birth outcomes generally declined from the first birth to the second. This decline

was statistically significant (P < = 0.05) for all birth outcomes among MM, SM,

and SS women, while among MS women the decline was significant only for SGA.

MS women were just as likely to have a low birthweight or preterm outcome in

the second birth as in the first.

Women who were married at the first birth and became single by the second

birth had a 30±40% increase in risk of low birthweight and SGA relative to women


Table 1. Demographic and pregnancy characteristics of Washington State women, bymarital status at two consecutive births, 1980±93

Married at 1st birth Single at 1st birthÐÐÐÐÐÐÐÐÐÐÐÐÐÐ ÐÐÐÐÐÐÐÐÐÐÐÐÐÐ

Married at Single at Married at Single at2nd birth 2nd birth 2nd birth 2nd birth

n = 113 905 n = 2843 n = 13 227 n = 12 936

n (%) n (%) n (%) n (%)Age (1st birth)

< 20 years 11 055 (9.7) 1063 (37.4) 6675 (50.5) 7574 (58.6)20±24 years 43 351 (38.1) 1329 (46.8) 4911 (37.1) 4035 (31.2)25±29 years 41 185 (36.2) 360 (12.7) 1230 (9.3) 1014 (7.8)30±34 years 15 808 (13.9) 84 (3.0) 352 (2.7) 262 (2.0)35+ years 2460 (2.2) 7 (0.3) 54 (0.4) 48 (0.4)Missing 46 0 5 3

Smoking (1st birth) 9368 (14.4) 658 (43.5) 3145 (38.3) 4598 (52.2)Missing 48 762 1330 5024 4135

Smoking (2nd birth) 13 573 (14.4) 1307 (51.6) 4395 (37.7) 6304 (55.3)Missing 19 850 308 1569 1545

Early prenatal care(1st birth)a 99 819 (89.0) 2076 (74.3) 8328 (64.4) 7411 (58.9)

Missing 1742 47 304 352Early prenatal care(2nd birth) 99 238 (88.8) 2041 (73.8) 9988 (77.3) 8101 (64.5)

Missing 2147 77 303 371Congenital malformations(2nd birth) 1838 (1.8) 46 (1.8) 267 (2.2) 259 (2.2)

Missing 11 345 362 1322 1399Interbirth interval

less than 2 years 38 701 (34.0) 396 (13.9) 3359 (25.4) 4896 (37.9)Same father both births

Yes 10 936 (96.0) 220 (7.7) 4018 (30.4) 1459 (11.3)No 4013 (3.5) 948 (33.4) 1493 (11.3) 752 (5.8)Not statedb 566 (0.5) 1675 (58.9) 7716 (58.3) 10 725 (82.9)

aOnset of prenatal care in the first trimester of pregnancy.bPaternity not stated on birth certificate for one or both births.


5 sup DISC

who remained married, adjusted for age at first birth and smoking status recorded

at the second birth (Table 2). Relative to those who remained single, women who

were single at the first birth and became married by the second birth had a 20±30%

lower adjusted relative risk of low birthweight, preterm delivery, and SGA.

Women who were single at the second birth were more likely than married

women to have adverse outcomes at the second birth whether or not they had an

adverse outcome at the first birth (Table 3). Among women with normal outcomes

in the first birth, the previously described patterns of increased adverse outcome

risk associated with change from married to single status and decreased risk

associated with change from single to married persisted, and MS women were 1.5

times as likely as MM women to have low birthweight deliveries in the second

birth, controlling for age at first birth and smoking status (95% confidence interval

1.2±1.8).

Discussion

This population-based study confirms previous findings of an association between

single marital status at one birth and concurrent higher rates of adverse birth


Table 2. Relative risks of adverse outcome at 2nd birth associated with marital statuschange between births, Washington State 1980±93

Adverse outcome Adverse outcomeat 1st birth at 2nd birth 95% confidence

n (%) n (%) RRa interval

Low birthweightMSb 135 (4.8) 146 (5.1) 1.4 1.2±1.7MMc 4162 (3.7) 2790 (2.5) 1.0 ref.SMd 723 (5.5) 472 (3.6) 0.7 0.6±0.8SSe 869 (6.7) 698 (5.4) 1.0 ref.

Preterm deliveryf

MS 176 (6.5) 1805 (6.6) 1.2 1.0±1.3MM 6737 (6.1) 5358 (4.9) 1.0 ref.SM 1061 (8.5) 715 (5.6) 0.8 0.7±0.9SS 1158 (9.5) 9335 (7.5) 1.0 ref.

SGAf

MS 262 (9.7) 205 (7.5) 1.3 1.1±1.5MM 6632 (6.0) 3937 (3.6) 1.0 ref.SM 1149 (9.2) 747 (5.9) 0.8 0.7±0.9SS 1358 (11.2) 990 (8.0) 1.0 ref.

aRR = relative risk of adverse outcome at 2nd birth of women who change marital statusbetween births compared with those who do not, by initial marital status. Adjusted formaternal age at first birth and smoking status at second birth.bMS = Married at first birth, single at second; cMM = Married at both births; dSM = Single atfirst birth, married at second; eSS = Single at both births; fexcludes 5679 women withmissing gestational length at 1st birth, and 5513 at 2nd birth.


5 sup DISC

outcomes.8,15,16,22±25 Additionally, we found that the rates of adverse outcomes at

the subsequent birth were higher for women who were single at that birth,

regardless of their marital status at the previous birth. We saw a fair amount of

difference in demographic and behavioural characteristics between groups with

the same marital status at first birth who had differing status at the next. The

relatively small subgroup of initially married women who became single by their

second birth were younger, much more likely to smoke, and less likely to get early

prenatal care during the first pregnancy than women who remained married.

Some differences also were seen among women who were initially single mothers:

those who subsequently married were less likely to smoke and more likely to get

early prenatal care than SS women, even during the first pregnancy when both

groups were single. This heterogeneity among unmarried women was also noted

by Skjaerven and Irgens in Norway, who found that initially unmarried women

married by the second birth were distinguishable from never-married women by

their relatively lower risk of perinatal mortality even at the first birth.18

After adjustment for maternal age at first birth and cigarette smoking, a change

from married to single status was associated with a 30±40% increase in risk of


Table 3. Frequency of adverse outcomes at 2nd birth by outcome of 1st birth, and relativerisks of adverse outcome at 2nd birth associated with marital status change among womenwith normal outcomes at 1st birth

Adverse outcome at 2nd birthÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐÐAdverse outcome Normal outcome

at 1st birth at 1st birth 95% confidencen (%) n (%) RRa interval

Low birthweightMS 26 (19.3) 120 (4.4) 1.5 1.2±1.8MM 636 (15.3) 2154 (2.0) 1.0 ref.SM 125 (17.3) 347 (2.8) 0.7 0.6±0.8SS 187 (21.5) 511 (4.2) 1.0 ref.

Preterm deliveryb

MS 29 (17.5) 138 (5.7) 1.1 1.0±1.4MM 1034 (16.1) 4131 (4.2) 1.0 ref.SM 142 (13.9) 535 (4.9) 0.8 0.7±0.9SS 189 (17.2) 682 (6.5) 1.0 ref.

SGAb

MS 51 (20.3) 145 (6.2) 1.4 1.1±1.6MM 1012 (16.0) 2773 (2.8) 1.0 ref.SM 199 (18.1) 501 (4.6) 0.8 0.7±0.9SS 264 (20.6) 667 (6.4) 1.0 ref.

aRR = relative risk of adverse outcome at 2nd birth of women who change marital statusbetween births compared with those who do not, by initial marital status, among womenwith normal outcome of first birth. Adjusted for maternal age at first birth and smokingstatus at second birth.bExcludes 5679 women with missing gestational length at 1st birth, and 5513 at 2nd birth.


5 sup DISC

LBW or SGA relative to remaining married. Similarly, a change from single to

married status was associated with a 20±30% reduction in risk of these outcomes

relative to remaining single. If these findings represent a true association between

marital status and birth outcome, at present we can only speculate as to the

mechanism. Unmarried status, especially being divorced, has been associated with

stress and depression, which have been linked to increased risk of adverse

pregnancy outcomes.26±28 Perhaps the clearest illustration of this association was a

randomised trial that found that infants of women with moderate to severe stress

who were provided with professional social support during pregnancy were

significantly less likely to have intrauterine growth retardation than infants of

those who received no such care.29 Two explanations have been proposed for the

possible link between unmarried status, stress, and adverse pregnancy outcomes.

The first posits a direct connection: stress leads to increased catecholamine release

and the consequent adverse effects of placental hypoperfusion and growth

impairment or uterine irritability and preterm labour.28 An indirect link has also

been hypothesised, as stress may lead to behaviours such as smoking, alcohol and

cocaine use, which are associated with growth retardation and preterm delivery.30

There are potential limitations common to studies relying on vital statistics

data. In this study, some desired data were not available, either because of im-

proper data collection within hospitals, or lack of certain variables in the database.

The smoking data included a large proportion of missing values, because this

information was not collected during the earlier years of the study period. Since

smoking was an important confounder, more precise knowledge about every

woman's status could have changed our risk estimates somewhat. In stratified

analyses, we controlled for the more completely recorded smoking status at the

second birth, and considered `unknown' smoking status as a separate stratum, to

partially control for the missing data. Our use of second birth smoking status

instead of first probably led to a conservative estimate of risks associated with

marital status change, since MS women were more likely to smoke at the second

birth than at the first.

Some potentially important variables, such as income and maternal educa-

tional level, were not contained in the Washington State birth certificate database.

Although poverty is considered a potential confounder in studies of low birth-

weight, lack of financial resources and its attendant consequences are part of the

causal pathway by which differences in risk for poor outcomes between marital

status groups in this study are postulated to occur, and adjustment for poverty

may not be appropriate in the present study. Similarly, while we would be inter-

ested to see the intergroup differences in alcohol and cocaine use and pregnancy

weight gain, these factors may be manifestations of the stress associated with

unmarried pregnancy, and should not be considered as potential confounders.

Our main exposure variable, marital status as recorded on the birth certificate,

may not have provided us with the most precise exposure information. `Married'



5 sup DISC

or `not married' does not capture the presence or absence of social and emotional

support and stability. The unmarried category can include women in stable,

supportive relationships while married can describe women who are separated or

estranged from their husbands and living alone. The possible misclassification

of such women would make our risk estimates more conservative than they might

be otherwise.

Our study identified a subgroup of previously married mothers whose infants

were at increased risk for adverse pregnancy outcomes relative to mothers who

remained married at the second of two consecutive livebirths. Although this

group of women differed in characteristics at the first birth from the women who

remained married, increased risks of poor pregnancy outcomes persisted even

after adjustment for confounding by smoking and age at first birth. While these

associations may be attributable to unmeasured variables in addition to marital

status change, public health policy and programmes directed at high-risk mothers

and infants should be aware of the increased risks and responsive to the special

needs of this subgroup of child-bearing women.

Acknowledgements

This research was supported in part by grant MCJ-4093 from the Maternal and

Child Health Bureau (Title V, Social Security Act), Health Resources and Services

Administration, US Department of Health and Human Services.

References

1 Wegman ME. Annual summary of vital statistics ± 1986. Pediatrics 1987; 80:817±827.2 Wegman ME. Annual summary of vital statistics ± 1989. Pediatrics 1990; 86:835±847.3 National Center for Health Statistics. Advance Report of Final Natality Statistics, 1980.Monthly Vital Statistics Report; Vol. 31, no. 8, suppl. Hyattsville, Maryland: Public HealthService, November 1982.4 National Center for Health Statistics. Advance Report of Final Natality Statistics, 1989.Monthly Vital Statistics Report; Vol. 40, no. 8, suppl. Hyattsville, Maryland: Public HealthService, 1991.5 National Center for Health Statistics. Advance Report of Final Natality Statistics, 1991.Monthly Vital Statistics Report; Vol. 42, no. 3, suppl. Hyattsville, Maryland: Public HealthService, 1993.6 Ventura SJ, Martin JA, Taffel SM, Mathews TJ, Clarke SC. Advance Report of FinalNatality Statistics, 1992. Monthly Vital Statistics Report; Vol. 43 no. 5, suppl. Hyattsville,Maryland: National Center for Health Statistics, 1994.7 National Center for Health Statistics. Vital Statistics of the United Sates, 1991, Vol. I,Natality. Washington: Public Health Service, 1995.8 Hein HA, Burmeister LF, Papke KR. The relationship of unwed status to infantmortality. Obstetrics and Gynecology 1990; 76:763±768.9 Arias F, Tomich P. Etiology and outcome of low birthweight and preterm infants.Obstetrics and Gynecology 1982; 60:277±281.



5 sup DISC

10 McCormick MC. The contribution of low birthweight to infant mortality and childhoodmorbidity. New England Journal of Medicine 1985; 312:82±90.11 Manning WG, Keeler EB, Newhouse JP, Sloss EM, Wasserman J. The taxes of sin: Dosmokers pay their way? Journal of the American Medical Association 1989; 262:901.12 Bennett T. Infant mortality by marital status of mother ± United States, 1983. Morbidityand Mortality Weekly Report 1990; 39:521±523.13 Ramsey CN, Abell TD, Baker LC. The relationship between family functioning, lifeevents, family structure, and the outcome of pregnancy. Journal of Family Practice 1986;22:521±527.14 Doucet H, Baumgarten M, Infante-Rivard C. Low birthweight and householdstructure. Journal of Developmental and Behavioral Pediatrics 1989; 10:249±252.15 Golding J, Henriques J, Thomas P. Unmarried at delivery. II. Perinatal morbidity andmortality. Early Human Development 1986; 14:217±227.16 Ahmed F. Unmarried mothers as a high-risk group for adverse pregnancy outcomes.Journal of Community Health 1990; 15:35±44.17 Defo BK, Partin M. Determinants of low birthweight: A comparative study. Journal ofBiosocial Science 1993; 25:87±100.18 Skjñrven R, Irgens LM. Perinatal mortality and mother's marital status in subsequentsiblings. Early Human Development 1988; 18:199±212.19 Herman AA, McCarthy BJ, Bakewell JM, Ward RH, Mueller BA, Maconochie NE, ReadAW et al. Data linkage methods used in maternally-linked birth and infant deathsurveillance data sets from the United States (Georgia, Missouri, Utah and WashingtonState), Israel, Norway, Scotland and Western Australia. Paediatric and Perinatal Epidemiology1997; 11(Suppl. 1):5±22.20 Williams RL, Creasy RK, Cunningham GC, Hawer WE, Norris FD, Tashiro M. Fetalgrowth and perinatal viability in California. Obstetrics and Gynecology 1982; 59:624±632.21 SAS Institute Inc. SAS/STAT User's Guide, Version 6, Fourth edn, Volume 2. Cary, NC:SAS Institute Inc., 1989.22 Kleinman JC, Kessel SS. Racial differences in low birthweight. New England Journal ofMedicine 1987; 317:749±753.23 Lee K, Ferguson RM, Ckorpuz M, Gartner LM. Maternal age and incidence of lowbirthweight at term: A population study. American Journal of Obstetrics and Gynecology 1988;158:84±89.24 Teberg AJ, Settlage R, King Y, Aquilar T. Maternal factors associated with delivery ofinfants with birthweight less than 2000 g in a low socioeconomic population. Journal ofPerinatology 1989; 9:291±295.25 Wen SW, Goldenberg RL, Cutter BR, Hoffman HJ, Cliver SP. Intrauterine growthretardation and preterm delivery: Prenatal risk factors in an indigent population. AmericanJournal of Obstetrics and Gynecology 1990; 162:213±218.26 Pearlin LI, Johnson JS. Marital status, life-strains and depression. American SociologicalReview 1977; 42:704±715.27 Newton RW, Webster PAC, Binu PS, Maskrey N, Phillips AB. Psychosocial stress inpregnancy and its relation to the onset of premature labour. British Medical Journal 1979;2:411±413.28 McAnarney ER, Stevens-Simon C. Maternal psychological stress/depression and lowbirthweight: Is there a relationship? American Journal of Diseases of Children 1990; 144:789±791.29 Rothberg AD, Lits B. Psychosocial support for maternal stress during pregnancy: Effecton birthweight. American Journal of Obstetrics and Gynecology 1991; 165:403±407.30 Henriques J, Golding J, Thomas P. Unmarried at delivery. I. The mothers and their care.Early Human Development 1986; 14:201±216.



5 sup DISC

the association of change in maternal marital status between births and adverse pregnancy outcomes...

Documents