Births to women who live in an intimate relationship with a partner but without legal marriage have become increasingly common and widely accepted in many Western societies. Jurisdictions differ in the definition and legal rights of common-law couples. The Canadian province of Quebec is one of the most prominent examples. The Quebec Civil Code defines a common-law union as a commitment by 2 persons aged 18 years or over who publicly consent to live together and respect the resulting rights and obligations.1 Common-law couples in Quebec are treated the same as legally married couples with respect to almost all legal and social consideration; for instance, common-law couples are entitled to the same family and child tax benefits as legally married couples. The increasing acceptance of this cohabitation status, combined with the same privileges in social and welfare benefits for common-law versus legally married couples, probably underlies the substantial rise in common-law unions in Quebec in recent years.
Unmarried women are known to have greater risks of adverse pregnancy outcomes than married women.2–5 However, little is known about pregnancy outcomes among mothers in common-law relationships because most previous studies have been unable to distinguish them from legally married mothers and those mothers living alone. Such a distinction is important, however, because common-law mothers may differ both from legally married and lone mothers, and social supports are believed to become similar among common-law and legally married couples with the increasing acceptance of common-law unions in modern Western societies.6 A recent European case–control study has reported that higher risks of adverse pregnancy outcomes for cohabiting but legally unmarried mothers may be limited to settings where such cohabitations are relatively uncommon.7 There have been no population-based studies of temporal trends in risks of adverse pregnancy outcomes among mothers in common-law versus traditional marriage relationships.
Studies using national linked birth/infant death data files should be able to offer the best population-based risk estimates, but these files usually contain only the legal marital status of the mother (married, divorced, separated, widowed, never married) and thus cannot identify births to common-law mothers. It is unclear whether disparities in pregnancy outcomes between common-law and legally married mothers persist if other individual- and community-level characteristics are adjusted for and whether those risks have diminished over time with the rise in common-law unions.
In the Canadian province of Quebec, the prevalence of common-law unions increased dramatically in the 1990s.8 Births to common-law mothers outnumbered births to legally married mothers in Quebec in 1998.9 Because this province has the best pregnancy outcomes in Canada10 and has favorable legal and social environments for births to common-law mothers, we hypothesized that disparities in pregnancy outcomes between married and common-law mothers should have diminished with the rise of common-law unions in recent years. We tested this hypothesis by assessing the risks of adverse pregnancy outcomes among births to mothers in traditional marriage, common-law unions, and lone-living arrangements in Quebec in recent years.
MATERIALS AND METHODS
We conducted a birth cohort-based study of all 720,586 births in Quebec during the period of 1990 (when relatively complete recording of the living arrangements of the mother first became available in Quebec birth registrations) through 1997 by using the most recent Statistics Canada's linked stillbirth, live birth, and infant death registration files. This probabilistic record linkage of vital records has been validated previously.11 Ethics approval was not sought for this study because it was based on anonymous birth registration data from Statistics Canada, which has agreements with all Canadian provinces on the confidentiality and use of data.11
Data on living arrangements and marital status were obtained from Institut de la Statistique du Québec, which compiles the provincial vital statistics. The birth registration forms contained questions on both legal marital status (married, never married, widowed, divorced, separated) and living arrangements (living or not living together as a couple) of the mother. Only mothers whose marital status was “married” were considered legally married, that is, in “traditional” marriage relationships. “Unmarried” (never married, widowed, divorced, separated) mothers whose “relationship with partner” was recorded as “living together as a couple” were considered to be in a common-law relationship. Mothers who were neither married nor in common-law relationships were considered “lone” mothers. The presence of information on the father—defined as the recording of father's age and/or place of birth on the birth registration—was used to further differentiate the lone mother's relationship with the father of the newborns.
Outcome measurements included preterm birth (less than 37 completed weeks of gestation), low birth weight (LBW, less than 2,500 g), small for gestational age (SGA, less than the 10th percentile based on a recently published Canadian standard),12 stillbirth (in Quebec, only those fetal deaths with birth weight of 500 g or greater are registered), and neonatal (0–27 days) and postneonatal (28–364 days) mortality rates. Outcomes were compared among four groups of mothers: legally married, common-law, lone with information on the father, and lone without information on the father. Of the 720,586 births recorded, 2,731 (0.4%) of the 717,478 live births and 1,111 (36%) of the 3,107 stillbirths were missing information on marital status and/or living arrangements and could not be classified by marital status and living arrangements. The high proportion of stillbirths missing data on marital status and living arrangements was mainly the result of 866 (28%) stillbirths recorded on older stillbirth registration forms, which did not include the question on living arrangements from which common-law status was deduced. However, only 7% of stillbirths were missing information on marital status. We therefore retained stillbirth as an outcome for comparisons between married versus unmarried only.
The crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for adverse pregnancy outcomes by using births to legally married mothers as the reference group. The adjusted ORs were estimated by using multilevel logistic regression analysis to control for infant sex (boy, girl), parity (primiparous, multiparous), plurality (1, 2, or more), maternal age (younger than 20, 20–34, 35 years or older), education (less than grade 11, 11 [high school ends with grade 11 in Quebec], 12–13, 14 years or more),13 mother tongue, that is, the first language spoken by the newborn's mother, a proxy for ethnicity (French, English, aboriginal [approximately 90% North American Indians, 10% Inuit], others [mostly immigrants]), and community size (1 million or greater, 500,000–999,999, 100,000–499,999, 10,000–99,999, fewer than 10,000 persons in their respective census metropolitan area or census agglomeration, according to the 1996 Canadian census), as well as community-level random effects (census metropolitan area or census agglomeration code as the random effect variable). The multilevel logistic regression analysis takes into account variations at both individual and cluster (census metropolitan area or census agglomeration) levels. The size of a community is likely associated with living environment, cultural, economic, and employment opportunities, and access to and quality of health care. Community size was treated as a fixed-effect variable at the community level, whereas infant sex, parity, plurality, maternal age, education, and mother tongue were treated as fixed-effect variables at the individual level. Likelihood ratio tests were used to assess interactions among explanatory variables; stratified analyses were used to quantify the effect of marital status within strata. The etiologic fraction (population-attributable risk, that is, the percentage of cases attributable to exposure in a given population) for common-law unions or lone mothers was calculated for outcomes by using the relative risk converted from the adjusted OR.14,15 We conducted all analyses by using SAS software (SAS Institute, Cary, NC), with the SAS GLMM800 macro used for multilevel logistic regression analysis.
A total of 720,586 births was registered in Quebec from 1990 to 1997, including 717,479 live births, 3,107 stillbirths (4.3 per 1,000 total births), 3,868 infant deaths (5.4 per 1,000 live births; 2,696 neonatal deaths [3.8 per 1,000 live births] and 1,172 postneonatal deaths [1.6 per 1,000 neonatal survivors]), 50,803 preterm births (7.2%), 75,111 SGA births (10.9%), and 43,752 LBW births (6.1%).
Of the 714,748 live births with information available on marital status and living arrangements, 380,962 (53.3%) births were to legally married mothers, 249,383 (34.9%) to common-law mothers, 55,436 (7.8%) to lone mothers with information on the father, and 28,967 (4.1%) to lone mothers without information on the father. During the years 1990 to 1997, a striking rise was observed in the proportion of births to common-law mothers, from 20% to 44% and a corresponding fall in the proportion of births to legally married mothers from 62% to 46% (Figure 1).
Marked significant differences in maternal characteristics were observed across the four groups by marital status and living arrangements (Table 1). Increasingly higher proportions of adolescent mothers (aged less than 20 years), mothers with less educational attainment, and primiparae were observed across the 4 marital status groups from legally married to common-law, to lone mothers with father information, to lone mothers without father information (Table 1). A relatively higher proportion of common-law mothers reported French mother tongue and were from small communities.
Increasingly higher risks of preterm birth, SGA, LBW, and neonatal and postneonatal death were observed across the 4 groups by marital status and living arrangements (Table 2). Pregnancy outcomes among common-law mothers were worse than among legally married mothers but much better than among lone mothers with or without information on the father. The magnitude of the excess risks increased from common-law, to lone with father information, to lone without father information. Stillbirth rates were 3.6 and 4.5 per 1,000 total births for married and unmarried mothers, respectively.
Modestly higher risks of all adverse pregnancy outcomes persisted for births to nonlegally married mothers after controlling for infant sex, parity, plurality, maternal age, education, mother tongue, community size, and community-level random effects by multilevel logistic regression (Table 3). The ORs were progressively higher from mothers in common-law, to lone mothers with father information, to lone mothers without father information. Compared with the crude ORs, the adjusted ORs were lower in magnitude after these adjustments. Adjusted ORs (95% CIs) for common-law versus legally married mothers were 1.14 (95% CI 1.11, 1.17) for preterm birth, 1.21 (95% CI 1.18, 1.25) for LBW, 1.18 (95% CI 1.16, 1.20) for SGA, 1.07 (95% CI 0.97, 1.19) for neonatal death, and 1.23 (95% CI 1.04, 1.44) for postneonatal death. Etiologic fractions for common-law unions were 4.2%, 5.1%, and 7.0% for preterm birth, SGA, and postneonatal death, respectively; the corresponding figures for lone mothers (with and without father information combined) were 3.3%, 3.5%, and 5.5%, respectively. The adjusted OR of stillbirth for married versus unmarried mothers was 1.18 (95% CI 1.06, 1.30).
Likelihood ratio tests suggested significant interactions between marital status and ethnicity and between marital status and maternal education (both at P < .01). Subsequent analyses stratified by ethnicity and maternal education revealed that adjusted ORs for all adverse pregnancy outcomes for common-law versus legally married were lower among mothers in the higher education stratum and much higher in the aboriginal group than in other ethnic groups. The adjusted ORs (95% CI) for common-law versus legally married among mothers with education less than 11 completed years were 1.21 (95% CI 1.14, 2.29) for preterm birth, 1.22 (95% CI 1.16, 1.29) for SGA, 1.37 (95% CI 1.06, 1.76) for neonatal death, and 1.56 (95% CI 1.09, 2.22) for postneonatal death, respectively; the corresponding ORs among mothers with education of more than 14 completed years were 1.06 (95% CI 1.02, 1.10), 1.10 (95% CI 1.06, 1.13), 0.91 (95% CI 0.78, 1.07), and 1.00 (95% CI 0.71, 1.40), respectively. The adjusted ORs (95% CI) for common-law versus legally married mothers of French mother tongue (the majority) were 1.12 (95% CI 1.09, 1.15) for preterm birth, 1.19 (95% CI 1.17, 1.22) for SGA, 1.07 (95% CI 0.97, 1.18) for neonatal death, and 1.15 (95% CI 0.99, 1.38) for postneonatal death, respectively; the corresponding ORs among aboriginal mothers were 1.44 (95% CI 1.13, 1.83), 1.42 (95% CI 1.00, 2.02), 3.83 (95% CI 1.57, 9.30), and 1.41 (95% CI 0.82, 2.45), respectively.
To assess temporal changes in the risks of adverse pregnancy outcomes for common-law versus legally married mothers over time, we estimated the crude and adjusted ORs for the periods 1990–1993 and 1994–1997 (during 4-year intervals rather than individual years to obtain more stable estimates). The crude and adjusted ORs were similar in the 2 periods for all adverse outcomes under study (Table 4); in fact, the ORs of neonatal death were slightly but nonsignificantly higher in the latter period. These ORs were also similar over 2-year intervals but with slightly wider confidence intervals (results not shown).
Our study provides a longitudinal population-based assessment of temporal trends in disparities in pregnancy outcomes among common-law versus legally married mothers. These observed disparities, although modest in magnitude, did not diminish in time despite the striking recent temporal rise and increasingly social acceptance of common-law unions. Lone mothers, particularly when no information on the father was recorded on the birth registration, had even worse pregnancy outcomes than mothers living in common-law arrangements, in agreement with the reported elevated risk among “lone” mothers.16 However, the elevated risks of adverse outcomes among common-law couples are of greater public health concern because of the high and increasing prevalence of births to mothers in common-law unions worldwide, as reflected by the greater etiologic fractions for adverse outcomes observed among common-law versus lone mothers.
The persistent disparities in pregnancy outcomes between married and common-law mothers in Quebec from 1990 to 1997 are in contrast to the results of a recent case–control study in Europe, which reported no excess risk of preterm birth in countries where out-of-marriage births were prevalent.7 The more than 40% prevalence of births to common-law mothers in Quebec in the late 1990s is much higher than the reported prevalence of births to unmarried “cohabitating” mothers in any European country.6 Despite the dramatic rise in common-law unions in Quebec, the modestly elevated ORs among common-law versus legally married mothers for all adverse outcomes persisted over time. This does not mean that the universal health care system in Quebec has no beneficial impact, because the disparities may be the result of differentials in unmeasured risk factors among common-law versus married mothers that are unrelated to health care. The differences in our findings from those recently reported in European countries may reflect greater income inequality and higher prevalence of poverty in North America.17 The proportion of mothers in lower socioeconomic strata may be higher among common-law mothers in Quebec than in Europe. We speculate that disparities in pregnancy outcome may be even larger in the United States and other jurisdictions where social and legislative environments are less favorable to common-law unions.
Ethnicity and maternal education modified the effect of marital status on adverse pregnancy outcomes. We observed much higher excess risks associated with common-law unions among aboriginal mothers, indicating a further social risk in ethnic groups already disadvantaged in pregnancy outcomes. Higher maternal education mitigated the adverse effect of common-law unions, a result consistent with a U.S. study reporting “unmarried” status did not affect the risk of infant mortality among infants born to college-educated women.4
We do not know the causal mechanisms underlying the observed persistent disparities in pregnancy outcomes among common-law versus legally married mothers. The limited literature offers very few clues. Mothers in common-law unions may experience greater stress during pregnancy owing to less stable relationships than mothers in traditional marriage,18,19 despite their similar legal status in modern Western societies. Our findings suggest common-law unions may be disadvantageous for pregnancy outcomes relative to traditional marriage. This does not necessarily indicate that legal marriage has “therapeutic” effects, because other unmeasured factors may mediate the observed relationships. The disparities in pregnancy outcomes may well be related to differences among women who enter varying personal and cohabitation relationships, or to the quality of the relationships themselves, rather than to a beneficial effect of the legal act of marriage. Unmeasured socioeconomic disadvantage among common-law mothers may lead to increased stress or (unmeasured) harmful behaviors during pregnancy.20
The validity of the Canadian linked vital data has been well documented.11 Moreover, the dramatic rise observed in common-law unions in Quebec is consistent with other recent reports.8,9 However, several limitations of our study require comments. Common to other studies using Canadian linked vital registration data, we had no information on maternal occupation, income, smoking, physical activity, maternal height, prepregnancy weight, gestational weight gain, the use of prenatal care, or maternal illness. Common-law unions represent a new social construct that clearly is not the direct cause of adverse pregnancy outcomes. Differentials in other unmeasured risk factors or behaviors may account for or underlie the causal mechanisms of the observed disparities. Our data did not allow controlling for income or occupation but did allow controlling for maternal education—the most important predictor of health outcomes among the 3 basic elements of socioeconomic status: education, income, and occupation.21 Maternal smoking has been linked with adverse pregnancy outcomes in many studies22 and may be one of the high-risk behaviors that is more frequent among common-law mothers and therefore may partly account for their higher risks of adverse pregnancy outcomes versus legally married mothers. According to the 1994 Canadian National Population Health Survey, the prevalence of current smokers among Quebec women aged 15–44 years was 32% for legally married women versus 40% for those in common-law relationships.23 Assuming these numbers are applicable to pregnant women and using the relative risks of smoking for preterm (1.4) and SGA (2.4) from a published meta-analysis,24 we estimate that smoking alone could account for a 2.8% excess risk in preterm birth (versus the observed 17% excess risk) and a 7.7% excess risk in SGA (vs the observed 29% excess risk) among common-law versus married mothers. It is unclear whether later or lesser use of prenatal care among common-law versus legally married mothers can account for the disparities in outcomes, given universal health insurance in Canada.25 However, the mere availability of care does not guarantee its use, particularly among women of low socioeconomic status,26 and the beneficial impact of routine prenatal care remains controversial.27 Maternal size, especially prepregnancy weight, also affects the risk of adverse pregnancy outcomes,28 and here as well no data are available to assess its potential confounding effect. More studies are required to understand the causal mechanisms underlying these disparities and to develop and test intervention strategies.
1. Anonymous. Bill 84, An Act instituting civil unions and establishing new rules of filiation, 2nd session, 36th Legislature, Quebec.
2. Hein HA, Burmeister LF, Papke KR. The relationship of unwed status to infant mortality. Obstet Gynecol 1990;76:763–8.
3. Cooperstock MS, Bakewell J, Herman A, Schramm WF. Association of sociodemographic variables with risk for very preterm birth in twins. Obstet Gynecol 1998;92:53–6.
4. Mathews TJ, MacDorman MF, Menacker F. Infant mortality statistics from the 1999 period linked birth/infant death data set. Natl Vital Stat Rep 2002;50:1–28.
5. Bennett T, Braveman P, Egerter S, Kiely JL. Maternal marital status as a risk factor for infant mortality. Fam Plann Perspective 1994;26:252–6, 271.
6. Brown S, Booth A. Cohabitation versus marriage: a comparison of relationship quality. J Marriage Fam 1996;58:668–78.
7. Zeitlin JA, Saurel-Cubizolles MJ, Ancel PY, for the EUROPOP Group. Marital status, cohabitation, and the risk of preterm birth in Europe: where births outside marriage are common and uncommon. Paediatr Perinat Epidemiol 2002;16:124–30.
8. Statistics Canada. 1996 Census: Marital status, common-law unions and families. The Daily (October 14, 1997). Available at: http://www.statcan.ca/Daily/English/971014/d971014.htm
. Retrieved March 22, 2004.
9. Jette M. The ELDEQ 1998–2002, a first annual longitudinal study of Quebec newborn. Available at: http://www.isuma.net/v01n02/jette/jette_e.pdf
. Accessed October 10, 2002.
10. Dzakpasu S, Joseph KS, Kramer MS, Allen AC. The Matthew effect: infant mortality in Canada and internationally. Pediatrics 2000;106:E5.
11. Fair M, Cyr M, Allen AC, Wen SW, Guyon G, MacDonald. An assessment of the validity of a computer system for probabilistic record linkage of birth and infant death records in Canada. The Fetal and Infant Health Study Group. Chronic Dis Canada 2000;21:8–13.
12. Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M, et al, for the Fetal/Infant Health Study Group of the Canadian Perinatal Surveillance System. A new and improved population-based Canadian Reference for birth weight for gestational age. Pediatrics 2001;108:E35.
13. Chen J, Fair M, Wilkins R, Cyr M. Maternal education and fetal and infant mortality in Quebec. The Fetal and Infant Mortality Study Group of the Canadian Perinatal Surveillance System [Erratum in: Health Rep 1998;10:77]. Health Rep 1998;10:53–64.
14. Zhang J, Yu KF. What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280:1690–1.
15. Kramer MS, Demissie K, Yang H, Platt RW, Sauve R, Liston R. The contribution of mild and moderate preterm birth to infant mortality. The Fetal and Infant Health Study Group of the Canadian Perinatal Surveillance System. JAMA 2000;284:843–9.
16. Whitehead M, Drever F. Narrowing social inequalities in health? Analysis of trends in mortality among babies of lone mothers (abridged version 2). BMJ 1999;318:912–4.
17. Lynch J, Smith GD, Hillemeier M, Shaw M, Raghunathan T, Kaplan G. Income inequality, the psychosocial environment, and health: comparisons of wealthy nations. Lancet 2001;358:194–200.
18. Wu Z. Cohabitation: an alternative form of family living. Don Mills, Ontario: Oxford University Press; 2000.
19. LeBourdais C, Neill G, Turcotte P, Vachon N, Archambault J. The changing face of conjugal relationships. Canadian Social Trends 2000;Spring:14–17. Statistics Canada: Catalogue No. 11–008.
20. Janssen PA, Holt VL, Sugg NK, Emanuel I, Critchlow CM, Henderson AD. Intimate partner violence and adverse pregnancy outcomes: a population-based study. Am J Obstet Gynecol 2003;188:1341–7.
21. Adler NE, Newman K. Socioeconomic disparities in health: pathways and policies. Inequality in education, income, and occupation exacerbates the gaps between the health “haves” and “have-nots” [review]. Health Aff 2002;21:60–76.
22. Wisborg K, Kesmodel U, Henriksen TB, Olsen SF, Secher NJ. Exposure to tobacco smoke in utero and the risk of stillbirth and death in the first year of life. Am J Epidemiol 2001;154:322–7.
23. Wilkins R. Special tabulations from the 1994 National Population Health Survey (NPHS). Ottawa: Statistics Canada; 2002.
24. Kramer MS. Determinants of low birth weight: methodological assessment and meta-analysis [review]. Bull World Health Organ. 1987;65:663–737.
25. Palley HA. Canadian federalism and the Canadian health care program: a comparison of Ontario and Quebec. Int J Health Serv 1987;17:595–616.
26. Gazmararian JA, Arrington TL, Bailey CM, Schwarz KS, Koplan JP. Prenatal care for low-income women enrolled in a managed-care organization. Obstet Gynecol 1999;94:177–84.
27. Groutz A, Hagay ZJ. Prenatal care: an update and future trends [review]. Curr Opin Obstet Gynecol 1995;7:452–60.
28. Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Prepregnancy weight and the risk of adverse pregnancy outcomes. N Engl J Med 1998;338:147–52.
Contributing members of the Fetal and Infant Health Study Group of the Canadian Perinatal Surveillance System include Robert Platt, Martha Fair, K. S. Joseph, Judith Lumley, and Jennifer Medves.