Does marriage lower depression risk?

'Married women are less likely to suffer depression than cohabiters or singletons,' according to the Daily Mail.

The claim is based on a large Canadian survey that assessed various factors in the lives of new mothers, including whether they had depressive symptoms that could indicate postnatal depression.

Researchers found that domestic violence and substance use were reported less by women who were married than unmarried cohabiting women, single and never married women, and separated or divorced women. However, the longer an unmarried woman had been living with her partner, the less difference there was when compared with a married woman.

The link with postnatal depression was less robust, as it was not statistically significant in the overall analysis. However, rates of postnatal depression were greater in mothers who had lived with their partners for up to two years – whether married or unmarried – compared with married women who had lived with their partners for more than five years.

The figures in this study only give us a snapshot of new mothers in Canada at one point in time, and the figures may not be representative of other countries at different times. Most importantly, the design of the study means that it is not possible to say that marriage is directly causing any of the differences seen between the groups.

Where did the story come from?

The study was carried out by researchers from St Michael's Hospital in Canada. Funding was not reported. The study was published in the peer-reviewed American Journal of Public Health.

The Daily Mail reporting focuses on postnatal depression, but this is the outcome with the least convincing findings of a difference between the groups of women. The Mail also implies that the results are due to a "positive effect" of marriage, when it is not possible to say from the study whether marriage itself is directly responsible for the findings. The study doesn't say whether the domestic violence, substance use, postnatal depression or marriage came first, so we can't say that one could cause the others. It is also possible that married mothers in this study differed in ways other than their marital status.

Although the researchers did try to take some of these factors into account, they or other factors could still affect the outcomes (that is, they could be potential confounders).

The Mail also committed a journalistic schoolboy error by misquoting the figures from the research paper. It says that 10.6% of married women, 20% of cohabiting women, 35% of single women and 67% of women separated or divorced in the year before they gave birth suffered postnatal depression.

These figures actually represented a much smaller proportion of women in the study who had any of the three psychosocial problems assessed: domestic violence, substance use in pregnancy (including tobacco or alcohol) or postnatal depression.

What kind of research was this?

This was a cross-sectional study looking at the association between marital status and the length of time couples had lived together, and women's experiences of domestic violence, substance use and postnatal depression.

The researchers say that the number of unmarried couples living together and births to unmarried mothers have been increasing. But it is uncertain what effect (if any) marital status has on the mother's wellbeing and birth outcomes.

The researchers were mostly interested in looking at this question to see if future research examining mother and child health should take into account any differences between these groups.

Cross-sectional studies assess factors at one point in time. If they assess more than one factor, they do not establish which came first, and therefore can only say that one factor is associated with another and not whether one factor causes the other.

What did the research involve?

The researchers used data collected as part of the 2006-2007 nationally representative Canadian maternity experiences survey. This analysis included 6,375 women over the age of 15 who had given birth to a single baby between February and May 2006 and had the relevant data available.

Most women (97%) were interviewed five to nine months after they had their baby. They were asked their marital status and whether they:

• were living with a partner, and if so for how long
• had experienced domestic violence (physical or sexual) in the past two years
• smoked 10 or more cigarettes daily in the last three months of pregnancy
• drank two or more drinks on one occasion during pregnancy
• used illegal drugs during pregnancy

They were also assessed for possible postnatal depression using an accepted screening questionnaire.

Women who reported experiencing any of these issues (domestic violence, substance use and postnatal depression) were categorised as having psychosocial problems. The researchers looked at whether the proportion of women who had psychosocial problems differed between women who were married and living with their partner and those who were:

• living with their partner but not married
• single or never married and not living with a partner
• divorced or separated and not living with a partner

They also looked at whether how long a woman had been living with her partner was related to the proportion of women who had psychosocial problems. The analyses took into account factors that could affect the results (called potential confounders), including:

• the woman's age
• how many children she had
• education
• household income
• whether she was born abroad
• ethnicity
• whether the pregnancy was wanted
• whether the partner disagreed with the pregnancy

Separate analyses were also carried out that excluded women who had a history of pre-pregnancy depression to see if this had an effect on the results.

What were the basic results?

Most of the mothers surveyed (92%) were living with a partner. The researchers found that the proportion of mothers experiencing at least one of the psychosocial problems was:

• 10.6% among married mothers living with their husband
• 20.0% among unmarried mothers living with their partner
• 35.0% among single, never married mothers
• 29.2% among mothers who were separated or divorced more than a year before the recent birth
• 67.1% among mothers who were separated or divorced in the year before the recent birth

After excluding women with pre-pregnancy depression and adjusting for all of the potential confounders, unmarried mothers who were single, recently separated or divorced, or cohabiting were more likely to have psychosocial problems than mothers who were married and living with their husband. 

When looking at the psychosocial problems individually, mothers who were not married and living with their partner, and women who were not living with a partner (either single and never married, or divorced or separated) were more likely to have experienced recent domestic violence or used substances in their pregnancy than women who were married and living with their husband.

But after taking into account the potential confounders, there was no difference between these groups of women in rates of postnatal depression.

The difference in psychosocial problems between unmarried mothers cohabiting with their partners and married mothers living with their husband tended to get smaller the longer a women had lived with her partner. In these analyses, the links between domestic violence and substance use in pregnancy and marital status were more consistent than the links with postnatal depression.

The proportion of women with postnatal depression was only significantly higher in unmarried mothers who had been cohabiting for up to two years, compared with married mothers who had been living with their partner for more than five years. There was no difference when mothers had been living with their partners for longer.

Women who had been married and living with their partner for up to two years were also more likely to have postnatal depression than married mothers who had been living with their partner for more than five years.

How did the researchers interpret the results?

The researchers concluded that their studies show that women who divorce or separate from their partners in the year before giving birth have the highest burden of psychosocial problems.

They also concluded that among unmarried mothers who live with their partners, couples who have lived together for shorter periods are more likely to experience psychosocial problems.

They also say that future research on maternal and child health would benefit from distinguishing between married and unmarried women living with their partners, and noting how long the couples had cohabited. Currently, the researchers suggest that many studies into maternal and child health do not record this fine level of detail.

Conclusion

This research has identified differences between new mothers in Canada who are married, those who are either unmarried and living with their partner, and those who not living with a partner in terms of common psychosocial problems such as domestic violence and substance use in pregnancy.

Although the news has suggested that married women are less likely to suffer from depression than women who cohabit or are single, this is a misleading impression of what the study found.

The study only looked at possible postnatal depression, rather than any type of depression. Also, there were actually few differences in postnatal depression between the groups of women. There were more consistent differences seen in domestic violence and the use of substances such as tobacco in pregnancy.

Postnatal depression was more common in women who had been living with their partner for up to two years – either married or unmarried – than those who were married and living with their partner for more than five years.

Among a number of limitations of the study, the most important is that this is a cross-sectional study. This means that all of the factors were assessed at the same time and it is therefore not possible to reliably determine which came first and therefore which factor may have influenced the other. To put it simply, cross-sectional studies cannot resolve "chicken and egg" situations.

Also, although the researchers attempted to control for factors that could have influenced the results, there may still have been other differences between the groups of women that contributed to the differences in psychosocial outcomes.

Therefore, it is not possible to say that marriage is directly causing any of the differences seen between the groups of women.