ArticlePDF Available

Maternity Leave Duration and Postpartum Mental and Physical Health: Implications for Leave Policies

Authors:
  • 24.25
  • National Institute on Minority Health and Health Disparities, National Institutes of Health (NIH)

Abstract and Figures

This study examines the association of leave duration with depressive symptoms, mental health, physical health, and maternal symptoms in the first postpartum year, using a prospective cohort design. Eligible employed women, eighteen years or older, were interviewed in person at three Minnesota hospitals while hospitalized for childbirth in 2001. Telephone interviews were conducted at six weeks (N=716), twelve weeks (N=661), six months (N=625), and twelve months (N=575) after delivery. Depressive symptoms (Edinburgh Postnatal Depression Scale), mental and physical health (SF-12 Health Survey), and maternal childbirth-related symptoms were measured at each time period. Two-stage least squares analysis showed that the relationship between leave duration and postpartum depressive symptoms is U-shaped, with a minimum at six months. In the first postpartum year, an increase in leave duration is associated with a decrease in depressive symptoms until six months postpartum. Moreover, ordinary least squares analysis showed a marginally significant linear positive association between leave duration and physical health. Taking leave from work provides time for mothers to rest and recover from pregnancy and childbirth. Findings indicate that the current leave duration provided by the Family and Medical Leave Act, twelve weeks, may not be sufficient for mothers at risk for or experiencing postpartum depression.
Content may be subject to copyright.
Maternity Leave Duration
and Postpartum Mental
and Physical Health:
Implications for Leave Policies
Rada K. Dagher
University of Maryland
Patricia M. McGovern
Bryan E. Dowd
University of Minnesota
Abstract This study examines the association of leave duration with depressive
symptoms, mental health, physical health, and maternal symptoms in the first post-
partum year, using a prospective cohort design. Eligible employed women, eighteen
years or older, were interviewed in person at three Minnesota hospitals while hospi-
talized for childbirth in 2001. Telephone interviews were conducted at six weeks
(N=716),twelve weeks (N=661), six months (N=625), and twelve months (N=575)
after delivery. Depressive symptoms (Edinburgh Postnatal Depression Scale), mental
and physical health (SF-12 Health Survey), and maternal childbirth-related symptoms
were measured at each time period. Two-stage least squares analysis showed that the
relationship between leave duration and postpartum depressive symptoms is U-shaped,
with a minimum at six months. In the first postpartum year, an increase in leave
duration is associated with a decrease in depressive symptoms until six months post-
partum. Moreover, ordinary least squares analysis showed a marginally significant
linear positive association between leave duration and physical health. Taking leave
from work provides time for mothers to rest and recover from pregnancy and child-
birth. Findings indicate that the current leave duration provided by the Family and
This work is based partly on the primary author’s doctoral dissertation, which received funding
from the National Institute for Occupational Safety and Health through the Midwest Center for
Occupational Safety and Health at the University of Minnesota: (1) Grant no. T42-CCT510422,
and (2) Grant no. T42-OH008434. All views expressed herein are solely the responsibility of the
authors. Preliminary versions of this article were presented at the 4th International Commission
on Occupational Health–Work Organization and Psychosocial Factors conference (‘‘The Chan-
ging World of Work’’) in Amsterdam on June 16, 2010, and at the International Biennial Congress
of the Marce
´Society (‘‘Taking Action around Childbirth Together: Mental Health Prevention and
Support to Parenthood’’) in Paris on October 4, 2012. This work was also presented at the Paid
Sick Days/Paid Family Leave Research Convening (organized by the Institute for Women’s
Policy Research and the Work-Family Strategy Council) in Washington, DC, on September 10,
2012.
Journal of Health Politics, Policy and Law, Vol. 39, No. 2, April 2014
DOI 10.1215/03616878-2416247 2014 by Duke University Press
Medical Leave Act, twelve weeks, may not be sufficient for mothers at risk for or
experiencing postpartum depression.
Introduction
Giving birth constitutes a critical life course transition for most women and
their families. At this stage of life, employed women experience a shift in
their household composition, family responsibilities, and work demands, as
well as changes in their childbirth-related health status. Frankenhaeuser’s
(1986) psychobiological model of stress posits that an individual’s well-being
is jeopardized when environmental demands exceed an individual’s resour-
ces. It is reasonable to expect this kind of stress-triggering situation or
imbalance between demands and resources to arise in the months after
childbirth, when women have to juggle multiple roles (worker, partner, and
mother) while physically recovering from childbirth, providing around-
the-clock infant care, and adapting to their ‘‘new normal.’’ Struggling to
fit the prevailing cultural ideals, the ‘‘good mother’’ and the ‘‘good worker’
(Moen and Orrange 2002), women might experience high levels of stress,
the persistence of which might jeopardize their health. Taking time off
from work after childbirth is one strategy that mothers may use to adjust
the imbalance between the demands and resources of this transitional period.
This study examines the impact of leave duration on maternal health during
the first year after childbirth.
Women’s share of the US labor force is expected to reach 47 percent
by the year 2020 (Toossi 2012). Mothers of infants have significantly
contributed to the dramatic rise in women’s labor force participation over
the past few decades. For example, in comparison with a labor force par-
ticipation rate of 38 percent in 1980, mothers of infants under three had a 61
percent participation rate in 2010 (US Bureau of Labor Statistics 2012).
Despite the relatively high participation rates of working mothers in the
workforce, the United States stands out among industrialized nations as
having a national childbirth-related leave policy that provides the shortest
duration of leave after childbirth and no wage replacement (Kamerman
2000). The Family and Medical Leave Act (FMLA) of 1993 is the only
federal policy providing job-protected leave after childbirth. However, it
is unpaid and provides only twelve weeks of unpaid leave to eligible employ-
ees working for covered employers. In a national US study of employed
first-time mothers who gave birth between 2005 and 2007, 58.6 percent
were back to work by three months postpartum (Laughlin 2011). Given
370 Journal of Health Politics, Policy and Law
these relatively short maternity leaves in the United States, it is important
to study their effects on maternal postpartum health. This study examines
the impact of leave duration on four health outcomes over the first year
after childbirth: postpartum depressive symptoms, maternal mental and
physical health, and maternal childbirth-related symptoms.
Policies Pertaining to Postpartum Women
While relatively few policies directly address maternal postpartum health,
family, parental, and maternity leave policies indirectly address it, as
childbirth-related leave from work provides new mothers with time at
home to rest and recover from childbirth and adjust to caring for their
infant. The nature and scope of childbirth-related leave policies varies, as
revealed in the definitions provided by the Clearinghouse on International
Developments in Child, Youth and Family Policies (2002: 2). It defines
maternity leaves as ‘‘job-protected leaves from employment for employed
women at the times they are due to give birth and following childbirth (or
adoption in some countries).’’ Parental leaves are ‘‘gender-neutral, job-
protected leaves from employment that usually follow maternity leaves
and permit either men or women to share the leaves or choose which of
them will use it’’ (ibid.). Paternity leaves are ‘‘job-protected leaves from
employment for fathers, for many of the same purposes as maternity and
parental leaves, but especially for reasons of gender equity’’ (ibid.). Finally,
family leaves are defined as ‘‘job- and benefit-protectedleaves for working
parents including maternity (birth or adoption), paternity, parental, child-
rearing, care for an ill-child, time to accompany a child to school for the
first time, or to visit a child’s school, personal leaves’’ (ibid.). These policies
will be discussed in relation to international and national provisions.
The United States in an International Context
According to article 11 of the 1981 United Nations Convention on the
Elimination of All Forms of Discrimination against Women, countries
must provide paid maternity leave with protection of seniority and benefits,
but the duration of leave that each country should guarantee was not
specified (Tinker 1981). A total of 185 countries have signed this con-
vention, but not the United States (Sundbye and Hegewisch 2011). In
1992, the European Union (EU) Pregnant Workers’ Directive was issued
(Council Directive 1992), under which each of the twenty-seven member
Dagher, McGovern, and Dowd -Maternity Leave Policies 371
states of the EU
1
must guarantee a minimum of fourteen weeks of paid,
job-protected maternity leave as a ‘‘health and safety measure,’ and com-
pensation has to be at least at the statutory sick pay rate.
2
In 1997, the
Amsterdam Treaty on European Union was signed by all EU member
states and included a directive that mandated all states to provide job- and
benefits-protected parental leave of three months duration to men and
women employees who had worked a minimum of one year with their
employers (Haas 2003). However, individual states were given the discre-
tion as towhether the leave is paid or not,part time or full time, whether one
parent is able to transfer it to the other, and whether small employers may be
exempt (Haas 2003).
In a comparative study of twenty-one member countries of the Orga-
nisation for Economic Co-operation and Development (OECD), Ray,
Gornick, and Schmitt (2010) found that the United States lags well behind
other industrialized countries in terms of the generosity of parental leave
policies. In fact, a recent international comparison spanning 181 countries
found that the United States is one of four countries in the world (the others
are Papua New Guinea, Swaziland, and Australia) that do not provide paid
maternity leave as a statutory right (Heymann and Earle 2010). In January
2011, Australia joined the rest of the industrialized countries (except the
United States) in providing paid leave after childbirth ( Mahon and Brennan
2012). The Australian law provides eighteen weeks of paid leave at the
minimum adult wage for parents who worked at least ten hours per week in
the preceding year (Mahon and Brennan 2012).
United States Context
The Family and Medical Leave Act. The primary federal policy that
provides support to working US mothers of infants is the FMLA. It was
signed into law by President Bill Clinton on February 5, 1993, and has
been in effect for most employees since August 5, 1993 (Wisensale 2003).
Under this act, the employee is provided a maximum of twelve weeks
unpaid, job-protected leave per year for giving birth; taking care of a new-
born, a newly adopted child, or a foster child; or attending to an immedi-
ate family member with a serious health condition or the employee’s own
1. The twenty-seven member states of the European Union are Austria, Belgium, Bulgaria,
Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary,
Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania,
Slovakia, Slovenia, Spain, Sweden, and the United Kingdom.
2. All European Union countries have mandatory paid sick leave as a statutory policy.
372 Journal of Health Politics, Policy and Law
serious health condition (Public Law 103-3, 107 Stat. 6 [29 U.S.C. 2601 et
seq.]). The law also provides for protected job benefits, including health
insurance, while the employee is on leave. However, only employees who
have worked in the public sector or for a private employer with fifty or more
employees for a minimum of 1,250 hours in the previous year are eligi-
ble for leave (Public Law 103-3, 107 Stat. 6 [29 USC 2601 et seq.]). These
eligibility criteria are satisfied by only 11 percent of private workplaces
in the United States and are applicable to only 58 percent of the worker
population in private establishments (Cantor et al. 2001). Moreover, females
mostly worked in the service industry, which is predominantly composed
of small companies (Elison 1997). This meant that a sizable proportion of
the female workforce, which the FMLAwas initially designed to protect,
would not be covered by the law. An additional provision of the FMLA
allows the employee to take intermittent leave or leave on a reduced
schedule basis under particular circumstances (Public Law 103-3, 107 Stat.
6 [29 U.S.C. 2601 et seq.], section 102 [b]). This leave can be taken in the
instances of a serious medical condition experienced by the employee or
an immediate family member, and upon certification by a medical pro-
vider. However, taking intermittent leave on a reduced-schedule basis after
childbirth, to care for a newborn, is allowed only if the employer agrees to
it. The FMLA is gender neutral and applies to both mothers and fathers
in the United States, in comparison with other industrialized countries,
where there are often separate laws for maternity and paternity leave.
The length of leave, specified in the FMLA as twelve weeks, was the
result of years of compromise between the stakeholders involved in the
FMLA debate. Initially, the bill was drafted to allow for a parental leave of
eighteen weeks and a medical leave of twenty-six weeks (Elving 1995). A
Harvard pediatrics professor, Berry Brazelton, testified that four months of
parental leave were needed for a healthy parent-child relationship because
it allowed adequate time for bonding (Elison 1997). However, the medical
viewpoint regarding the duration of leave as it affects the mother’s recovery
from childbirth did not enter into the main discussions of the bill. Rather,
the central arguments focused on whether businesses would suffer from a
longer duration of leave, especially given the strong opposition of businesses
to the passage of such a law (Elving 1995). This contrasts with the policy
debates in European countries, where the discussions around the duration
of maternity leave were mainly concerned with the physical health of the
mother and needs of the child (Galtry and Callister 2005; Kamerman 2000).
The unpaid leave benefit provided by the FMLA was a concession
made by sponsors of the bill to ensure its survival in Congress, especially
in the Reagan era of ‘‘small government’’ and ‘‘neo-laissez-faire’’ politics
Dagher, McGovern, and Dowd -Maternity Leave Policies 373
(Elving 1995). The sponsors knew that a mandate on private businesses to
grant leave with compensation would be unacceptable to employers. Thus
the bill was drafted from the start to provide for unpaid leave even though
feminist groups such as the Women’s Legal Defense Fund (WLDF) and the
National Organization for Women (NOW) would have preferred a paid
leave law (Elison 1997; Elving 1995). Women, rather than men, are still the
primary child caregivers when time off work is needed (Bianchi 2011; US
Department of Labor 2005). Moreover, women contribute 35 percent of
combined household income for households with median income earners
(US Department of Commerce 2001); thus taking unpaid leave could seri-
ously threaten the family finances. Women also head about 80 percent of
single-parent households (which constitute 27 percent of all households),
and the loss of paid work could result in no family income at all (Naples
2001). McGovern and colleagues (2000) found that women’s duration of
maternity leave was primarily determined by access to employers’ vol-
untary paid leave benefits or higher partner earnings. Thus, many working
women may not take the duration of leave they need or prefer given that the
FMLA benefits are unpaid. This, in turn, may adversely affect maternal
recovery from pregnancy and childbirth and may affect mental and phys-
ical postpartum health. Many states have laws that expand on some aspect
of leave associated with childbirth, such as the Parental Leave Act of
Minnesota, the state in which our study was conducted.
Parental Leave Act of Minnesota. Minnesota leads the nation in female
labor force participation at 71.2 percent, compared with the national esti-
mate of 59.6 percent according to the Institute for Women’s Policy
Research (IWPR) (2004). It was one of the first states to promulgate a
parental leave law and thus may serve as a model for other states on work-
family issues. Additionally, similarities exist between the distribution of
Minnesota’s employed women and national estimates of occupation, edu-
cational preparation, and earnings (IWPR 2004), increasing the relevance
of the proposed study findings to other states. For example, the propor-
tion of Minneapolis–St. Paul metropolitan area women in managerial and
professional occupations at 33 percent is comparable with that for the nation,
at 34 percent. On average, Minnesota women earned $31,900 per year rel-
ative to $30,100 nationally. Twenty-six percent of Minnesota women earned
a college degree, versus national estimates at 23 percent (IWPR 2004).
The Minnesota Parental Leave Act of 1987 provided eligible employ-
ees with six work weeks of unpaid parental leave for care of a newborn or
an adopted child and ensured restoration to the same position on return
to work (Minn. Stat. Sec. 181.941). This law applies to employers with
374 Journal of Health Politics, Policy and Law
twenty-one or more employees, but the employee had to have worked at
least on a part-time basis in the twelve consecutive months that immedi-
ately precede the leave request. The Minnesota Parental Leave Act requires
the employer to reinstate the benefits received before the employee’s leave.
Although this includes continued coverage for health care premiums, the
employer has the discretion to require the employee to pay full costs; by
contrast, the FMLA does not give the employer such an option. A state-
specific study of a disproportionate random sample of 654 women obtained
from the Minnesota Department of Health in 1991 and 1992 found that
75 percent of Minnesota mothers had returned to work at twelve weeks
postpartum, and 95 percent had returned to work at six months postpartum
(McGovern et al. 1997). Among these Minnesota mothers the average
duration of total time off after childbirth for women with paid leave
(any combination of paid maternity, vacation, or sick leave) benefits was
approximately 10.5 weeks, and for those without any paid leave benefits
it was approximately 6.5 weeks (McGovern et al. 2000). Moreover, the
mothers lacking paid leave benefits were younger and poorer than their
peers with paid leave benefits, suggesting their economic vulnerability
to unpaid leave. The discussions during the legislative hearings of the
Minnesota Parental Leave Act mainly focused on the impact of leave on
employer costs, the importance of job protection for the mother, and
only briefly on the need for bonding with the baby ( McGovern and Segal
1987). However, there was no mention of the mother’s health in any of
these discussions.
In 1997, Minnesota enacted an At-Home Infant Care (AHIC) program,
which provides partial wage replacement to low-income, working parents
for the first year of a child’s life (National Partnership for Women and
Families 2005). This program provides funding to eligible caretaking
parents (i.e., those who are working, looking for work, or going to school in
the nine months preceding the application) whose earnings do not exceed
175 percent of the federal poverty level, to stay at home and care for their
infants. The maximum amount of subsidy is usually around 90 percent of
Minnesota’s maximum rate of payment to a licensed family child care
provider for providing full-time infant care. Only one parent qualifies for
this cash assistance, and the child should be less than one year old (National
Partnership for Women and Families 2005). These AHIC programs (also
enacted in Montana and New Mexico) are typically limited to employed
parents, making AHIC less like the Temporary Assistance for Needy
Families (TANF) program and more like paid family and medical leave
(National Partnership for Women and Families 2005). In its first thirteen
months of implementation, the Minnesota AHIC program served fifty-four
Dagher, McGovern, and Dowd -Maternity Leave Policies 375
families for five months on average (Minnesota Department of Children,
Families, and Learning 1999). Preliminary findings from the evaluation of
the Montana AHIC program showed that it resulted in child care cost
savings of $114,388 for the state, $90,351 of which stemmed from
enhanced ability of parents to care for their other children while at home
taking care of their infants (Annie E. Casey Foundation 2003).
The availability of paid leave policiesmaternity, family, and parental
leave— may influence women’s decisions regarding the timing of return to
work after childbirth, but there is a need for studies to determine whether
such an association exists. Studies that examined the impact of paid and
unpaid leave policies on leave duration are described next.
Impact of Leave Policies on Leave Duration
A national US study of employed first-time mothers who had their child
delivered between 2005 and 2007 found that 58.6 percent were back at
work by three months postpartum, 14.3 percent between three and six
months, and 6.3 percent between six and twelve months (Laughlin 2011).
A study of the impact of the FMLA and other unpaid state leave policies
on leave taking of parents after childbirth, using 1991–1999 data from the
Survey of Income and Program Participation, found that these unpaid
leave laws (number of job-protected leave weeks allowed by each state)
were associated with increased probabilityof leave taking as well as longer
leaves among mothers (these associations disappeared when controlling
for state fixed effects), but not among fathers (Han and Waldfogel 2003).
By contrast, another study of the impact of the FMLA on mothers’ employ-
ment patterns, using longitudinal data (1984–1997) from the Panel Study of
Income Dynamics, found that working mothers who gave birth after the
passage of the FMLA returned to work more quickly and were more likely
to go back to the same employer than those who delivered in the pre-FMLA
period (Hofferth and Curtin 2006). This finding was interpreted by the
authors as implying that instead of women taking longer leaves and
searching later for new jobs, as they did before the FMLAwas passed, they
have an incentive to take shorter leaves and return to their same job.
National survey data of covered establishments show that 52 percent of
employees who utilized FMLA leave in 2000 did so because of personal
ill health and that only 18.5 percent took leave to care for a newborn or a
newly adopted child (Cantor et al. 2001). Over half of the leaves taken
for child care were for ten days or less. The unpaid nature of the FMLA
appears to make it less feasible for parents to take longer periods of leave.
376 Journal of Health Politics, Policy and Law
Next is a discussion of the literature examining the impact of leave dura-
tion on women’s health after childbirth.
Impact of Leave Duration on Postpartum Health
Taking time off after childbirth is one of the ‘‘adaptive strategies’ ( Moen
and Wethington 1992) that women may use in response to the mismatch
between resources and demands typical of this transition period. The first
year after childbirth holds a high risk of depression for women. The most
frequently cited postpartum depression prevalence is from a 1996 meta-
analysis by O’Hara and Swain, suggesting an average of 13 percent among
all pregnancies. This disorder is characterized by a number of debilitating
symptoms similar in clinical presentation to other major depressive dis-
orders (Wisner, Parry, and Piontek 2002). According to the Diagnostic
and Statistical Manual of Mental Disorders-IV (DSM-IV), postpartum
depression begins within four weeks after childbirth (American Psychia-
tric Association 2000). However, multiple studies have shown that the
first three months after childbirth carry the highest risk of postpartum
depression (Horowitz and Goodman 2004; Stowe, Hostetter, and Newport
2005), that depression may start as late as three to six months postpartum
(Beeghly et al. 2002; Goodman 2004; Stuart et al. 1998), and that it may
last beyond the first year after childbirth if untreated (Cooper and Murray
1998). Moreover, a number of longitudinal studies contested the traditional
medical perspective of a six-week postpartum period, the time required
for women’s reproductive organs to go back to their nonpregnant state,
as they found that most women continue to experience several minor to
moderate physical discomforts for many weeks and months after child-
birth (e.g., fatigue, back or neck pain, respiratory symptoms, headaches,
breast soreness, cesarean section or episiotomy discomfort, constipation,
and sexual concerns) (Brown and Lumley 1998; Killien, Habermann, and
Jarrett 2001; McGovern et al. 2011). Thus women may need longer periods
than the conventional six weeks to recover from delivery and childbirth.
A number of studies investigated the impact of leave taken from work
after childbirth and women’s mental health outcomes. A study of 436
white, married, first-time mothers who delivered in one of two hospitals in
St. Paul, Minnesota, found, using the Mental Health Inventory, that taking
more than twenty-four weeks of leave after childbirth, in comparison with
nine weeks of leave or less, was associated with better mental health at nine
and twelve weeks postpartum (Gjerdingen and Chaloner 1994). Another
study of a sample of 266 partnered and mostly white women from Wis-
consin followed through four months postpartum found, using the Center
Dagher, McGovern, and Dowd -Maternity Leave Policies 377
for Epidemiologic Studies Depression Scale (CES-D), that taking six weeks
of leave versus twelve weeks was related to higher depression scores only
among mothers reporting high marital concerns and unrewarding jobs
(Hyde et al. 1995). A similar study that used follow-up data at one year
postpartum of the sample studied by Hyde and colleagues (1995) found
that longer leaves were associated with higher depression scores among
mothers with higher ranking on work role involvement than family role
involvement (Klein et al. 1998). Another study of 123 partnered/married,
primiparous, and mostly white women from the Pacific Northwest found
an association between longer leave duration and increased gratification
with maternal role at eight months after childbirth. However, no associations
were found between leave duration and parental stress or with maternal
separation anxiety (Killien 1998). A study by McGovern and colleagues
(1997) that used a disproportionate random sample of 654 women from
the Minnesota Department of Health data found that leave duration of
more than fifteen weeks after childbirth was associated with better mental
health at seven months postpartum, as measured by the Mental Health
Index (Short-form) from the Medical Outcomes Study (Stewart, Hays,
and Ware 1988). In contrast to the previously mentioned studies, this study
controlled for the potential endogeneity of maternity leave duration using
instrumental variables (McGovern et al. 1997). A study that used a large
and nationally representative sample of 1,762 mothers from the National
Maternal and Infant Health Survey of 1988 reported that a one-week
increase of maternity leave duration was associated with a 6–7 percent
decrease in depressive symptoms (CES-D) six to twenty-four months
postpartum, using state-level leave policies and labor market conditions as
identifying instruments for leave duration (Chatterji and Markowitz 2005).
However, the two-stage least squares (2SLS) coefficient for leave was only
marginally significant. A more recent study by the same authors (Chatterji
and Markowitz 2012) using national data from the Early Childhood
Longitudinal Study Birth Cohort (ECLS_B) on 3,350 mothers found in
ordinary least squares (OLS) regression results that taking less than twelve
weeks of maternity leave increases postpartum depressive symptoms (as
measured by the CES-D) by 15 percent, and increases the likelihood of
being categorized as severely depressed by two points, at nine months after
childbirth. However, there were no significant associations between leave
duration and depressive symptoms in 2SLS analyses. In addition, a sec-
ondary analysis that limited the sample to 2,200 married mothers showed
that fathers’ leave taking was not significantly associated with maternal
depressive symptoms or depression severity and that including paternity
378 Journal of Health Politics, Policy and Law
leave taking in the models did not change the relationship between
maternity leave and depressive symptoms (Chatterji and Markowitz 2012).
The above-mentioned studies point to a potential association between
duration of leave taken from work after childbirth and maternal mental
health outcomes. However, these studies have limited generalizability, as
most of them focus on married and first-time mothers. In addition, only
four of these studies have examined the association between duration of
maternity leave and postpartum depression; these were at four months
(Hyde et al. 1995), one year (Klein et al. 1998), six to twenty-four months
(Chatterji and Markowitz 2005), and nine months (Chatterji and Marko-
witz 2012) after childbirth. All four studies used the Center for Epide-
miological Studies Depression Scale, which is not specifically designed
to measure postpartum depressive symptoms. Moreover, only two of these
studies controlled for the potential endogeneity of leave duration (Chatterji
and Markowitz 2005, 2012). The present article investigates the impact of
duration of leave taken from work after childbirth on employed women’s
postpartum depressive symptoms over the first year after childbirth, using
the Edinburgh Postnatal Depression Scale and controlling for the endo-
geneity of leave duration. In addition, this study examines the impact of
leave duration on women’s mental health, using the Short Form-12 (SF-12)
Mental Summary Score (Ware et al. 2002).
Very little literature examines the association between leave duration
and physical health after childbirth. In general, the few studies that exam-
ined these relationships found no association between taking time off after
childbirth and physical health measures. For example, Killien, Habermann,
and Jarrett (2001) studied 149 employed, partnered women residing in an
urban area in the northwestern United States and found no associations
between week of return to employment and a summary indicator of health
status (including clinical health/symptom experiences, role performance,
and global perception of health) at four months, eight months, and twelve
months postpartum. Similarly, Romito, Saurel-Cubizolles, and Cuttini
(1994) studied 141 employed Italian first-time mothers and found no
associations between length of leave and extreme tiredness, backache, or
lack of sleep. Along the same line, Chatterji and Markowitz (2005) did not
find in their analyses of the National Maternal and Infant Health Survey any
associations between leave duration and other health indicators (number of
outpatient physician or clinic visits) in the first six months postpartum. In
contrast, the study by McGovern and colleagues (1997) showed a U-shaped
association between time off work and maternal health: women who took
twelve weeks of leave or more had greater vitality at seven months after
childbirth, and those who took twenty weeks of leave or more had fewer
Dagher, McGovern, and Dowd -Maternity Leave Policies 379
limitations to their daily roles. Thisstudy examines the association ofleave
duration with two measures of physical health, the SF-12 Physical Sum-
mary Score (Ware et al. 2002) and maternal childbirth-related symptoms,
over the first postpartum year.
Theoretical Framework
The theoretical framework guiding this study is a hybrid model of health
and workforce participation adapted from Becker’s (1965) household
production theory and Grossman’s (1972) health production function
theory. Becker’s theory emphasizes the dynamics of women’s choices to
allocate time between market work and work at home, and Grossman’s
theory establishes that an individual’s level of health partly depends
on the resources allocated to its production and the efficiency with which
it is produced. The main assumption of the hybrid model is that health is
determined by genetic endowment, other ‘‘predetermined factors,’’ and per-
sonal choices. For example, women’s postpartum health status (e.g., postpartum
depression) is a function of predetermined factors, such as demographics,
in addition to personal choices. Women may choose the duration of leave
from work after childbirth as an input to the production of their postpartum
health subject to constraints such as income and employer leave policies.
Methods
Study Design and Population
This study utilizes data from the Maternal Postpartum Health Study, a
nonrandomized, prospective cohort study (McGovern et al. 2006). The
study population consisted of all women, eighteen years or older, who
delivered a live, singleton infant at three community hospitals in the Twin
Cities (Minneapolis and St. Paul) metropolitan area of Minnesota in 2001.
Two of these hospitals were private, nonprofit, tertiary-care hospitals from
the St. Paul area; one was urban and the other was suburban. The third was
a private, nonprofit hospital, located in a Minneapolis suburb characterized
by an ethnically diverse, working-class population.
Analyses comparing the demographics and birth factors for the 3,465
birth mothers at study hospitals to the 22,470 birth mothers at other hos-
pitals in 2001, using vital statistics data, showed the following factors to be
comparable between the study population and all birth mothers, age
eighteen years and older, residing in the general Twin Cities metropolitan
community: age at childbirth, marital status, average number of previous
380 Journal of Health Politics, Policy and Law
live births, the average gestational age for infants, and birth weight.
However, the study hospitals included more Asian mothers, a slightly
lower proportion of African American mothers, and a lower proportion of
mothers having cesarean deliveries relative to other hospitals in the Twin
Cities. While statistically significant, group differences were negligible in
practical terms.
Sample Selection and Data Collection
Sample selection criteria included the following. Women must have given
birth to a live infant at the participating hospitals and have kept the infant
rather than giving it up for adoption. Other infant-related criteria included
having a generally healthy infant— that is, having a pregnancy that lasted
no less than thirty-two weeks, and an infant who weighed more than 1,500
grams and had no reported serious neonatal complications or congenital
anomalies. Moreover, women had to have been continuously employed a
minimum of twenty hours per week for three months in the year before
childbirth and had to have planned to return to work after childbirth.
Women also had to be able to speak English because of the prohibitive
costs of translating the study instruments.
All mothers from the three selected hospitals who met the sample
selection criteria were invited to participate in the study. A total of 2,736
women who gave birth at these hospitals in 2001 were approached.
Among those, 581 women were ineligible because they did not meet demo-
graphic or health criteria of the study, and 998 were ineligible because of
employment-related criteria. Out of an eligible population of 1,157, a sample
of 817 women agreed to participate and were interviewed during early
stages of labor or prior to hospital discharge, constituting a 71 percent
enrollment rate. Refusals were mainly because of women’s time constraints
or lack of interest.
This study was approved by the institutional review boards for the
protection of human subjects at all three participating hospitals and the
University of Minnesota before data collection began. Hospital labor and
delivery nurses screened the medical charts for all maternal admissions.
For those women who met study selection criteria, the nurses abstracted
selected prenatal, labor, and delivery information from the medical chart.
Out of 340 nonparticipants, 295 women consented to the use of their
medical chart data for comparison purposes. Comparisons of participants
to refusals using chi-square analyses and t-tests showed no differences in
regard to infant birth weight, duration of gestation, maternal age, marital
Dagher, McGovern, and Dowd -Maternity Leave Policies 381
status, duration of employment for the participants’ main employer, and
number of hours worked per week before childbirth.
Telephone interviews were conducted with the 817 women enrollees at
six weeks, twelve weeks, six months, and twelve months after childbirth,
with cooperation rates of 88 percent (N=716), 81 percent (N=661), 77
percent (N=625), and 70 percent (N=575) for each time period, respec-
tively. The interviews were conducted using four-week and three-month
windows (i.e., four to eight weeks postpartum; ten to fourteen weeks post-
partum; five to seven months postpartum; and eleven to thirteen months
postpartum); the windows of time for measurement reflected clinically
meaningful intervals for new mothers. The interviewers used structured
questionnaires that included reliable and valid measures for the different
concepts. Each telephone interview lasted around forty-five minutes.
The rationale for collecting data at the selected time periods related to
changes in women’s work patterns after childbirth and the potential impact
of these changes on their health. First, parental leave law in the state of
Minnesota grants six weeks of unpaid leave to eligible women working at
qualifying firms, and most temporary disability insurance policies provide
six weeks of paid leave to covered individuals. Next, the federal Family
Medical and Leave Act grants eligible women at qualifying firms the right
to twelve weeks of unpaid leave. Thus both six weeks and twelve weeks
after childbirth constitute critical intervals for women concerning return
to work and changing workloads. In addition, six months after childbirth
is another critical interval, as previous research in Minnesota revealed that
95 percent of all women returned to work at six months after childbirth
(McGovern et al. 1997). Data collection at twelve months aimed to eval-
uate whether Gjerdingen and colleagues’ (1993, 1994) findings of health
effects associated with postpartum recovery and return to work at twelve
months after childbirth for a sample of married, first-time mothers could be
replicated in more diverse samples such as this sample, which includes
single and multiparous mothers.
Analytical Sample
We limited our analyses to study participants who filled out the six-week
questionnaire in order to have baseline survey data on all women included
in these analyses. Thus out of 817 enrollees in the study, 101 cases were
dropped from the analyses because they had no survey data at six weeks.
Statistical comparisons of the analytical sample (N=716) to the dropped
cases (N=101) showed no differences in regard to parity, duration of
gestation, and number of hours worked per week before childbirth. Dropped
382 Journal of Health Politics, Policy and Law
cases were significantly younger (27.7 vs. 29.9 years), had lower infant
birth weight(3,390 g vs. 3,506 g), had a shorter duration of employment for
their main employer (2.9 vs. 4.2 years), and had a lower household income
($65,055 vs. $73,244). They also were significantly more likely to be single
(48.5 percent vs. 25.8 percent), less likely to have a college degree (22.8
percent vs. 46.2 percent), and less likely to be white (65.3 percent vs. 85.9
percent). Moreover, dropped cases compared with six-week study partic-
ipants were significantly less likely to return to work at six weeks (2.0
percent vs. 7.2 percent), twelve weeks (18.4 percent vs. 39.5 percent), six
months (11.2 percent vs. 41.5 percent), and twelve months (4.1 percent vs.
4.8 percent); moreover, they were more likely to never return to work (64.3
percent vs. 7.1 percent). However, there were no significant differences
between dropped cases
3
and six-week participants in postpartum depres-
sion scores, mental health, physical health, and childbirth-related symptoms
at twelve weeks, six months, and twelve months postpartum. The size of
the analytical sample at each period was 716 at six weeks, 638 at twelve
weeks, 603 at six months, and 554 at twelve months after childbirth.
Measures
Data on the dependent and explanatory variables were collected using
telephone interviews at six weeks, twelve weeks, six months, and twelve
months after childbirth. Data for potentially confounding covariates (con-
trol variables) were collected at enrollment or five weeks postpartum as
detailed below.
Maternal Postpartum Depressive Symptoms. This is one of the dependent
variables in this study, measured at each of the study periods using the
Edinburgh Postnatal Depression Scale (EPDS; Cox, Holden, and Sagovsky
1987). This scale consists of ten short statements about how the mother felt
during the past seven days, with four response categories ranging from 0
to 3 according to increased severity of the symptoms. At six weeks, twelve
weeks, six months, and twelve months this measure had internal con-
sistency reliabilities (Cronbach’s alpha) of 0.82, 0.83, 0.84, and 0.86,
respectively. The EPDS has been found to have satisfactory validity in
identifying mothers with postpartum depression, where a threshold score
of 12 to 13 identified women with Definite Major Depressive Illness
3. We had data on health outcomes for only 30 out of 101 of the dropped cases; thus our
reporting of no differences in health outcomes between dropped cases and six-week participants
is based on available data on only 29.7 percent of the dropped cases.
Dagher, McGovern, and Dowd -Maternity Leave Policies 383
according to Research Diagnostic Criteria (Cox, Holden, and Sagovsky
1987). A recent study comparing identification rates of maternal postpar-
tum depression using the EPDS scale versus routine clinical evaluation
found a significantly higher likelihood of identification by the former
(Fergerson, Jamieson, and Lindsay 2002).
Maternal Physical and Mental Health. Physical and mental health were
two dependent variables measured at each of the study periods using the
SF-12 (Version 2), a twelve-item measure of general health that does not
target a specific age or disease group. This measure includes the Physical
Component Summary (PCS) and Mental Component Summary (MCS)
scales (Ware et al. 2002). The PCS consists of six items that address four
content domains of general physical health: physical functioning (two
items), bodily pain (one item), general health (one item), and role limita-
tions because of physical health (two items). The MCS consists of six
items that address four content domains of general mental health: social
function (one item), mental health (two items), vitality (one item), androle
limitations because of emotional health (two items) (Ware et al. 2002).
The MCS and PCS item sets can each be summed into a score that ranges
from 0 to 100. These measures are internationally recognized for their
validity and reliability (Ware et al. 2002). Because the SF-12 is standard-
ized and norm-based, scores above and below 50 are above and below the
average, respectively, in the general US population, and higher scores
denote better health (Ware et al. 2002).
Maternal Childbirth-Related Symptoms. In this study maternal childbirth-
related symptoms comprised a dependent variable that was measured at
each of the study periods using a physical problem checklist adapted from
Gjerdingen and colleagues (1993). It asks women to report the presence of
problems such as breast symptoms, vaginal discomfort, hemorrhoids, and
fatigue. It is scored as a simple summation of symptoms experienced in the
preceding four weeks, consistent with the recall period of the SF-12 v.2.
The summary score evaluated the presence of twenty-eight symptoms
frequently experienced during the postpartum period, representative of the
major body systems or constitutional in nature (e.g., skin, respiratory, car-
diovascular, gastrointestinal, genito-urinary, musculoskeletal, and endocrine).
Leave Duration
Number of days on leave after childbirth was measured using the ques-
tion ‘‘How long were you on leave after the baby was born, including any
384 Journal of Health Politics, Policy and Law
part-time leave?’’ This question was asked at each of the four periods after
childbirth. The measure was originally adapted from Cantor and colleagues
(2001). This is an endogenous variable, as it is a choice made by the woman
contingent on her employer’s policy, her financial status, and her health
status. This variable was coded in the analysis in a cumulative fashion. For
example, someone who had been on leave for sixty days was coded as
having had forty-two days of leave at six weeks after childbirth, even
though this person’s leave extends to the twelve-week period, where they
would be coded as having had sixty days of leave. The squared form of this
leave variable was also included in the analyses to determine whether the
relationships between leave duration and the dependent variables were
nonlinear.
Control Variables. Based on a priori causal assumptions, several covari-
ates that could confound the relationship between leave duration and
maternal postpartum health outcomes were included as control variables.
These control variables included maternal sociodemographic character-
istics and prenatal moods of depression and anxiety.
Maternal Sociodemographic Characteristics. Age was abstracted from
the medical chart. The measure of race was adapted from a measure of
the US Census Bureau (2001) and subsequently coded as (1 =white;
0=nonwhite). Measures of education (1 =college graduate; 0 =not a
college graduate), annual household income, and parity (1 =multiparous;
0=primiparous) were adapted from the National Health Interview Survey
(Ries 1991). Household income was defined as income from all sources
in the year before childbirth. Data on all of these variables were collected
in person at enrollment. The measure of marital status was adapted from
the National Health Interview Survey (Ries 1991) and assessed at six
weeks postpartum. It was subsequently coded as (married =1, else =0;
partnered =1, else =0; single =reference). Occupation was classified
with reference to Alphabetical Index of Industries and Occupations (1990)
three-digit occupational codes. Ultimately, occupation was coded as a dummy
variable in the analyses (blue-collar/service =1, else =0; clerical =1,
else =0; professional =reference). All of these variables are considered
predetermined.
Prenatal Moods. This predetermined variable was created by Dr. Dwenda
Gjerdingen and validated by McGovern and colleagues (1997). It consists
of one item: ‘‘During this pregnancy didyou ever have a problem with your
Dagher, McGovern, and Dowd -Maternity Leave Policies 385
mood, such as feeling depressed or anxious?’’ Data on this variable were
collected in person at enrollment.
Estimation and Empirical Models
There are two potential sources of bias in our estimates of the effect of leave
on the dependent variables: (1) omitted variables and (2) reverse causality.
Duration of leave is, in part, a choice made by the mother; however, this
choice may be influenced by a number of factors, including the employer’s
leave policy, the mother’s financial status, and how much support she has
with child care. If one or more of these variables affects both leave duration
and any of the dependent variables in the analysis and is not controlled, the
result is ‘‘omitted variable bias’’ (spurious correlation). Reverse causality
arises if a dependent variable (e.g., postpartum depression) exerts a causal
influence on leave duration.
We first tested for the endogeneity of leave duration using the Hausman
test. The Hausman test requires a variable or ‘‘instrument’’ that predicts
leave duration but is not correlated with the error term in the dependent
variable equation. We allowed for a nonlinear relationship between leave
duration and women’s postpartum health by adding the square of leave dura-
tion to each equation. Since squared leave may also be endogenous, two
instruments are required.
We have two instruments in our data: the maximum available duration
of all paid leave according to employer policy including vacation, sick,
maternity, or disability leave (Mean
days
=46.38, SD =39.73, Range =
0–273 days) and the maximum available duration of job-protected
leave according to employer policy (Mean
days
=124.97, SD =134.80,
Range =0–1,825 days). These two variables have been shown to predict
duration of leave (McGovern et al. 1997, 2000). We conducted a joint
significance test for these two instruments, and the F-statistic was 50.37
(p=0.000). Thus these instruments are good predictors of leave duration.
Moreover, these instruments are not theoretically related to women’s post-
partum health except through the actual duration of leave after childbirth.
These instruments would be invalid if women chose workplaces because of
leave policies and based on prior health status. However, these women have
worked an average of four years ( Mean =4.2, SD =4.0) with their employer
before childbirth (87 percent of the women worked one year or more), so it is
unlikely that they were making short-term employment decisions.
The Durbin-Wu-Hausman test rejected the null hypothesis of exogenous
leave duration and exogenous squared leave in the postpartum depression
386 Journal of Health Politics, Policy and Law
equation (Chi-sq (2) =17.15; p=0.00019); however, it was not signifi-
cant for the mental health, physical health, and childbirth-related symp-
toms equations. Thus we estimated the postpartum depression equation
with two-stage least squares (2SLS) regression using the two instruments
discussed above, and used ordinary least squares (OLS) regression to
estimate the remaining health outcome equations. However, we show both
2SLS and OLS regression coefficients in all the tables to allow the reader to
compare the results for all outcome variables.
Results
Descriptive Statistics
Table 1 shows the demographic characteristics of the participants for all
study periods. These variables were measured once at baseline, and the
statistics represent the demographics of those who answered the survey at
each time period.
4
The majority of women in the study sample (N=716)
were white (85.9 percent) and married (73.2 percent), with annual house-
hold incomes of $50,000 or higher (72 percent). The sample mean age was
thirty years (SD =5.3), 46.2 percent had a college degree or higher, 46.5
percent were first-time mothers, and 46.5 percent reported prenatal moods.
Women had worked an average of 4.2 years (SD =4.0; Range =0.04–21)
for their current employer, and 38.1 hours per week (SD =8.5; Range =
20–80) before delivery. In the first year after childbirth, 7.1 percent
(N=51) of the women had returned to work by six weeks, 46.2 percent
(N=331) by twelve weeks, 87.3 percent (N=625) by six months, and
92.0 percent (N=659) by twelve months postpartum. The mean duration
of leave taken over the twelve-month period after childbirth was 72.51
days (SD =40.57, Range =2–365).
Table 2 shows the mean scores for each of the dependent variables. On
average, over all the study periods (N=2,511), the mean postpartum
depression score was 4.36 (SD =4.03, Range =0–24). The proportion
of women who met the threshold of 12.5 for the Edinburgh Postnatal
Depression Scale was 5.6 percent (N=40) at six weeks, 4.7 percent (N=30)
4. Logistic regression, with response status at each time period as the dependent variable and
demographics as the explanatory variables, was used to estimate attrition from the sample. At
twelve weeks postpartum, attrition did not differ by demographic characteristics, while at six
months postpartum those who responded to the survey were significantly more likely to be in a
professional occupation and to be married or partnered. At twelve months postpartum, responders
were significantly more likely to be in a professional occupation and to have reported no prenatal
moods.
Dagher, McGovern, and Dowd -Maternity Leave Policies 387
at twelve weeks, 4.1 percent (N=25) at six months, and 5.8 percent (N=32)
at twelve months postpartum. On average, over all the study periods
(N=2,713), the mean mental health score was 50.32 (SD =7.49,
Range =9.24–71.27), and the mean physical health score was 54.38
(SD =6.31, Range =9.31–72.90). Moreover, the mean maternal symp-
toms score averaged 4.72 (SD =3.47, Range =0–18) over all the study
periods (N=2,511).
Multivariate Analyses
Ordinary least squares and two-stage least squares regressions were con-
ducted to examine the impact of leave duration on each of the dependent
Table 1 Participant Demographics at Each of the Study Periods
6 Weeks 12 Weeks 6 Months 12 Months
Variable (N=716) (N=638) (N=603) (N=554)
Age
Mean (SD) 29.90 (5.28) 30.05 (5.26) 30.13 (5.13) 30.20 (5.13)
Range 18–45 18–45 18–45 18–45
Marital Status (%)
Single 10.5 10.0 8.1 8.4
Partnered 16.3 15.4 15.3 14.3
Married 73.2 74.6 76.6 77.3
College Degree (%)
1=Yes 46.2 48.7 49.6 51.3
Income (%)
$0–24,999 6.4 6.3 5.8 4.5
$25,000–49,999 21.6 20.4 19.4 20.0
$50,000–74,999 25.3 24.6 24.9 25.5
$75,000–99,999 29.1 30.4 30.7 31.4
>$100,000 17.6 18.3 19.2 18.6
Race (%)
1=White 85.9 87.0 87.6 88.4
Occupation (%)
Blue-Collar 14.4 13.5 12.6 13.4
Clerical 39.2 37.8 37.3 34.8
Professional 46.4 48.7 50.1 51.8
Parity (%)
1=Multiparous 53.5 54.2 53.9 53.8
Prenatal Moods (%)
1=Yes 46.5 46.1 44.8 44.4
Source: Authors’ calculations
388 Journal of Health Politics, Policy and Law
variables: postpartum depressive symptoms, mental health, physical
health, and maternal childbirth-related symptoms (tables 4 to 7, respec-
tively). Table 3 presents the results of first-stage regressions of instru-
mental variables predicting leave duration.
Dependent Variable: Postpartum Depressive Symptoms. OLS analyses
showed no significant association between leave duration and postpartum
depression scores; however, 2SLS
5
analyses showed that the estimated
relationship between leave duration and postpartum depressive symptoms
is significant and U-shaped, with a minimum at 180 days, or approximately
six months (table 4). In general, significanteffects of leave duration in OLS
analyses become nonsignificant in 2SLS analyses because the estimated
coefficients become smaller and the associated standard errors become
larger. The postpartum depression equation results showed an exception to
this pattern: both the estimated coefficients and the associated standard
errors became larger, suggesting the presence of an omitted confounder in
the OLS analyses that was positively associated with leave duration and
Table 2 Descriptive Characteristics of the Dependent Variables
at Each of the Study Periods
Health Outcome Variables
(Theoretical scores) 6 Weeks 12 Weeks 6 Months 12 Months
Postpartum Depressive Symptoms (0–30)
Mean (SD) 4.86 (3.91) 4.19 (3.95) 4.21 (4.03) 4.04 (4.27)
Range 0–21 0–23 0–23 0–24
Maternal Mental Health (0–100)
Mean (SD) 49.42 (7.58) 50.43 (7.33) 50.42 (7.47) 51.08 (7.51)
Range 15.05–65.97 9.24–63.02 14.80–71.27 19.11–64.78
Maternal Physical Health (0–100)
Mean (SD) 51.38 (7.20) 55.79 (5.16) 55.16 (5.86) 55.41 (5.69)
Range 19.16–67.68 19.82–72.90 9.31–67.78 20.01–68.80
Maternal Symptoms (0–28)
Mean (SD) 6.11 (3.46) 4.17 (3.20) 4.65 (3.39) 3.62 (3.28)
Range 0–18 0–16 0–17 0–17
Source: Authors’ calculations
5. A three-stage least squares analysis was also conducted, as it provides more efficient
estimates when both omitted variables and simultaneity bias exist. The estimated coefficients and
standard errors of leave duration and leave squared were the same as those estimated by 2SLS for
postpartum depressive symptoms.
Dagher, McGovern, and Dowd -Maternity Leave Policies 389
Table 3 Results of the First-Stage Regressions of Instrumental Variables Predicting Leave Duration
Explanatory Variables b
(PPD)
SE
(PPD)
95% CI
(PPD)
b
(MPC)
SE
(MPC)
95% CI
(MPC)
Total Paid Leave
Policy (days)
0.0919*** 0.0174 0.0577 0.1261 0.1124*** 0.0169 0.0793 0.1455
Total Job-Protected
Leave Policy (days)
0.0403*** 0.0048 0.0310 0.0497
Prenatal Moods 0.8577 1.2919 -1.6756 3.3911 0.8923 1.2770 -1.6118 3.3963
Age (years) 0.6369*** 0.1482 0.3463 0.9275 0.6738*** 0.1442 0.3911 0.9564
Education (1 =College
educated)
-0.9784 1.5811 -4.0787 2.1220 0.5185 1.5657 -2.5517 3.5886
Race (1 =White) 3.2855 2.0712 -0.7760 7.3470 3.6681{2.0331 -0.3185 7.6546
Income
$0–24,999 (Reference)
$25,000–49,999 -3.9938 3.1312 -10.1339 2.1462 -3.3555 3.0450 -9.3264 2.6154
$50,000–74,999 -1.4292 3.4718 -8.2373 5.3788 0.4396 3.3769 -6.1821 7.0613
$75,000–99,999 -1.1314 3.6386 -8.2665 6.0037 -0.3311 3.5371 -7.2669 6.6046
>$100,000 0.4698 3.8364 -7.0532 7.9929 -0.2100 3.7415 -7.5465 7.1265
Occupation
Professional (Reference)
Blue-Collar -8.9098*** 2.1715 -13.1679 -4.6518 -10.6971*** 2.1445 -14.9022 -6.4920
Clerical -2.3775 1.5699 -5.4559 0.7010 -4.1170** 1.5433 -7.1432 -1.0907
Table 3 (continued )
Explanatory Variables b
(PPD)
SE
(PPD)
95% CI
(PPD)
b
(MPC)
SE
(MPC)
95% CI
(MPC)
Marital Status
Single (Reference) — — — —
Married 3.6448 2.7006 -1.6509 8.9404 4.2533 2.6276 -0.8991 9.4057
Partnered 3.0148 2.6711 -2.2231 8.2527 3.2062 2.5893 -1.8712 8.2835
Parity (1 =Multiparous) -3.0465* 1.3408 -5.6758 -0.4172 -2.8261* 1.3205 -5.4154 -0.2366
Time Period
6 weeks (Reference) — — — —
12 weeks 23.9440*** 1.7051 20.6003 27.2877 23.6189*** 1.7154 20.2553 26.9826
6 months 33.7924*** 1.7317 30.3968 37.1881 33.5483*** 1.7366 30.1431 36.9536
12 months 37.0234*** 1.7625 33.5673 40.4795 36.5526*** 1.7319 33.1566 39.9486
Source: Authors’ calculations
Notes:b=unstandardized coefficient; SE =standard error;
(PPD)
=first equation estimates generated for the postpartum depression equation;
(MPC)
=first
equation estimates generated for Mental health, Physical health, and Childbirth-related symptoms equations.
{p<0.10; *p£0.05; **p£0.01; ***p£0.001
N
(PPD)
=2,455, F
(PPD)
=48.57, Prob >F
(PPD)
=0.0000, adjusted R-squared
(PPD)
=0.2587; N
(MPC)
=2,645, F
(MPC)
=47.28, Prob >F
(MPC)
=0.0000,
adjusted R-squared
(MPC)
=0.2293
Table 4 Results of OLS and 2SLS Analyses of the Association between Leave Duration and Postpartum Depressive
Symptoms (N=2,455)
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Leave Duration -0.0018 0.0054 -0.0123 0.0088 -0.1075*** 0.0311 -0.1685 -0.0465
Leave Squared -3.93e
-06
0.00002 -0.00004 0.00003 0.0003* 0.0001 0.0001 0.0005
Prenatal Moods 2.0563*** 0.1597 1.7431 2.3695 2.1878*** 0.1787 1.8374 2.5382
Age (years) 0.0182 0.0181 -0.0173 0.0538 0.0543* 0.0223 0.0107 0.0980
Education (1 =College
educated)
-0.0931 0.1944 -0.4743 0.2882 0.0286 0.2130 -0.3892 0.4463
Race (1 =White) -1.0317*** 0.2555 -1.5327 -0.5307 -0.9175*** 0.2775 -1.4617 -0.3733
Income
$0–24,999 (Reference)
$25,000–49,999 1.0103** 0.3857 0.2541 1.7666 1.1285** 0.4222 0.3007 1.9563
$50,000–74,999 -0.6026 0.4284 -1.4426 0.2375 -0.1841 0.4865 -1.1381 0.7698
$75,000–99,999 -0.2288 0.4479 -1.1071 0.6495 0.2714 0.5161 -0.7406 1.2834
>$100,000 -0.5509 0.4720 -1.4764 0.3747 0.0744 0.5524 -1.0089 1.1577
Occupation
Professional (Reference)
Blue-Collar 0.1336 0.2701 -0.3961 0.6633 -0.6571{0.3584 -1.3598 0.0457
Clerical -0.1249 0.1934 -0.5042 0.2544 -0.2487 0.2123 -0.6650 0.1676
Table 4 (continued )
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Marital Status
Single (Reference) — — — —
Married -0.4066 0.3339 -1.0614 0.2482 -0.3207 0.3648 -1.0360 0.3946
Partnered -0.9989** 0.3297 -1.6455 -0.3524 -0.8854* 0.3606 -1.5926 -0.1783
Parity (1 =Multiparous) 0.2671 0.1652 -0.0569 0.5910 0.0416 0.1876 -0.3263 0.4096
Time Period
6 weeks (Reference) — — — —
12 weeks -0.5800* 0.2271 -1.0254 -0.1347 1.0830* 0.5017 0.0992 2.0668
6 months -0.4615{0.2391 -0.9304 0.0074 1.5811** 0.5931 0.4181 2.7441
12 months -0.6631** 0.2410 -1.1357 -0.1905 1.1716* 0.5629 0.0678 2.2754
Source: Authors’ calculations
Notes:b=unstandardized coefficient; SE =standard error;
(OLS)
=estimates generated using ordinary least squares regression;
(2SLS)
=estimates generated
using two-stage least squares regression; adjusted R-squared
(OLS)
=0.1147
{p<0.10; *p£0.05; **p£0.01; ***p£0.001
negatively associated with postpartum depression scores or vice versa,
and using 2SLS corrected the bias. The results of 2SLS analyses can be
illustrated in the following equation, where PPD stands for the depen-
dent variable postpartum depressive symptoms, and LD stands for leave
duration: PPD =-0.108 LD +0.0003 LD-squared. To find the postpar-
tum depressive symptoms minimizing value of leave duration (LD), we
need to differentiate PPD with respect to LD and set the derivative equal to
zero. This results in the equation -0.108 +0.0006 LD =0. Thus LD =
180 days, which is approximately six months. This is a minimum because
the second derivative of PPD with respect to LD is positive (0.0006).
Therefore, on average, in the first postpartum year every additional day
of leave results in a decrease in postpartum depressive symptoms until
six months postpartum. After six months, the relationship reverses and
every additional day of leave results in an increase in postpartum depres-
sive symptoms. However, we acknowledge that while we measured leave
continuously, we measured postpartum depression at discrete intervals;
thus there is a possibility that the relationship between leave duration and
postpartum depression symptoms may not be continuous. To illustrate this
relationship we depicted the variation of postpartum depressive symptoms
by leave duration over the first year after childbirth, using predicted values
from the regression, in a graph (fig. 1). T-tests showed that women who
were back to work in the first six months after childbirth (i.e., by six weeks,
twelve weeks, or six months) had higher depressive symptoms than those
who were still on leave ( Mean
6 weeks
=6.51 vs. 4.74, p-value =0.000;
Mean
12 weeks
=5.24 vs. 3.30, p-value =0.000; Mean
6months
=4.41 vs.
3.43, p-value =0.0012). However, at twelve months postpartum this
relationship reversed: those who were back to work had lower depressive
symptoms than those who were still on leave (Mean
12 months
=3.97 vs.
4.90, p-value =0.0012). However, these results should be interpreted
with caution, as 87.3 percent (N=625) of the women returned to work
by six months postpartum, and 92.0 percent (N=659) were back to work
by twelve months. Thus it is difficult to make accurate predictions of the
impact of leave durations that are longer than six months on women’s
postpartum depression scores.
Significant findings in the 2SLS regression also emerged from control
variables (see table 4).Womenwho experienced prenatal moodsof anxiety
and depression had an increase of 2.2 points in their postpartum depression
score (worse outcome) compared with those who did not experience those
moods. Older age was associated with increased depression scores. White
394 Journal of Health Politics, Policy and Law
women had lower depression scores (almost one point lower) than non-
whites. Moreover, partnered women had lower depression scores than
single women. Women with household incomes of $25,000 to $50,000 had
significantly higher depression scores than those with incomes less than
$25,000, but there were no differences in postpartum depression scores
for women with incomes higher than $50,000 as compared with those
with incomes less than $25,000. Time was significantly associated with
increased postpartum depressionscores. Each time period— twelve weeks,
six months, and twelve months— was associated with significantly higher
depression scores relative to the reference period at six weeks postpartum,
with the highest increase in depression scores at six months after childbirth
(see table 4).
Dependent Variable: Maternal Mental Health. Neither 2SLS regression
analyses nor OLS analyses showed a significant association between leave
duration
6
and maternal mental health (table 5). This is consistent with
the results of the Hausman test, which failed to reject the hypothesis of
Figure 1 Employment Status (On Leave/Working Again) and Predicted
Values of Postpartum Depressive Symptoms over Four Time Periods in
the First Year after Childbirth
6. We dropped leave squared from the regression analyses because it was not significant in
OLS or in 2SLS analyses.
Dagher, McGovern, and Dowd -Maternity Leave Policies 395
Table 5 Results of OLS and 2SLS Analyses of the Association between Leave Duration
and Postpartum Mental Health (N=2,645)
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Leave Duration 0.0063 0.0044 -0.0022 0.0149 0.0280 0.0213 -0.0138 0.0697
Prenatal Moods -3.8560*** 0.2874 -4.4195 -3.2926 -3.8667*** 0.2889 -4.4332 -3.3002
Age (years) 0.0195 0.0322 -0.0436 0.0826 0.0013 0.0368 -0.0709 0.0734
Education (1 =College
educated)
0.1385 0.3520 -0.5517 0.8286 0.1123 0.3545 -0.5829 0.8074
Race (1 =White) -0.3149 0.4565 -1.2101 0.5803 -0.3700 0.4617 -1.2754 0.5354
Income
$0–24,999 (Reference)
$25,000–49,999 -1.6372* 0.6829 -2.9762 -0.2982 -1.6063* 0.6867 -2.9528 -0.2597
$50,000–74,999 -0.0061 0.7581 -1.4927 1.4805 -0.0559 0.7632 -1.5525 1.4406
$75,000–99,999 -0.0100 0.7916 -1.5622 1.5423 -0.0611 0.7968 -1.6236 1.5013
>$100,000 0.0180 0.8367 -1.6226 1.6586 -0.0429 0.8426 -1.6952 1.6094
Occupation
Professional (Reference)
Blue-Collar 0.6418 0.4847 -0.3086 1.5921 0.8941 0.5442 -0.1730 1.9612
Clerical 0.2298 0.3479 -0.4525 0.9121 0.3130 0.3586 -0.3902 1.0163
Table 5 (continued )
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Marital Status
Single (Reference)
Married 0.3001 0.5921 -0.8608 1.4611 0.1983 0.6029 -0.9839 1.3804
Partnered 1.3045* 0.5823 0.1626 2.4463 1.2103* 0.5920 0.0494 2.3712
Parity (1 =Multiparous) -1.0308*** 0.2965 -1.6121 -0.4495 -0.9509** 0.3076 -1.5541 -0.3476
Time Period
6 weeks (Reference)
12 weeks 0.9344* 0.3999 0.1503 1.7186 0.4229 0.6357 -0.8236 1.6694
6 months 0.7351{0.4175 -0.0836 1.5538 0.0104 0.8144 -1.5865 1.6072
12 months 1.4637*** 0.4214 0.6373 2.2900 0.6727 0.8715 -1.0361 2.3816
Source: Authors’ calculations
Notes:b=unstandardized coefficient; SE =standard error;
(OLS)
=estimates generated using ordinary least squares regression;
(2SLS)
=estimates generated
using two-stage least squares regression; adjusted R-squared
(OLS)
=0.0894
{p<0.10; *p£0.05; **p£0.01; ***p£0.001
exogenous leave duration, so correction for endogeneity by utilizing 2SLS
did not have an effect. To illustrate this relationship we depicted the var-
iation of maternal mental health scores by leave duration over the first
year after childbirth, using predicted values from the regression, in a graph
(fig. 2). T-tests showed no differences in predicted mental health scores
between women who were back to work and those still on leave at twelve
weeks, six months, and twelve months. However, women who were back
to work at six weeks after childbirth had worse mental health scores than
those who were still on leave (Mean
6weeks
=48.69 vs. 49.41, p-value =
0.026).
A review of the control variables in the OLS model reveals that women
who experienced prenatal moods of depression and anxiety scored approx-
imately four points lower on the mental health measure (poorer mental
health) than women with no prenatal moods (see table 5). Women with
household incomes of $25,000 to $50,000 had better mental health scores
than those with incomes less than $25,000. Partnered women had better
mental health than single women. Multiparous women had poorer mental
health scores than primiparous women. Time was significantly associated
with better mental health. For the time periods twelve weeks and twelve
Figure 2 Employment Status (On Leave/ Working Again) and Predicted
Values of Maternal Mental Health Scores over Four Time Periods
in the First Year after Childbirth
398 Journal of Health Politics, Policy and Law
months, time was associated with significantly better mental health relative
to the reference period at six weeks postpartum, with the highest improve-
ment in mental health at twelve months after childbirth (see table 5).
Dependent Variable: Maternal Physical Health. Both OLS and 2SLS
analyses showed a marginally significant linear association between leave
duration
7
and maternal physical health, where longer leaves were asso-
ciated with better physical health (table 6). These similar findings between
OLS and 2SLS analyses are in line with the Hausman test results, which
failed to reject the hypothesis of exogenous leave duration with respect to
physical health. To illustrate this relationship we depicted the variation of
maternal physical health scores by leave duration over the first year after
childbirth, using predicted values from the regression, in a graph (fig. 3).
T-tests showed that women who were back to work at six weeks or twelve
weeks after childbirth had slightly worse, although statistically significant,
physical health scores than those who were still on leave during those periods
(Mean
6 weeks
=51.00 vs. 51.39, p-value =0.004; Mean
12 weeks
=55.54 vs.
55.95, p-value =0.000). However, there were no differences in predicted
physical health scores between women who were back to work and those
still on leave at six months and twelve months after childbirth.
In terms of control variables, OLS regression showed that women with
prenatal moods of depression and anxiety had poorer physical health than
those with no prenatal moods (see table 6). Moreover, older age was
associated with poorer physical health. White women had better physical
health scores (1.62 points higher) than nonwhites. Women with household
incomes of $100,000 or higher had better physical health scores than
those with incomes less than $25,000. Time was significantly associated
with better physical health. Each time periodtwelve weeks, six months,
and twelve months— was associated with significantly better physical
health relative to the reference period at six weeks postpartum, with the
highest improvement in physical health at twelve weeks after childbirth
(see table 6).
Dependent Variable: Maternal Childbirth-Related Symptoms. OLS
analyses showed no significant association between leave duration
8
and
maternal childbirth-related symptoms, while 2SLS analyses showed a
7. We dropped leave squared from the regression analyses because it was not significant in
OLS or in 2SLS analyses.
8. We dropped leave squared from the regression analyses because it was not significant in
OLS or in 2SLS analyses.
Dagher, McGovern, and Dowd -Maternity Leave Policies 399
Table 6 Results of OLS and 2SLS Analyses of the Association between Leave Duration
and Postpartum Physical Health (N=2,645)
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Leave Duration 0.0063{0.0037 -0.0008 0.0135 0.0297{0.0179 -0.0055 0.0648
Prenatal Moods -0.5180* 0.2413 -0.9911 -0.0449 -0.5295* 0.2433 -1.0065 -0.0525
Age (years) -0.0895*** 0.0270 -0.1425 -0.0366 -0.1092*** 0.0310 -0.1699 -0.0484
Education (1 =College
educated)
0.2491 0.2955 -0.3304 0.8285 0.2208 0.2985 -0.3646 0.8062
Race (1 =White) 1.6248*** 0.3833 0.8732 2.3764 1.5655*** 0.3888 0.8030 2.3279
Income
$0–24,999 (Reference) — — — —
$25,000–49,999 -0.0604 0.5733 -1.1846 1.0638 -0.0271 0.5783 -1.1610 1.1068
$50,000–74,999 1.2126{0.6365 -0.0355 2.4607 1.1589{0.6427 -0.1014 2.4191
$75,000–99,999 0.5793 0.6646 -0.7239 1.8825 0.5241 0.6710 -0.7916 1.8399
>$100,000 1.4863* 0.7024 0.1089 2.8637 1.4206* 0.7095 0.0292 2.8120
Occupation
Professional (Reference) — — — —
Blue-Collar 0.0025 0.4069 -0.7954 0.8003 0.2745 0.4583 -0.6241 1.1731
Clerical 0.1783 0.2921 -0.3946 0.7511 0.2680 0.3020 -0.3242 0.8602
Table 6 (continued )
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Marital Status
Single (Reference) — — — —
Married -0.1237 0.4971 -1.0984 0.8511 -0.2335 0.5077 -1.2290 0.7620
Partnered -0.3361 0.4889 -1.2947 0.6226 -0.4376 0.4985 -1.4152 0.5400
Parity (1 =Multiparous) -0.0172 0.2489 -0.5053 0.4708 0.0690 0.2591 -0.4390 0.5770
Time Period
6 weeks (Reference) — — — —
12 weeks 4.2288*** 0.3358 3.5704 4.8872 3.6773*** 0.5353 2.6276 4.7270
6 months 3.5550*** 0.3506 2.8676 4.2424 2.7735*** 0.6858 1.4288 4.1182
12 months 3.7519*** 0.3538 3.0581 4.4456 2.8991*** 0.7339 1.4600 4.3381
Source: Authors’ calculations
Notes:b=unstandardized coefficient; SE =standard error;
(OLS)
=estimates generated using ordinary least squares regression;
(2SLS)
=estimates generated
using two-stage least squares regression; adjusted R-squared
(OLS)
=0.0994
{p<0.10; *p£0.05; **p£0.01; ***p£0.001
marginally significant linear association; as leave duration increased,
maternal childbirth-related symptoms decreased (table 7). Whereas the
signs and magnitude of the coefficients did not change much across OLS
and 2SLS analyses, there was a minor change in statistical significance;
thus these findings are consistent with the results of the Hausman test,
which failed to reject the hypothesis of exogenous leave duration in the
equation for maternal childbirth-related symptoms. To illustrate this rela-
tionship we depicted the variation of maternal childbirth-related symptoms
by leave duration over the first year after childbirth, using predicted values
from the regression, in a graph (see fig. 4). T-tests showed no differences in
predicted childbirth-related symptoms between women who were back to
work and those still on leave at six weeks, six months, and twelve months.
However, women who were back to work at twelve weeks after childbirth
had a lower score on childbirth-related symptoms than those who were still
on leave (Mean
12 weeks
=4.07 vs. 4.23, p-value =0.019).
In OLS regression, control variables significantly associated with
fewer childbirth-related symptoms (better health) included an absence of
prenatal mood problems, a household income of $50,000–74,999 or
Figure 3 Employment Status (On Leave/ Working Again) and Predicted
Values of Maternal Physical Health Scores over Four Time Periods
in the First Year after Childbirth
402 Journal of Health Politics, Policy and Law
Table 7 Results of OLS and 2SLS Analyses of the Association between Leave Duration and Maternal
Childbirth-Related Symptoms (N=2,455)
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Leave Duration -0.0013 0.0021 -0.0054 0.0028 -0.0165{0.0099 -0.0360 0.0030
Prenatal Moods 1.4030*** 0.1358 1.1366 1.6693 1.4079*** 0.1373 1.1386 1.6772
Age (years) 0.0548*** 0.0154 0.0245 0.0850 0.0677*** 0.0176 0.0331 0.1023
Education (1 =College
educated)
0.3636* 0.1654 0.0393 0.6880 0.3765* 0.1674 0.0482 0.7048
Race (1 =White) 0.6125** 0.2175 0.1860 1.0389 0.6536** 0.2214 0.2195 1.0878
Income
$0–24,999 (Reference) — — — —
$25,000–49,999 -0.0716 0.3282 -0.7151 0.5720 -0.0971 0.3321 -0.7484 0.5542
$50,000–74,999 -0.8622* 0.3641 -1.5762 -0.1483 -0.8476* 0.3682 -1.5695 -0.1257
$75,000–99,999 -0.6608{0.3805 -1.4069 0.0853 -0.6426{0.3848 -1.3972 0.1119
>$100,000 -0.9646* 0.4007 -1.7504 -0.1789 -0.9265* 0.4058 -1.7223 -0.1308
Occupation
Professional (Reference) — — — —
Blue-Collar -0.1217 0.2286 -0.5699 0.3266 -0.2913 0.2552 -0.7917 0.2092
Clerical -0.1195 0.1647 -0.4424 0.2033 -0.1689 0.1694 -0.5011 0.1632
(continued)
Table 7 Results of OLS and 2SLS Analyses of the Association between Leave Duration and Maternal
Childbirth-Related Symptoms (N=2,455) (continued )
Explanatory Variables b
(OLS)
SE
(OLS)
95% CI
(OLS)
b
(2SLS)
SE
(2SLS)
95% CI
(2SLS)
Marital Status
Single (Reference) — — — —
Married 0.1157 0.2842 -0.4416 0.6732 0.1816 0.2904 -0.3878 0.7510
Partnered -0.0964 0.2807 -0.6468 0.4540 -0.0251 0.2873 -0.5886 0.5383
Parity (1 =Multiparous) 0.2289 0.1405 -0.0466 0.5044 0.1738 0.1463 -0.1131 0.4607
Time Period
6 weeks (Reference) — — — —
12 weeks -1.9991*** 0.1863 -2.3644 -1.6339 -1.6349*** 0.2993 -2.2218 -1.0480
6 months -1.4525*** 0.1954 -1.8357 -1.0694 -0.9389* 0.3830 -1.6899 -0.1879
12 months -2.5177*** 0.2010 -2.9118 -2.1236 -1.9528*** 0.4141 -2.7649 -1.1407
Source: Authors’ calculations
Notes:b=unstandardized coefficient; SE =standard error;
(OLS)
=estimates generated using ordinary least squares regression;
(2SLS)
=estimates generated
using two-stage least squares regression; adjusted R-squared
(OLS)
=0.1338
{p<0.10; *p£0.05; **p£0.01; ***p£0.001
$100,000 or higher (vs. $0–24,999), younger age, and being nonwhite
(see table 7). Women with a college degree experienced more childbirth-
related symptoms (worse health) than those who did not have a college
degree. At each time period, time was associated with significantly decreased
maternal childbirth-related symptoms (better health) relative to the refer-
ence period at six weeks postpartum, with the highest decrease in symp-
toms at twelve months after childbirth (see table 7).
Discussion
Taking time off from work after childbirth may help mothers to recover
emotionally from pregnancy, labor, and delivery. This study found a sig-
nificant nonlinear relationship between the duration of leave from work
after childbirth and postpartum depressive symptoms. Additional days of
leave up to six months after childbirth decreased postpartum depres-
sive symptoms. Thus taking less than six months of leave (approximately
twenty-six weeks) may increase the risk of postpartum depression for
Figure 4 Employment Status (On Leave/Working Again) and Predicted
Values of Maternal Childbirth-Related Symptoms over Four Time Periods
in the First Year after Childbirth
Dagher, McGovern, and Dowd -Maternity Leave Policies 405
some working women. This finding of a beneficial effect of longer leave
duration in relation to postpartum depressive symptoms is consistent with
the findings of other studies that examined leave duration and postpartum
depressive symptoms (Chatterji and Markowitz 2005, 2012; Hyde et al.
1995; Klein et al. 1998). However, the present study adds to the literature by
showing a nonlinear relationship between leave duration and postpartum
depressive symptoms and specifying a maximum period, six months after
childbirth, when taking leave can protect against depressive symptoms.
Moreover, in this study postpartum depressive symptoms were measured
with the Edinburgh Postnatal Depression Scale, a more specific measure of
depressive symptoms experienced after childbirth than the more general
Center for Epidemiological Studies Depression Scale used in previous
studies (Chatterji and Markowitz 2005, 2012; Hyde et al. 1995; Klein et al.
1998).
The present study also found a marginally significant positive linear
association between leave duration and maternal physical health, espe-
cially during the first twelve weeks after childbirth. This finding of a
beneficial effect of longer leave duration in relation to postpartum physical
health is consistent with McGovern and colleagues’ (1997) findings in
relation to vitality and role function, except in the present study we did not
find a U-shaped association between leave duration and physical health.
Moreover, our findings are not consistent with a few other studies that
examined leave duration and physical health after childbirth and found no
associations (Chatterji and Markowitz 2005; Killien, Habermann, and
Jarrett 2001; Romito, Saurel-Cubizolles, and Cuttini 1994). However,
these studies did not use the SF-12 Physical Summary score to measure
physical health; instead they used a summary indicator of health status
(Killien, Habermann, and Jarrett 2001), proxy measures such as extreme
tiredness, backache, and lack of sleep (Romito, Saurel-Cubizolles, and
Cuttini 1994), and other health indicators such as number of outpatient
physician or clinic visits (Chatterji and Markowitz 2005).
Future studies should examine the mechanisms through which leave
duration after childbirth influences postpartum depressive symptoms, includ-
ing postpartum fatigue, a risk factor for postpartum depression (Corwin
et al. 2005). Moreover, additional studies should ascertain the relation-
ship between leave duration and maternal physical health and determine
which mechanisms mediate this association. Finally, future research should
investigate the impact of paternity leave on maternal mental and physical
health outcomes and find out whether it acts as a mediator or moderator
in the relationship between maternity leave duration and maternal health
outcomes.
406 Journal of Health Politics, Policy and Law
Limitations
The findings of this study should be interpreted in light of the study limi-
tations. Although study findings are internally valid, they can mainly be
generalized to employed women of similar racial and ethnic backgrounds,
income levels, and comparable health insurance coverage. The study
population for this study was limited to Minnesota mothers, aged eighteen
years and older. In comparison with available national data, Minnesota
mothers had a lower probability of being married (77 percent) than new
mothers aged twenty-five to twenty-nine years (83 percent) (Bachu and
O’Connell 2001), and a lower probability of giving birth by cesarean
section (17 percent vs. 24 percent, respectively) (Martin et al. 2002).
Moreover, Minnesota mothers had a higher probability of having com-
pleted high school (94 percent) in comparison with data on new mothers
aged twenty-five to thirty-four years nationwide (84 percent to 90 percent)
(National Center for Health Statistics 2005). In terms of racial composi-
tion, the study population contained similar proportions of white and
Native Americans as national data (78 percent and 1 percent, respectively)
but had relatively fewer African American mothers (9 percent vs. 15 per-
cent) and more Asian mothers (11 percent vs. 5 percent) (National Center
for Health Statistics 2005). Because of the demographic distribution of the
residents of the Twin Cities area, Minnesota, and the restriction of the study
sample to English-speaking women, the racial and ethnic representation of
study participants may underrepresent racial minorities and immigrant
populations relative to many urban areas in the United States. Future
research should include a more racially and ethnically diverse sample and
should be replicated in other states in the United States to assess the gen-
eralizability of study findings to other populations of women.
Postpartum depressive symptoms were measured by self-report and were
not validated by medical diagnoses or clinical evaluations. Thus inter-
pretation of the findings should be specific to postpartum depressive
symptoms and not to postpartum depression diagnosis. However, Mir-
owsky and Ross (1989: 12) recommended ‘‘eliminating diagnosis from
research on the nature, causes, and consequences of mental, emotional,
and behavioral problems. These authors consider diagnoses of mental
disorder ill suited to research on the impact of social and interpersonal
arrangements.
Other limitations pertain to the data. The present study had a minor
limitation related to the distribution of women’s responses by time at each
data collection period, which resulted in small inaccuracies regarding the
Dagher, McGovern, and Dowd -Maternity Leave Policies 407
precision of how we reported the time periods in this article. Women
generally completed their telephone interviews early in the windows of
potential response periods for each data collection period. For example, on
average, women completed the six-week interview at 4.8 weeks (SD =
0.9); the twelve-week interview at 11.2 weeks (SD =1.2); the six-month
interview at 23.5 weeks (SD =1.8); and the twelve-month interview at
50.1 weeks (SD =2.1). Thus the reporting of health outcomes at six and
twelve weeks and six and twelve months was slightly earlier than the
nominal period of data collection and should be noted by investigators who
wish to compare their results with ours. In addition, 87.3 percent (N=625)
of the women were back to work by six months postpartum, and 92.0
percent (N=659) returned by twelve months postpartum. Therefore, it is
difficult to get an accurate estimate of the impact on women’s postpartum
health of leave durations that are longer than six months. Finally, our
analytical sample excluded 101 cases that did not fill out the six-week
interview. However, there were no significant differences in the four health
outcomes between 30 percent of the excluded cases we had data on and
the six-week respondents at twelve weeks, six months, and twelve months
postpartum. The data on the excluded cases suggest that these women did
not participate in the six-week interview because of economic factors
rather than health concerns: the average excluded case was less educated,
had lower family income and shorter employment history with the current
employer, and was less likely to return to work during the first year after
childbirth than mothers retained in the study. It may be that the costs of
working (substitute infant care, transportation, etc.) were greater than the
costs of unemployment and staying home to care for their infant, but
additional research is needed to determine why women leave the work-
force in the year after childbirth.
Study Implications for US Leave Policy
The primary federal policy that provides support to working US mothers of
infants is the FMLA, which has been in effect since 1993. Under this act,
the employee is provided a maximum of twelve weeks of unpaid, job-
protected leave per year forgiving birth; taking care of a newborn, a newly
adopted child, or a foster child; or attending to an immediate family
member with a serious health condition or to the employee’s own serious
health condition. Our findings on leave duration and postpartum depres-
sive symptoms have implications for two key aspects of the FMLA: (1)
408 Journal of Health Politics, Policy and Law
the twelve-week duration benefit it provides and (2) the lack of a wage
replacement benefit.
The main policy debates surrounding the FMLA did not consider the
effects of leave duration on maternal health. The twelve-week leave
duration provided by this law was a result of years of compromise between
sponsors of the bill, including women’s organizations, and its opponents
from business groups. Stakeholders such as the US Chamber of Commerce
and the Economic Policy Council insisted that longer leaves would harm
businesses. Thus the discussions regarding duration of leave focused on
what businesses could afford rather than the health and productivity of
the employees benefiting from the law. Additionally, in an effort to make
the policy relevant to a larger proportion of the population— for equity
considerations as well as to garner the political support of a larger lobby-
ing body — the objectives of the FMLA were to address leave for a variety
of reasons (e.g., illness of a spouse, a child, or parent). Thus, the historical
context of providing leave as a benefit for women in association with
pregnancy and childbirth, as is the case in most European countries, was
diminished.
9
The findings of effects of leave duration on postpartum depressive
symptoms and physical health in this study provide additional evidence
consistent with other studies (Chatterji and Markowitz 2005, 2012; Gjer-
dingen and Chaloner 1994; Hyde et al. 1995; McGovern et al. 1997) that
future leave policy debates should consider maternal health implications.
The finding that taking additional leave days up to six months after
childbirth may help decrease postpartum depressive symptoms provides
a specific leave duration that proponents of longer leave durations can
present in their future testimonies on leave policies. The current leave
duration provided by the FMLA, twelve weeks, may not be sufficient for
mothers who are at risk for postpartum depression. Ideally, leave policies
should acknowledge health protection for postpartum women; however,
this special treatment of women may risk jeopardizing political support
that has usually existed when the policy is applicable to a larger population.
Thus an alternative approach that presents the advantages to postpar-
tum women’s health as well as to the welfare of other populations, such
as newborns, may be more politically feasible. For example, the World
Health Organization recommends exclusive breastfeeding of infants for
the infant’s first six months of life, suggesting another reason for a longer
9. For example, the 1992 directive issued by the European Union that requires member states
to provide fourteen weeks of paid maternity leaves was mainly based on health and safety
considerations pertaining to mothers and their newborns.
Dagher, McGovern, and Dowd -Maternity Leave Policies 409
period of leave (Eidelman and Schanler 2012). The health effects of
breastfeeding for the infant are well documented by the American Acad-
emy of Pediatrics in its Policy Statement on Breastfeeding and the Use
of Human Milk and include a decreased risk of respiratory, ear, and gas-
trointestinal tract infections; a decreased risk of clinical asthma, atopic
dermatitis, and eczema; and a reduced risk of Sudden Infant Death Syn-
drome (SIDS), in addition to health benefits for the mother (Eidelman
and Schanler 2012).
A related issue is the paid/unpaid nature of the leave law. Although the
FMLA provided twelve weeks of leave for eligible women, it did not
require employers to compensate women while they are on leave. The
unpaid leave benefit provided by the FMLA was a concession made by the
bill’s sponsors to ensure its survival in Congress, especially in the Reagan
era of ‘‘small government’’ and ‘‘neo-laissez-faire’’ politics (Elving 1995).
National survey data of covered establishments show that 52.4 percent of
employees who utilized FMLA leave in 2000 did so because of personal ill
health, and only 18.5 percent took leave to care for a newborn or a newly
adopted child (Cantor et al. 2001). Over half of the leaves taken for child
care were for ten days or less. The unpaid nature of the FMLA appears to
make it harder for parents to take longer periods of leave. In fact, the few
studies conducted on this issue have shown that unpaid leave policies
result in women taking short durations of leave (Hofferth and Curtin 2006;
McGovern et al. 2000). In this study, the total number of days of paid leave
provided by employer policy predicted the number of leave days women
took after childbirth.
10
More studies are needed to ascertain this associa-
tion, especially studies that compare duration of leave between women
with access to only unpaid leave policies and women with access to paid
leave policies and that include larger samples of women of color and low-
income women. This is important given the possibility that longer leave
durations may provide protective effects for new mothers who are at risk
for postpartum depression.
However, the politics of any policy change are challenging. The 2012
Republican presidential debates showed that a significant and vocal por-
tion of the public favors less regulation and increased state or local (ver-
sus federal) control over public policy decisions, in addition to the current
concern about the country’s national debt and a general resistance to tax
increases. The public sentiment suggests the need for advocates of leaves
10. This finding came about in the first-stage equation testing number of days of paid leave
provided by employer policy as an instrument for the actual number of days of leave taken by
women in this study.
410 Journal of Health Politics, Policy and Law
that are longer and paid to employ creative political strategies that favor
incremental change at the state or local level with financing mecha-
nisms that do not involve public taxes but could involve tax incentives for
employers and contributions from employees and employers. Broad coa-
litions motivated by a values-based ideology will be needed to develop a
legislative strategy and a readiness to seize opportune moments when the
political winds favor these issues.
Conclusion
This is the first longitudinal study to investigate the effects of leave duration
on women’s postpartum mental and physical health over the first twelve
months after childbirth. New mothers taking leaves from work less than six
months after childbirth appear to have an increased risk of postpartum
depressive symptoms. Longer leaves also hada protective effect on general
maternal physical health in the first twelve weeks after childbirth. While
the national political discourse favors less governmental regulation, and
policy implementation at state and local levels as opposed to the national
level, the study findings suggest the importance of leave as a protective
factor for women’s mental and physical health at a critical time in the life
course of women and their families. Given the twentieth anniversary of the
FMLA this year, our study findings can inform family leave discussions
among policy makers, employers, and families with the ultimate goal of
positively influencing the health of new mothers and their choices about
work and family.
nnn
Rada K. Dagher is assistant professor of health services administration at the
Universityof Maryland School ofPublic Health and a faculty associate at the Maryland
Population Research Center. She teaches courses on health care management and
health services research methods. Her research interests and areas of expertise
include maternal postpartum depression; work organization and employee health
and health expenditures; family leave policies and postpartum health; and gender,
racial, and ethnic disparities in mental health and health services use. Dagher has a
particular interest in studying government and work policies affecting maternal
health and women’s health more generally. She earned a PhD in health services
research, policy, and administration from the University of Minnesota and an MPH
from the American University of Beirut.
Dagher, McGovern, and Dowd -Maternity Leave Policies 411
Patricia M. McGovern is the Bond Professor of Environmental and Occupational
Health Policy and Deputy Director of the Midwest Center for Occupational Health
and Safety at the University of Minnesota School of Public Health. She teaches
courses on environmental and occupational health policy and on occupational and
environmental health nursing. Her research addresses women’s perinatal health in
association with employment policies, health services utilization, work and personal
factors, and children’s environmental health exposures in association with develop-
mental outcomes. She has received grants from the National Institutes of Health,
the National Institute for Occupational Safety and Health, and the Emma B. Howe
Foundation. McGovern earned an MPH and a PhD in health services research and
policy from the University of Minnesota.
Bryan E. Dowd is Mayo Professor in the Division of Health Policy and Management
(HPM), School of Public Health, at the University of Minnesota. Dowd’sresearch inter ests
include public and private health insurance and econometric methods. He is co-chair of
the Program in Human Rights and Health at the University of Minnesota, chair of the
Methods Council of AcademyHealth, and senior associate editor of the journal Health
Services Research. He teaches courses in advanced research methods, health policy
analysis, and writing for research in the HPM doctoral program. He has over 140 pub-
lications in refereed journals and three ‘‘Article of the Year’’ awards. He is a licensed
architect and holds a PhD in public policy analysis from the University of Pennsylvania.
References
Alphabetical Index of Industries and Occupations. 1992. 1990 Census of Population.
Washington, DC: US Department of Commerce, Bureau of the US Census.
American Psychiatric Association. 2000. Diagnostic and Statistical Manual of Mental
Disorders. 4th ed., Text Revision. Washington, DC: American Psychiatric Association.
Annie E. Casey Foundation. 2003. 2003 KIDS COUNT Resource Kit: Countering the
Costs of Being Poor. www.aecf.org/upload/publicationfiles/2003%20resource%
20kit.pdf.
Bachu, Amara, and Martin O’Connell. 2001. Fertility of American Women: June 2000.
Currrent Population Reports P20-543 revised. www.census.gov/prod/2001pubs
/p20-543rv.pdf.
Becker, Gary. 1965. ‘‘A Theory of the Allocation of Time.’’ Economic Journal 75:
493–517.
Beeghly, Marjorie, M. Katherine Weinberg, Karen L. Olson, Henrietta Kernan, Joan
Riley, and Edward Z. Tronick. 2002. ‘‘Stability and Change in Level of Maternal
Depressive Symptomatology during the First Postpartum Year.Journal of Affective
Disorders 71: 169–80.
Bianchi, Suzanne M. 2011. ‘‘Family Change and Time Allocation in American
Families.’’ Annals of the American Academy of Political and Social Science 638:
21–44.
412 Journal of Health Politics, Policy and Law
Brown, Stephanie, and Judith Lumley. 1998. ‘‘Changing Childbirth: Lessons from an
Australian Survey of 1,336 Women.’’ British Journal of Obstetrics and Gynecology
90: 1176–77.
Cantor, David, Jane Waldfogel, Jeffrey Kerwin, Mareena M. Wright, Kerry Levin,
John Rauch, Tracey Hagerty, and Martha S. Kudela. 2001. Balancing the Needs
of Families and Employers: Family and Medical Leave Surveys, 2000 Update.
Washington DC: US Department of Labor.
Chatterji, Pinka, and Sara Markowitz. 2005. ‘‘Does the Length of Maternity Leave
Affect Maternal Health?’Southern Economic Journal 72, no. 1: 16–41.
Chatterji, Pinka, and Sara Markowitz. 2012. ‘‘Family Leave after Childbirth and the
Mental Health of New Mothers.’Journal of Mental Health Policy and Economics
15: 61–76.
Clearinghouse on International Developments in Child, Youth and Family Policies
at Columbia University. 2002. Mother’s Day: More Than Candy and Flowers,
Working Parents Need Time Off. Issue Brief, Spring. New York: Clearinghouse.
www.childpolicyintl.org.
Cooper, Peter J., and Lynne Murray. 1998. ‘‘Fortnightly Review: Postnatal Depres-
sion.’British Medical Journal 316: 1884–86.
Corwin, Elizabeth J., Jean Brownstead, Nichole Barton, Starlet Heckard, and Karen
Morin. 2005. ‘‘The Impact of Fatigue on the Development of Postpartum Depres-
sion.’’ Journal of Obstetric, Gynecologic, and Neonatal Nursing 34: 577–86.
Council Directive. 1992. ‘‘Council Directive 92/85/EEC of 19 October 1992 Con-
cerning the Implementation of Measures to Encourage Improvements in the Safety
and Health at Work of Pregnant Workers and Workers Who Have Recently Given
Birth or Are Breast-Feeding.’Official Journal of the European Communities,L.
348, November 28.
Cox, John L., Jeni M. Holden, and Ruth Sagovsky. 1987. ‘‘Detection of Postnatal
Depression: Development of the 10-Item Edinburgh Postnatal Depression Scale.
British Journal of Psychiatry 150: 782–86.
Eidelman, Arthur I., and Richard J. Schanler. 2012. ‘‘American Academy of Pediatrics
Policy Statement on Breastfeeding and the Use of Human Milk.’Pediatrics 129,
no. 3: e827–e841.
Elison, Sonja K. 1997. ‘‘Policy Innovation in a Cold Climate: The Family and Medical
Leave Act of 1993.’Journal of Family Issues 18, no. 1: 30–55.
Elving, Ronald D. 1995. Conflict and Compromise: How Congress Makes the Law.
New York: Simon and Schuster.
Fergerson, Sarah S., Denise J. Jamieson, and Michael Lindsay. 2002. ‘‘Diagnosing
Postpartum Depression: Can We Do Better?’American Journal of Obstetrics and
Gynecology 186, no. 5: 899–902.
Frankenhaeuser, Marianne. 1986. ‘‘A Psychobiological Framework for Research on
Human Stress and Coping.’’ In Dynamic of Stress: Physiological, Psychological,
and Social Perspectives, edited by Mortimer H. Appley and Richard Trumbull,
101–16. New York: Plenum.
Galtry, Judith, and Paul Callister. 2005. ‘‘Assessing the Optimal Length of Parental
Leave for Child and Parental Well-Being: How Can Research Inform Policy?’
Journal of Family Issues 26, no. 2: 219–46.
Dagher, McGovern, and Dowd -Maternity Leave Policies 413
Gjerdingen, Dwenda K., and Kathryn M. Chaloner. 1994. ‘‘The Relationship of
Women’s Postpartum Mental Health to Employment, Childbirth, and Social Sup-
port.’Journal of Family Practice 38, no. 5: 465–72.
Gjerdingen, Dwenda, et al. 1993. ‘‘Changes in Women’s Physical Health during the
First Postpartum Year.’’ Archives of Family Medicine 2: 277–83.
Goodman, Janice. 2004. ‘‘Postpartum Depression beyond the Early Postpartum
Period.’Journal of Obstetric, Gynecologic, and Neonatal Nursing 33: 410–20.
Grossman, Michael. 1972. The Demand for Health: A Theoretical and Empirical
Investigation. New York: Columbia University Press.
Haas, Linda L. 2003. ‘‘Parental Leave and Gender Equality: Lessons from the Euro-
pean Union.’Review of Policy Research 20, no. 1: 89–114.
Han, Wen-Jui, and Jane Waldfogel. 2003. ‘‘Parental Leave: The Impact of Recent
Legislation on Parents’ Leave Taking. Demography 40, no. 1: 191–200.
Heymann, Jody, and Alison Earle. 2010. Raising the Global Floor: Dismantling the
Myth That We Can’t Afford Good Working Conditions for Everyone. Stanford:
Stanford University Press.
Hofferth, Sandra L., and Sally C. Curtin. 2006. ‘‘Parental Leave Statutes and Maternal
Return to Work after Childbirth in the United States.’’ Work and Occupations 33,
no. 1: 73–105.
Horowitz, June A., and Janice H. Goodman. 2004. ‘‘A Longitudinal Study of Maternal
Postpartum Depression Symptoms.’Research and Theory for Nursing Practice 18:
149–63.
Hyde, Janet S., Marjorie H. Klein, Marilyn J. Essex, and Roseanne Clark. 1995.
‘Maternity Leave and Women’s Mental Health.’Psychology of Women Quarterly
19: 257–85.
IWPR (Institute for Women’s Policy Research). 2004. The Status of Women in the
United States. Washington, DC: Author.
Kamerman, Sheila B. 2000. ‘‘From Maternity to Parental Leave Policies: Women’s
Health, Employment, and Child and Family Well-Being.’Journal of American
Medical Women’s Association 55, no. 2: 96–99.
Killien, Marcia G. 1998. ‘‘Postpartum Return to Work: Mothering Stress, Anxiety, and
Gratification.’Canadian Journal of Nursing Research 30, no. 3: 53–66.
Killien, Marcia G., Barbara Habermann, and Monica Jarrett. 2001. ‘‘Influence of
Employment Characteristics on Postpartum Mothers’ Health.’ Women and Health
33, nos. 1–2: 63–81.
Klein, Marjorie H., Janet S. Hyde, Marilyn J. Essex, and Roseanne Clark. 1998.
‘Maternity Leave, Role Quality, Work Involvement, and Maternal Health One
Year after Delivery.’’ Psychology of Women Quarterly 22: 239–66.
Laughlin, Lynda. 2011. Maternity Leave and Employment Patterns: 1961–2008.
Washington, DC: US Census Bureau.
Mahon, Rianne, and Deborah Brennan. 2012. ‘‘Federalism and the ‘New Politics’ of
Welfare Development: Childcare and Parental Leave in Australia and Canada.’
Publius: The Journal of Federalism (first published online April 13): 1–19.
doi:10.1093/publius/pjs015.
414 Journal of Health Politics, Policy and Law
Martin, Joyce A., Brady E. Hamilton, Stephanie J. Ventura, Fay Menacker, Melissa M.
Park, and Paul D. Sutton. 2002. ‘‘Births: Final Data for 2001.’National Vital
Statistics Report 51, no. 2. Hyattsville, MD: National Center for Health Statistics.
McGovern, Patricia M., Rada K. Dagher, Heidi Roeber Rice, Dwenda Gjerdingen,
Bryan Dowd, Laurie K. Ukestad, and Ulf Lundberg. 2011. ‘‘A Longitudinal Anal-
ysis of Total Workload and Women’s Health after Childbirth.’Journal of Occu-
pational and Environmental Medicine 53, no. 5: 497–505.
McGovern, Patricia M., Bryan Dowd, Dwenda Gjerdingen, Cynthia R. Gross, Sally
Kenney, Laurie Ukestad, David McCaffrey, and Ulf Lundberg. 2006. ‘‘Postpartum
Health of Employed Mothers Five Weeks after Childbirth.Annals of Family
Medicine 4, no. 2: 159–67.
McGovern, Patricia M., Bryan Dowd, Dwenda Gjerdingen, Ira Moscovice, Laura
Kochevar, and William Lohman. 1997. ‘‘Time Off Work and the Postpartum Health
of Employed Women.’’ Medical Care 35, no. 5: 507–21.
McGovern, Patricia M., Bryan Dowd, Dwenda Gjerdingen, Ira Moscovice, Laura
Kochevar, and Sarah Murphy. 2000. ‘‘The Determinants of Time Off Work after
Childbirth.’Journal of Health Politics, Policy and Law 25, no. 3: 527–64.
McGovern, Patricia M., and M. Segal. 1987. ‘‘Minnesota Parental Leave Weighs
Social Reform against Costs.’Business and Health 5, no. 2: 42–43.
Minnesota Department of Children, Families, and Learning. 1999. At Home Infant
Care Program Fact Sheet. Roseville, MN: Minnesota Department of Children,
Families, and Learning.
Mirowsky, John, and Catherine E. Ross. 1989. ‘‘Psychiatric Diagnosis as Reified
Measurement.’Journal of Health and Social Behavior 30, no. 1: 11–25.
Moen, Phyllis, and Robert M. Orrange. 2002. ‘‘Careers and Lives: Socialization,
Structural Lag, and Gendered Ambivalence.’’ In Advances in Life-Course Research:
New Frontiers in Socialization, edited by Richard A. Settersten Jr. and Timothy J.
Owens, 231–60. London: Elsevier Science.
Moen, Phyllis, and Elaine Wethington. 1992. ‘‘The Concept of Family Adaptive
Strategies.’Annual Review of Sociology 18: 233–51.
Naples, Michele I. 2001. Family Leave for Low-Income Working Women: Providing
Paid Leave through Temporary Disability Insurance—the New Jersey Case.Pub-
lication A128. Washington, DC: Institute for Women’s Policy Research.
National Center for Health Statistics. 2005. Birth, Percentages of Births by Race,
2000–2002. PeriStats. www.marchofdimes.com/peristats/Peristats.aspx.
National Partnership for Women and Families. 2005. ‘‘At-Home Infant Care (AHIC):
A Side-by-Side Comparison of Federal and State Initiatives.’’ www.national
partnership.org/site/DocServer/AHICchartOct05.pdf?docID =1048 (accessed October
8, 2013).
O’Hara, Michael W., and Annette M. Swain. 1996. ‘‘Rates and Risk of Postpartum
Depression: A Meta-analysis.’International Review of Psychiatry 8: 37–54.
Ray, Rebecca, Janet C. Gornick, and John Schmitt. 2010. ‘‘Who Cares? Assess-
ing Generosity and Gender Equality in Parental Leave Policy Designs in 21
Countries.’Journal of European Social Policy 20, no. 3: 196–216.
Dagher, McGovern, and Dowd -Maternity Leave Policies 415
Ries, Peter. 1991. ‘‘Educational Differences in Health Status and Health Care.’Vital
and Health Statistics 10, no. 179: 1–66.
Romito, Patrizia, and Marie-Josephe Saurel-Cubizolles, and Marina Cuttini. 1994.
‘Mother’s Health after the Birth of the First Child: The Case of Employed Women
in an Italian City.’’ Women and Health 21, nos. 2–3: 1–22.
Schroeder, Patricia. 1989. Champion of the Great American Family. New York:
Random House.
Stewart, Anita L., Ron D. Hays, and John E. Ware. 1988. ‘‘The MOS Short-Form
General Health Survey.’’ Medical Care 26: 724–35.
Stowe, Zachary N., Amy L. Hostetter, and Jeffrey D. Newport. 2005. ‘The Onset of
Postpartum Depression: Implications for Clinical Screening in Obstetrical and
Primary Care.’American Journal of Obstetrics and Gynecology 192, no. 2: 522–
26.
Stuart, Scott, Greg Couser, Kelly Schilder, Michael O’Hara, and Lori Gorman. 1998.
‘Postpartum Anxiety and Depression: Onset and Comorbidity in a Community
Sample.’Journal of Nervous and Mental Disease 186: 420–24.
Sundbye, Annamaria, and Ariane Hegewisch. 2011. Maternity, Paternity, and
Adoption Leave in the United States. Publication A143. Washington, DC: Institute
for Women’s Policy Research.
Tinker, Catherine. 1981. ‘‘Human Rights for Women: The U.N. Convention on the
Elimination of All Forms of Discrimination against Women.’’ Human Rights
Quarterly 3, no. 2: 32–43.
Toossi, Mitra. 2012. ‘‘Labor Force Projections to 2020: A More Slowly Growing
Workforce.’’ Monthly Labor Review Online 135, no. 1: 43–64.
US Bureau of Labor Statistics. 2012. Employment Characteristics of Families 2010.
Washington, DC: US Bureau of Labor Statistics. www.bls.gov/news.release/famee
.nr0.htm.
US Census Bureau. 2001. Overview of Race and Hispanic Origin Census 2000
(C2KBR/01-1). Washington, DC: US Department of Commerce.
US Department of Commerce, Census Bureau. 2001. Statistical Abstract for the
United States. Washington, DC: US Government Printing Office.
US Department of Labor, Women’s Bureau. 2005. Employment Status of Women and
Men in 2005. Washington, DC: US Government Printing Office. www.dol.gov/wb
/factsheets/Qf-ESWM05.htm.
Ware, John E., Mark Kosinski, Diane M. Turner-Bowker, and Barbara Gandek. 2002.
How to Score Version 2 of the SF-12 Health Survey. Lincoln, RI: QualityMetric.
Wisensale, Steven K. 2003. ‘‘Two Steps Forward, One Step Back: The Family and
Medical Leave Act as Retrenchment Policy.’’ Review of Policy Research 20, no. 1:
135–52.
Wisner, Katherine L., Barbara L. Parry, and Catherine M. Piontek. 2002. ‘‘Postpartum
Depression.’New England Journal of Medicine 347, no. 3: 194–99.
416 Journal of Health Politics, Policy and Law
... It has been suggested that family, reproductive health, employment, and political participation policies that maintain gender inequalities worsen health outcomes for women through mechanisms such as stress, discrimination, violence, financial difficulties, poverty, double burden of work, and time pressure (Borrell et al. 2013). For instance, short and unpaid maternity leave have been found to have negative consequences for women's physical and mental health (Dagher, McGovern, and Dowd 2014). Hence, it could be assumed that women living in countries with a higher degree of gender inequality are more constrained in their preventive health behavior. ...
Article
Full-text available
The current study examines whether the extent of macrolevel gender inequality affects the association between women’s educational attainment and their participation in cervical and breast cancer screening and how this relationship is moderated by a country’s cancer screening strategy (organized vs. opportunistic). A multilevel design with women (N cervical = 99,794; N breast = 55,021) nested in 30 European countries was used to analyze data from the European Health Interview Survey (2013–2015). Results of multilevel logistic regression models demonstrate that higher macrolevel gender inequality is associated with (a) a lower overall likelihood that women have had a mammography and Pap smear and (b) a larger gap in participation between women with low and high levels of education, regardless of a country’s screening strategy (i.e., no moderation by a country’s screening strategy was found). We conclude that macrolevel gender stratification should not be neglected when designing cancer screening policy.
... When paid family leave is not available, it can contribute to the social and economic costs to the family as well as the society. When mothers return to work less than 12 weeks after giving birth, they are more likely to report stress and depression (Dagher, McGovern, & Dowd, 2014). They experience difficulty in handling the demands of work and family life, and, as a result, many choose to leave their job. ...
Article
During the past four decades, most OECD (Organisation for Economic Co-operation and Development) countries have adopted or expanded paid family leave, which offers leave to workers following the birth or adoption of a child as well as care for ill family members. While the effects of paid maternity leave on child health have been the subject of a large body of research, little is known about fathers’ leave-taking and the effects of paid paternity leave. This is a limitation, since most of the recent expansion in paid family leave in OECD countries has been to expand leave benefits to fathers. Mothers’ and fathers’ leave-taking may improve child health by decreasing postpartum depression among mothers, improving maternal mental health, increasing the time spent with a child, and increasing the likelihood of child medical checkup. The purpose of this paper is to examine the effects of paid family leave on the wellbeing of children, extending what we know about the effects of maternity leave and establishing new evidence on paternity leave. The paper examines the effects of paid family leave expansions on country-level neonatal mortality rates, infant mortality rates, under-five mortality rates, and the measles immunization rates in 35 OECD countries, during the time period of 1990 to 2016. Using an event study design, an approximately 1.9 to 5.2 percent decrease in the infant, neonatal, and under-five mortality rates has been found following the adoption of paid maternity leave. However, the beneficial impact is not as visible for extension of paid leave to fathers. The implications and potential reasons behind the larger protective effects of maternity leave over paternity leave on child health outcomes are discussed.
... We refrain from further contrasting U.S. and Western European studies of postnatal MDS trajectories because of stark differences between these nations in a mother's social ecology after childbirth. The United States is the only industrialized country without a national policy for paid maternal leave, with most mothers restricted to no more than 12 weeks of unpaid leave after giving birth (Ruhm, 2017), despite evidence of an inverse relation between length of maternity leave and MDS (Andres, Baird, Bingenheimer, & Markus, 2016) up to 6 months postpartum (Dagher, McGovern, & Dowd, 2014). In France (Sutter-Dallay et al., 2012;van der Waerden et al., 2015b), Finland (Luoma et al., 2015), and the Netherlands (Cents et al., 2012), mothers are entitled to paid maternal leave, extended periods of unpaid maternal leave, and/or childcare subsidies after giving birth (Ruhm, 2017). ...
Article
Infants are uniquely vulnerable to maternal depression's noxious effects, but few longitudinal studies have tried to identify discrete postnatal trajectories of maternal depressive symptoms (MDS) beginning in infancy. This study extends evidence of heterogeneous change in postnatal MDS by examining their cross-contextual antecedents in infancy and their consequences for children's early behavior problems and language skills in late toddlerhood. A community sample of mother-child dyads (N = 235, 72% Caucasian) was assessed when children were 7, 15, and 33 months old. Mothers reported their socioeconomic status (SES), social support, marital relationship quality, family dysfunction, parenting stress, and infants' functional regulatory problems at 7 months postpartum, and children's internalizing and externalizing symptoms at 33 months. Children completed a receptive vocabulary assessment at 33 months in the lab. Latent class growth analysis identified three postnatal MDS trajectory classes that fit the data best: low-decreasing, moderate, and increasing. Psychosocial measures at seven months postpartum primarily predicted membership to these postnatal trajectory classes, which subsequently differed in children's internalizing, externalizing, and receptive vocabulary in late toddlerhood, controlling for family SES and functional regulatory problems in infancy. We discuss salient antecedents and consequences of postnatal depression for mothers and their offspring.
... In accordance with the Bureau of Labor Statistics, only 13% of workers employed by private companies have access to paid family leave in the United States [30]. The provision of up to twelve weeks of unpaid medical leave and job protection as defined and entitled by the Family Medical Leave Act has remained essentially unchanged since its inception in 1993 [31] and is not accessible to all employed persons in the United States. Although leave and job protection are provided for up to twelve weeks for American workers meeting minimum length of employment, hours worked, and company size requirements, existing research indicates that twelve weeks may not be a sufficient length of time for women who have given birth to rest, recover [32], and adjust nor affordable for many women. ...
Article
Full-text available
Background Childbirth represents a significant transition for women, with physical and psychological sequelae. Reentry to the workplace during the postpartum period is understudied, with implications for maternal well-being and job-related outcomes. This study's aim was to examine selected pregnancy, childbirth, and return-to-work correlates of overall self-rated health within the first month of work reentry after maternity leave. Methods Between December 2016 and January 2017, we surveyed women employed at a large, public Midwestern university who had given birth in the past five years (N = 249) to examine self-rated overall health in the first month of work reentry. Using ordinal logistic regression, we examined whether physical or psychological health problems during pregnancy, childbirth complications, length of maternity leave, and depression and anxiety at work reentry were related to overall health. Results Women who experienced depression (odds ratio [OR] = 0.096 [95% confidence interval {CI} = 0.019 to 0.483, p = 0.004]) and anxiety (OR = 0.164, [95% CI = 0.042 to 0.635, p = 0.009]) nearly every day reported worse health at work reentry than those with no symptoms. Controlling for demographics and mental health, women who experienced medical problems during pregnancy (OR = 0.540 [95% CI = 0.311 to .935, p = 0.028]) were more likely to report poor health, while taking a longer maternity leave (OR = 14.552 [95% CI = 4.934 to 42.918, p < 0.001]) was associated with reporting better health at work reentry. Conclusion Women who experience medical complications during pregnancy, return to the workplace too soon after birth, and experience mental health symptoms are vulnerable physically as they return to work.
... 24 25 While depression during pregnancy is difficult to alleviate, 26 it has been reported that postpartum depression resolves until 6 months after delivery then relapses. 27 The difference in the clinical course of depression in pregnant women may affect follow-up rate to the questionnaires for 6-month olds and 1-year olds. ...
Article
Full-text available
Objectives We examined the factors related to lost-to-follow-up of a birth cohort study during the first year after delivery. Design Longitudinal cohort study. Setting Questionnaires were provided by mail. Mothers answered the questionnaires about the children twice: at 6 months and 1 year. Participants Of 103 062 pregnancies who consented to participate in the Japan Environment and Children’s Study (JECS), 93 417 mothers were included in the study after excluding those with multiple births, miscarriages or stillbirths and those who withdrew from the study within 1 year after providing informed consent. Primary and secondary outcome measures Participants’ socioeconomic status, medical history, health status, health-related behaviours, their children’s health conditions and living situations were collected by self-administered questionnaires during pregnancy or 1 month after delivery as the baseline survey. In addition, two self-administered questionnaires were distributed 6 months and 1 year after delivery. Using the response status of the two questionnaires after delivery, participants’ follow-up status was divided into four groups. The related factors were examined using logistic regression analysis. Results Factors positively correlated with lost-to-follow-up to the questionnaires were postpartum physical conditions, psychological distress during pregnancy, the child’s health status at birth, the child’s primary caregiver and the number of siblings of the child. Partners’ active participation in JECS was associated with a lower lost-to-follow-up rate to the two questionnaires, whereas inactive participation was positively associated with a higher lost-to-follow-up rate. Conclusion The response rate to the questionnaires seems to be related to the interest and understanding of participants’ partners. In addition, the response rates are related to participants’ physical conditions and living conditions. To decrease lost-to-follow-up rates in consecutive questionnaire surveys within a cohort study, it may be important for investigators to recognise that participants and their motivation in research can be influenced by perceptions they may have regarding the objectives of the research.
Article
We examined whether using paid leave is directly linked to employed mothers’ depressive symptoms. In addition, we examined the moderating effect of employed mothers’ perceived fairness of division of household labor (housework and childcare) on the above association. We collected data from 92 employed mothers who were eligible to take paid leave and returned to the workplace after childbirth in Upstate New York State area. Using hierarchical regression analyses, we found that employed mothers’ paid leave use was not significantly associated with their levels of depressive symptoms. However, perceived fairness of division of household labor moderated the above association. Specifically, when employed mothers took paid leave, they reported lower levels of depressive symptoms than those who did not take leave when they perceived that the division of household labor was fair to them.
Article
Introduction In January 2016, the U.S. Department of Defense increased the duration of paid maternity leave for active duty service members from 6 weeks to 12 weeks. Our study aims to determine the impact of maternity leave length on breastfeeding duration and postpartum depression rates in active duty service members. Materials and Methods An institutional review board-approved survey of 9 questions was given to patients at the Brooke Army Medical Center Obstetrics and Gynecology Clinic. Patients were offered voluntary participation if they had delivered a baby while on active duty in the preceding 20 years. The survey included questions about length of maternity leave, duration of breastfeeding, age at delivery, route of delivery, and whether the patient suffered from postpartum depression. Results A total of 214 surveys were collected. Fisher exact test was used to compare rates of breastfeeding and postpartum depression between the 2 groups. A total of 87% of service members initiated breastfeeding. Among women who had 6 weeks versus 12 weeks of maternity leave, 51.64% versus 56.96% breastfed up to 6 months, p = 0.45. The overall rate of postpartum depression was 13.5%. Among women who had 6 weeks versus 12 weeks of maternity leave, 16.1% versus 9.5% reported postpartum depression, p = 0.11. Conclusion The ideal maternity leave duration is unknown. With recent changes to the Department of Defense maternity leave policy, we aimed to evaluate the effect this had on breastfeeding and postpartum depression rates. No statistically significant difference was seen when we compared rates of breastfeeding in women who had 6 weeks versus 12 weeks of maternity leave. Further research is required to determine the ideal maternity leave duration and best practices to promote breastfeeding. When looking at postpartum depression, our study shows that postpartum depression was noted in 16% of patients who took 6 weeks versus 9% of those who took 12 weeks of maternity leave. No statistically significant difference was seen; however, this was likely because of the small sample size. Only 29 out of 214 women suffered from depression regardless of length of maternity leave. More research is needed to determine if maternity leave length does indeed impact postpartum depression rates.
Article
Purpose To explore the current status and determine the uniformity of parental leave policies among US radiology residency programs. Methods An electronic survey was developed and sent to 222 radiology residency program directors (PDs) in June 2019 to assess their policies and attitudes toward parental leave. The survey was administered via the Internet Qualtrics Research Suite (Qualtrics, Provo, Utah) format with four reminders sent over the course of 2 months before closing the data collection. Results In all, 74 PDs responded to the survey. Of those, 88% claimed to have a maternal leave policy (88% explicitly written and 77% paid); 80% had a paternal leave policy (88% explicitly written and 75% paid). The average length of maternal and paternal leaves was 7.4 ± 3.9 and 3.7 ± 3.7 weeks, respectively. Parental leaves were allocated at least every other year in over 70% of programs. Approximately 60% of the PDs required residents to make up call rotations for parental leaves. About 90% of responsive programs adjusted angiography and fluoroscopy rotations for trainees throughout the pregnancy or according to the trainee’s request. Policies did not generally address issues of breastfeeding and nontraditional parenthood. Conclusion The proportion of radiology programs with explicit maternal leave policies remained constant compared with previous surveys during the last two decades. However, there was a dramatic rise in the adoption of paternal leave policies. Overall, there was a lack of national uniformity in radiology residency programs’ policies concerning parental leave, leaving open the possibility of national guidance in addressing the issue.
Article
The U.S. is the only high-income country without a national paid family leave (PFL) policy. While a handful of U.S. states have implemented PFL policies in recent years, there are few studies that examine the effects of these policies on health. In this study, we tested the hypothesis that California's PFL policy—implemented in 2004—improved parent health outcomes. Data were drawn from the 1993–2017 waves of the Panel Study of Income Dynamics, a large diverse national cohort study of U.S. families (N = 6690). We used detailed longitudinal sociodemographic information about study participants and a quasi-experimental difference-in-differences analytic technique to examine the effects of California's PFL policy on families who were likely eligible for the paid leave, while accounting for underlying trends in these outcomes among states that did not implement PFL policies in this period. Outcomes included self-rated health, psychological distress, overweight and obesity, and alcohol use. We found improvements in self-rated health and psychological distress, as well as decreased likelihood of being overweight and reduced alcohol consumption. Improvements in health status and psychological distress were greater for mothers, and reductions in alcohol use were greater for fathers. Results were robust to alternative specifications. These findings suggest that California's PFL policy had positive impacts on several health outcomes, providing timely evidence to inform ongoing policy discussions at the federal and state levels. Future studies should examine the effects of more recently implemented state and local PFL policies to determine whether variation in policy implementation and generosity affects outcomes.
Article
Full-text available
How does federal state architecture affect the design of welfare? We explore the development of childcare and parental leave in Canada and Australia to address this question. Both countries are considered liberal welfare regimes, but their federal institutions operate in quite different ways, providing an opportunity for comparative analysis. We consider the ways in which federal institutions have affected mobilization around childcare and parental leave and have helped to shape policy outcomes. There is a plausible connection between the institutional configuration of each federation and policy design. It is not definitive, but interacts with variables such as the nature and scale of federal fiscal and policy capacity, the gendered assumptions embedded in the structures of the welfare state, political party strategies, and feminist mobilization.
Article
The projected labor force growth over the next 10 years will be affected by the aging of the baby-boom generation; as a result, the labor force is projected to grow at a slower rate than in the last several decades
Article
Breastfeeding and human milk are the normative standards for infant feeding and nutrition. Given the documented short-and long-term medical and neurodevelopmental advantages of breastfeeding, infant nutrition should be considered a public health issue and not only a lifestyle choice. The American Academy of Pediatrics reaffirms its recommendation of exclusive breastfeeding for about 6 months, followed by continued breastfeeding as complementary foods are introduced, with continuation of breastfeeding for 1 year or longer as mutually desired by mother and infant. Medical contraindications to breastfeeding are rare. Infant growth should be monitored with the World Health Organization (WHO) Growth Curve Standards to avoid mislabeling infants as underweight or failing to thrive. Hospital routines to encourage and support the initiation and sustaining of exclusive breastfeeding should be based on the American Academy of Pediatrics-endorsed WHO/UNICEF "Ten Steps to Successful Breastfeeding." National strategies supported by the US Surgeon General's Call to Action, the Centers for Disease Control and Prevention, and The Joint Commission are involved to facilitate breastfeeding practices in US hospitals and communities. Pediatricians play a critical role in their practices and communities as advocates of breastfeeding and thus should be knowledgeable about the health risks of not breastfeeding, the economic benefits to society of breastfeeding, and the techniques for managing and supporting the breastfeeding dyad. The "Business Case for Breastfeeding" details how mothers can maintain lactation in the workplace and the benefits to employers who facilitate this practice. Pediatrics 2012; 129:e827-e841
Article
Thesis (Ph. D.)--Columbia University, 1970. Includes bibliographical references (leaves 167-170). Photocopy.
Article
Delayed marriage and childbearing, more births outside marriage, the increase in women’s labor force participation, and the aging of the population have altered family life and created new challenges for those with caregiving demands. U.S. mothers have shed hours of housework but not the hours they devote to childrearing. Fathers have increased the time they spend on childcare. Intensive childrearing practices combine with more dual-earning and single parenting to increase the time demands on parents. Mothers continue to scale back paid work to meet childrearing demands. They also give up leisure time and report that they “are always rushed” and are “multitasking most of the time.” Time-stretched working couples reduce the time they spend with each other. A large percentage of both husbands and wives also report they have “too little time” for themselves. Delayed childbearing and the aging population also increase the likelihood that both (adult) children and elderly parents need support and care from workers later in life.