Published by Michael Silverman
Department of Psychiatry, Icahn School of Medicine at Mount Sinai
These findings are described in the article entitled Is depression more likely following childbirth? A population-based study, recently published in the journal Archive of Women’s Mental Health (Archive of Women’s Mental Health 21 (2018) 1-6). This work was conducted by Michael E. Silverman and Abraham Reichenberg from the Icahn School of Medicine at Mount Sinai, Paul Lichtenstein from the Karolinska Institutet, and Sven Sandin from the Icahn School of Medicine at Mount Sinai and Karolinska Institutet.
The period immediately following the birth of a child is a critical time for a number of important early developmental events and the lack of energy and ability of a mother to cope effectively with the demands of this period can constitute a serious threat to the infant’s well-being. While major depressive episodes result in the well-known attendant morbidity, including marked functional impairment, distress and increased risk of suicide, early maternal depression also puts a second individual, the child, at significant risk during what many believe to be the most critical of developmental periods.
Early maternal depression is also associated with diminished enrichment behaviors, shortened durations of breastfeeding, as well as the renewal of maternal smoking. Interactions between depressed mothers and their children are known to be of lower quality than those of non-depressed mothers which have been shown to adversely affect physical growth and neurobehavioral development. Other research links maternal depression to the increased risk of conduct disorders and psychiatric disturbances among children, as well as greater child insecurity and anxiety- diagnoses that have been shown to persist into adulthood.
Historically, physicians have noted depressive symptoms following childbirth since the time of Hippocrates, sometime around 300BC. More recently, Hollywood celebrities including Brook Shields, Gwyneth Paltrow, Alanis Morissette, Chrissy Teigen, and Adele, have each offered their voice towards reducing the stigma associated with postpartum depression.
The recent attention in both the professional and lay literature has resulted in the folk assumption that childbirth may independently, and regularly, cause depression. However, from a scientific standpoint, an important question that remains to be answered is whether the postpartum period actually represents a time of greater vulnerability for depressive episodes.
To address this question, an international group of researchers conducted the largest and most rigorous study of postpartum depression to date. Some of these findings were recently published in the peer-reviewed journal, Archives of Women’s Mental Health.
Utilizing the entire national population of first births in Sweden between the years 1997 and 2008 and accounting for more than 700,000 individual women, findings suggest that the postpartum period is not a time of increased vulnerability for depression for most women. To conduct the study, researchers first needed to determine the overall risk of developing depression in the first year following child delivery based on each woman’s personal medical record. The result of this analysis provided a woman’s overall risk of developing postpartum depression. The next step was to assign each woman in the study with a computer-generated randomized date of child delivery. This is referred to as a phantom delivery date because it is not in any way related to the actual birth of a child. Most importantly, this phantom delivery date had to be selected from a time period in which each woman would normally be expected to deliver a child. Once the phantom delivery date was assigned, researchers determined the woman’s risk of developing depression, again based on an actual depression diagnosis, for the full calendar year following that random date. Surprisingly, the risk for postpartum depression was markedly lower than the risk of developing depression at some other point during her reproductive lifetime.
So why do these recent findings differ from what is commonly reported to be the most common complication of child delivery?
1) The vast number of past studies have explored postpartum depression symptomatology, whereas this study relied on clinically relevant postpartum depression diagnoses.
The difference between feeling depressed and requiring treatment is not an insignificant issue. For example, according to the diagnostic manual used by clinicians (DSM-5) who diagnose depression, a woman must have 5 of 9 symptoms consistently for two weeks to receive the diagnostic specifier associated with postpartum depression. Importantly, many of these same symptoms overlap with the common non-pathologic discomforts associated with childbirth – tearfulness, sleep difficulty, irritability, appetite and weight change, fatigue, worry, feelings of shame or guilt, indecisiveness, and difficulty concentrating. These intermittent symptoms normal to the early maternal experience are precisely why the accurate diagnosis of postpartum depression necessitates a two-stage process – the assessment of depression symptoms followed by a structured diagnostic interview. Unfortunately, for logistical reasons, both administrative and financial, the vast majority of research exploring postpartum depression has relied solely on self-reported symptom inventories such as the Edinburgh Postnatal Depression Scale (EPDS), which as a single stage process sans the prerequisite clinical assessment, can result in a gross overestimation of prevalence. This is something even the authors of the EPDS have pointed out.
2) Most studies of postpartum depression have been conducted on small convenience samples that come from local clinics or regional hospitals and as such cannot represent every woman within a population.
The current study, which represents the largest most rigorous study of clinically relevant postpartum depression to date used a nationally inclusive dataset that included every woman who delivered their first child in the county of Sweden between January 1, 1997, and December 31, 2008. Only a handful of similar methodologically rigorous, population-based studies of postpartum depression using a two-stage diagnostic process like the current study exist. These studies, which used medical diagnoses to determine the rates of clinically significant postpartum depression in the United States, Sweden, Denmark, Finland and Great Brittan, repeatedly point to considerably lower prevalence rates than what is commonly reported. Unfortunately, because stories of lower risk do not capture attention, these studies generally go unnoticed by the public and are frequently orphaned within the professional literature.
3) The idea of depression covarying with childbirth may actually represent an illusory correlation.
There is a difference between correlation and causation and the well-established observation of depression following childbirth does not necessarily equate to causation. Indeed the history of medicine is littered with erroneous relationships such as vaccinations causing autism and full moons causing lunacy. Our tendency to perceive two events as causally related, when their relationship is coincidental (or even non-existent) is a form of cognitive bias known as illusory correlation. The overutilization of screenings with high sensitivity and poor specificity has the potential to produce illusory correlations– a situation when clinicians may believe events or risk factors, such as depression and childbirth, are related even when they are not. Underscoring this possibly is the fact that pregnant and postpartum women represent a medically captured population, and are therefore regularly available to be screened for depressive symptomatology. It is, therefore, possible that the contemporary characterization of postpartum depression, as a common complication associated with childbirth, confuses the understanding of trigger and cause – simply because that is when they are available for screening.
Supporting this possibility is the finding that the risk postpartum depression is almost entirely explained by a woman’s history of depression. In a previous study published in the peer-reviewed journal Depression and Anxiety, using the same sample of women in Sweden, the overall risk for postpartum depression was 62 in 10,000. However, if a woman had a prior history of depression her risk for postpartum depression rose dramatically to 1,154 in 10,000 – a 21-fold increase in risk.
Notably, the rate of postpartum depression is often stated to be about 1 in 8, but sometimes as high as 1 in 4. While these rates have garnered considerable attention in both the professional and lay literature, empirical evidence suggest that these rates are misrepresented, clinically unreliable and potentially dangerous to continue promoting. Without question, they have led some to believe that the postpartum period is a time of unique vulnerability for mood disturbance.
Because postpartum depression impacts both the mother and her child, it is a public health concern. As such, the increased awareness of early maternal depression and the subsequent destigmatization of a very real psychological illness is welcomed progress. Notably, the increased attention has also led to initiatives directed at enhancing postpartum depression detection through universal screening programs. However, despite considerable evidence suggesting that some women are susceptible to intense mood changes associated with the unprecedented hormonal fluctuations that occur during and immediately after pregnancy, pathologizing what for most women may represent the normal discomforts of a healthy adjustment period can be equally deleterious. By continuing to misrepresent a woman’s actual risk, saddling ill-equipped clinicians with responsibilities outside their expertise, and implying symptom screening tools alone adequately diagnose clinical depression, increases the likelihood of misdiagnoses, over-treatment and the unwarranted medicalization of an already vulnerable population.
Dr. Silverman’s editorial, Childbirth and Depression: Is Postpartum Depression an Illusory Correlation?, sparked a lively, critical response from the Marcé Society of North America, an organization that promotes research and mental health services to support pregnant and post partum women. Citing no difference in diagnosed depression in 700.000+ Swedish women who have or have not given birth within 12 months, the editorial impugns the reality of post partum mental illness. The author trumpets as new the long recognized fact that maternal depression extends before and after the post partum period. He draws the baseless conclusion that post partum depression may thus be “illusory.” Such “mansplaining,” using massaged and highly abstracted facts to downplay women’s lived experience, offends women who have experienced crippling levels of post-partum distress, the researchers who have studied it, and the many agencies who provide support and care.
Members of the perinatal research and advocacy community have long recognized that post partum depression overlaps with depression in other contexts. A U.S. study that screened 10,000 women between 4 and 6 weeks postpartum and thoroughly evaluated 800+ screen positive cases for diagnosis found that episodes of post partum onset were modestly higher (40.1%) than those that began in pregnancy (33.%) or before pregnancy (26.3%). The rates of non-puerperal episodes, combined, are consistent with the Swedish study’s main findings of no differences in depression between a year post partum and periods unrelated to childbirth. The legitimate interpretation of these findings is that depression identified in the post partum period is heterogeneous, not that it is “illusory.”
Moreover, the twelve month period chosen to define post partum vs non postpartum episodes in the Swedish study is too long to isolate the impact of the drastic hormonal changes and circadian rhythm disturbances that characterize the first four to six weeks after birth. Key neural systems implicated in mood are targets of sex steroids. Estrogen receptors in the brain are concentrated in the limbic system. Pregnancy exposes these receptors to very high estrogen levels, which abruptly decline with delivery of the placenta. An elegant study demonstrates that women with perinatal depression are differentially sensitive to the mood destabilizing effects of neuronal withdrawal from gonadal steroids. The hormonal changes of giving birth are universal, but Sweden’s enlightened policies may buffer this instability during a critical time, reducing the rates of development of new diagnosable disorder to a level undetectable by the study’s methods.
Observations from one country cannot be extrapolated to the whole world. As Carly Snyder, a contributor to the Huffington Post notes, Sweden has the world’s lowest rate of post partum depression. Sweden provides women with free pre- and post- natal care, visits by midwives, generous employment benefits and mandated leave for all new parents. Parental leave bolsters maternal-infant attachment, improves health for the maternal-infant pair and supports breastfeeding. One cannot generalize from the Swedish data to the United States, the only industrialized nation with no federal law supporting paid parental leave.
Of greatest concern, the editorial speculates that considering post partum depression a discreet type of psychopathology “medicalizes” (and therefore stigmatizes) a normal behavioral adaptation, at a disservice to women. But directing medical and psychiatric attention to the post partum period has confirmed the critical effects of unacknowledged and untreated maternal mental illness on the fetus or developing child. Research based in the assumption that maternal depression during pregnancy or post partum is worthy of study has advanced our knowledge of how hormones interact with neurotransmitters and influence behavior. “Medicalization” has prompted countries like England, Norway, and Australia, and states as diverse politically as Massachusetts, Minnesota, New Jersey, Illinois and Texas, to promote or mandate screening for depression during pre- and post- natal care, to educate the public about a variety of mental illnesses and mental distress related to pregnancy, and create task forces to study the issue and develop new service. The USPSTF and the American Psychiatric Association have recently become advocates for attending to perinatal mental health and mental illness.
Big data and the sophisticated statistical methods needed to analyze it have many virtues, along with the capacity to do significant harm. It would be a serious miscarriage of science for policy makers to cite this study to divert research and clinical resources away from women with special needs for and support based on the special circumstances surrounding pregnancy, birth and infant care.
As a relationship counsellor, I spent 15 years working backwards with distressed parents. One of the key questions we ask in a first session is “when did things start to change between you?” Almost without fail, most said it was after the birth of their first baby. Laura’s comment about “everything they ever wanted” is key. Our society sets expecting parents up with the expectation that having a baby will make them happier and more fulfilled than ever – and yet new parents will tell you that those first few days, weeks, months, even years…are gritty.
Research here in Australia found 60% of women put their postnatal depression down to unrealistically high expectations of parenthood. Older longitudinal studies in the US found 92% of parents report increased conflict in the first year after baby and 67% a decline in relationship satisfaction – at a time they expect to be more in love and closer together than ever.
My experience tells me that dashed expectations and unexpected relationship issues are major contributing factors to perinatal anxiety and depression in both mothers and fathers. I believe we need to be resourcing expecting couples with the skills to cope with all the changes and challenges that come with new parenthood – we know what they are – so parents can work together and actually create the family they hope for.
As a perinatal psychiatrist I will simply speak from experience. There is a significant percentage of moms that have no risk factors (history or current circumstances) and find themselves at a loss as to why they feel so bad when they have “everything they ever wanted”. Generally, many of these women get better quickly with low doses of a SSRI. If you add the support of a group format most of them get better more quickly than expected. The others generally get better in accordance with what would be expected for a person with depression. Those ladies usually have a history of depression/ bipolar disorder or significant bio/psycho/social factors that explain the depression.