Who will screen for postpartum depression?

Writing in the New York Times about her experience with postpartum depression, filmmaker Sarah Gosner suggests that the answer of most medical providers is, “Not I”. Gosner quotes a licensed clinical social worker as saying, “Obstetricians and pediatricians won’t touch it with a 10-foot pole.”

As Gosner reports, the American Congress of Obstetricians and Gynecologists has given their professionals a hall pass:

Depression is very common during pregnancy and the postpartum period. At this time there is insufficient evidence to support a firm recommendation for universal antepartum or postpartum screening.

In contrast, the American Academy of Pediatrics sees a role for pediatricians:

The primary care pediatrician, by virtue of having a longitudinal relationship with families, has a unique opportunity to identify maternal depression and help prevent untoward developmental and mental health outcomes for the infant and family. Screening can be integrated, as recommended by Bright Futures and the AAP Mental Health Task Force, into the well-child care schedule and included in the prenatal visit.

The AAP recommends that pediatricians administer the Edinburgh Postnatal Depression Scale during the 1-, 2-, 4-, and 6-month visits. In mild to moderate cases, psychotherapy and/or medication could be indicated. In severe cases involving suicidality or psychosis, immediate intervention is advised.

While important, this recommendation does not translate easily into universal screening. Gosner quotes a psychologist at the University of Pennsylvania’s School of Medicine who says that universal screening would require a “radical redesign” of the system.

On this point, the Kenyan and U.S. experience is comparable. We find that Kenyan nurses, who provide most of the primary care to pregnant women and new mothers in the country, recognize the need for depression screening, but feel uncertain of what they can do. I believe this reluctance is partly cultural, but largely pragmatic. There is almost no capacity to treat women who screen positive for depression.

In the rural dispensary where we work (think primary care center), there are no personnel trained in psychotherapy. The nearest referral hospital is 30 minute drive, a prohibitively expensive trip for most women suffering from something not immediately life threatening. Even if they could make it there once (let alone every week for regular psychotherapy), they would find an overworked psychiatric nurse trained primarily to dispense medications, and maybe some quick advice.

These treatment barriers aside, we are trying to bring universal pre- and postnatal depression screening to the most isolated women via Baby Monitor. While we recognize the potential benefits of  a locally-derived screening measure, we are starting with a Swahili version of the Edinburgh Postnatal Depression Scale. Other groups have used the somewhat comparable Patient Health Questionnaire (PHQ-9) in Western Kenya.

In our first study outside of Nairobi, we found that women endorsed more depression symptoms over the phone using our automated system compared to in-person assessments when asked by a nurse. This is consistent with other research on interactive voice responses (IVR) systems that has found people more willing to admit to potentially embarrassing conditions given the perceived anonymity of a recorded service.

In our current work in Bungoma East District, we are replicating the first study and incorporating a new type of depression assessment: voice-based screening. Current methods of screening for depression, assessing severity, and monitoring response to treatment rely on either patient-report or clinician judgement, both of which are subjective and error prone. The search for more objective biomarkers of depression has led researchers to study how depression affects speech. Findings from a recent randomized controlled trial in the U.S. demonstrated that it is feasible to obtain valid measures of depression severity and response to treatment via the analysis of vocal recordings captured via IVR. We are exploring whether women’s speech patterns and vocal qualities are associated scores on our perinatal depression screening.

Once we feel like we have a valid and reliable remote screening instrument, we will tackle the next big challenge: treatment. There are several interesting lay-health worker models we can build upon that “task-shift” treatment responsibilities to personnel with minimal training, thus making scale-up more feasible. It will be interesting to explore what role our Baby Monitor system can play in monitoring response to treatment and possibly serving as a vehicle for some aspects of service delivery.

So, “Who will screen for postpartum depression?” Baby Monitor will.