Recently, I wrote an article about how I felt a study that challenged a New Jersey-based postpartum depression program's effectiveness was flawed in its conclusions .
In response to my article, I received a reply from the authors of the New Jersey PPD initiative Medicaid study: Katy Backes Kozhimannil, Alyce Adams
Stephen Soumerai, Alisa Busch and Haiden Huskamp.
They took great care to respond to each of my points, so I felt it should be published as a new article.
The following is what the authors wrote:
To: Susan Dowd Stone, MSW, LCSW
Date: 2/16/11
"We read your post on February 7, 2011 regarding our recent publication in Health Affairs with interest. It raised important issues about the challenges of studying policy change, and we welcome the opportunity to respond to your concerns. First, we agree with your emphasis on the importance of understanding contextual details about the law and its implementation in New Jersey and wanted to reiterate how much we appreciated the assistance we received from New Jersey’s Division of Medical Assistance and Health Services, the Division of Family Health Services, the Division of Mental Health Services, and the Maternal and Child Health Consortia. Indeed, the tremendous and wide-ranging efforts on the part of these organizations, and others, motivated and informed our research efforts from the start.
We agree with your comment that some groups who may have benefited from the initiatives were not included in our analysis, and we provide information about which groups were excluded in the published paper. However, we would like to provide additional detail about our rationale for these exclusions, which were based on a need for data to measure study outcomes and to maintain comparable groups of women who had more or less equal exposure to the policy.
Our study used administrative enrollment and billing data from New Jersey’s Medicaid program, which has several eligibility categories, including pregnant women and low-income parents. Each category has a different income eligibility threshold. We drew our initial sample (103,414) from those whose eligibility was a result of their pregnancy. New Jersey’s Medicaid program covers all pregnant women whose incomes are <300% of the Federal Poverty Level during their pregnancy and for 60 days following delivery.
However, coverage ends 60 days after childbirth for women with incomes between 115% and 300% of the Federal Poverty Level. Mothers with incomes less than 115% of the Federal Poverty Level remain eligible for coverage as low-income parents.
Since our analysis focused on postpartum depression, where onset, illness recognition, diagnosis, and treatment may well occur more than 60 days following childbirth, we were limited to examining impacts among those for whom data were available more than 60 days after delivery. To assess the impact of this decision on the analysis, we conducted sensitivity analyses around the criteria for continuous eligibility, and while relaxing the criteria did allow us to include some more women, the results of our analysis were unchanged.
Women with diagnoses of bipolar disorder and schizophrenia represent a highly vulnerable population. Although they may experience bouts of depression, we excluded them from this analysis for two reasons. One is that pharmacologic treatment for bipolar disorder and schizophrenia varies significantly from the treatment of depression, making their utilization not comparable to that of other women in the study. 1-7
For example, in the case of bipolar disorder, prescribing an antidepressant in absence of an antimanic medication is contraindicated and can destabilize the illness and precipitate mania.1
Also, there is recent evidence that antidepressants may be ineffective in treating bipolar depression. 8 While psychotherapy would be appropriate for persons with bipolar disorder and schizophrenia who were in the midst of a depressive episode, it would be impossible to distinguish visits for depressive episodes in these populations.
While women who suffer depression during pregnancy are at increased risk for postpartum depression, we did not include these women in our analysis because women in treatment prior to delivery are not the focus of screening efforts required by the New Jersey law; we focused on new onset and new recognition of illness. The study outcomes include treatment initiation, follow up, and continuation of care.
Without knowing when treatment began, we were unable to assess these outcomes in the population of women who were already receiving mental health care during pregnancy. Additionally, we excluded women who had prenatal treatment for anxiety disorders, since depression and anxiety often can have overlapping symptoms and depressive disorders are therefore more likely to be misclassified as anxiety disorders.
In the article, we point out that our analysis is limited to outcomes that were observable in Medicaid administrative claims data. Importantly, this meant that we were unable to measure any changes in rates of screening (which is not paid for by New Jersey’s Medicaid program and thus not measurable in claims data), rates and levels of depressive symptomology as well as awareness, stigma, and support (all of which are not recorded in administrative data). What we could examine was initiation and continuation of medical treatment for postpartum depression among a limited, but important, group of low-income women.
We agree that New Jersey’s initiatives to address postpartum depression have indeed inspired myriad efforts to advance understanding and improve detection, treatment, and care for women who experience perinatal mental illness. Especially given the fact that tremendous effort underlies New Jersey’s initiatives, it is critical to ensure that policy efforts to address such a compelling and important issue result in tangible benefits – in terms of treatment access – to all women who suffer postpartum depression, including those who are most vulnerable."
Katy Backes Kozhimannil
Alyce Adams
Stephen Soumerai
Alisa Busch
Haiden Huskamp
The additional perspective is welcome, but it is hoped that future studies will analyze available data across all populations, including the most vulnerable who may be uninsured, but benefit greatly from programs and services available to all. Finally, modifying the study title in a way that reflects limitations helps avoid global conclusions that could influence or discourage policymakers from pursuit of program development.
1. American Psychiatric Association. Practice guidelines for the treatment of patients with bipolar disorder (revision). American Journal of Psychiatry. 2002;159(4):Supplement 1-50.
2. Lehman AF, Steinwachs DM, Dixon LB, et al. Translating research into practice: the schizophrenia Patient Outcomes Research Team (PORT) treatment recommendations. 24. 1998;1(1-10).
3. Lehman AF, Kreyenbuhl J, Buchanan RW, et al. The Schizophrenia Patient Outcomes Research Team (PORT): Updated treatment recommendations 2003. Schizophrenia Bulletin. 2004;30(2):193-217.
4. Buchanan RW, Kreyenbuhl J, Kelly DL, et al. The 2009 schizophrenia PORT psychopharmacological treatment recommendations and summary statements. Schizophrenia Bulletin. 2009;36(1):71-93.
5. Hirschfeld RM. Guideline Watch: Practice guideline for the treatment of patients with bipolar disorder: American Psychiatric Association; 2005:1-9.
6. Sachs GS, Printz D, Kahn D, Carpenter D, Docherty JP. The Expert Consensus Guideline Series: Medication Treatment of Bipolar Disorder 2000. Postgraduate Medicine. 2000;1-104.
7. Suppes T, Dennehy EB, Hirschfeld RMA, et al. The Texas implementation of medication algorithms: Update to the algorithms for treatment of bipolar I disorder. Journal of Clinical Psychiatry. 2005;66(7):870-886.
8. Sachs GS, Nierenberg AA, Calabrese JR, et al. Effectiveness of adjunctive antidepressant treatment for bipolar depression. New England Journal of Medicine. 2007;356(17):1711-1722.