Research That Matters (January 17 - 20, 2008)

Background: While there is widespread support for perinatal depression screening in health care and community settings, consumer involvement in developing culturally informed approaches to screening and psychosocial risk management is a new area where social work may play a vital role in translating research to practice. The aim of this study is to assess whether community-informed perinatal depression screening enhances relevance, responsiveness, and rigor when compared with traditional screening alone.

Methods: Community-based research was conducted within an organization serving low-income, predominantly African-American pregnant and post-partum women. Five depression screening instruments were identified that had respectable validity and reliability in the research literature. Staff and project participants expressed that none of the published screeners adequately reflected the experience of perinatal depression in the target population. Through a series of focus groups, the community identified the CES-D as the preferred screening instrument and modified it to include five “tag on” questions derived from community-specific expressions of depression. Content analysis informed community-worded items (examples: “feeling angry,” fear of harm, “everyone's against me”) that were coded in a format consistent with the CES-D. Agency staff were trained simultaneously on depression screening and cultural sensitivity regarding mental health and depression. The screener was administered at baseline (entry to program or first assessment post-training) and at 3 month follow-up intervals. Psychometric testing compared results from the CES-D items alone and the CES-D in combination with the community-informed “tag on” questions.

Results: Psychometric testing at baseline (N=139) revealed that the community-informed screener with additional “tag on” items had a higher internal consistency than the CES-D alone (alpha = .87 vs. .80). All tag-on items had significant inter-item correlation with the CES-D items, with the exception of “I'm afraid something will happen to myself or my baby;” further analysis revealed this item to be more predictive of general psychosocial risk than depression. Two criterion validity tests were conducted. The first used the published CES-D cut-off score; factor loadings for the community-informed items were in the range of .61-.74. The second validity test used a psychosocial “high risk” category based on agency assessment; in this analysis, the community informed items presented with factor loadings ranging from .61-.87. In both analyses, factor loadings were relatively consistent between the community-informed and CES-D items. A final test-retest reliability analysis was conducted using three month follow-up data (N=29); there were no statistically significant differences in item response on either the CES-D or community informed questions.

Implications: This study reinforces the value of consumer and staff engagement into the design and implementation of relevant and culturally responsive approaches to screening and risk assessment for perinatal depression. A deliberate process of community involvement and informed question design resulted in a valid, internally consistent, and reliable screening instrument in this community-based sample. Additionally, community engagement in development, implementation and evaluation had a secondary benefit of promoting agency awareness regarding the cultural expressions and experiences of depression, resulting in screening approaches that may be more culturally relevant and responsive as well as rigorous.