The ROSE Scale-up Study: Informing a Decision About ROSE as Universal PPD Prevention

In 2019, the US Preventive Services Task Force (USPSTF) issued a recommendation indicating that it is possible to prevent postpartum depression through counseling interventions. The Reach Out, Stand Strong, Essentials for New Mothers (ROSE) program was one of the two interventions mentioned by name serving as a basis for the recommendation. ROSE was tested in 4 randomized trials in community prenatal settings with racially/ethnically diverse low-income adult pregnant women and prevented ~50% postpartum depression (PPD) cases. ROSE is scalable in that: (1) it does not require mental health clinicians or licensed providers of any kind; (2) it has a flexible delivery structure (e.g., group vs. individual, timing during pregnancy); (3) it is the shorter (5 vs. 8-17 sessions) of the two interventions described in the USPSTF report. PIs Johnson and Zlotnick (the developer of ROSE) are in Year 4 of an implementation trial examining how much technical support is needed for various kinds of agencies to implement and sustain ROSE over time (R01 MH114883). That trial enrolled 98 agencies serving low income women nationally (i.e., settings where women receive perinatal care), including prenatal clinics, home visiting, Healthy Start, and doula groups, among others. It answers many of the questions posed in this RFA, including identifying: (1) a "service-ready efficacious preventive intervention" that is "scalable and can be delivered with fidelity by setting providers;" (2) "strategies to train providers and support delivery of evidence-based approaches with fidelity across health care and community settings;" (3) implementation barriers and facilitators; and (4) effective implementation strategies. However, that trial raised a question about the USPSTF recommendation.

Universal vs. selective or indicated prevention. All effectiveness trials of ROSE and of the other recommended PPD prevention intervention included only low-income women, a single risk factor that doubles incidence of PPD. Three of the 4 ROSE trials included women who were screened for additional risk for PPD per the Cooper survey to maximize trial power. Because the existing evidence base consisted of women at increased risk for PPD, the USPSTF recommendation suggested that "clinicians provide or refer… persons who are at increased risk of perinatal depression to counseling interventions," and called for "more studies on the use of tools for risk assessment." However, most of the healthcare and community agencies in our implementation trial (78%) provide or offer ROSE to every woman as part of their standard work flow (i.e., provide it as universal prevention), finding this more feasible than creating a screening and referral process for indicated or selective prevention. Most of the other agencies use provider judgment to refer to ROSE. Only 7% use a screening tool (typically the Edinburgh) for ROSE referral. In addition to being more feasible, universal prevention may be better because: (1) the cost of a screening false negative (resulting in a preventable case of PPD; $32,000) far exceeds the cost of ROSE delivery ($50-300/woman); (2) women who already face societal stigma (racial/ethnic minorities, undocumented) can be fearful about consequences and feel further stigmatized when a screener singles them out as needing PPD prevention. This project assesses effectiveness, cost-outcome, equity, and scalability of using ROSE as universal vs. selective or indicated prevention for PPD using commonly available/existing screening tools. We expect universal prevention to be optimal. Effectiveness of ROSE among low-income women at risk for PPD is known (ROSE prevents ~50% of PPD cases). To inform a recommendation about using ROSE as universal vs. selective or indicated prevention, we need to determine the effectiveness of ROSE among all pregnant persons, including those screening negative for PPD risk (defined in various ways; Section A6). We will enroll 2,320 people receiving prenatal care in a large regional health system (based in Detroit, MI), assess them with PPD risk tools (chosen for ease of use, wide use, or availability in electronic health records), randomize everyone to ROSE or to enhanced care as usual (ECAU), and collect information needed to inform a decision about ROSE as universal PPD prevention. Each proposed aim gathers a piece of information missing in the USPSTF report to guide decision-making. We will assess ROSE as universal, selective, and indicated prevention in terms of:

1. ROSE effectiveness relative to ECAU for each prevention approach for (i) SCID5 major depressive episode (primary) and (ii) functioning in the 6 months after birth. We will explore effects between baseline and birth.
2. Cost-outcome comparison of universal vs. selective or indicated prevention in terms of PPD and functioning. Because it prevents costly false negatives, universal prevention is almost always cost-beneficial.
3. Equity: (a) Percent false negative screens for minority vs. non-Hispanic white women; (b) minority status as a moderator of outcome across risk levels; (c) PPD cases that would be prevented by ROSE as universal prevention that would not be prevented under selective/indicated for minority vs. non-Hispanic white women.
4. Scalability as indicated by: (a) cost-outcome comparison of universal vs. indicated prevention (Aim 2); (b) a scalability measure and qualitative interviews with the 98 agencies in the past ROSE implementation trial about how they decided on universal vs. some other kind of prevention and how it went in their settings.
5. Mechanisms of ROSE effects across PPD risk levels. We will integrate results to advise about ROSE as universal prevention. This study fits the RFA by assessing how best to get an evidence-based program to those who need it in settings where they receive perinatal care, by addressing a pragmatic and novel question (should ROSE be universal prevention?), and by examining equity.
6. Innovation B1. This will be the first fully-powered RCT to assess universal vs. selective or indicated prevention of an evidence-based intervention for PPD in terms of prevention of PPD cases. The USPSTF recommendation on prevention of postpartum depression (i.e., that "clinicians provide or refer pregnant and postpartum women who are at increased risk of perinatal depression to counseling interventions") is new (2019) and was limited to women at elevated risk for PPD because that is what the evidence to date consisted of (i.e., trials among women at elevated risk for PPD). However, our current 98-agency implementation trial and other factors suggest that universal prevention may be more scalable, cost-effective, and equitable than selective or indicated prevention for PPD. This trial will be the first to examine the advisability of universal vs. selective or indicated PPD prevention (a) at all, \& (b) with respect to equity, scalability, and comparative cost.

B2. This definitive trial will be the largest PPD prevention trial to date. Among other advantages, this provides better power than previous studies to examine race/ethnicity by intervention interactions.

B3. This trial will be the first to explore ROSE as antenatal depression (AD) prevention.

B4. EPDS as a predictor of future PPD. The EPDS is a well-validated and almost ubiquitously used measure of perinatal depressive symptoms and it makes sense that elevated symptoms likely predict future disorder. However, no previous study has ever tested the EPDS as a predictor of future development of full PPD (only its correlation with current disorder), despite being used this way in clinical practice. A bonus additional feature of this trial is that we can use the control condition (n~1,160) to evaluate the predictive validity of this commonly used measure. Currently, no AD symptom measure has been validated as a predictor of PPD.

B5. Scalability is a relatively understudied area of implementation science. This study will provide the first (quantitative and qualitative) analysis of scalability of universal, selective, and indicated PPD prevention.

Approach: Research Design This project informs use of ROSE as universal vs. selective or indicated prevention for PPD by assessing: (1) ROSE effectiveness for each prevention approach among 2,320 women in terms of (i) SCID5 major depressive episode in the 6 months after birth (primary), and (ii) functioning. (2) Cost-outcome comparison of universal vs. selective or indicated prevention in terms of PPD and functioning (equation in Section D2.11). Savings from preventing false negatives (PPD cases among "low risk" women) drives this comparison. (3) Equity: (a) percent false negative screens for minority vs. non-Hispanic white women; (b) minority status as a moderator of outcome across risk levels; (c) PPD cases that would be prevented by ROSE as universal prevention that would not be prevented under selective/indicated for minority vs. non-Hispanic white women. (4) Scalability of universal vs. other kinds of prevention as indicated by: (a) cost-outcome analyses (Aim 2); (b) a scalability measure and qualitative interviews with the 98 agencies in the ROSE implementation trial. (5) Mechanisms of ROSE effects across PPD risk levels.