Study protocol for the ROSE Scale-Up Study: Informing a decision about ROSE as universal postpartum depression prevention

Jennifer E Johnson; Amy M Loree; Alla Sikorskii; Ted R Miller; Laura Carravallah; Brandon Taylor; Caron Zlotnick

doi:10.1016/j.cct.2023.107297

. Author manuscript; available in PMC: 2024 Sep 1.

Published in final edited form as: Contemp Clin Trials. 2023 Jul 18;132:107297. doi: 10.1016/j.cct.2023.107297

Study protocol for the ROSE Scale-Up Study: Informing a decision about ROSE as universal postpartum depression prevention

Jennifer E Johnson ¹, Amy M Loree ², Alla Sikorskii ³, Ted R Miller ⁴, Laura Carravallah ⁵, Brandon Taylor ⁶, Caron Zlotnick ⁷

PMCID: PMC10528027 NIHMSID: NIHMS1922709 PMID: 37473848

Abstract

Purpose.

To examine the effectiveness, cost-outcome, equity, scalability, and mechanisms of the Reach Out, Stay strong, Essentials for mothers of newborns (ROSE) postpartum depression prevention (PPD) program as universal versus selective or indicated prevention.

Background.

The United States Preventive Services Task Force (USPSTF) currently recommends PPD prevention for pregnant people at risk of PPD (i.e., selective/indicated prevention). However, universal prevention may be more scalable, equitable, and cost-beneficial.

Design.

Effectiveness of ROSE for preventing PPD among people at risk is known. To assess ROSE as universal prevention, we need to determine the effectiveness of ROSE among all pregnant people, including those screening negative for PPD risk. We will enroll 2,320 pregnant people, assess them with commonly available PPD risk prediction tools, randomize everyone to ROSE or enhanced care as usual, and assess ROSE as universal, selective, and indicated prevention in terms of: (1) effectiveness (PPD prevention and functioning), (2) cost-benefit, (3) equity (PPD cases prevented by universal prevention that would not be prevented under selective/indicated for minority vs. non-Hispanic white people), (4) quantitative and qualitative measures of scalability (from 98 agencies previously implementing ROSE), (5) ROSE mechanisms across risk levels. We will integrate results to outline pros and cons of the three prevention approaches (i.e., universal, selective, indicated).

Conclusion.

This will be the first trial to assess universal vs. selective/indicated PPD prevention. Trial design illustrates a novel, efficient way to make these comparisons. This trial, the largest PPD prevention trial to date, will examine scalability, an understudied area of implementation science.

Keywords: Implementation science, prevention, postpartum depression, scale-up, health equity, pragmatic trials, clinical trial protocol

1. Introduction

In 2019, the United States Preventive Services Task Force (USPSTF) recommended preventing postpartum depression (PPD; see Table 1) through counseling interventions. The Reach Out, Stay strong, Essentials for mothers of newborns (ROSE) program was one of two interventions USPSTF mentioned by name.^1–3 To maximize trial power, effectiveness trials of these two preventive interventions included only people at risk for PPD.^1–11 Because the evidence base consisted of people at increased risk for PPD, the USPSTF suggested that “clinicians provide or refer… persons… at increased risk of perinatal depression to counseling interventions,” and called for more research on risk assessment.^1–3

Table 1.

Glossary of Terms and Abbreviations

CSQ: Cooper Survey Questionnaire, a measure of risk for postpartum depression

ECAU: Enhanced Care as Usual

EHR: Electronic health record

EPDS: Edinburgh Postnatal Depression Scale, a commonly used measure of current perinatal depressive symptoms

False negatives: Cases originally identified as not at risk by a screening tool that go on to develop PPD

HFH: Henry Ford Health, a large regional healthcare system in Southeast Michigan, serving Detroit and surrounding areas

ISAT: Intervention Scalability Assessment Tool, a measure of provider perceptions of scalability
LIFE: Longitudinal Interview Follow-Up Evaluation, which provides a structured way to assess presence or absence of major depressive episode over time.

MDE: Major depressive episode

PPD: Postpartum depression, a major depressive episode in the postpartum period

Prevention approaches

Universal prevention: ROSE is provided to all pregnant people.

Selective prevention: ROSE is provided only to pregnant people at risk for developing PPD. This trial will compare ways of operationalizing selective prevention (i.e., determining risk for developing PPD), including: Medicaid enrollment, CSQ score, past MDE assessed via structured clinical interview (SCID-5), and past MDE assessed using EHR diagnoses (see Section 2.2)
Indicated prevention: ROSE is provided only to pregnant people with already elevated depressive symptoms (operationalized in this trial as having elevated EPDS scores as study baseline).

QALY: Quality adjusted life year

ROSE: Reach Out, Stand Strong, Essentials for New Mothers, a 5-session educational intervention offered during pregnancy that prevents ~50% of PPD cases. It can be offered by nurses, health educators, and other non-mental health providers.

SCID-5: Structured Clinical Interview for DSM-5 Axis I disorders, the gold standard method of assessing major depressive disorder

USPSTF: United States Preventive Services Task Force

VR-12: Veterans Rand 12-item Health Survey (VR-12) variant of the Short Form Health Survey, a widely used measure of physical and mental health functioning

Open in a new tab

However, most of the 98 healthcare and community agencies in an ongoing national ROSE implementation trial¹² provide or offer ROSE to every pregnant person 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 selective or indicated prevention (see Table 1 for definitions). The remaining agencies typically used provider judgment (rather than a screening tool). Experience in the implementation trial also suggests that people who face societal stigma (racial/ethnic minorities, undocumented) are sometimes fearful about consequences (such as loss of parental rights, visibility leading to deportation) and feel further stigmatized when singled out as needing PPD prevention. Universal prevention also may be better because the cost of a screening false negative (resulting in a preventable case of PPD; $32,000¹³) far exceeds the cost of ROSE delivery ($50-$300/pregnancy).

The current project assesses effectiveness, cost-outcome, equity, and scalability of using ROSE as universal vs. selective or indicated prevention for PPD. Effectiveness of ROSE for preventing PPD among people at risk for PPD is known (ROSE prevents ~50% of PPD cases).^4–7 To inform a recommendation about using ROSE as universal prevention, we need to determine the effectiveness of ROSE among all pregnant people, including those screening negative for PPD risk. We will enroll 2,320 people receiving prenatal care, assess them with commonly available PPD risk tools, randomize everyone to ROSE or enhanced care as usual (ECAU), and assess ROSE as universal, selective, and indicated prevention in terms of:

ROSE effectiveness relative to ECAU for each prevention approach for (i) major depressive episode and (ii) functioning in the 6 months after birth.
Cost-outcome comparison of universal vs. selective/indicated prevention in terms of PPD and functioning.
Equity: (a) Percent false negative screens for minority vs. non-Hispanic white people; (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 people.
Scalability: (a) cost-outcome comparison of universal vs. selective/indicated prevention; (b) a scalability measure and qualitative interviews with the 98 agencies in the ongoing national ROSE implementation trial about how they decided on universal vs. selective/indicated prevention and how it went in their settings.
Mechanisms of ROSE effects across PPD risk levels.

Results will inform decision-making about recommending ROSE as universal, selective, or indicated prevention.

2. Methods

2.1. Rationale for trial design.

2.1.1. Aim 1. ROSE effectiveness for each PPD prevention approach.

To inform a recommendation about ROSE as universal prevention, we need to determine the effectiveness of ROSE among general pregnant populations, including those screening negative for PPD risk. Hypotheses are shown in Table 2.

Table 2.

Hypotheses

Aim 1: Effectiveness	Prevalence of MDD in the 6 months after birth will be lower in ROSE than ECAU for:
	1a. people not at increased risk of PPD (defined using selective prevention)
	1b. people at increased risk of PPD (defined using selective prevention)
	1c. people not at increased risk of PPD (defined using indicated prevention)
	1c. people not at increased risk of PPD (defined using indicated prevention)
	1e. the sample as a whole (universal prevention)
	We will also explore whether ROSE prevents antenatal depression cases (universal)
Aim 2: Cost-outcome comparison	Universal prevention will have lower net cost than selective or indicated prevention
	2a. From a societal perspective
	2b. From a health system perspective
Aim 5: Mechanisms	The effect of ROSE on PPD in the sample as a whole will be mediated by (a) instrumental support and (b) perceived stress.

Open in a new tab

Note: Aims 1, 2, and 5 have hypotheses (shown above). Aims 3–4 focus on parameter estimation (false negative rates, scalability indicators).

2.1.2. Aim 2.

Cost-outcome comparisons will describe costs and cost-savings of ROSE as universal vs. selective vs. indicated prevention from societal and health services perspectives in terms of PPD and functioning. Primary analyses will determine the net incremental (i.e. relative to selective or indicated) costs of universal prevention (cost of providing ROSE to everyone who screened negative minus [the probability of a prevented case of PPD among those who screen negative*cost of PPD]). This trial will quantify the unknown variables in this equation for various screening tools: the probability of false negatives, the associated percent screening positive, and the percent of PPD cases prevented by ROSE among those screening negative. Secondary analyses will compare incremental benefit-cost and cost-utility.

Universal intervention may be less costly, given that the cost of a screening false negative resulting in a preventable case of PPD ($32,000 societal excluding quality of life loss, $17,100 of health care¹³) is much larger than the cost of ROSE delivery ($50-$300 per woman). Our preliminary analysis (shown in Table 3) suggest that the lower net cost for universal than selective/indicated prevention is driven by avoiding the high cost of false-negative screens (the percent of people screening negative for PPD risk who go on to develop PPD without intervention). If ROSE reduces PPD rates among false negative screens by even 1–2%, universal prevention costs less than selective/indicated prevention.

Table 3.

Cost savings per negative-screened woman of universal PPD prevention relative to selective or indicated PPD prevention from societal and health system perspectives^*

(Prob[PPD] in ECAU minus prob[PPD] in ROSE) among negative risk screens	Provider Cost per ROSE Participant
	$50 (all low-cost group)	$100	$150	$225	$300 (all individual)
Universal, 0%	Societal: (78) Health: (3)	Societal:(128) Health: (53)	Societal:(178) Health: (103)	Societal:(253) Health: (178)	Societal:(328) Health: (253)
Universal, 1%	Societal: 242 Health: 168	Societal: 192 Health: 118	Societal: 142 Health: 68	Societal: 67 Health: (7)	Societal: (8) Health: (82)
Universal, 3%	Societal: 882 Health: 510	Societal: 832 Health: 460	Societal: 782 Health: 410	Societal: 707 Health: 335	Societal: 632 Health: 260
Universal, 10%	Societal: 3122 Health: 1707	Societal: 3072 Health: 1657	Societal: 3022 Health: 1607	Societal: 2947 Health: 1532	Societal: 2872 Health: 1457

Open in a new tab

Societal/health system costs for selective/indicated prevention per screened negative woman range from $33–60. We used the midpoint ($47) for illustrative purposes. Red #s are negative.

2.1.3. Aim 3.

Equity. Selective or indicated prevention can lead to inequities if screening tools have different rates of false negatives (eventual cases of PPD not identified by a screening tool) across race/ethnicity. Furthermore, although ROSE trials among high-risk people included strong representation of racial/ethnic minority people and ROSE effects did not differ by race/ethnicity, this trial will include screened negative people and will be much larger, allowing for more robust tests of any race/ethnicity effects for ROSE. These two considerations together (false negative rates, ROSE effectiveness rates across risk levels by race/ethnicity) will allow us to estimate rates of 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 people, reflecting equity of ROSE as universal vs. selective/indicated prevention.

2.1.4. Aim 4.

Scalability of prevention approaches. Comparative cost-outcome analyses of Aim 2 will inform scalability. We will also return to the 98 agencies in the ongoing national ROSE implementation study¹² and evaluate their perspectives on universal, selective, and indicated prevention using quantitative and qualitative measures. Drawing from these agencies’ experiences using ROSE (and the benefits and challenges they encountered) will contextualize and illuminate pros and cons of each prevention approach.

2.1.5. Aim 5.

Mechanisms of ROSE effects across PPD risk levels. We selected instrumental support and perceived stress potential mechanisms of ROSE because ROSE targets both and both predict PPD (Supplementary Material 1).^14,15 These mediators will be evaluated in the entire population. Then we will explore whether factors defining selective or indicated prevention change the magnitude of the effect of ROSE on the mediators. We do not expect substantial variations in ROSE effectiveness across subgroups, but if any are found, moderated mediation analysis could provide an explanation.

2.1.6. Why not randomize agencies to prevention approaches?

The current design is more feasible and allows us to assess selective/indicated prevention using a variety of screening tools, which would not be possible with agency-level randomization.

2.2. Measures to be used for selective and indicated prevention in this pragmatic trial

are commonly available/existing screening tools (chosen for ease of use, wide use, availability in electronic health records [EHRs], and/or use in previous ROSE trials [to allow comparison of current and past findings]). Participants’ PPD risk will be categorized once during pregnancy, at study intake, using baseline values of these measures. Measures used to assess selective prevention (i.e., risk for developing PPD) include Medicaid enrollment (a single factor that doubles rates of PPD,^16,17 a positive screen (score of 27+) on the Cooper Survey Questionnaire (CSQ; which triples risk of developing PPD),^18,19 past major depressive episode (MDE) assessed using structured clinical interview (generalizable to systems using provider referral/judgment), and past MDE assessed using EHR diagnoses (scalable). We will use the Edinburgh Postnatal Depression Scale (EPDS)²⁰ to assess indicated prevention. In indicated prevention, elevated symptoms are thought to indicate risk of developing full disorder. Although the EPDS is a well-validated and widely-used²¹ measure of current perinatal depressive symptoms available in EHRs, its validity when used during pregnancy to predict future development of PPD is unknown. As a bonus feature, we will use the trial’s control condition (n~1,160) to evaluate the predictive validity of this measure.

2.3. Intervention conditions.

To determine the naturalistic effects of adding ROSE for PPD prevention across a range of PPD risk levels, participants in both conditions can receive any other treatment available to them. We will characterize other treatment received for descriptive and dose-response analyses.

2.3.1. Control: Enhanced Care as Usual (ECAU)

consists of usual care plus monitoring and emergency referral, as is required to fulfill ethical obligations to trial participants. The trial will take place within Henry Ford Health (HFH), a large regional health system in Southeast Michigan, with 9,500 births per year. As is the case nationally,^21–24 care as usual in HFH-affiliated prenatal clinics does not include PPD prevention. Rather, the clinics screen for PPD symptoms using the EPDS and refer for mental health care.

2.3.2. ROSE.

The goal of ROSE is to prevent rather than treat PPD. ROSE is administered to pregnant people in a group or individually (4 sessions and 1 postpartum booster). It is structured and contains psychoeducation and skills for improving relationships and building social support, including roleplays and homework with feedback. ROSE overcomes barriers to attendance by having a flexible delivery structure (e.g., group vs. individual, office vs. telehealth, timing during pregnancy, sessions combined or split). This trial will use telehealth.

Facilitators such as health educators, nurses, and paraprofessionals have delivered ROSE with good fidelity.²⁵ For this trial, facilitators will be community health workers, counselors, or nurses who work with pregnant people at HFH. They will be trained using a standardized half-day training,^4–7,12 and will demonstrate proficiency prior to offering ROSE to participants. The study supervisor will review facilitators’ audio recorded sessions and meet every other week for an hour for group feedback and case discussion.

2.4. Provider participants

(Aim 4b) will be representatives of the 98 agencies from 32 states (i.e., external to the HFH system) enrolled in the ongoing national ROSE implementation trial.¹² That study compared three doses of implementation support and evaluated sustainment over 30 months. Agencies naturalistically chose a variety of strategies to invite pregnant people to participate in ROSE, including providing or offering ROSE to all, clinician judgment, or a screening tool. For the current study, we will re-consent agency respondents and ask if they would participate in an additional survey and qualitative interview about scalability of universal, selective, and indicated prevention.

2.5. Patient participants

(Aims 1–3, 4a, and 5) will be: (1) 12–32 weeks pregnant; (2) receiving prenatal services at HFH; (3) aged 18+; (4) be willing/able to provide contact information of 2 locator persons; (5) able to understand English well enough to understand questionnaires when read aloud; and (6) have access to a device capable of video calls. Participants will be excluded and referred for treatment if they have current MDE, mania/hypomania, or psychosis (assessed via structured clinical interview) at baseline. We will reach out to all pregnant people presenting at HFH for prenatal services using email and HFH MyChart app notification twice. A randomly selected 50 each week will receive additional intensive outreach (spaced out texts, emails, phone calls, and a final mailed letter; see Supplementary Material 2).

2.6. Allocation, randomization, and masking.

Unstratified randomization of patient participants to ROSE or ECAU in a 1:1 ratio will occur after the baseline assessment. The study statistician will prepare the randomization schedule before the first participant is enrolled. Immediately after randomization, study staff will review the study follow-up schedule, means of contacting the research staff, and participants’ contacts. Different study staff, masked to intervention assignment, will perform telephone follow-up assessments.

2.7. Sample size and power.

Table 4 includes sample size requirements using the projected rates of PPD (primary outcome) based on previous ROSE trials,^4–7 prevalence of risk factors at HFH and in the literature,²⁶ and power >.80 in two-sided tests at α=.05. For Hypotheses 1a and 1b (Table 2), three criteria for selective implementation are considered: Medicaid, CSQ-positivity, and past MDE (2 different assessments, interview and EHR, considered separately). For Hypothesis 1c, EPDS positivity is used as a factor for indicated PPD prevention. Required sample size ranges from 576 to 1856 for Hypotheses 1a–1e, based on projected differences between ROSE and ECAU and prevalence of the risk factor. We selected the largest, n=1856 to ensure power of .80 - .997 for all hypotheses involving the primary outcome. This sample size is sufficient for power of >.80 for secondary outcome and mediation analyses. Previous ROSE trials had 86% to 97% follow-up rates.^4–7 We conservatively estimate an 80% follow-up rate in the current study. To account for 20% attrition, we will enroll n=2,320.

Table 4.

Power and sample size estimates (white lines reflect previous ROSE trials; pink lines are more conservative effects for which the current trial is powered)

Hypothesis and subgroup	Est. % of overall population	Est. PPD rate in ECAU (%)	Est. PPD rate in ROSE (%)	Required n for 80% power in the subgroup	Required total n given est.% in population	With 20% attrition
1a: CSQ- screen	82%	13	7	784	957	1,196
1a: CSQ- screen	82%	10	5	870	1,061	1,326
1b: CSQ+ screen	18%	30	16	282	1,567	1,959
1b: CSQ+ screen	18%	25	13	334	1,856	2,320
1a: Medicaid-	51%	13	7	784	1,537	1,921
1a: Medicaid-	51%	12	6	712	1,396	1,745
1b: Medicaid+	49%	30	16	282	576	720
1b: Medicaid+	49%	25	13	334	682	853
1a: No past MDE	87%²⁶	6²⁶	3	1496	1,719	2,149
1b: Past MDE	13%²⁶	33²⁶	16	198	1,523	1,904
1c: EPDS- screen	82%	13	7	784	1,089	1,361
1c: EPDS- screen	82%	10	5	870	1,209	1,511
1d: EPDS+ screen	18%	30	16	282	1,567	1,959
1d: EPDS+ screen	18%	25	13	334	1,856	2,320
1e: Whole sample	100%	15	8	650	650	813

Open in a new tab

2.8. Assessments.

Assessments (Table 5) will take place at baseline (during pregnancy; 12–32 weeks) and at 6 months after birth, mainly by phone or occasionally at HFH offices. Birth outcomes will be collected via EHR.

Table 5.

Schedule of patient assessments

	Intake (during pregnancy)	6 months after birth
Patient assessments

Demographics	X
PPD Risk
Medicaid	X
Cooper (CSQ)	X
Past MDE (SCID, EHR)	X
Edinburgh (EPDS)	X	X
Outcomes
PPD (LIFE)		X
Functioning (VR-12)	X	X
Mechanisms
Instrumental support	X	X
Perceived stress	X	X

Open in a new tab

2.8.1. Aim 1. ROSE effectiveness for each PPD prevention approach.

PPD risk: selective prevention.

Medicaid status at baseline will be abstracted from EHR data. In Michigan, all pregnant people (included undocumented) who meet income requirements are Medicaid eligible. For sensitivity analyses, we will (1) collect self-reported household size/income to simulate Medicaid eligibility in other states; and (2) track how many enroll in Medicaid after baseline to estimate stability/validity of this predictor over time. Scores on the CSQ, a 17-item measure of PPD risk,^18,27 range from 0 to 63; a score of 27+ indicates PPD risk.¹⁸ We will separately test two measures of past MDE: the Structured Clinical Interview for DSM-5 Axis I disorders (SCID-5; gold standard)²⁸ and EHR documentation of MDE diagnosis (more scalable).

PPD risk: indicated prevention.

EPDS scores range from 0 to 30; a score of 10+ will be considered “at risk.”^20,29

Outcomes. PPD and antenatal depression cases.

The Longitudinal Interview Follow-up Examination (LIFE)^30,31 will be used to determine whether the participant experienced a MDE at any time between birth and 6 months postpartum (primary) and between study baseline and birth (exploratory). The LIFE provides a structured way to assess SCID-defined MDE over time. Functioning will be measured using the Veterans Rand 12-item Health Survey (VR-12) variant of the Short Form Health Survey,³² a brief, widely used measure of physical and mental health functioning that provides our secondary cost-effectiveness measure.

ROSE attendance records will be maintained by ROSE facilitators. Other mental health care received will be characterized using the Treatment History Interview³³ and the EHR.

2.8.2. Aim 2. Comparative cost-outcome analysis of prevention approaches.

Outcome measures will include: PPD cases and the VR-12 with Sengupta’s HUI3 scoring,³⁴ which measures functional status in quality-adjusted life years (QALYs).

2.8.3. Aim 3.

Indicators of the equity of ROSE universal vs. selective or indicated prevention include: (a) Percent false negative screens for minority vs. non-Hispanic white; (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 people.

2.8.4. Aim 4. Scalability of ROSE universal vs. selective or indicated prevention.

Indicators will include: (a) cost-outcome analyses of universal vs. selective/indicated prevention (Aim 2); and (b) a scalability measure and qualitative interviews with the 98 agencies in the ongoing ROSE implementation trial. We will use the Intervention Scalability Assessment Tool (ISAT)³⁵ to assess perceived scalability of universal, selective, and indicated prevention for each agency based on their experiences implementing ROSE, and then compare scores for each kind of prevention. Domains are shown in Table 6. Virtual qualitative interviews with the 98 agencies will ask about (1) scalability of universal, selective, and indicated prevention, (2) why and how they chose the approach they did; (3) successes or challenges with their chosen approach in their context, and (4) suggestions for other agencies.

Table 6.

Dimensions of scalability of universal, selective, and integrated prevention to be assessed in qualitative interviews

ISAT domains	• the problem (is it of sufficient concern to warrant scale-up to the proposed level?) • the program/intervention (how well does it address needs of target group/problem?) • strategic/political context (is problem consistent with policy/funding/strategic directions/priorities?) • evidence of effectiveness (of ROSE as selective, indicated, or universal prevention) • program/intervention costs (of ROSE as selective, indicated, or universal prevention) • fidelity and adaptation (can program fidelity be monitored/maintained if implemented at scale?) • reach and acceptability (of selective, indicated or universal prevention, esp to people from marginalized groups) • delivery setting/workforce (feasibility/acceptability of selective, indicated, universal prevention in existing structures) • are implementation infrastructure requirements of selective, indicated, universal prevention feasible for scale-up? • sustainability (are integration, resourcing, workforce needed for selective, indicated, universal prevention sustainable at scale?)
IHI Model for going to Full Scale^48,49	• support systems (learning systems, data systems, infrastructure, human capacity, capability for scale-up, sustainability) • adoption mechanisms (leadership, communication, social networks, culture of urgency and persistence) • content (develop and validate change package, replicate/adapt it across contexts)
Zamboni definition	• the ability of a health intervention shown to be efficacious on a small scale or under controlled conditions to be expanded under real-world conditions to reach a greater proportion of the eligible population, while retaining effectiveness^50–52

Open in a new tab

2.8.5. Aim 5:

Mechanisms of ROSE effects across PPD risk levels. We will use the well-validated^36,37 8-item Patient Reported Outcomes Measurement Information System Instrumental Support (Short Form 8a),³⁸ and the well-validated,³⁹ frequently used 10-item Perceived Stress Scale.^40,41 For both measures, we will assess both the month after birth and the current month.

2.9. Analyses.

2.9.1. Aim 1: ROSE effectiveness for each PPD prevention approach.

Primary analyses will be intent-to-treat with two-sided tests at α=0.05. Secondary analyses will examine dose-response effects (for ROSE and for other treatment received) and effects of weeks gestation at baseline. P-values will be supplemented with measures of clinical significance.

Missing Data.

Baseline variables predicting differences in attrition between conditions will be covaried in primary analyses. If patterns of missing data indicate potential not missing at random mechanisms, then models describing missing mechanisms will be considered (e.g., pattern-mixture models) in sensitivity analyses.

Outcomes.

Primary.

We will compare presence of LIFE-assessed MDE (i.e., PPD) across the 6 months after birth between ROSE and ECAU, using generalized linear models with binomial error distribution, with trial arm as a predictor, covarying baseline levels of the dependent variable, risk prediction measures (y/n SCID-5 past MDE, EHR past MDE, Medicaid enrollment, CSQ and EPDS scores), and factors identified in attrition analyses. Based on study hypotheses (Table 2), this model will be fit for: (ai) people who are CSQ- at baseline; (aii) people who are non-Medicaid at baseline; (aiii) those with no past MDE at baseline (assessed with SCID5); (aiv) those with no past MDE at baseline (EHR); (bi) people who are CSQ+ at baseline; (bii) people who on Medicaid at baseline; (biii) people with past MDE at baseline (assessed with SCID5); (biv) people with EHR record of past MDE at baseline; (c) people who are EPDS- at baseline; (d) people who are EPDS+ at baseline; and (e) the sample as a whole. Any differences in findings of ROSE effectiveness across these subsets will inform decisions about universal, indicated, or selective prevention. Secondary. We will test the effects of ROSE vs. ECAU on functioning (VR-12 score) at 6 months, using general linear models with baseline functioning (VR-12 score), risk prediction measures, and factors identified in attrition analyses as covariates. The model will be fit for the same groups as for the primary outcome. Exploratory: We will test the effectiveness of ROSE for preventing antenatal depression cases in the sample as a whole (i.e., universal prevention) only.

2.9.2. Aim 2:

Cost-outcome analysis of ROSE as universal vs. selective vs. indicated prevention in terms of PPD costs and functioning (VR-12). We will compute cost-outcome measures across the prevention strategies and screening tools from societal and health systems perspectives. Net cost analysis will compare screening costs against ROSE delivery costs to those screening negative net of cost savings from reducing PPD among false negatives. Societal costs will sum participant costs of time spent on the intervention plus the health systems costs of screening and ROSE delivery. In the universal arm, cost savings from reduced PPD will include medical cost savings from the health systems perspective plus wage and quality of life losses avoided. We will use our existing data on health system cost of screening and of ROSE delivery. Screening costs will be allocated to those screening negative by dividing screening costs per patient by the percent of screenees screening negative. Quality of life savings will quantify the changes in VR-12 scores as monetized QALYs⁴² in net cost and benefit-cost measures and unmonetized in a cost/QALY saved. We will report the costs and benefits of each alternative approach, showing how much each saves or costs the health system and society, and which approach has the highest return per dollar spent. We will determine the 95% uncertainty interval around outcome measures and conduct sensitivity analyses.

2.9.3. Aim 3:

Analysis of equity of ROSE as universal, selective, or indicated prevention will examine whether any of these approaches disadvantage people from racial/ethnic minority groups. Aim 3a. Using the control group, we will estimate percent false negative screens for minority vs. non-Hispanic white people for each risk indicator (Medicaid, CSQ+, SCID past MDE, EHR past MDE, EPDS+). False negative rates for non-Hispanic white people vs. minority people generally (primary) and African American people specifically (secondary) will be compared using chi-square tests (along with any group with sufficient representation in the sample: anticipated to ~10% Hispanic, 33% African American, 50% non-Hispanic white). Aim 3b. Analysis of minority status as a moderator of ROSE PPD outcome for each group listed in Table 2 will follow the analysis plan from Aim 1 with minority status and trial arm by minority status interaction added to the list of explanatory variables. We will combine findings in Aims 3a and 3b in Aim 3c. The ultimate reflection of equity for each indicator (i.e., CSQ+, EPDS+, etc.) will be characterized as missed opportunity (rates of PPD cases that would have been prevented under universal but not selective or indicated prevention) for those who screen false negative under selective or indicated prevention, whose PPD would have been prevented if ROSE was universal. For each indictor, we will estimate the proportion of these cases as (PPD reduction rate with ROSE among screened negative)*(no risk rate)*(false negative rate based on the control arm) for non-Hispanic white vs. all minority (primary), and vs. African American (secondary) people and compare them.

2.9.4. Aim 4: Scalability of ROSE universal vs. indicated prevention.

Qualitative and quantitative data will be integrated to describe which kind of prevention is most scalable, in which settings, and why. Cost-outcome comparisons of ROSE as universal, selective or indicated prevention are described in Aim 2. Quantitative. The three ISAT scores (reflecting perceived scalability of universal, selective or indicated prevention) from each from 98 agencies who previously implemented ROSE¹² will be compared using linear mixed effects models with the random effect corresponding to agency. Qualitative. Recordings of qualitative interviews¹² will be transcribed and coded. Deductive codes will be drawn from interview question topics and scalability dimensions listed in Table 6. Inductive codes capturing emergent themes will arise from team-level review of the transcripts. Once codes are developed, coding team members will independently code remaining transcripts; 20% of the transcripts will be double coded and reviewed to ensure coding fidelity. Team members will perform thematic analysis⁴³ in which all of the passages assigned to codes will be read in aggregate to identify key themes. Mixed methods. We will sort qualitative data by prevention approach implemented (universal, selective, indicated), agency type (e.g., OBGYN, Healthy Start), and/or scores on ISAT domains to examine response patterns and provide context and triangulation.

2.9.5. Aim 5: Mechanisms of ROSE effects across PPD risk levels.

Overall mechanisms.

The hypothesis of mediation with universal prevention will be tested using the Preacher and Hayes⁴⁴ approach to estimate direct and indirect (through the mediator) effects of the trial arm (ROSE v. ECAU) on the outcome at month 6. The test will be performed using a bias corrected bootstrapping analytic strategy^45,46 based on 5,000 bootstrap samples. Do mechanisms operate differently at different PPD risk levels? To explore whether Medicaid, CSQ positivity, past MDE (SCID5, EHR) or EPDS positivity moderate ROSE v. ECAU effects on hypothesized mediators (i.e., instrumental support, perceived stress), we will estimate conditional indirect effects.

2.10. Recommendations.

We will integrate results from all aims into a policymaker-friendly table describing pros and cons of universal, selective, and indicated PPD prevention.

3. Conclusion

This will be the first trial to assess universal vs. selective or indicated PPD prevention. Trial design illustrates a novel and efficient way to make these comparisons, including across multiple potential risk screening tools. This low-touch, high volume pragmatic trial will be the largest PPD prevention trial to date, and will examine scalability, an understudied area of implementation science. It will also be the first study to explore ROSE as antenatal depression prevention and to examine whether EPDS scores during pregnancy predict future PPD.

Research usually moves along the translational spectrum from effectiveness to implementation research. However, this line of research demonstrates a case in which effectiveness research led to implementation research and then back to a hybrid effectiveness-implementation trial. Specifically, after randomized trials demonstrating effectiveness of ROSE among those at risk for PPD,^4–7 an ongoing implementation trial¹² suggested a potentially different course (i.e., universal prevention) from the one in USPSTF recommendations^1–3 and National Institutes of Health guidance⁴⁷ on next research steps (i.e., find better risk prediction tools for selective prevention). This trial (which blends effectiveness and implementation science approaches) is the resulting next step.

Supplementary Material

NIHMS1922709-supplement-1.docx^{(12.7KB, docx)}

NIHMS1922709-supplement-2.docx^{(16.4KB, docx)}

Acknowledgments.

We would like to acknowledge the trial research staff (Alanna Foulon, Haneen Hammad, Takeya Harris, Shardae Herriford, Rowyda Kazan, Quincy Kittle, Ellen Nixon, and Nichole Rekowski) and qualitative analysts (Garrett Brown and Aisling Nolan). We would also like to thank Raven Miller, project coordinator in the Rose Sustainment Study¹² and the 98 agencies who worked to implement ROSE in that trial. We would also like to thank our HFH clinical partners and their patients. Finally, we would like to acknowledge Dr. James Dearing for helping to frame our dissemination approach.

Funding, oversight, and trial registration.

This pragmatic randomized trial was funded by the National Institute of Mental Health (NIMH; R01 MH130948; Principal Investigators Johnson and Zlotnick). NIMH did not participate in study design; in data collection, analysis, or interpretation; in writing the report; or in the decision to submit the article for publication. This study is overseen by the Michigan State University Biomedical Institutional Review Board (FWA #00004556). The trial is registered at clinicaltrials.gov (NCT05700760).

Dissemination policy.

Standard academic venues. Data sharing will occur on National Institutes of Health-designated timelines. Publications will be publicized using university communication services, Twitter, and professional organizations. Community dissemination. We will share results with Henry Ford Health (the trial setting) through written reports and community talks. We will offer to share final study results with study participants (both patient participants and representatives of the 98 agencies participating in the previous ROSE implementation trial).

Dissemination to policy-makers. We will integrate results from all aims into a policymaker-friendly table and accompanying document that outlines ROSE effectiveness, cost-benefit, equity, scalability, and mechanisms across prevention approaches (universal, selective, indicated). It may look something like a 3×5 tabular “pros and cons” list. It will also include pros and cons of each tested screening tool relative to each other and to universal prevention. We will circulate this policy brief to professional organizations, to the USPSTF, to advocacy groups, and to relevant legislative advocates.

Effective dissemination targets key influencers with concept-based messages. Policy entrepreneurs are boundary-spanning individuals who advocate with high-status persons for the adoption of evidence-based practices and policies. They identify solutions to salient problems, work to exploit political windows of opportunity, frame solutions to problems in politically palatable ways, and join together disparate individuals, groups and networks to diffuse policies. Individuals who function as policy entrepreneurs may be state representatives, nonprofit leaders, researchers, chief executive officers, or experts within a profession. We will communicate findings and approaches with individuals that we identify as policy entrepreneurs for their consideration in their interactions with elected officials, legislative staff and other experts.

Footnotes

Competing interests. The authors have no competing interests.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Contributor Information

Jennifer E. Johnson, Charles Stewart Mott Department of Public Health, Michigan State University College of Human Medicine, 200 East 1^st St Room 366, Flint, MI 48502.

Amy M. Loree, Center for Health Policy & Health Services Research, Henry Ford Health, 1 Ford Place, Suite 5E, Detroit, MI 48220.

Alla Sikorskii, Department of Psychiatry, Michigan State University College of Osteopathic Medicine, 909 Wilson Rd, East Lansing, MI 48824, USA..

Ted R. Miller, Pacific Institute for Research and Evaluation, 11720 Beltsville Drive Suite 900, Calverton, MD 20705.

Laura Carravallah, Department of Pediatrics and Human Development, Michigan State University College of Human Medicine, 200 East 1^st St, Flint, MI 48502..

Brandon Taylor, Charles Stewart Mott Department of Public Health, Michigan State University College of Human Medicine, 200 East 1^st St, Flint, MI 48502, USA..

Caron Zlotnick, Butler Hospital and Women and Infants Hospital, 345 Blackstone Blvd, Providence, RI 02906, USA. University of Cape Town, Cape Town, South Africa..

References

1.US Preventive Services Task Force. Intervention to prevent perinatal depression: US Preventive Services Task Force Recommendation Statement. JAMA. 2019;321(6):580–587. [DOI] [PubMed] [Google Scholar]
2.O’Connor E, Senger CA, Henniger ML, Coppola E, Gaines BN. Interventions to prevent perinatal depression: Evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2019;321(6):588–601. [DOI] [PubMed] [Google Scholar]
3.O’Connor E, Senger CA, Henniger ML, Gaynes BN, Coppola E, Weyrich MS. Interventions to prevention perinatal depression: A systematic evidence review for the U.S. Preventive Services Task Force. 2018. [Google Scholar]
4.Zlotnick C, Johnson SL, Miller IW, Pearlstein T, Howard M. Postpartum depression in women receiving public assistance: Pilot study of an interpersonal-therapy-oriented group intervention. American Journal of Psychiatry. Apr 2001;158(4):638–640. [DOI] [PubMed] [Google Scholar]
5.Zlotnick C, Miller I, Pearlstein T, Howard M, Sweeney P. A preventive intervention for pregnant women on public assistance at the risk for postpartum depression. American Journal of Psychiatry. 2006;163(8):1443–1445. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Crockett K, Zlotnick C, Davis M, Payne N, Washington R. A depression preventive intervention for rural low-income African-American pregnant women at risk for postpartum depression. Arch Womens Ment Health. 2008;11(5–6):319–25. [DOI] [PubMed] [Google Scholar]
7.Phipps MG, Raker CA, Ware CF, Zlotnick C. Randomized controlled trial to prevent postpartum depression in adolescent mothers. Research. American Journal of Obstetrics & Gynecology March 2013 2013;208(3):192.e1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Tandon SD, Leis JA, Medelson T, Perry DF, Kemp K. Six-month outcomes from a randomized controlled trial to prevent perinatal depression in low-income home visiting clients. Matern Child Health J. 2014;18(4):873–881. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Tandon SD, Perry DF, Mendelson T, Kemp K, Leis JA. Preventing perinatal depression in low-income home visiting clients: a randomized controlled trial. J Consult Clin Psychol. 2011;79(5):707–712. [DOI] [PubMed] [Google Scholar]
10.Le HN, Perry DF, Stuart EA. Randomized controlled trial of a preventive intervention for perinatal depression in high-risk Latinas. J Consult Clin Psychol. 2011;79(2):135–141. [DOI] [PubMed] [Google Scholar]
11.Muñoz R Le HN, Ippen C, Diaz M, Urzar G, Soto J. Prevention of postpartum depression in low-income women: development of the Mamás y Bebés/Mothers and Babies course. Cogn Behav Pract. 2007;14:70–83. [Google Scholar]
12.Johnson JE, Wiltsey-Stirman S, Sikorskii A, Miller T, King A, Blume JL, Pham X, Moore Simas TA, Poleshuck E, Weinberg R, Zlotnick C. Protocol for the ROSE sustainment (ROSES) study, a sequential multiple assignment randomized trial to determine the minimum necessary intervention to maintain a postpartum depression prevention program in prenatal clinics serving low-income women. Implementation Science. 2018/08/22 2018;13(1):115. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Luca DL, Margiotta C, Staatz C, Garlow E, Christensen A, Zivin K. Financial Toll of Untreated Perinatal Mood and Anxiety Disorders Among 2017 Births in the United States. Am J Public HEalth. 2020;110(6):888–896. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.O’Hara MW, Rehm LP, Campbell SB. Postpartum depression: A role for social network and life stress variables. Journal of Nervous Mental Diseases. 1983;171:336–41. [PubMed] [Google Scholar]
15.Meltzer-Brody S, Kanes SJ. Allopregnanolone in postpartum depression: Role in pathophysiology and treatment. Neurobiol Stress. 2020;12(2020 Feb 3):100212. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Hobfoll SE, Ritter C, Lavin J, Hulsizer M, Cameron R. Depression prevalence and incidence among inner-city pregnant and postpartum women. Journal of Consulting and Clinical Psychology. 1995;63(3):445–453. [DOI] [PubMed] [Google Scholar]
17.Scholle SH, Haskett RF, Hanusa BH, Pincus HA, Kupfer DJ. Addressing depression in obstetrics/gynecology practice. General Hospital Psychiatry. Mar-Apr 2003;25(2):83–90. [DOI] [PubMed] [Google Scholar]
18.Cooper PJ, Murray L, Hooper R, West A. The development and validation of a predictive index for postpartum depression. Psychological Medicine. 1996;26:628–634. [DOI] [PubMed] [Google Scholar]
19.Cooper PJ, Murray L, Hooper R, West A. The development and validation of a predictive index for postpartum depression. Psychological Medicine. 1996;26:628–634. [DOI] [PubMed] [Google Scholar]
20.Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: Development of the Edinburgh Postnatal Depression Scale. British Journal of Psychiatry. 1987;150:782–786. [DOI] [PubMed] [Google Scholar]
21.Long MM, Cramer RJ, Bennington L, Morgan FG Jr, Wilkes CA, Fontanares AJ, Sadr N, Bertolino SM, Paulson JF. Psychometric assessment of the Edinburgh Postnatal Depression Scale in an obstetric population. Psychiatry Research. 2020;291:113161. [DOI] [PubMed] [Google Scholar]
22.Center for Women’s Mental Health. MGH Center for Women’s Mental Health. Accessed 10–26-21, 2021. https://womensmentalhealth.org/
23.Earls M, Committee on Psychosocial Aspects of Child and Family Health, American Academy of Pediatrics. Incorporating recognition and management of perinatal and postpartum depression into pediatric practice. Pediatrics. 2010;126(5):1032–1039. [DOI] [PubMed] [Google Scholar]
24.Postpartum Support International. PSI Lauds USPSTF Report Calling for Perinatal Screenings. Accessed 10–25-21, 2021. https://www.postpartum.net/psi-lauds-report-calling-for-screenings/
25.Zlotnick C, Tzilos G, Miller I, Seifer R, Stout R. Randomized controlled trial to prevent postpartum depression in mothers on public assistance. Journal of Affective Disorders. January 1, 2016 2016;189:263–268. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Johansen S, Stenhaug BA, Robakis TK, Williams KE, Cullen MR. Past psychiatric conditions as risk factors for postpartum depression: A nationwide cohort study. J Clin Psychiatry. 2020;81(1):e1–e7. [DOI] [PubMed] [Google Scholar]
27.O’Hara M. Postpartum depression: what we know. J Clin Psychol 2009;65:1258–69. [DOI] [PubMed] [Google Scholar]
28.First M, Williams JBW, Karg RS, Spitzer RL. Structured Clinical Interview for DSM-5. American Psychiatric Association; 2015. [Google Scholar]
29.Harris B HP, Thomas R, Johns S, Fung H. The validation of the Edinburgh Post-natal Depression Scale on a community sample. Br J Psychiatry. 1989;157(2):288–290. [DOI] [PubMed] [Google Scholar]
30.Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J, McDonald-Scott NC, Andreasen NC. The longitudinal interval follow-up evaluation:A comprehensive method for assessing outcome in prospective longitudinal studies. Archives of General Psychiatry. 1987;44(6):540–548. [DOI] [PubMed] [Google Scholar]
31.Past Keller M., present, and future directions for defining optimal treatment outcome in depression: Remission and beyond. JAMA. 2003;289(3152–3160) [DOI] [PubMed] [Google Scholar]
32.Burdine JN, Felix MR, Abel AL, Wiltraut CJ, Musselman YJ. The SF-12 as a population health measure: An exploratory examination of potential for application. Health Serv Res. Oct 2000;35(4):885–904. [PMC free article] [PubMed] [Google Scholar]
33.Linehan M, Heard H. Treatment History Interview-4 (THI-4). University of Washington; 1996. [Google Scholar]
34.Sengupta N, Nichol MB, Wu J, Globe D. Mapping the SF-12 to the HUI3 and VAS in a managed care population. Med Care. 2004;42(9):927–937. [DOI] [PubMed] [Google Scholar]
35.Milat A, Lee K, Conte K, Grunseit A, Wolfenden L, Nassau F, Orr N, Sreeram P, Bauman A. Intervention Scalability Assessment Tool: A decision support tool for health policy makers and implementers. Health Research Policy and Systems. 2020;18:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Cella D, Choi S, Garcia S, Cook KF, Rosenbloom S, Lai JS, Tatum DS, Gerson R. Setting standards for severity of common symptoms in oncology using the PROMIS item banks and expert judgment. Qual Life Res. 2014. Dec;23(10):2651–61 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Cella D, Riley W, Stone A, et al. Initial adult health item banks and first wave testing of the Patient-Reported Outcomes Measurement Information System (PROMIS) network: 2005–2008. Journal of Clinical Epidemiology. 2010;63:1179–1194. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.PROMIS Health Organization and PROMIS Cooperative Group. Instrumental Support - Short Form 8a. Accessed 10–11-21, 2021. https://www.healthmeasures.net/index.php?option=com_instruments&view=measure&id=205&Itemid=992
39.Lee E. Review of the psychometric evidence of the Perceived Stress Scale. Asian Nursing Research. 2012;6(121–127):121–127. [DOI] [PubMed] [Google Scholar]
40.Cohen S, Williamson G. Perceived stress in a probability sample in the United States. In: Spacapan S, Oskamp S, ed. The Social Psychology of Health. Sage; 1988. [Google Scholar]
41.Cohen S. Perceived Stress Scale. Accessed 10–18-21, 2021. https://www.northottawawellnessfoundation.org/wp-content/uploads/2018/04/PerceivedStressScale.pdf
42.Cutler DM, Richardson E. The value of health: 1970–1990. American Economic Review. 1989;88(s):97–100. [Google Scholar]
43.Braun V, Clark V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3:77–101. [Google Scholar]
44.Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav Res Methods Instrum Comput 2004;36(4):717–31. [DOI] [PubMed] [Google Scholar]
45.Preacher KJ, Selig JP. Advantages of Monte Carlo confidence intervals for indirect effects. Communication Methods and Measures. 2012;6:77–98. [Google Scholar]
46.Biesanz JC, Falk CF, Savalei V. Assessing mediational models: Testing and interval estimation for indirect effects. Multivariate Behavioral Research. 2010;45:661–701. [DOI] [PubMed] [Google Scholar]
47.National Institute of Mental Health. Prevention of Perinatal Depression: Improving Intervention Delivery for At-Risk Individuals (R01 Clinical Trial Required). Accessed 8–15-21, 2021. https://grants.nih.gov/grants/guide/rfa-files/RFA-MH-21-240.html
48.Barker PM, Reid A, Schall MW. A framework for scaling up health interventions: lessons from large-scale improvement initiatives in Africa. Implementation Science. 2015;11:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Parry G. Evolving the IHI Scale-up Framework. Accessed 10–2-2020, 2020. ihi.org/communities/blogs/evolving-the-ihi-scale-up-framework
50.Zamboni K, Schellenberg J, Hanson C, Betran AP, Dumont A. Assessing scalability of an intervention: Why, how, and who? Health Policy Plan. 2019;34(7):544–552. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Milat A, Bauman A, Redman S. The concept of scalability: increasing the scale and potential adoption of health promotion interventions into policy and practice. Health Promotion International. 2013;28:285–298. [DOI] [PubMed] [Google Scholar]
52.Milat A, Bauman A, Redman S. Narrative review of models and success factors for scaling up public health interventions. Implementation Science. 2015;10:113. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS1922709-supplement-1.docx^{(12.7KB, docx)}

NIHMS1922709-supplement-2.docx^{(16.4KB, docx)}

[R1] 1.US Preventive Services Task Force. Intervention to prevent perinatal depression: US Preventive Services Task Force Recommendation Statement. JAMA. 2019;321(6):580–587. [DOI] [PubMed] [Google Scholar]

[R2] 2.O’Connor E, Senger CA, Henniger ML, Coppola E, Gaines BN. Interventions to prevent perinatal depression: Evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2019;321(6):588–601. [DOI] [PubMed] [Google Scholar]

[R3] 3.O’Connor E, Senger CA, Henniger ML, Gaynes BN, Coppola E, Weyrich MS. Interventions to prevention perinatal depression: A systematic evidence review for the U.S. Preventive Services Task Force. 2018. [Google Scholar]

[R4] 4.Zlotnick C, Johnson SL, Miller IW, Pearlstein T, Howard M. Postpartum depression in women receiving public assistance: Pilot study of an interpersonal-therapy-oriented group intervention. American Journal of Psychiatry. Apr 2001;158(4):638–640. [DOI] [PubMed] [Google Scholar]

[R5] 5.Zlotnick C, Miller I, Pearlstein T, Howard M, Sweeney P. A preventive intervention for pregnant women on public assistance at the risk for postpartum depression. American Journal of Psychiatry. 2006;163(8):1443–1445. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Crockett K, Zlotnick C, Davis M, Payne N, Washington R. A depression preventive intervention for rural low-income African-American pregnant women at risk for postpartum depression. Arch Womens Ment Health. 2008;11(5–6):319–25. [DOI] [PubMed] [Google Scholar]

[R7] 7.Phipps MG, Raker CA, Ware CF, Zlotnick C. Randomized controlled trial to prevent postpartum depression in adolescent mothers. Research. American Journal of Obstetrics & Gynecology March 2013 2013;208(3):192.e1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Tandon SD, Leis JA, Medelson T, Perry DF, Kemp K. Six-month outcomes from a randomized controlled trial to prevent perinatal depression in low-income home visiting clients. Matern Child Health J. 2014;18(4):873–881. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Tandon SD, Perry DF, Mendelson T, Kemp K, Leis JA. Preventing perinatal depression in low-income home visiting clients: a randomized controlled trial. J Consult Clin Psychol. 2011;79(5):707–712. [DOI] [PubMed] [Google Scholar]

[R10] 10.Le HN, Perry DF, Stuart EA. Randomized controlled trial of a preventive intervention for perinatal depression in high-risk Latinas. J Consult Clin Psychol. 2011;79(2):135–141. [DOI] [PubMed] [Google Scholar]

[R11] 11.Muñoz R Le HN, Ippen C, Diaz M, Urzar G, Soto J. Prevention of postpartum depression in low-income women: development of the Mamás y Bebés/Mothers and Babies course. Cogn Behav Pract. 2007;14:70–83. [Google Scholar]

[R12] 12.Johnson JE, Wiltsey-Stirman S, Sikorskii A, Miller T, King A, Blume JL, Pham X, Moore Simas TA, Poleshuck E, Weinberg R, Zlotnick C. Protocol for the ROSE sustainment (ROSES) study, a sequential multiple assignment randomized trial to determine the minimum necessary intervention to maintain a postpartum depression prevention program in prenatal clinics serving low-income women. Implementation Science. 2018/08/22 2018;13(1):115. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Luca DL, Margiotta C, Staatz C, Garlow E, Christensen A, Zivin K. Financial Toll of Untreated Perinatal Mood and Anxiety Disorders Among 2017 Births in the United States. Am J Public HEalth. 2020;110(6):888–896. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.O’Hara MW, Rehm LP, Campbell SB. Postpartum depression: A role for social network and life stress variables. Journal of Nervous Mental Diseases. 1983;171:336–41. [PubMed] [Google Scholar]

[R15] 15.Meltzer-Brody S, Kanes SJ. Allopregnanolone in postpartum depression: Role in pathophysiology and treatment. Neurobiol Stress. 2020;12(2020 Feb 3):100212. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Hobfoll SE, Ritter C, Lavin J, Hulsizer M, Cameron R. Depression prevalence and incidence among inner-city pregnant and postpartum women. Journal of Consulting and Clinical Psychology. 1995;63(3):445–453. [DOI] [PubMed] [Google Scholar]

[R17] 17.Scholle SH, Haskett RF, Hanusa BH, Pincus HA, Kupfer DJ. Addressing depression in obstetrics/gynecology practice. General Hospital Psychiatry. Mar-Apr 2003;25(2):83–90. [DOI] [PubMed] [Google Scholar]

[R18] 18.Cooper PJ, Murray L, Hooper R, West A. The development and validation of a predictive index for postpartum depression. Psychological Medicine. 1996;26:628–634. [DOI] [PubMed] [Google Scholar]

[R19] 19.Cooper PJ, Murray L, Hooper R, West A. The development and validation of a predictive index for postpartum depression. Psychological Medicine. 1996;26:628–634. [DOI] [PubMed] [Google Scholar]

[R20] 20.Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression: Development of the Edinburgh Postnatal Depression Scale. British Journal of Psychiatry. 1987;150:782–786. [DOI] [PubMed] [Google Scholar]

[R21] 21.Long MM, Cramer RJ, Bennington L, Morgan FG Jr, Wilkes CA, Fontanares AJ, Sadr N, Bertolino SM, Paulson JF. Psychometric assessment of the Edinburgh Postnatal Depression Scale in an obstetric population. Psychiatry Research. 2020;291:113161. [DOI] [PubMed] [Google Scholar]

[R22] 22.Center for Women’s Mental Health. MGH Center for Women’s Mental Health. Accessed 10–26-21, 2021. https://womensmentalhealth.org/

[R23] 23.Earls M, Committee on Psychosocial Aspects of Child and Family Health, American Academy of Pediatrics. Incorporating recognition and management of perinatal and postpartum depression into pediatric practice. Pediatrics. 2010;126(5):1032–1039. [DOI] [PubMed] [Google Scholar]

[R24] 24.Postpartum Support International. PSI Lauds USPSTF Report Calling for Perinatal Screenings. Accessed 10–25-21, 2021. https://www.postpartum.net/psi-lauds-report-calling-for-screenings/

[R25] 25.Zlotnick C, Tzilos G, Miller I, Seifer R, Stout R. Randomized controlled trial to prevent postpartum depression in mothers on public assistance. Journal of Affective Disorders. January 1, 2016 2016;189:263–268. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Johansen S, Stenhaug BA, Robakis TK, Williams KE, Cullen MR. Past psychiatric conditions as risk factors for postpartum depression: A nationwide cohort study. J Clin Psychiatry. 2020;81(1):e1–e7. [DOI] [PubMed] [Google Scholar]

[R27] 27.O’Hara M. Postpartum depression: what we know. J Clin Psychol 2009;65:1258–69. [DOI] [PubMed] [Google Scholar]

[R28] 28.First M, Williams JBW, Karg RS, Spitzer RL. Structured Clinical Interview for DSM-5. American Psychiatric Association; 2015. [Google Scholar]

[R29] 29.Harris B HP, Thomas R, Johns S, Fung H. The validation of the Edinburgh Post-natal Depression Scale on a community sample. Br J Psychiatry. 1989;157(2):288–290. [DOI] [PubMed] [Google Scholar]

[R30] 30.Keller MB, Lavori PW, Friedman B, Nielsen E, Endicott J, McDonald-Scott NC, Andreasen NC. The longitudinal interval follow-up evaluation:A comprehensive method for assessing outcome in prospective longitudinal studies. Archives of General Psychiatry. 1987;44(6):540–548. [DOI] [PubMed] [Google Scholar]

[R31] 31.Past Keller M., present, and future directions for defining optimal treatment outcome in depression: Remission and beyond. JAMA. 2003;289(3152–3160) [DOI] [PubMed] [Google Scholar]

[R32] 32.Burdine JN, Felix MR, Abel AL, Wiltraut CJ, Musselman YJ. The SF-12 as a population health measure: An exploratory examination of potential for application. Health Serv Res. Oct 2000;35(4):885–904. [PMC free article] [PubMed] [Google Scholar]

[R33] 33.Linehan M, Heard H. Treatment History Interview-4 (THI-4). University of Washington; 1996. [Google Scholar]

[R34] 34.Sengupta N, Nichol MB, Wu J, Globe D. Mapping the SF-12 to the HUI3 and VAS in a managed care population. Med Care. 2004;42(9):927–937. [DOI] [PubMed] [Google Scholar]

[R35] 35.Milat A, Lee K, Conte K, Grunseit A, Wolfenden L, Nassau F, Orr N, Sreeram P, Bauman A. Intervention Scalability Assessment Tool: A decision support tool for health policy makers and implementers. Health Research Policy and Systems. 2020;18:1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Cella D, Choi S, Garcia S, Cook KF, Rosenbloom S, Lai JS, Tatum DS, Gerson R. Setting standards for severity of common symptoms in oncology using the PROMIS item banks and expert judgment. Qual Life Res. 2014. Dec;23(10):2651–61 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Cella D, Riley W, Stone A, et al. Initial adult health item banks and first wave testing of the Patient-Reported Outcomes Measurement Information System (PROMIS) network: 2005–2008. Journal of Clinical Epidemiology. 2010;63:1179–1194. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.PROMIS Health Organization and PROMIS Cooperative Group. Instrumental Support - Short Form 8a. Accessed 10–11-21, 2021. https://www.healthmeasures.net/index.php?option=com_instruments&view=measure&id=205&Itemid=992

[R39] 39.Lee E. Review of the psychometric evidence of the Perceived Stress Scale. Asian Nursing Research. 2012;6(121–127):121–127. [DOI] [PubMed] [Google Scholar]

[R40] 40.Cohen S, Williamson G. Perceived stress in a probability sample in the United States. In: Spacapan S, Oskamp S, ed. The Social Psychology of Health. Sage; 1988. [Google Scholar]

[R41] 41.Cohen S. Perceived Stress Scale. Accessed 10–18-21, 2021. https://www.northottawawellnessfoundation.org/wp-content/uploads/2018/04/PerceivedStressScale.pdf

[R42] 42.Cutler DM, Richardson E. The value of health: 1970–1990. American Economic Review. 1989;88(s):97–100. [Google Scholar]

[R43] 43.Braun V, Clark V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3:77–101. [Google Scholar]

[R44] 44.Preacher KJ, Hayes AF. SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behav Res Methods Instrum Comput 2004;36(4):717–31. [DOI] [PubMed] [Google Scholar]

[R45] 45.Preacher KJ, Selig JP. Advantages of Monte Carlo confidence intervals for indirect effects. Communication Methods and Measures. 2012;6:77–98. [Google Scholar]

[R46] 46.Biesanz JC, Falk CF, Savalei V. Assessing mediational models: Testing and interval estimation for indirect effects. Multivariate Behavioral Research. 2010;45:661–701. [DOI] [PubMed] [Google Scholar]

[R47] 47.National Institute of Mental Health. Prevention of Perinatal Depression: Improving Intervention Delivery for At-Risk Individuals (R01 Clinical Trial Required). Accessed 8–15-21, 2021. https://grants.nih.gov/grants/guide/rfa-files/RFA-MH-21-240.html

[R48] 48.Barker PM, Reid A, Schall MW. A framework for scaling up health interventions: lessons from large-scale improvement initiatives in Africa. Implementation Science. 2015;11:12. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Parry G. Evolving the IHI Scale-up Framework. Accessed 10–2-2020, 2020. ihi.org/communities/blogs/evolving-the-ihi-scale-up-framework

[R50] 50.Zamboni K, Schellenberg J, Hanson C, Betran AP, Dumont A. Assessing scalability of an intervention: Why, how, and who? Health Policy Plan. 2019;34(7):544–552. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Milat A, Bauman A, Redman S. The concept of scalability: increasing the scale and potential adoption of health promotion interventions into policy and practice. Health Promotion International. 2013;28:285–298. [DOI] [PubMed] [Google Scholar]

[R52] 52.Milat A, Bauman A, Redman S. Narrative review of models and success factors for scaling up public health interventions. Implementation Science. 2015;10:113. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Study protocol for the ROSE Scale-Up Study: Informing a decision about ROSE as universal postpartum depression prevention

Jennifer E Johnson

Amy M Loree

Alla Sikorskii

Ted R Miller

Laura Carravallah

Brandon Taylor

Caron Zlotnick

Abstract

Purpose.

Background.

Design.

Conclusion.

1. Introduction

Table 1.

2. Methods

2.1. Rationale for trial design.

2.1.1. Aim 1. ROSE effectiveness for each PPD prevention approach.

Table 2.

2.1.2. Aim 2.

Table 3.

2.1.3. Aim 3.

2.1.4. Aim 4.

2.1.5. Aim 5.

2.1.6. Why not randomize agencies to prevention approaches?

2.2. Measures to be used for selective and indicated prevention in this pragmatic trial

2.3. Intervention conditions.

2.3.1. Control: Enhanced Care as Usual (ECAU)

2.3.2. ROSE.

2.4. Provider participants

2.5. Patient participants

2.6. Allocation, randomization, and masking.

2.7. Sample size and power.

Table 4.

2.8. Assessments.

Table 5.

2.8.1. Aim 1. ROSE effectiveness for each PPD prevention approach.

PPD risk: selective prevention.

PPD risk: indicated prevention.

Outcomes. PPD and antenatal depression cases.

2.8.2. Aim 2. Comparative cost-outcome analysis of prevention approaches.

2.8.3. Aim 3.

2.8.4. Aim 4. Scalability of ROSE universal vs. selective or indicated prevention.

Table 6.

2.8.5. Aim 5:

2.9. Analyses.

2.9.1. Aim 1: ROSE effectiveness for each PPD prevention approach.

Missing Data.

Outcomes.

Primary.

2.9.2. Aim 2:

2.9.3. Aim 3:

2.9.4. Aim 4: Scalability of ROSE universal vs. indicated prevention.

2.9.5. Aim 5: Mechanisms of ROSE effects across PPD risk levels.

Overall mechanisms.

2.10. Recommendations.

3. Conclusion

Supplementary Material

Acknowledgments.

Funding, oversight, and trial registration.

Dissemination policy.

Footnotes

Contributor Information

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases