Abstract
Importance:
Although safe sleep practices reduce the risk of sudden unexpected infant death (SUID), US rates of adherence remain low, particularly among low-income families.
Objective:
Evaluate the effectiveness of prenatal and postnatal text-delivered video safe sleep interventions for WIC participants.
Design, Setting, and Participants:
Unblinded 4-group RCT of pregnant mothers recruited 05/2022–12/2024 at WIC and Federally Qualified Health Centers in 18 states, with 2 stages of randomization (at 34 weeks’ gestation and at delivery) to safe sleep or control interventions.
Interventions:
Short educational videos promoting safe sleep (intervention) or breastfeeding (control) were delivered via text messages in the prenatal period from 34 weeks’ gestation, the postnatal period until 60 days after birth, or both.
Main Outcomes and Measures:
Primary outcomes were maternal report of 4 infant sleep practices (usual supine sleep position, usual roomsharing without bedsharing, soft bedding nonuse, any pacifier use) at 60+ days after birth. Secondary outcomes included exclusive supine position and roomsharing without bedsharing (collected at 60+ days), prenatal planned exclusive supine position and roomsharing without bedsharing (reported at 36 weeks’ gestation), and time-to-initial-report of non-exclusive supine, non-exclusive roomsharing without bedsharing, and soft bedding use (collected through weekly text queries).
Results:
Of 1790 (86%) participants randomized, 1383 (77%) completed the 60+ day survey and were included in the analysis. Of these, 818 (59.1%) were 24–34 years old; 598 (43.2%) White, 323 (23.4%) Black, and 325 (23.5%) Hispanic. Participants were assigned to prenatal and postnatal intervention (n=337), prenatal intervention alone (n=344), postnatal intervention alone (n=334), or control (n=348); 12% of all participants received the intervention in Spanish.
Among the primary outcomes, depending on the group, 87.6% - 92.2% of mothers reported usual supine position, 83.8%−89.5% of mothers reported usual roomsharing without bedsharing, 75.6%−84.5% reported no soft bedding use, and 72.5–76.0% reported pacifier use. There were no significant differences among groups for the primary outcomes. Among the 7 secondary outcomes, reported exclusive supine positioning at 60 days was higher among participants receiving the prenatal intervention than controls (70.6% (240/343) vs. 60.9% (212/348); adjusted risk difference 9.7%, 95% CI 3.8–15.0).
Conclusions and Relevance:
A text-delivered video intervention to promote safe sleep practices among low-income families in the prenatal and postnatal periods did not have significant effects on the primary outcomes.
Clinicaltrials.gov registration
# NCT04387552; date of registration March 2, 2023
INTRODUCTION
Sudden Unexpected Infant Death (SUID), which includes Sudden Infant Death Syndrome (SIDS), ill-defined causes of death, and accidental suffocation and strangulation in bed, is a leading cause of postneonatal mortality in the US, with >3500 deaths annually.1 Compared with higher-income families, low-income families have higher SUID rates1–3 and lower rates of adherence4–8 to safe sleep practices, including supine (back) positioning, roomsharing (with parents) without bedsharing, no soft bedding, and pacifier use, which decrease SUID risk and are recommended by the American Academy of Pediatrics (AAP).9 There is overlap of race, ethnicity, and income with regard to these differences.10,11
Safe sleep education and interventions typically are focused on the postnatal period. However, the Theory of Planned Behavior states intention (planned behavior), influenced by attitudes, social norms, and perceived control, is a critical step before actual behavior.12 This suggests that delivery of parent education before birth may increase intention regarding infant safe sleep and result in greater adherence. In addition, the higher prevalence of unsafe sleep practices4–8 and SUID1–3 among low-income families suggests the need for interventions tailored to pregnant mothers with these demographics.
The current study, Social Media and Risk-Reduction Training-Enhanced Reach (SMARTER), was a randomized clinical trial to evaluate the individual and combined impact of prenatal and postnatal text-delivered safe sleep interventions among low-income families.
METHODS (Trial protocol in Supplement 1)
Study Population
Eligibility:
The target population was low-income pregnant individuals <32 weeks’ gestation who were clients at Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) centers, or who had Medicaid or no insurance. WIC eligibility (gross income ≤185% of US Poverty Income Guidelines13) is an easily verifiable proxy for income. Individuals were excluded if they were non-English- or Spanish-speaking; did not live in the US, have access to text messaging, or plan to live with the infant after birth; or had a prenatal diagnosis precluding supine infant positioning or breastfeeding.
Recruitment:
From May 2022 to December 2024, pregnant clients at participating WIC centers and federally qualified health centers (FQHCs) in 17 states and Washington DC were informed about the study by blast messages, advertisements, or flyers (see Acknowledgments for participating programs). Interested individuals scanned a QR code to learn about the study. Those meeting eligibility criteria were asked to provide informed consent and enroll in the study before 32+1/7 weeks gestation. Between May and December 2024, the same system was used to simultaneously recruit for this and another study (Figure 1). The institutional review boards at WashU Medicine, 15 WIC programs, and 1 FQHC approved the study (the remaining WIC programs and FQHC IRBs ceded to the WashU IRB). The trial was conducted in accordance with the Consolidated Standards of Reporting Trials reporting guidelines.
Figure 1:
Recruitment, randomization, and follow-up in the SMARTER trial
Randomization:
Between enrollment and 34 weeks’ gestation, participants received weekly messages about healthy pregnancy topics. Participants were randomized twice (at 34 weeks ‘gestation, with re-randomization immediately upon birth) using computer-generated assignments to receive safe sleep or control (breastfeeding) education, creating four groups (Safe Sleep prenatal/Safe Sleep postnatal, Safe Sleep/Control, and Control/Safe Sleep, and Control/Control). Randomization was stratified by WIC/FQHC site, preferred language, and pre-randomization engagement level (any video viewing and/or responses to text messages versus not), and blocked (so that the number of mothers from each WIC/FQHC site assigned to each group balanced after 4 or 8 mothers).
Intervention
The safe sleep and control videos (all branded as “TodaysBaby: Helping You Make the Best Choices for Your Baby”) were delivered via links within text messages in English or Spanish, beginning after prenatal randomization. Mothers randomized to safe sleep received videos encouraging and explaining the rationale for safe sleep. Those in the control group received videos providing guidance on breastfeeding. Both groups received videos about feeding within a safe sleep environment. Videos were delivered 1–2 times daily in the first week, then 2–3 times weekly thereafter, each <3 minutes long, addressing common concerns and barriers to adherence, and including both health care professional advice and parent testimonials.
Participants also received 2–4 text queries weekly to maintain engagement, provide immediate feedback and reinforce desired practices (based on behavioral change theories14,15), and collect near-real-time data on planned (prenatal) and actual (postnatal) infant sleep and feeding practices. Intervention group respondents received positive feedback to reinforce desired practices if their response aligned with AAP recommendations and feedback promoting AAP recommendations if it did not; control respondents received neutral acknowledgment without feedback. Participants were entered into a monthly US$100 lottery for each query response.
Families received standard counseling and educational materials about safe sleep and breastfeeding from clinicians and WIC staff.
Data Collection
Participants completed 3 surveys, one at 32+0/7–33+6/7 weeks’ gestation (“baseline,” before prenatal education), one between 36+0/7 weeks’ gestation and infant birth (following 2–3 weeks of prenatal education), and one sent on Day 60 of the postnatal intervention (“60+ day survey”), with validated questions about planned and actual infant care practices. On the baseline survey, participants were asked to self-report demographic information, including race/ethnicity using US Census Bureau categories, because there are longstanding differences among groups in rates of SUID1–3 and safe sleep adherence.4–8 Participants received gift cards after completing each survey.
Weekly postnatal text query responses were collected and analyzed for the time-to-event analyses.
Outcomes
Primary Outcomes:
Four usual safe sleep practices (supine position, roomsharing without bedsharing, no soft bedding use, pacifier use) were chosen as primary outcomes to facilitate comparison with prior studies. For each, mothers were asked on the 60+ day survey to report the practice that they used the most during the past 2 weeks. “Usually supine” was defined as a “back” response to the question about the position used most for sleeping or napping. “Usually roomsharing without bedsharing” required that the infant slept most often in the same room as a sleeping adult without sharing a sleep surface. “No Soft Bedding Use” was defined as a report of “none” to any items in their baby’s sleep space. “Any Pacifier Use” was defined as a “yes” to use of a pacifier when placing their baby to sleep.
Secondary Outcomes:
For safe sleep practices, exclusive supine sleep positioning is the current recommendation9 and thus the most important area for potential improvement in practices. For sleep position, 3 secondary outcomes were selected: 1) Exclusively supine, defined as a response of solely “back”, on the 60+ day survey, when asked for all positions they placed their baby for sleeping/napping in the last 2 weeks;) Time-to-initial-report of non-exclusive supine position, defined as any response other than “back” to a weekly text query: and 3) Planned use of “exclusive” supine sleep during the first month at home, defined as a solely “back” response on the 36-week gestation survey. Similar secondary outcomes were defined for sleep location: Exclusively Roomsharing without bedsharing; Time-to-initial-report of NOT using “exclusive” Roomsharing without bedsharing; and Prenatal planned use of “exclusive” Roomsharing without bedsharing during the first month at home. For soft bedding use, Time-to-initial-report of using soft bedding, was defined as the first time there was a response other than “none” to a weekly text query about all items in their baby’s sleep space.
Statistical Methods
The target sample size of 1284 participants completing the 60+ day postnatal survey was calculated based on prior study results16 to detect an intervention OR of 1.7, corresponding to a 10-percentage point increase from a baseline prevalence of 70% at 93% power (adjusting for multiple comparisons and testing at the family-wise 2 tailed p<0.05 level). Details are in the statistical analysis plan (Supplement 2).
For each primary outcome, multivariable analyses tested the hypothesis that, relative to control, those receiving a safe sleep intervention would have higher uptake of our safe sleep outcomes of interest, and that receipt of both safe sleep interventions would be more effective than either intervention alone. Each dichotomous outcome was modelled using logistic generalized estimating equations (GEE), adjusting for clustering by 92 WIC/FQHC sites. Models adjusted for infant age at the 60+ day survey, program language, and pre-randomization engagement level, which were all prespecified. The Hochberg procedure was used to adjust for multiple comparisons, maintaining a family-wise 2-tailed alpha level of 0.05 when testing for intervention effects (e.g., adjusted thresholds, in descending order 0.05, 0.025, 0.0167, 0.0125, 0.01, 0.0083, 0.0071 and 0.00625 for 8 primary hypotheses (4 primary hypotheses × 2 interventions).
Regarding secondary outcomes, the exclusive supine positioning and roomsharing without bedsharing outcomes were analyzed using similar methodology as the primary outcomes. Time-to event secondary outcomes were modeled using time-varying Cox proportional hazards models and controlled for clustering by WIC locations using a marginal approach with robust sandwich estimates. Because of departures from the proportionality assumption, particularly at early timepoints (e.g., in the first 3–4 days following initiating daily text messages), likely due to the early intervention effects for participants who were randomized to safe sleep prenatally, we present adjusted risk differences at day 60, averaging estimated survival probabilities over the observed covariate distributions, for each intervention group. We calculated risk differences across groups and used bootstrapping methods resampling with replacement (500 replicates) to obtain 95% confidence intervals for each risk difference.
To account for loss to follow-up, in a post-hoc sensitivity analysis, data were multiply imputed to account for missingness. Missing data were imputed 20 times to produce 20 complete datasets where parameters of interest were estimated for each dataset and estimated were pooled to create a single summary estimate for each parameter of interest. Analyses were performed using SAS V9.4.17 To provide information for planning future research, intraclass correlations were estimated using the GLIMMIX procedure in SAS.18
RESULTS
Study Population
A total of 1790 participants (86% of those enrolled), including 30 recruited from FQHCs, were randomized. Of those randomized, 1383 (77%) completed the 60+ day survey, defined as any response to at least one primary study outcome, and were included in the primary analysis (Figure 1). Survey non-respondents (excluded from the primary analysis) were more likely to be younger, Hispanic or non-Hispanic Black, not US-born, have lower education and income, be without a current partner and to have received the program in Spanish (Supplement 3 Table 1). Of the 60+ day survey respondents, approximately half (52.5%) completed it within 3 days; 93.1% completed it within 90 days. Sample size varied among the primary outcomes due to missing values for individual items.
Of the 1383 participants in the primary analysis, 818 (59.1%) were 24–34 years old; 598 (43.2%) were White, 323 (23.4%) Black, and 325 (23.5%) Hispanic (Table 1). Participants were assigned to prenatal and postnatal intervention (n=337), prenatal intervention alone (n=344), postnatal intervention alone (n=334), or control (n=348); 12% of all participants received the intervention in Spanish.
Table 1:
Baseline Characteristics by Intervention Group
| Intervention group (Prenatal/Postnatal) | ||||
|---|---|---|---|---|
| Characteristic | Safe Sleep/Safe Sleep (N=346) |
Safe Sleep/Control (N=350) |
Control/Safe Sleep (N=338) |
Control/Control (N=349) |
| Sex of infant (N=1356) a | ||||
| Male | 169 (50.1) | 163 (47.5) | 167 (50.6) | 185 (53.5) |
| Female | 168 (49.9) | 180 (52.5) | 163 (49.4) | 161 (46.5) |
| Parity (N=1287) a | ||||
| Nulliparous | 96 (29.9) | 113 (34.7) | 88 (28.0) | 90 (27.6) |
| 1 child | 86 (26.8) | 79 (24.2) | 91 (29.0) | 85 (26.1) |
| 2+ children | 139 (43.3) | 134 (41.1) | 135 (43.0) | 151 (46.3) |
| Maternal Age, y (N=1383) | ||||
| < 24 | 69 (19.9) | 77 (22.0) | 72 (21.3) | 66 (18.9) |
| 24–29 | 112 (32.4) | 113 (32.3) | 101 (29.9) | 117 (33.5) |
| 30–34 | 94 (27.2) | 98 (28.0) | 97 (28.7) | 86 (24.6) |
| 35+ | 71 (20.5) | 62 (17.7) | 68 (20.1) | 80 (22.9) |
| Maternal Race/ethnicity (N=1355) | ||||
| Hispanic, any race | 96 (28.4) | 74 (21.6) | 78 (23.6) | 77 (22.4) |
| Non-Hispanic, Black | 83 (24.6) | 76 (22.2) | 78 (23.6) | 86 (25.1) |
| Non-Hispanic, Otherb | 28 (8.3) | 28 (8.2) | 22 (6.6) | 31 (9.0) |
| Non-Hispanic, White | 131 (38.8) | 165 (48.1) | 153 (46.2) | 149 (43.4) |
| Maternal Country of birth (N=1369) | ||||
| USA | 271/346 (78.8) | 280/350 (81.2) | 272/338 (80.5) | 280/349 (81.9) |
| Maternal education (N=1375) | ||||
| Less than High school | 30 (8.7) | 33 (9.5) | 28 (8.3) | 36 (10.4) |
| High school graduate | 112 (32.5) | 112 (32.3) | 109 (32.3) | 105 (30.3) |
| Some college | 150 (43.5) | 144 (41.5) | 141 (41.8) | 133 (38.4) |
| Graduated 4-year College + | 53 (15.4) | 58 (16.7) | 59 (17.5) | 72 (20.8) |
| Annual household income, $ (N=1145) | ||||
| <20,000 | 128 (44.6) | 113 (39.5) | 119 (42.5) | 128 (43.8) |
| 20,000–49,999 | 128 (44.6) | 141 (49.3) | 128 (45.7) | 125 (42.8) |
| 50,000 or more | 31 (10.8) | 32 (11.2) | 33 (11.8) | 39 (13.4) |
| Current marital status (N=1383) | ||||
| Married | 114 (32.9) | 100 (28.6) | 94 (27.8) | 115 (33.0) |
| Living with a partner, but not married | 96 (27.7) | 136 (38.9) | 122 (36.1) | 126 (36.1) |
| No current partner | 90 (26.0) | 68 (19.4) | 68 (20.1) | 63 (18.1) |
| Other | 20 (5.8) | 20 (5.7) | 27 (8.0) | 19 (5.4) |
| Prefer not to answer | 26 (7.5) | 26 (7.4) | 27 (8.0) | 26 (7.4) |
| Program language (N=1383) | ||||
| English | 303/346 (87.6) | 312/350 (89.1) | 305/338 (90.2) | 309/349 (88.5) |
| Planned infant feeding (N=1270) a | ||||
| Breast milk only | 148 (45.7) | 145 (46.9) | 158 (50.2) | 154 (47.8) |
| Some breast milk and some formula | 151 (46.6) | 134 (43.4) | 129 (41.0) | 134 (41.6) |
| Formula only | 25 (7.7) | 30 (9.7) | 28 (8.9) | 34 (10.6) |
| Time between postnatal randomization and maternal follow-up survey (N=1383) | ||||
| < 63 days (<9 wks) | 192 (55.5) | 191 (54.6) | 159 (47.0) | 184 (52.7) |
| 63–76 days (9–10 wks) | 96 (27.7) | 119 (34.0) | 122 (36.1) | 112 (32.1) |
| 77–90 days (11–12 wks) | 27 (7.8) | 24 (6.9) | 32 (9.5) | 29 (8.3) |
| > 90 days (>12 wks) | 31 (9.0) | 16 (4.6) | 25 (7.4) | 24 (6.9) |
Collected after enrollment, either at the baseline survey or after infant birth
Includes those who reported one or more of the following: American Indian/Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, More than one race, Other
Between prenatal randomization and 60+ survey completion, mothers received 53 videos. Opening videos was measured as a proxy for video viewing. Approximately half (48.4%) viewed ≥6 videos, with 24.6% viewing >20 videos. There was no difference in video viewing rates among the 4 groups, and no correlation between the number of videos viewed and safe sleep outcomes.
Infant Safe Sleep Outcomes
No multiplicative interactions were observed between the prenatal and postnatal interventions in the main and time-to-event analyses; separate effects of each intervention are reported.
Primary Outcomes
Depending on the study group, 87.6% - 92.2% of mothers reported usual supine position, 83.8%−89.5% of mothers reported usual roomsharing without bedsharing, 75.6%−84.5% reported no soft bedding use, and 72.5–76.0% reported pacifier use (Table 2). In covariate-adjusted models, the prenatal and postnatal interventions did not have significant effects on any of the 4 primary outcomes after Hochberg adjustment for multiple comparisons.
Table 2:
Primary Safe Sleep Outcomes (Percentages) by Intervention Group
| Intervention groups (Prenatal/Postnatal) | Adjusted Risk for Pre-natal intervention only, %a | Adjusted Risk for Post-natal intervention only, % | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Outcomes in the Past 2 Weeks | Safe Sleep/Safe Sleep N (%) |
Safe Sleep/ Control N (%) |
Control/Safe Sleep N (%) |
Control/Control N (%) |
Safe Sleep | Control | Difference (95% CI)b | p-valuec | Safe Sleep | Control | Difference (95% CI)b | p-valuec | p for interactiond |
| Usually supine sleep position | 308/337 (91.4) | 317/344 (92.2) | 299/334 (89.5) | 305/348 (87.6) | 91.1 | 87.6 | 3.5 (−0.4,6.3) | 0.07 | 88.3 | 87.6 | 0.7 (−3.2,3.8) | 0.70 | 0.50 |
| Usually roomsharing without bedsharing | 297/332 (89.5) | 292/334 (87.4) | 287/327 (87.8) | 280/334 (83.8) | 86.7 | 83.8 | 2.9 (−1.5,6.4) | 0.18 | 87.4 | 83.8 | 3.6 (−1.0,7.1) | 0.12 | 0.72 |
| No soft bedding usee | 284/336 (84.5) | 274/344 (79.7) | 251/326 (77.0) | 260/344 (75.6) | 81.4 | 75.6 | 5.8 (1.7,9.3) | 0.01 | 79.1 | 75.6 | 3.5 (−2.6,8.6) | 0.25 | 0.33 |
| Pacifier usef | 249/333 (74.8) | 260/342 (76.0) | 240/331 (72.5) | 255/339 (75.2) | 76.7 | 75.2 | 1.5 (−3.5,5.9) | 0.54 | 73.4 | 75.2 | −1.8 (−6.3,2.2) | 0.39 | 0.64 |
Adjusted for infant age at survey, program language, pre-randomization study participation and accounting for clustering by WIC/FQHC site.
Calculated from odds ratios and 95% CIs from logistic regression
No p-values were < 0.05, after accounting for multiple comparisons.
Calculated from test for multiplicative interaction in the logistic regression model.
Defined as nonuse of bumper pads, sleep positioners, stuffed toys, pillows, pacifier with stuffed doll attached, towel or burp cloth, or loose blanket (regular or weighted).
Included responses of “usually use pacifier” and “sometimes use pacifier”
Secondary Outcomes
For sleep position, covariate-adjusted models estimated that 70.6% of those receiving the prenatal intervention were adherent to the exclusive supine sleep position, compared with 60.9% of the control group (adjusted risk difference [aRD] 9.7%, 95% CI 3.8%−15.0%) (Table 3). The prenatal intervention group had a 91.8% rate of planned exclusive supine sleep position, compared with 74.5% in the control group (aRD 17.3%, 95% CI 13.9%−19.7%). In time-to-event analyses (Supplement 3 Figure), the time-to-initial-report of non-exclusive supine sleep was longer for infants whose parents received either the prenatal or the postnatal safe sleep intervention, with a lower hazard of not reporting exclusive supine sleep at 60 days for those receiving only prenatal and only postnatal safe sleep interventions (adjusted hazard ratios [aHR] 0.74 [0.65–0.84] and 0.70 [0.60–0.82], respectively), compared with control participants.
Table 3:
Secondary Safe Sleep Outcomes (Percentages) by Intervention Group
| Intervention groups (Prenatal/Postnatal) | Adjusted Risk for Pre-natal intervention only, %a | Adjusted Risk for Post-natal intervention only, % | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Outcomes | Safe Sleep/Safe Sleep N (%) |
Safe Sleep/ Control N (%) |
Control/Safe Sleep N (%) |
Control/Control N (%) |
Safe Sleep | Control | Difference (95% CI)b | Safe Sleep | Control | Difference (95% CI)b |
| Sleep Position | ||||||||||
| Prenatal - Planned Exclusive Supined | 566/635 (89.1) | 476/639 (74.5) | 91.8 | 74.5 | 17.3 (13.9,19.7) | -- | -- | -- | ||
| Exclusively supine in last 2 weeksc | 254/337 (75.4) | 240/343 (70.0) | 220/334 (65.9) | 212/348 (60.9) | 70.6 | 60.9 | 9.7 (3.8,15.0) | 66.4 | 60.9 | 5.5 (−0.0,10.7) |
| Time to event - Initial report of non-exclusive supine sleep positione | See Supplement Figure Ag | 69.1 | 55.5 | 13.6 (8.0,18.8)h | 70.0 | 55.5 | 14.5 (8.8,19.7)h | |||
| Sleep Location | ||||||||||
| Prenatal - Planned Exclusive Roomsharing without bedsharingd | 535/635 (84.3) | 497/639 (77.8) | 86.0 | 77.8 | 8.2 (3.1,12.2) | -- | -- | -- | ||
| Exclusively Roomsharing without bedsharing in last 2 weeksc | 243/336 (72.3) | 240/344 (69.8) | 232/324 (71.6) | 219/338 (64.8) | 67.8 | 64.8 | 3.0 (−3.2,8.6) | 69.9 | 64.8 | 5.1 (−1.6,11.1) |
| Time to event - Initial report of non-exclusive Roomsharing without bedsharinge | See Supplement Figure Cg | 61.2 | 54.1 | 7.1 (1.7,12.0)h | 60.9 | 54.1 | 6.8 (1.6,12.0)h | |||
| Soft bedding f | ||||||||||
| Time to event - Initial report of soft bedding usee | See Supplement Figure Cg | 75.9 | 70.5 | 5.4 (0.3,10.1)h | 74.9 | 70.5 | 4.5 (−0.4,9.0)h | |||
Adjusted for infant age at survey, program language, pre-randomization study participation and accounting for clustering by WIC/FQHC site.
Calculated from odds ratios and 95% CIs from logistic regression.
Measured at 60-day survey.
Prenatal survey at 36 weeks.
Time to event analyses are among all twice randomized subjects and are based on text queries sent after birth.
Defined as nonuse of bumper pads, sleep positioners, stuffed toys, pillows, pacifier with stuffed doll attached, towel or burp cloth, or loose blanket (regular or weighted).
For time to event analyses, adjusted risk differences and 95% CIs are adjusted average survival at day 60 calculated from adjusted Cox proportional hazards models for entire followup period.
Confidence intervals for risk differences at day 60 calculated using bootstrap resampling with replacement (n=500 replicates).
For sleep location, neither prenatal nor postnatal intervention was associated with increased reported adherence to exclusive roomsharing without bedsharing. For soft bedding use, the prenatal intervention was associated with longer time-to-initial-report of any soft bedding use, with a 5.4 (0.3–10.1) percentage point difference in non-use at day 60, relative to controls (aHR 0.83 [0.71–0.97]) (Supplement 3 Figure).
In the post-hoc imputation analysis (Supplement 3), intervention effects for primary outcomes were consistent with the main analysis. Estimated intra-class correlations (ICCs) for the four primary outcomes were: supine 0.054; roomsharing without bedsharing 0.030; soft bedding 0.019; and pacifier use 0.006, and for the secondary exclusive outcomes: supine 0.012; roomsharing without bedsharing 0.014.
DISCUSSION
In this RCT of prenatal and postnatal video interventions delivered via text messaging to improve infant safe sleep practices among families with low income, there were no significant improvements in the primary outcomes. Given that approximately 90% of mothers reported usual supine positioning, and rates of usual roomsharing without bedsharing approached 90%, the study’s power to identify effects in these primary outcomes may have been limited by ceiling effects, as there is a small proportion of families who for cultural or philosophical reasons are unlikely to change their practices.
Among the secondary outcomes, receipt of the prenatal safe sleep intervention beginning at 34 weeks’ gestation, compared to control, was associated with higher proportions of mothers reporting exclusive supine placement and no soft bedding use. The postnatal safe sleep education did not result in a significant increase in exclusive supine placement; however, those receiving both safe sleep interventions had the highest rates of exclusive supine placement. Additionally, the estimated 91.8% with planned use of “exclusive” supine positioning between 36 weeks’ gestation and birth and delayed time-to-initial-report of non-supine sleep among those receiving the prenatal safe sleep intervention suggest that starting safe sleep education before the infant’s birth may hold promise.
One key finding among the secondary outcomes was that only 60.9% and 75.4% of the mothers in the control group and the group receiving both safe sleep interventions, respectively, reported exclusive supine sleep position in the last 2 weeks. This suggests this and the related outcomes may be appropriate targets for further study. Clinically, early establishment of exclusive supine positioning is important because infants who are not consistently placed supine are at extremely high relative risk (19-fold higher) of dying suddenly and unexpectedly on the occasions when they are placed or roll into the prone position.19
A randomized trial by our group previously demonstrated that a postnatal video safe sleep intervention in an economically-diverse population resulted in higher rates of usual supine sleep position.16 In the current study, the rates of usual supine sleep position (87.6–91.4%) in all groups were similar to the 92.5% we had achieved with the prior intervention.16 The prior study also resulted in a 14.9 percentage point increase (95% CI 10.1–19.1) in exclusive supine positioning (unpublished data. Michael Corwin, MD, Boston University, email communication, January 13, 2026). The fact that postnatal intervention had no significant impact on either usual or exclusive supine positioning in the current study may underscore the importance of prenatal safe sleep education.
Regarding sleep location, 86% of mothers in the prior trial reported usual roomsharing without bedsharing,16 and the current trial observed rates similar to this in all groups. For exclusive roomsharing without bedsharing, the current study did not observe group differences in reported practice. It is possible that, among families with fewer financial resources, sleep location is less dictated by the desire to adhere to safe sleep guidelines and more by the household circumstances – for instance, if no infant sleep space separate from the parental bed is available.
Limitations
This study had several limitations. First, this trial had the limitations inherent in self-reporting. Mothers who received the safe sleep interventions may have been more reluctant to report practices inconsistent with the messages received, which would overestimate intervention effects. Second, the 4 primary outcomes, 7 secondary outcomes, and 2 intervention conditions limited study power. For the secondary outcomes, hypothesis testing was not conducted and it was not possible to draw definitive conclusions. Third, recruitment largely used blast messages and advertisements. While this is highly time-efficient and while the demographics of this study population are reflective of the demographics of WIC recipients nationally,20 those who chose to enroll may have been more receptive to safe sleep messaging than the broader WIC population. Fourth, non-respondents to the 60+ day survey were younger, Hispanic or non-Hispanic Black, non-US born, with lower education and income, without a current partner and received the program in Spanish. Other strategies may be required to reach those who do not voluntarily enroll in or complete interventions such as these. Fifth, enrollment was limited to English- and Spanish-speakers. Although this captures >90% of the US population,21 results may not be generalizable to other groups. Additionally, mothers delivering infants before 34 weeks’ gestation were excluded from the study, so that study messaging would not conflict with the messages given by clinicians during the infant’s hospitalization (often in intensive care). It is important for safe sleep messages to reach this group, as SUID rates are higher among preterm infants22,23 and preterm birth occurs disproportionately among infants who are not non-Hispanic White.24 An intervention that specifically addresses the needs of preterm infants is needed. Sixth, interactions between the intervention and race, ethnicity, language, parity, and intended breast milk feeding duration were not evaluated. Finally, sample size estimates did not account for clustering by WIC site, potentially limiting study power. Intraclass correlation coefficients are reported for future study planning.
CONCLUSION
A text-delivered video safe sleep intervention to promote safe sleep practices among low-income families in the prenatal and postnatal periods did not have significant effects on the prespecified primary outcomes.
Supplementary Material
KEY POINTS:
Question:
Can prenatal or postnatal text-delivered video safe sleep interventions improve safe infant sleep practices among low-income families?
Findings:
Prenatal and postnatal interventions did not affect any of the 4 primary outcomes (usual supine position, usual roomsharing without bedsharing, soft bedding nonuse, or pacifier use, based on maternal report).
Meaning:
A text-delivered video intervention to promote safe sleep practices among low-income families in the prenatal and postnatal periods did not have significant effects on primary outcomes.
ACKNOWLEDGMENTS:
We received information to facilitate study enrollment from the following WIC Programs and Health Centers in the States listed below, and the results reported do not necessarily represent their views. IRBs of state WIC Programs and Health Centers approved the study, as required. Opinions expressed by the authors are their own and this material should not be interpreted as representing the official viewpoint of WIC Programs, Health Centers, U.S. Department of Health and Human Services, or the National Institutes of Health.
State DPH WIC Program Partners: Alabama, Connecticut, District of Columbia, Florida, Hawaii, Kansas, Maine, Massachusetts, Michigan, Minnesota, Missouri, New Hampshire, Ohio, Pennsylvania, Rhode Island, Virginia, and Washington.
Federally Qualified Health Center Partners: St. Louis County Department of Public Health, Missouri; Squirrel Hill Health Centers, Pennsylvania; Ohio Better Birth Outcomes, Ohio; Ohio Health, Ohio; and Copeland Clinic, Florida.
The authors declare no potential conflicts of interest. There are no financial relationships with the company that produced the videos used in the intervention.
This study was funded by a grant from the National Institute of Child Health and Human Development (NICHD) (1R01HD072815). The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Dr. Bellamy and Dr. Corwin (Slone Epidemiology Center) had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr. Bellamy (Department of Biostatistics, Boston University School of Public Health) and Mr. Kerr (Slone Epidemiology Center, Boston University, Boston, MA) conducted and are responsible for the data analysis.
Data sharing statement is in supplement 4.
REFERENCES
- 1.United States Department of Health and Human Services (US DHHS), Centers of Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS), Office of Analysis and Epidemiology (OAE), Division of Vital Statistics (DVS), Linked Birth / Infant Death Records on CDC WONDER Online Database. http://wonder.cdc.gov/lbd.html
- 2.Parks SE, DeSisto CL, Kortsmit K, Bombard JM, Shapiro-Mendoza CK. Risk Factors for Suffocation and Unexplained Causes of Infant Deaths. Pediatrics. Jan 1 2023;151(1)doi: 10.1542/peds.2022-057771 [DOI] [Google Scholar]
- 3.Branche T, Shapiro J, Najera C, Matoba N, Rankin K, Collins JWJ. Rates of Sudden Unexpected Infant Death (SUID) and its Subcategories in the U.S.: The Effect of Maternal Race, Ethnicity and Nativity. Matern Child Health J. May 2025;29(5):660–668. doi: 10.1007/s10995-025-04084-9 [DOI] [PubMed] [Google Scholar]
- 4.Edmondson C, Shenoy RD, Herrera JH, Williams K, Acuna JM. Maternal adherence to safe infant sleep recommendations in the USA: trends and disparities from PRAMS 2016–2022. BMJ Paediatr Open. Mar 12 2026;10(1)doi: 10.1136/bmjpo-2025-004382 [DOI] [Google Scholar]
- 5.Bombard JM, Kortsmit K, Warner L, et al. Vital Signs: Trends and Disparities in Infant Safe Sleep Practices - United States, 2009–2015. MMWR Morb Mortal Wkly Rep. Jan 12 2018;67(1):39–46. doi: 10.15585/mmwr.mm6701e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Shapiro-Mendoza CK, Colson ER, Willinger M, Rybin DV, Camperlengo L, Corwin MJ. Trends in infant bedding use: national infant sleep position study, 1993–2010. Pediatrics. Jan 2015;135(1):10–17. doi: 10.1542/peds.2014-1793 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Colson ER, Rybin D, Smith LA, Colton T, Lister G, Corwin MJ. Trends and factors associated with infant sleeping position: the national infant sleep position study, 1993–2007. Arch Pediatr Adolesc Med. Dec 2009;163(12):1122–8. doi:163/12/1122 [pii] 10.1001/archpediatrics.2009.234 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Colson ER, Willinger M, Rybin D, et al. Trends and factors associated with infant bed sharing, 1993–2010: the National Infant Sleep Position study. JAMA Pediatr. October 2013;167(11):1032–1037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Moon RY, Carlin RF, Hand I, Task Force On Sudden Infant Death S, The Committee On F, Newborn. Sleep-Related Infant Deaths: Updated 2022 Recommendations for Reducing Infant Deaths in the Sleep Environment. Pediatrics. Jul 1 2022;150(1)doi: 10.1542/peds.2022-057990 [DOI] [Google Scholar]
- 10.Boyer BT, Lowell GS, Roehler DR, Quinlan KP. Racial and ethnic disparities of sudden unexpected infant death in large US cities: a descriptive epidemiological study. Inj Epidemiol. Mar 25 2022;9(1):12. doi: 10.1186/s40621-022-00377-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wolf ER, Rivara FP, Sen A, Woolf SH. Sudden Unexpected Infant Death and Disparities in Infant Mortality in the US, 1999–2022. JAMA Pediatr. Mar 1 2025;179(3):344–346. doi: 10.1001/jamapediatrics.2024.6200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ajzen I The theory of planned behavior. Organizational Behav Human Decision Proc. 1991;50:179–211. [Google Scholar]
- 13.WIC eligibility income requirements. Accessed May 25, 2007, www.fns.usda.gov/wic/howtoapply/incomeguidelines.htm
- 14.Bandura A Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice Hall; 1986. [Google Scholar]
- 15.Berkowitz AD. An overview of the social norms approach. In: Lederman LP, Stewart LP, eds. Challenging the culture of college drinking: a socially situated health communication campaign. Hampton Press; 2005:chap 13. [Google Scholar]
- 16.Moon RY, Hauck FR, Colson ER, et al. The Effect of Nursing Quality Improvement and Mobile Health Interventions on Infant Sleep Practices: A Randomized Clinical Trial. JAMA. Jul 25 2017;318(4):351–359. doi: 10.1001/jama.2017.8982 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.SAS Institute Inc., SAS/STAT User’s Guide, Version 9.3. SAS Institute; 2011. [Google Scholar]
- 18.Ene M, Leighton EA, Blue GL, Bell B. Multilevel models for categorical data using SAS PROC GLIMMIX: The basics. 2015:
- 19.Mitchell EA, Thach BT, Thompson JMD, Williams S. Changing infants’ sleep position increases risk of sudden infant death syndrome. Arch Pediatr Adolesc Med. 1999;153:1136–1141. [DOI] [PubMed] [Google Scholar]
- 20.U.S. Department of Agriculture Food and Nutrition Service. National and State Level Estimates of WIC Eligibility and Program Reach in 2020. Washington, DC: US Department of Agriculture, 2023. [Google Scholar]
- 21.United States Census Bureau. Detailed Languages Spoken and Home and Ability to Speak English for the Population 5 Years and Over: 2017–2021. US Census Bureau. Accessed February 6, 2026. https://www.census.gov/data/tables/time-series/demo/language-use/2017-2021-lang-tables.html [Google Scholar]
- 22.Hwang SS, Barfield WD, Smith RA, et al. Discharge timing, outpatient follow-up, and home care of late-preterm and early-term infants. Comparative Study. Pediatrics. Jul 2013;132(1):101–8. doi: 10.1542/peds.2012-3892 [DOI] [PubMed] [Google Scholar]
- 23.Malloy MH. Prematurity and sudden infant death syndrome: United States 2005–2007. J Perinatol. Jun 2013;33(6):470–5. doi: 10.1038/jp.2012.158 [DOI] [PubMed] [Google Scholar]
- 24.Brown J, Chang X, Matson A, et al. Health disparities in preterm births. Front Public Health. 2023;11:1275776. doi: 10.3389/fpubh.2023.1275776 [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.