Effects of a Mindfulness-Based Intervention Delivered by Mobile Phone on Perinatal-Specific Stress and Neonatal Outcomes: A Randomized Controlled Trial

This study employed an 8-week, two-armed, randomized controlled trial design, with four data collection time points: baseline (T0, before the intervention), post-intervention (T1, immediately after the completion of the 8-week intervention), the 37th week of gestation (T2), and a follow-up conducted 5 weeks postpartum (T3). Throughout the study, we strictly adhered to the CONSORT guidelines pertaining to nonpharmacological treatments. Prior to their participation, all participants provided written informed consent, signifying their understanding and agreement to be part of this research endeavor. This study is not preregistered.

The study participants were a cohort of pregnant women of Chinese ethnicity in Hong Kong communities. The recruitment occurred in three waves––through advertisement on relevant websites, maternal discussion forums, and postings on Facebook. The research associates facilitated an online application questionnaire to ascertain the eligibility of perinatal women applicants for the intervention programs. Subsequently, applicants fulfilling the preliminary inclusion criteria were contacted and invited to participate in a brief intake one-to-one video screening, which ranged from 15 to 20 min. To verify each applicant’s eligibility and introduce the study in detail, the screening interviews were conducted by a social worker registered and experienced in providing mental health services. Before being randomly assigned, eligible participants were requested to complete a questionnaire as a baseline assessment via a Qualtrics link.

The inclusion criteria for the study required that eligible participants must (1) be 18 years of age or older; (2) be in the 6th to 28th week of gestation; (3) be able to read and hear Chinese; and (4) demonstrate adequate proficiency in using a mobile phone and have access to the Internet.

Conversely, the exclusion criteria precluded applicants who (1) were identified as having a high-risk pregnancy, including preterm labor risk, placental abnormalities, multiple gestations, or morbid obesity; (2) had a history of psychiatric disorders necessitating attention and care, such as bipolar disorder, psychosis, organic mental disorders, schizoaffective disorder, drug/alcohol use disorder, or pervasive developmental delay; or (3) had suicidal ideation and behaviors.

In accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines for nonpharmacological interventions, we conducted a rigorous process of participant enrollment, screening, and intervention process. As shown in Fig. 1, initially, a total of 240 Chinese pregnant women were assessed for eligibility. Of these, 50 were excluded from the sample pool. Specifically, 42 participants were excluded due to gestational weeks not aligning with established criteria. Two individuals reported severe depressive symptoms and suicidal ideation during the one-to-one video screening interview and were subsequently referred to specialized psychiatric care. Six participants were excluded due to other reasons, for instance, experiencing a miscarriage between the program registration and the screening interview. This resulted in a final sample of 190 pregnant women who were enrolled in the intervention study.

Simple randomized sampling was conducted by the research associate via an online random grouping generator. A total of 190 pregnant women were enrolled in the study and randomly assigned to one of two groups: the MBI group, consisting of 95 participants; and the educational control group, also consisting of 95 participants. To ensure equitable distribution of participants’ gestational weeks and their ages across the two groups, stratified randomization was employed (Suresh, 2011).

Due to several personal reasons (e.g., changed mind; overfull schedule), 10 participants (2 in the MBI group and 8 in the control group) declined to participate before the program commenced. Thus, in the final total, 178 pregnant women were enrolled in the intervention programs, with 93 participants in the MBI group and 85 in the control group. Participants were informed of their group allocation information 1 to 2 weeks prior to the program commencement.

The aim of the study was concealed from the participants to minimize potential biases. Furthermore, the statistical analysis was conducted in a blinded manner, ensuring that the group allocation information was not disclosed to the analysts.

The dataset utilized in the study excluded participant’s personal identifiers to ensure confidentiality and anonymity. Prior to their participation, participants were informed that they would be randomly assigned to one of two groups, both of which were presented as being equally efficacious and beneficial. The assessments were conducted through self-report questionnaires and submitted via secure online Qualtrics links, facilitating ease of data collection and management. Furthermore, neonatal health outcomes were collected through an online self-report questionnaire after each participant’s delivery.

Participants in the mobile-delivered MBI group received structured mindfulness training with one session per week over 8 weeks. Each participant was suggested to finish the eight MBI lessons in 8 weeks and practice mindfulness for approximately 30 min to 1 hr daily. The eight-session MBI was designed by integrating and modifying elements from Mindfulness-Based Cognitive Therapy (MBCT) (Dhillon et al., 2017). Our MBI program integrated the four-session advanced practice of the Four Immeasurables Meditation (FIM), which encompasses the development of loving-kindness (one’s goodwill and kindness), self-compassion (engaging in compassionate thoughts and actions with oneself and others), appreciative joy (one’s good fortunate and joy), and equanimity (wholesome attitude and profound state of balance) (Lv et al., 2020). FIM emphasizes self-awareness and the achievement of enlightenment dedicated to caring for others and the world (Chan, 2010). Several studies argued that cultivating the “Four Immeasurables” could effectively enhance pregnant women’s self-efficacy for being a mother, develop a compassionate connection with themselves and others (e.g., incoming baby, partners, and significant others), generate positive emotions, decrease the risk of experiencing perinatal specific distress, and influence infant health status (Chan, 2010, 2014). Our MBI was conducted in a secular way (e.g., using easily understood terms instead of Buddhist concepts), aiming to make it more accessible to pregnant women from varied religious beliefs. During the interventions, different themes that were tailored to the needs and concerns of perinatal women were introduced to the participants as customized mindfulness practices, such as connecting with your body and baby and inner inspection of being a mother. Each participant received prompts and guidance for daily mindfulness practices sent by our research associates via email or a social media platform. Each session contained two compulsory components: (1) participants attended mobile-delivered self-learning didactic mindfulness training through video platforms and PowerPoint slides; (2) participants engaged in mindfulness practices for 30 to 60 min, which were supported by recorded audio guides. In addition, as extra homework, the participants were encouraged to reflect on each course’s theme. For more information regarding the MBI protocol, please see the supplementary materials (S1).

As an optional component, participants were able to seek practical guidance and technical support from a registered social worker with experience in mindfulness practice through email when needed. Those communications mainly focused on participants’ experiences or difficulties in their mindfulness practices.

The participants in the control group received a mobile-delivered perinatal education program that adhered to an equivalent timeline as the MBI group, consisting of eight structured educational sessions focused on perinatal knowledge, such as “A guide to pregnancy changes, common discomforts, and health care” and “Dietary considerations during pregnancy.” The knowledge contents did not include mindfulness elements or counseling support. The educational program was developed utilizing curricular materials sourced from the systematic perinatal education classes provided by a renowned local medical institution. A mental health practitioner was available for online support throughout the eight-session program period whenever participants in the control group had questions or concerns about their perinatal knowledge. For more information regarding the psycho-education intervention protocol, please see the supplementary materials (S1).

Participants who finished 100% of the 8-week intervention program were considered perinatal women who had completed the intervention. Those who completed fewer than 8 weeks of sessions did not receive the postpartum assessment or the follow-up assessment.

Baseline differences in the groups were calculated using the independent sample t-test for continuous variables and the χ² test for categorical variables. The between-group differences in outcomes were examined using repeated-measures MANCOVAs. Linear regression models for the continuous-outcome variables were utilized for analyzing the mediation model and the intervention’s effects on neonatal outcomes. Several serial-trend analyses from T0 to T3 were conducted on the outcome variables to investigate the within-group intervention effect. The mediation analysis was conducted utilizing the bootstrapping method (with a 95% confidence interval, CI) to calculate the indirect effects. The analysis controlled for the participant’s age, pre-pregnancy body mass index (BMI), gestation weeks at baseline, and daily mindfulness practice frequencies. The statistical tests were conducted utilizing SPSS version 27. The mediation test was conducted via the PROCESS package (Hayes, 2017).

A total of 178 pregnant women who completed the baseline assessment were included in the baseline data analysis. Their mean age was 33.60 (SD = 7.59), and their mean gestational age was 20.35 weeks (SD = 5.38). The majority of participants were married or cohabiting (n = 177, 99.4%), most had a bachelor’s degree or above (n = 174, 86.5%), most had a full-time job (n = 145, 81.5%), and a majority had a monthly income of 20,000 HKD or above (n = 137, 77%). There was no significant difference between the intervention group (n = 93) and the control group (n = 85) in the demographic and outcome variables, except for their BMIs and daily practice of mindfulness. More information about the participants’ characteristics and baseline measurements can be found in Table 1.

Those who completed the intervention were invited to complete the post-intervention assessments. A total of 140 pregnant women who completed the intervention sessions were included in the final data analysis, with 69 in the intervention group and 71 in the control group. Independent sample t-tests were conducted for comparing the continuous variables, and chi-squared tests were conducted for the categorical variables when comparing completers and non-completers. Results revealed that there was no significant difference between the completer (n = 69) and non-completer (n = 24) in each of the demographic characteristic and assessment variable, except for their age and gestational weeks. In the control group comparison tests, there was no significant difference between the completer (n = 71) and non-completer (n = 14) in each of the demographic characteristic and assessment variable. More information about the comparisons of completers and non-completers can be found in Tables S1–S2 (Supplementary Information).

As shown in Table 2, the analysis revealed a statistically significant interaction effect between time and group across three assessment time points (F(2,65) = 4.98, p < 0.01, partial eta squared = 0.08) regarding the PSS change. Interaction effects analysis showed that the time*group effect across T0 to T1 was not significant (F(1, 138) = 0.29, p = 0.59, partial eta squared = 0.01). The time*group effect across T0 to T2 was significant (F(1,65) = 4.067, p < 0.05, partial eta squared = 0.06). Specifically, at the post-intervention assessment (T1), there was no significant difference in the reduction of PSS levels between women in the MBI group and those in the control group (t (138) = 0.47, p = 0.641, Cohen’s d = 0.08, 95% CI = [− 0.25, 0.41]). In addition, during their 37th week of gestation (T2), results showed a significant reduction in PSS among women in the MBI group, which was significantly greater than that reduction among women in the control group (t (69) = 2.07, p < 0.05, Cohen’s d = 0.50, 95% CI = [0.02, 0.96]).

In addition to PSS, we examined the effect of the intervention on mindfulness; the overall interaction effect between time and group across all four assessment time points was not significant (F(3, 60) = 2.00, p = 0.116, partial eta squared = 0.04). However, the time*group effect across T0 to T1 was significant (F(1, 137) = 4.58, p < 0.05, partial eta squared = 0.03), and the time*group effect across T0 to T2 was marginally significant (F(1, 65) = 3.91, p = 0.052, partial eta squared = 0.06). The time*group effect across T0 to T3 was not significant (F(1, 133) = 2.99, p = 0.086, partial eta squared = 0.02). Specifically, at the immediate post-intervention assessment (T1), women in the MBI group showed a significant increase in mindfulness level compared to the control group (t (138) =  − 2.02., p < 0.05, Cohen’s d =  − 0.35, 95% CI = [− 0.69, − 0.01]). In addition, the effect of the MBI on the women’s mindfulness enhancement was marginally sustained in the 37th gestation week (t (69) =  − 1.86, p = 0.055, Cohen’s d =  − 0.42, 95% CI = [− 0.90, 0.02]), thus suggesting that the MBI program had potential sustained effects. However, in the 5th week postpartum, we found no significant between-group effect size difference in the participants’ mindfulness (t (69) =  − 0.90, p = 0.185, Cohen’s d =  − 0.15, 95% CI = [− 0.49, 0.18]).

As is shown in Table 3, the hierarchical linear regression analysis showed a significant group difference in the neonates’ Apgar scores. Infants of mothers in the MBI group had significantly higher Apgar scores compared to infants of mothers in the control group (β = 0.20, t (135) = 2.22, p < 0.05), with an estimated between-group effect difference of 0.34 (95% CI = [0.00, 0.68]). No statistically significant between-group differences were found in the levels of infants’ maturity (i.e., born before or after 37 weeks of gestation) (β =  − 0.11, t (136) =  − 1.24, p = 0.217) or infants’ birth weights (β =  − 0.05, t (136) =  − 0.57, p = 0.573).

For mediation analysis, results revealed that the improvement in mindfulness levels (the difference between baseline (T0) and post-intervention (T1)) among women in the MBI group significantly mediated the intervention’s effect on reducing their PSS levels at their 37th week of gestation (the difference between T0 and T2) (indirect effect: β =  − 0.51, 95% CI = [− 1.37, − 0.01]), which accounted for 25.5% of the total intervention effect. Additionally, the total intervention effect on the participants’ PSS in the 37th week of gestation remained significant when their mindfulness enhancement at T1 was entered into the mediation model (direct effect: β =  − 0.24, t (69) =  − 2.07, p < 0.05, 95% CI = [− 2.94, − 0.05]). This suggests that mindfulness improvement had a partial mediating effect. The paths and the standardized coefficients for the mediation model are shown in Fig. 2.

The current study had several strengths and limitations. The study’s major strengths included the relatively high completion rate, a relatively large sample size at baseline, and its rigorous RCT design with an active education-oriented control group. Sixty-nine of the 93 (74.2%) participants in the MBI group fully completed the eight-session intervention (including the post-test (T1) and the follow-up test (T3, 5 weeks after delivery)). Thus, the rate of completion in our study aligns with the rates documented in previous intervention studies that have implemented in-person MBI for perinatal individuals, with reported completion rates averaging around 80% (Felder et al., 2017; Goodman et al., 2014). In addition, our completion rate was comparable to those reported for similar online delivery modes of MBIs, which reported even lower compliance rates of 15 to 57% (Goetz et al., 2020; Hulsbosch et al., 2023; Krusche et al., 2018). For instance, Hulsbosch et al. (2023) reported a completion rate of 15% among a cohort of 109 participants who engaged in an online MBI program. The researchers attributed the low completion rate to the absence of real-time support and the lack of instructor availability for consultation on issues encountered by the participants. In our study, the mobile-delivered MBI was designed to ameliorate these limitations from prior research, partly offering a viable avenue for perinatal women to access mindfulness practices and educational content through digital platforms. It also facilitated ongoing interactions between mindfulness professionals and the participants when needed. Perinatal women are easily able to gain access to a mobile MBI. In addition, our mobile-delivered MBI provided our perinatal participants with a more feasible and less stigmatized mode for participating in mindfulness practices and improving their mental health status.

Despite these study strengths, our findings have several limitations. First, although roughly 75% of the participants in the MBI group completed the whole session and the post-tests at T1 and T3, only 35 of 69 (50.7%) participants in the MBI group completed the first follow-up test (at T2, in the 37th week of gestation). One reason for the low participation at T2 can be explained by the women’s busy status before delivery, as disclosed by the participating women “not having the time or right mood to finish the follow-up test.” Another reason given for the missing data at T2 was that, for some participants who were in the later stages of pregnancy (e.g., 27 to 28 weeks) at the baseline timepoint (T0), the time of the post-testing after completing the 8-week course occurred almost at the 36th to 37th week of gestation. Therefore, the T2 follow-up was not conducted for these participants, and they were only asked to complete the post-test (T1) and the final follow-up (T3, at 5 weeks postpartum).

The study’s second limitation was that, although we facilitated the online MBI by using email support so that participants could obtain access to support from mental health professionals, we did not consider the moderating role of such support. We suggest future intervention studies to resolve the issue by keeping records of participants’ subjective experiences, of both intervention involvement and professional support.

Third, our sample was predominantly low-risk, middle-income, well-educated women, and that could restrict the generalizability of the MBI’s effectiveness. To increase the recruitment scope and social impact of the programs, we did not conduct screening of PSS among the applicants. However, as a result, the participant sample may include pregnant women with relatively lower levels of PSS, which could lead to an increase in the attrition rate. Future research should consider re-examining the effects of the MBI in a more diversified sample (e.g., recruiting women in low-socioeconomic status families and screening for women with relatively high PSS scores).

Fourth, in the final data analysis, we utilized a completer analysis in which we only included participants who had fully reached the intervention endpoint although the completer analysis allowed us to investigate the efficacy of intervention under a rigorous and ideal situation (Tripepi et al., 2020). However, the completer analysis may fail to demonstrate the comprehensive picture of intervention settings. For instance, it is challenging to draw conclusive and rigorous conclusions about whether the incomplete intervention is due to the ineffectiveness of the intervention itself or the personal choices of the non-completed participants. Hence, to improve the internal validity of the intervention studies and maintain the benefits of randomization, future studies should consider utilizing intent-to-treat analysis.

Fifth, while many studies have demonstrated the challenges that postpartum women face (e.g., Ruderman et al., 2021; Walker & Murry, 2022), our intervention program did not address the potential benefits of MBI during postpartum. We recommend that future research consider a two-phase MBI, with the first phase focusing on perinatal-specific distress and the second phase targeting post-partum distress. This approach may provide pregnant and postpartum women with more effective, sustained support and guidance.

Sixth, although we found that participants’ increased mindfulness level mediated the relation between intervention effect and decreased PSS, the current study did not explore the specific factor of mindfulness that plays the most crucial role underneath such mechanism. Future studies could utilize more nuanced measurement tools, such as conducting qualitative interviews and measuring sub-concepts of four Immeasurables (e.g., self-compassion, joy) to better understand the potential mechanisms of mobile-delivered MBIs.

In summary, this study investigated the effectiveness of a mobile-delivered MBI program for reducing PSS and improving pregnant women’s mindfulness and their neonatal outcomes. With a reasonably low dropout rate (25.8%), the current MBI delivered by mobile phone proved acceptable and effective for reducing PSS at the 37th week of gestation (T2, the first follow-up assessment). In addition, the mobile-delivered MBI proved effective in improving mindfulness immediately after the intervention. The association between the intervention effect and changes in the participants’ PSS was mediated by changes in their mindfulness. Moreover, the study’s mobile-delivered MBI also effectively enhanced the participants’ newborn infants’ Apgar scores. The findings contribute significantly to the evidence supporting the benefits, feasibility, and acceptability of a mobile-delivered MBI during pregnancy. We, therefore, suggest that perinatal care providers include mindfulness interventions as an essential component of psychological care for women during the perinatal period.