Evaluation of an e-health integrated intervention for pregnant women with co-occurring mental health and substance use symptoms: an exploratory randomized clinical trial Authors: Rodrigo Carmona Camacho, Nayara López Carpintero, María Luisa Barrigón, Lucia Albarracín García, Igor Barahona, Enrique Baca-García, Cristina Ruiz Nogales, Inés Menéndez, Montserrat Sánchez Alonso, Carmen Vidal Mariño, Lucía Rodríguez Blanco, Silvia María Campos Soler, Montserrat Díaz Rosell, Benjamin Le Cook, Margarita Alegría, Irene Caro Cañizares Published: 2026 ABSTRACT AbstractBackgroundThe co-occurrence of substance use and mental health problems in pregnant women implies higher maternal-foetal morbidity, and treatments often do not meet care needs. This study aims to generate preliminary evidence and hypothesis by exploring the potential effect of an integrated intervention based on third-generation therapies by comparing a telephone intervention with a digital intervention and treatment as usual.MethodsIn an exploratory randomised trial, 111 pregnant women (out of 2014 screened) with co-occurring mental health and substance use symptoms were assigned to either telephone-delivered intervention, digitally-delivered intervention, or treatment as usual groups. Interventions, based on third-generation therapy, lasted eight weeks. Outcomes (alcohol, tobacco use, depression, anxiety, and trauma) were measured at baseline and after two, four, eight, and twelve months.ResultsThe telephone intervention group performed significantly better with regard to alcohol, depression, and trauma symptoms after two and four months compared to the digital intervention and treatment as usual groups. The digital intervention group showed limited benefit and higher attrition. All groups had returned to baseline by the 12 month point.ConclusionsStandardised clinical interventions delivered by telephone can improve the treatment outcomes of pregnant women with co-occurring mental health and substance use problems. New approaches are needed to effectively implement these interventions via digital support.Trial registrationClinicalTrials.gov ID NCT06965270 (retrospectively registered in May 11, 2025).Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09436-z.Keywords: Dual disorders, Pregnancy, E-health, Mental health, Substance use, Perinatal care Background The use of substances during pregnancy constitutes a serious social problem. High maternal-foetal morbidity [1] is associated with both the toxic effects of the substances used and complications arising from potentially risk behaviours associated [2]. Meanwhile, pregnancy represents a vital change that can contribute to the appearance or aggravation of mental health problems, the most frequent being emotional disturbances, changes in anxiety states, and post-traumatic stress disorder [3– 5]. The concurrence of substance use problems with mental health problems implies a worse prognosis and more challenges to intervention [6]. The co-occurring mental health and substance use symptoms in pregnant women has important implications for their clinical care, as the situation is associated with a higher risk of non-compliance with prenatal medical follow-up [7], health problems, and various obstetric complications such as premature birth, low birth weight and delayed child development [8–13]. Despite this, existing treatments for co-occurring mental health and substance use problems generally do not meet care needs, especially when these are related to pregnancy and issues other than depression need to be addressed [4]. The scarcity of pharmacological resources that can be used during pregnancy, as well as the scarcity of specific psychotherapeutic programmes, leads professionals to adapt resources used for the general population; as a result, the specific needs that arise during pregnancy are sometimes not considered [14]. However, pregnancy and maternity are associated with a greater motivation to seek treatment and a greater predisposition to contact health services [15, 16]. During this period, an appropriate intervention with the pregnant woman can reduce substance use symptoms and its consequences for foetal development, as well as improving the conditions of childbirth [7, 16]. Prior literature [17] discusses three different approaches to the treatment of co-occurring mental health and substance use symptoms: Sequential treatment, whereby one condition is addressed before the other; this has little support since it requires prioritisation, depending on the severity of the problems, and makes forecasting difficult. Sequential treatment, whereby one condition is addressed before the other; this has little support since it requires prioritisation, depending on the severity of the problems, and makes forecasting difficult. Parallel treatment, whereby each condition is addressed by different health resources and different professionals simultaneously; this is the most widespread model, although it entails difficulties in coordinating different resources and hinders therapeutic adherence. Parallel treatment, whereby each condition is addressed by different health resources and different professionals simultaneously; this is the most widespread model, although it entails difficulties in coordinating different resources and hinders therapeutic adherence. Integrated treatment, whereby both conditions are addressed using the same resource by the same professionals at the same time; this is the optimal approach, although it is less widespread given the complexity of offering integrated health system resources. Integrated treatment, whereby both conditions are addressed using the same resource by the same professionals at the same time; this is the optimal approach, although it is less widespread given the complexity of offering integrated health system resources. Developing integrated treatments from different specialists to address the complexity of these cases improves symptom remission. The literature highlights the necessity of multidisciplinary and comprehensive healthcare for these patients due to the complexity of the medical, psychological and social problems generally involved [6, 18, 19]. In addition, there is a need to involve and train different professionals and coordinate interventions by different institutions to improve the care provided to women who face substance use problems and to their children [20]. It is also necessary to evaluate e-health alternatives that could potentially enhance the interventions’ scope. To date, the interventions developed and validated to address substance use or psychopathology in pregnant women focus specifically on only one of these two problems, with limited attention paid to integrated interventions for co-occurring conditions [14]. This situation highlights a critical research gap. Against this backdrop, and within the context of the WOMAP (Woman Mental Health and Addictions on Pregnancy) initiative, an exploratory randomised clinical trial (ClinicalTrials.gov ID NCT06965270; retrospectively registered on May 2025) was conducted to explore the effect of an integrated intervention based on third-generation therapies for pregnant women with substance use problems and mental health symptoms. The effect of a telephone-based intervention was contrasted with a digital-based intervention and treatment as usual (TAU). The objective was to generate preliminary evidence and hypotheses by assessing the direction of the effect and the feasibility of each condition. Methods Setting and design The exploratory clinical trial was conducted from July 2016 to December 2019 in five hospitals across the Madrid metropolitan area in Spain. The participating hospitals, which included Jiménez Díaz Foundation, Infanta Leonor Hospital, Tajo University Hospital, General Hospital of Villalba, and Infanta Elena Hospital, cover a health area of over 1.3 million inhabitants from diverse socioeconomic backgrounds. During their routine obstetric visits, we invited all pregnant women over 18 years old who had been pregnant for less than 26 weeks to participate in the study. Participants were approached either in person after their obstetric visits or by phone with prior consent. Research assistants explained the study and conducted screening interviews with the women who agreed to participate. We conducted a comprehensive screening of 2014 women using the AC-OK screen for mental health, substance use, and co-occurring mental health/substance use problems [21, 22]. Women who reported a death desire or suicidal behaviour on the AC-OK were further evaluated with the Paykel Suicide Scale [23], with immediate psychiatric referral for those indicating suicidal plans or attempts. The assessment also included questions about cigarette smoking habits in the past year, current frequency of smoking, and utilisation of mental health or substance use services (programmed appointments in the following month or the previous three months). Additionally, we collected socio-demographic information including age, country of origin, racial group, and education level. Women were included in the WOMAP exploratory clinical trial if they met the following criteria: (1) they scored two or more positive responses on the AC-OK Mental Health subscale; (2) they scored one or more positive responses on the AC-OK Substance Abuse subscale and/or reported smoking more than once a month, (3) they were not users of specialised services, and (4) the Paykel Suicide Scale (if administered) did not indicate they had any suicidal plans or had made any suicide attempts. The exclusion criteria were (1) having a diagnosis of psychotic or bipolar-related disorders, and (2) having a lack of capacity to provide consent, determined by an inability to comprehend the study’s purpose or procedures. Assessment The eligible women (n = 120) underwent a comprehensive assessment using validated mental health and substance use questionnaires. To assess mental health, we used the Patient Health Questionnaire (PHQ-9), which addresses major depressive disorder symptoms based on DSM-IV criteria [24], the General Anxiety Disorder seven-item screener (GAD-7) for assessing anxiety [25], and the Post-Traumatic Stress Disorder Checklist (PCL-5), a self-report measure for the 17 DSM-IV symptoms of Post-Traumatic Stress Disorder (PTSD) [26]. For substance use, we administered the Alcohol Use Disorders Identification Test (AUDIT), a screen developed by the World Health Organisation [27], the Drug Abuse Screening Test (DAST), a brief self-report instrument designed for drug abuse and dependence disorders detection [28], and the Fagerström Test for Nicotine Dependence [29]. Additionally, participants reported the frequency of their use of substances in the past month, including alcohol, opioids, barbiturates, benzodiazepines, cocaine, amphetamines, cannabis, hallucinogens, and inhalants. Assessments were conducted at baseline and subsequently at the two, four, eight, and 12 month points by research assistants who were blinded to the participants’ study condition. Exploratory clinical trial phase The exploratory clinical trial explored and compared the effect of two different integrated behavioural treatments, namely a standardised clinical intervention delivered by phone (telephone-delivered intervention, TDI) and a computer/smartphone-based intervention (digitally-delivered intervention, DDI), with the treatment-as-usual (TAU) approach provided to pregnant women at risk. A study investigator (BLC) generated a stratified block-randomisation scheme that assigned eligible participants to the enhanced TAU or intervention groups, using a 1:1 ratio for each three-person block. Stratified by recruitment site and then clinician, each patient had a 33.3% chance of being assigned to any of the groups. The project coordinator, who was not involved in data collection, randomised the patients after obtaining their informed consent and baseline assessment. Patients undergoing the TDI were treated by clinicians who had undergone specific training. This training was provided by a clinical psychologist (ICC) with more than six years’ experience and included two supervised practice cases. After completing the training, the clinicians handled two additional pilot cases to ensure they were fully prepared to deliver the intervention. Patients undergoing the DDI were contacted via telephone on two occasions (weeks 1 and 7) by a care manager in order to enhance the DDI. Care managers and clinicians were not blinded to the participants’ study condition. Patients who remained highly symptomatic after participation in any of the interventions were referred to their primary care physician for assessment of whether additional treatment was needed. Technical issues The validated and previously developed digital tool MeMind (www.memind.net) was used to facilitate the exploratory trial [30]. This platform, which can be accessed via the web or a mobile application, enables tailoring of available content for each user based on their previously programmable profiles. The electronic health record interface was used for data collection, and the treatment interface for implementation of the DDI. Intervention To develop the therapeutic intervention, three clinical psychologists (CVM, LRB, ICC) with experience in collaborating on adapting another mental health/substance use disorder intervention (IIDEA) worked together to adapt this previous manualised intervention to the objective of the clinical trial. The previous IIDEA intervention consisted of 10 sessions of evidence-based treatments for mental health and substance abuse problems [31, 32]. Some of the topics were not directly relevant to our aims, especially those related to human immunodeficiency virus (HIV) or sexually-transmitted disease (STD), and were thus excluded. The adaptation first shortened the intervention by eliminating the sessions focused on HIV/STD, resulting in a total of eight sessions, ideally held weekly. Then, we changed the theoretical approach. The IIDEA intervention mainly consists of cognitive-behavioural therapy (CBT), with a focus on decreasing symptoms by reducing discomfort, whereas the WOMAP intervention builds similar skills based on the third-generation therapy approach, a new wave in behaviourism that includes mindfulness-based therapies and acceptance and commitment therapy, among others, with a focus on discomfort acceptance (see Table 1). Table 1Intervention program outlineSessionContentsSession 1Introduction to the program, motivation for change, introduction to mindfulnessSession 2Psychoeducation on substance use and strategies for managing the desire to useSession 3Work on attachment, exploration of the bond with the babySession 4Psychoeducation on mental health disorders, introduction to the triple response system and common thinking stylesSession 5Introduction to cognitive restructuring: “evaluating my thoughts”; cognitive restructuring practiceSession 6Assertiveness, characteristics and components of effective communication, basic communication techniquesSession 7Reinforcement of substance use management strategies, relapse preventionSession 8Self-care plan review and farewell Intervention program outline Finally, the entire intervention was formalised into two twin manuals for the TDI: one for clinicians and the other for patients. In parallel, a self-applied online-based intervention protocol was developed, following the same structure, theoretical framework and content as the protocol developed for the telephone-based intervention. The intervention was implemented using the MeMind digital tool. The online platform was adapted to offer weekly content to the registered patients and was organized into eight sessions, covering the same content as the eight telephone intervention sessions. Access to a new session’s content was unlocked each week to ensure the intervention’s progression. Each session included an audio guide for a mindfulness exercise, a review of the objectives achieved in the previous week, a space to write the objectives for the current week, and a psychoeducational document on the topic to be addressed in the session (such as instructions for auto-cognitive restructuring, consumption self-registration, etc.) that could be auto-applied, along with suggested activities to complete during the week to implement the changes. In addition, Short Message Service (SMS), text messages and pop-up notifications were programmed to improve adherence. These messages and notifications encompassed three types of content. One type of notification informed participants when content was available for a new session. Another type of notification asked participants about their progress toward the goal set for that week. And the third type of notification was motivational, encouraging participants to track their progress and complete the assigned activities. Finally, this intervention was reinforced with two concierge calls. In the first week after the start, a call was made to verify access to the application, provide guidance in case of difficulties, and encourage its use. In week number 7 from the start, a second call was made to review the clinical components of the program and to provide motivation to continue. Ethical considerations and data protection The study was carried out in compliance with the Declaration of Helsinki and approved by the Jiménez Díaz Foundation Ethics Committee for Clinical Investigation (Ref. 2015/43). After a complete description of the study, all screened participants gave informed consent. To ensure data protection, access to the MeMind clinician user interface was restricted to researchers. Only the principal investigator, clinicians and researchers, using usernames and passwords, had access to identifiable information. The data provided by the clinician were encrypted via Secure Socket Layer/Transport Layer Security (SSL/TLS) between the investigator’s computer and the server. Data were stored in an external server created for research purposes, and were encrypted using the industry-standard AES-256 algorithm. Furthermore, an external auditor guaranteed that the security measures met the Organic Law for Data Protection (Spanish Government, Ley Orgánica 3/2018 de Protección de Datos Personales y garantía de los derechos digitales, 2018) standards at a high protection level. Statistical analysis The participants’ responses were assessed using five scales: AUDIT (alcohol use), Fagerström (for nicotine dependence), PHQ-9 (depression), GAD-7 (anxiety), and PCL-5 (trauma symptoms). Given the minimal number of patients reporting drug use, the DAST scale was excluded from the analyses. The scores were compared across the three treatment groups: telephone-delivered intervention (TDI), digitally-delivered intervention (DDI), and treatment as usual (TAU). Measurements were conducted at five specific time points: baseline, two months, four months, eight months, and 12 months. Therefore, the trial has a 3 × 5 factorial structure as illustrated below: In the previous formulation is the main effect of the group treatment with i = 1,…,3 and is the main effect of the specific time point with j = 1,…,5. Finally is the error. This framework based in fixed effects facilitated the assessment of within-subject changes over time while also accounting for potenti al differences among the three interventions. Interaction plots were generated to illustrate any time-by-treatment interaction, thus providing a visual depiction of how the outcomes for each treatment evolved across the 12-month period. Two-tailed student’s t-test (with ) were employed to test, there not exist statistical differences for the treatments across time periods: for . Statistical power is defined as the probability of correctly rejecting the null hypothesis when a true difference between the population means exists. It is calculated as illustrated below. For the hypothesis tests presented above, statistical power was calculated separately for each time period j = 1,…,5. The power calculations for the corresponding t-statistics were based on a two-sided significance level (α = 0.05), the observed sample sizes in each comparison group , and the estimated population standard deviation (σ). Results Of the 2,014 women screened (31.9 ± 5.8 years old), 120 met the inclusion criteria for the WOMAP clinical trial. From this group, 111 women ultimately participated and were randomised as follows: 37 to TDI, 36 to DDI, and 38 to TAU. Figure 1 shows the CONSORT flow diagram. CONSORT flow diagram Sample description The mean age of the sample was 31.9 years (standard deviation 5.8); 71.2% (79) were of Spanish origin. The participants’ employment status is described in Table 2. Table 2Employment statusFrequencyPercentageCumulative PercentageUnemployed without benefits2018.018.0Looking for a job21.819.8Unemployed with benefits76.326.1Permanently disabled10.927.0Temporarily disabled109.036.0Employed6558.694.6Domestic caregiver54.599.1Student10.9100.0Total111100.0100.0 Employment status Substance use outcomes in the three randomized controlled trial groups Alcohol The scores for AUDIT across follow-up in the total sample and sample sizes—in total and for each group—are presented in Table 3; Fig. 2. The retention rate was around 40%, with 40–45 participants providing data at each follow-up and the digital intervention group falling to seven participants 7 (8.75%) at the end of the year. AUDIT scores were low for all groups across the study duration, with a maximum score of 1.45 at the end of the follow-up for the TAU group. Lower values were found in the second and fourth months for the DDI group compared with the TAU group. The TDI group’s scores were significantly lower than the TAU group’s at two, four and eight months. No differences were found between the DDI and TDI groups. By the one-year point, all groups exhibited higher AUDIT scores compared with the baseline. Values for the statistical power ranged between 0.25 and 0.77 for all conducted tests. Table 3Descriptive statistics and sample sizes for the AUDIT score during follow-upBaseline2 Months4 Months8 Months12 MonthsDescriptive statisticsMin / Max0 / 10.00 / 8.00 / 11.00 / 6.00 / 7.0Med [IQR]0 [0;1.0]0 [0;0]0 [0;0]0 [0;2.0]0 [0;2.5]Mean (std)0.8 (1.9)0.3 (1.3)0.4 (1.7)1.0 (1.6)1.3 (2.0)Sample sizeTAU36 (34.95%)15 (14.56%)14 (13.59%)18 (17.48%)20 (19.42%)Telephone Therapy37 (37.76%)13 (13.27%)16 (16.33%)16 (16.33%)16 (16.33%)Digital Therapy36 (45.00%)12 (15.00%)15 (18.75%)10 (12.50%)7 (8.75%)N (total)109*40454443*Two patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 109 Descriptive statistics and sample sizes for the AUDIT score during follow-up *Two patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 109 AUDIT scores over time by intervention Tobacco The scores for the Fagerström Test for Nicotine Dependence across follow-up for participants who reported tobacco use at baseline (n = 42) and sample sizes—in total and for each group—are presented in Table 4; Fig. 3. Attrition rates were high across follow-up points, with 12 participants completing the one-year follow-up, of whom only two (6.06%) were in the DDI group. Nicotine dependence level tended to be low, with median figures between 3 and 4 across the groups. With statistical power values ranging from 0.069 to 0.072, no statistical differences were found across the groups. Table 4Descriptive statistics and sample sizes for the Fagerström score during follow-upBaseline2 Months4 Months8 Months12 MonthsDescriptive statisticsMin / Max0 / 8.00 / 7.00 / 6.00 / 5.00 / 6.0Med [IQR]3.0 [1.0;4.0]4.0 [0;5.5]3.0 [1.0;5.0]3.0 [0;3.0]3.5 [1.5;4.2]Mean (std)2.9 (2.3)3.1 (2.7)2.7 (2.2)2.2 (1.9)3.0 (2.1)Sample sizeTAU14 (37.84%)8 (21.62%)6 (16.22%)4 (10.81%)5 (13.51%)Telephonetherapy11 (47.83%)1 (4.35%)2 (8.70%)4 (17.39%)5 (21.74%)Digitaltherapy17 (51.52%)5 (15.15%)5 (15.15%)4 (12.12%)2 (6.06%)N (total)4214131212 Descriptive statistics and sample sizes for the Fagerström score during follow-up Telephone therapy Digital therapy Fagerström test scores over time by intervention Mental health outcomes in the three randomized controlled trial groups Anxiety Reflecting anxiety symptoms, the scores for GAD-7 across follow-up and sample sizes—in total and for each group—are presented in Table 5; Fig. 4. The retention rate was around 40%, with 40–43 participants providing follow-up data at each time point, and the DDI group declined to seven participants (8.97%) at the end of the year. Mean scores across all groups ranged between 5.2 and 6.2. Statistically significant intervention effects were found for anxiety symptoms, with a significant improvement effect among the TDI group at months 2, 4 and 8 as compared to the other groups. The DDI group performed worse than the TAU group from months 4 to 12. The results in the hypothesis test were supported with a moderate statistical power equal to 0.797. Table 5Descriptive statistics and sample sizes for the GAD-7 score during follow-upBaseline2 Months4 Months8 Months12 MonthsDescriptive statisticsMin / Max0 / 21.00 / 21.00 / 20.00 / 17.00 / 19.0Med [IQR]5.0 [2.0;9.0]4.0 [1.0;6.0]3.0 [0;7.5]3.0 [1.0;7.0]4.0 [1.0;8.5]Mean (std)6.2 (4.8)5.2 (5.3)4.5 (4.8)4.5 (4.2)5.5 (5.2)Sample sizeTAU36 (34.95%)15 (14.56%)14 (13.59%)18 (17.48%)20 (19.42%)Telephone Therapy37 (38.54%)13 (13.54%)15 (15.62%)15 (15.62%)16 (16.67%)Digital Therapy35 (44.87%)12 (15.38%)14 (17.95%)10 (12.82%)7 (8.97%)N (total)108*40434343*Three patients had missing data on items of this scale, impeding correction, which reduced thesample size from 111 to 108 Descriptive statistics and sample sizes for the GAD-7 score during follow-up *Three patients had missing data on items of this scale, impeding correction, which reduced thesample size from 111 to 108 GAD-7 scores over time by intervention Depression For depression, the scores for PHQ-9 across follow-up and sample sizes—in total and for each group—are presented in Table 6; Fig. 5. The retention rate was about 40%, with 40–44 participants completing each follow-up and the DDI group decreasing to seven participants (8.86%) at the end of the year. The mean scores across groups ranged between 10.6 and 12.9. Significantly lower values were found in the fourth and eighth months for the TDI group compared with the TAU and DDI groups (see Fig. 5). The statistical power ranged between 0.23 and 0.98 for all conducted hypothesis test. Table 6Descriptive statistics and sample sizes for the PHQ-9 score during follow-upBaseline2 Months4 Months8 Months12 MonthsDescriptive statisticsMin / Max6.8 / 26.37.5 / 24.86.8 / 23.36.8 / 19.56.8 / 21.0Med [IQR]12.0[9.8;15.0]10.5[9.0;12.0]9.8[8.3;12.0]10.5 [8.3;12.2]10.5 [8.3;12.0]Mean (std)12.9 (3.8)11.4 (3.5)10.6 (3.3)10.9 (3.2)11.1 (3.7)Sample sizeTAU36 (34.95%)15 (14.56%)14 (13.59%)18 (17.48%)20 (19.42%)Telephone therapy37 (37.76%)13 (13.27%)16 (16.33%)16 (16.33%)16 (16.33%)Digital therapy36 (45.57%)12 (15.19%)14 (17.72%)10 (12.66%)7 (8.86%)N (total)109*40444443*Two patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 109 Descriptive statistics and sample sizes for the PHQ-9 score during follow-up *Two patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 109 PHQ-9 scores over time by intervention Post-traumatic symptoms For symptoms of PTSD, the scores for PCL-5 across follow-up and sample sizes—in total and for each group—are presented in Table 7; Fig. 6. The retention rate was about 40%, with 40–44 participants completing each follow-up, and the digital intervention group decreased to seven participants (9.09%) at the end of the year. The mean scores across groups ranged between 23.5 and 29.4. Significantly lower values were found at the four and eight months for the TDI compared with the TAU and DDI groups. The DDI group showed significantly higher scores than the TAU group at the one-year point (see Fig. 6). Statistical power for the conducted tests ranged between 0.136 and 0.409. Table 7Descriptive statistics and sample sizes for PCL-5 score during follow-upBaseline2 Months4 Months8 Months12 MonthsDescriptive statisticsMin / Max16.0 / 75.216.0 / 61.616.0 / 45.616.0 / 63.216.0 / 77.6Med [IQR]24.8[19.6;35.2]20.8[16.0;27.6]17.2[16.0;28.6]18.9[16.0;29.4]20.0[16.0;29.6]Mean (std)29.4 (13.4)24.8 (10.8)22.9 (9.2)23.5 (10.4)24.9 (12.9)Sample sizeTAU35 (34.31%)15 (14.71%)14 (13.73%)18 (17.65%)20 (19.61%)Telephone Therapy37 (38.14%)13 (13.40%)15 (15.46%)16 (16.49%)16 (16.49%)Digital Therapy35 (45.45%)12 (15.58%)13 (16.88%)10 (12.99%)7 (9.09%)N (total)107*40424443*Four patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 107 Descriptive statistics and sample sizes for PCL-5 score during follow-up *Four patients had missing data on items of this scale, impeding correction, which reduced the sample size from 111 to 107 PCL-5 scores over time by intervention Discussion Main findings This study presents the findings of the WOMAP (Woman Mental Health and Addictions on Pregnancy) exploratory randomised clinical trial, which explores effect direction of an integrated intervention for pregnant women with co-occurring substance use symptoms (alcohol and tobacco use) and mental health symptoms (depression, anxiety, and post-traumatic stress disorder). The exploratory trial compared two delivery modalities—a telephone-based intervention (TDI) and a digitally-delivered intervention (DDI), both based on the IIDEA intervention—against treatment as usual (TAU), with the intention of exploring whether there were differential trend effects depending on the treatment condition. Given the small effective sample sizes at follow-up and the extremely high and differential attrition (especially in the digital intervention group), the findings should be interpreted with caution. Nevertheless, the results obtained offer an approximation of the trends in the effects for the different clinical variables evaluated, depending on the intervention modality. We found that, compared to the other two conditions, TDI appeared to achieve greater short- and medium-term reductions in alcohol consumption and in symptoms of depression and trauma among pregnant women with concurrent mental health and substance use symptoms. However, the DDI did not reduced anxiety symptoms and in some cases even worsened it. Irrespective of this, the potential beneficial effects of both interventions tended to disappear by one year, with outcomes converging across groups. In addition, overall retention rates dropped to approximately 40% at the end of the follow-up; the DDI group showed the lowest retention rate, with only about 9% of its initial participants remaining. Notably, most attrition occurred within the first two months, after which retention rates remained relatively stable. Substance use outcomes analysis Regarding substance use outcomes, baseline scores on the AUDIT and Fagerström scales indicated that most participants were in the low-risk range for alcohol-related problems and showed very low to low levels of nicotine dependence. The telephone-delivered intervention achieved the most pronounced short-term reduction in risk scores for alcohol use at the two and four month points, but this effect was not sustained at the one-year point, with all groups returning to similar risk levels. For smoking, our interpretation of the outcomes is compromised by the small initial sample size and the very high attrition rates (ranging from 4.5% in the TAU group to 88.2% in the DDI group). Importantly, the attrition rates for both alcohol and tobacco assessments were higher than for the mental health scales, with less than a third of participants completing the final assessment and the DDI group showing the lowest retention, which limits the interpretation of the results. Our findings provide preliminary supporting evidence for the potential usefulness of digital interventions in reducing alcohol consumption in early pregnancy and suggest that combining both approaches may lead to better outcomes [33]. The available evidence suggests that early motivational interventions can effectively reduce alcohol consumption amongst pregnant women [34]. This is in line with our results, which point to greater benefits of both therapeutic approaches during the second and fourth months. Regarding tobacco use, we should exercise caution in drawing definitive conclusions given the high rate of attrition. Interestingly, in our sample no women reported using substances other than alcohol or tobacco. Furthermore, the low AUDIT scores and low Fagerström test scores depict a population at low risk for substance use problems. We consider two possible explanations for these findings. First, participants may have been reluctant to disclose tobacco or alcohol use due to the stigma associated with maintaining these habits during pregnancy [7, 35]. Second, pregnancy itself increased awareness among the participants about the harmful effects of substance use, thus promoting greater self-care. This is consistent with the observed increase in abstinence rates at the end of the follow-up period. Nevertheless, it is essential to emphasise that any alcohol or tobacco use during pregnancy is clinically significant, underscoring the importance of implementing strategies that support cessation and reduce substance-related risks throughout gestation. Mental health outcomes analysis Regarding mental health outcomes, most participants presented with symptoms in the mild to moderate range for both depression and anxiety, and below the clinical threshold for trauma symptoms. The TDI achieved the greatest initial reduction in all three mental health outcomes, with the most pronounced improvements observed at four and eight months for depression and two and four months for trauma. However, these potential benefits were not sustained at one year, with scores returning close to baseline levels across all groups. In contrast, the DDI seemed to have limited therapeutic benefit, with anxiety symptoms remaining in the mild to moderate range throughout follow-up and no meaningful improvement in depression or trauma symptoms; in some cases, there was even a tendency toward higher anxiety scores compared to the other groups. A possible effect resulting from the weekly contact with patients in the TAU group by the care coordinator to monitor their progress may have contributed to mitigating anxiety in this group as well. Regarding anxiety and depression, our results seem to be in line with previous systematic reviews conducted in both the general population and specifically amongst pregnant women. In the general population, digital CBT interventions have been shown to be effective for the short-term improvement of depression and anxiety symptoms, especially when accompanied by human support such as telephone or therapist contact [36]. Similarly, in pregnant women, several reviews have concluded that e-health interventions have limited impact on preventing and treating symptoms of anxiety and depression unless human support is included, highlighting the need for more research [14, 33, 37]. Regarding PTSD, studies in pregnant women have not examined digital interventions in depth; however, evidence from traditional approaches suggests that the best outcomes are achieved with early intervention, which is consistent with our findings pointing greater improvement during the initial months of the trial [38]. Regarding digital interventions, recent research in the general population demonstrates the promising potential of e-health approaches [39, 40]. Nevertheless, two points should be highlighted in relation to our study: first, most existing research has focused primarily on youth populations [40], whereas pregnant women represent a distinct and unique group; and second, the baseline levels of PTSD symptoms in our sample were sub-threshold, indicating that this was not an ideal population for assessing the effectiveness of interventions targeting trauma symptoms. Finally, it is worth mentioning that the mean scores of each group remained within the moderate depression range throughout the entire year. This finding is consistent with evidence from multiple studies showing that the prevalence of clinically significant depressive symptoms during pregnancy ranges from 11% to 24%, and for major depressive disorder from 4% to 15%, depending on population, diagnostic method, and socioeconomic context [41– 43]. These data highlight the clear need for systematic depression screening during pregnancy. Adherence to treatment A significant finding of our study is that adherence to treatment was low, with up to 60% of participants discontinuing follow-up within the first few months and attrition rates reaching approximately 80% by the end of the study. This issue was particularly pronounced in the DDI group. These findings showed preliminary feasibility data that are consistent with previous research, which emphasises that improving adherence to psychosocial treatments during pregnancy is a persistent challenge. A recent systematic review found that attrition from psychotherapy is frequent among women with mental disorders, although reported rates vary and are generally not systematically presented as a primary outcome [44]. Observational studies have reported dropout rates of up to approximately 40%, depending on the context [45, 46]. Similarly, for substance use treatment in pregnant women, adherence rates are usually low, with attrition rates up to 50% for alcohol and tobacco use [47, 48]. Retention rates for digital interventions are highly variable. The limited literature available suggests high dropout rates; the most frequently reported range is between 30% and 50% for digital mental health interventions targeting pregnant women [49]. Similar figures have been observed for substance use treatment, with attrition rates typically ranging from 19% to 50% [33]. Notably, our figures even surpass this range. In this regard, it is worth considering whether the designed intervention might have caused resistance among the participating women, perhaps due to the stigma of seeing themselves reflected in the problems described. On the other hand, it is also possible that the participants did not fully identify themselves with the intervention because it highlighted somewhat more serious problems than those they were experiencing. The high attrition rates observed in our study may also be explained by the extended duration of follow-up, which lasted one year. Most interventions tested in this field are typically of shorter duration. Notably, the best outcomes tend to occur early in the intervention, suggesting that efforts to improve retention may be most effective if focused on the initial phase of treatment. Additionally, our sample was characterised by low initial risk scores for substance use, which may have reduced the perceived need or motivation to remain engaged in treatment. However, retention remained low despite high scores on mental health scales, particularly for depression. This underscores the need for further research to identify strategies that can effectively engage and retain this population in treatment. Implications for clinical practice We did not find evidence of an effect in the direction of superiority of digital interventions in our study. On the contrary, our findings suggest that telephone-based interventions may be associated with greater short- and medium-term reductions in alcohol use and improvements in symptoms of depression and trauma among pregnant women. Globally, our findings point to two key implications that may be of use for clinical practice: first, intervention efforts may be of more use if concentrate in the early stages of pregnancy; and second, digital interventions alone appear insufficient, and human support—whether by telephone or in person—remains necessary. Also, it is important to identify the patient profiles most likely to benefit from digital approaches. Adherence continues to be a major challenge in this vulnerable population. Additionally, these results indicate heterogeneous patterns across outcomes, with short-term benefits primarily concentrated in the telephone-delivered intervention (TDI) and limited effects for the digitally delivered intervention (DDI). Retention declined to approximately 40% overall, with the DDI group dropping to ~ 9% at 12 months, markedly lower than TDI and TAU. Such asymmetric loss to follow-up may introduce attrition bias if participants remaining in DDI differ systematically in symptom severity or motivation, potentially inflating or attenuating observed differences. This concern is amplified across scales with distinct baseline distributions and variances (e.g., low-range AUDIT and Fagerström scores versus moderate PHQ-9 and broader PCL-5 dispersion), where small group sizes in later waves reduce precision and undermine statistical power. Consequently, while early TDI effects appear for anxiety and depressive symptoms, long-term convergence across groups should be interpreted cautiously, as reduced and uneven sample composition limits the stability and generalizability of follow-up comparisons. Strengths and limitations As an exploratory study, some limitations must be noted. First, the clinical trial registry was retrospective. This may raise concerns regarding outcome pre-specification, analytical flexibility, and selective reporting. However, this study is part of a project funded by a national organization. A proposal was prepared to obtain the grant, which included the study design and the data analysis plan. In other words, although the registry was retrospective, the data analysis plan was specified and decided before the analyses began. Furthermore, the relevant variables and the primary and secondary measures to be considered were defined in the research design before the analyses were conducted. Finally, the results were interpreted within the context of an exploratory, not confirmatory, study. Methodological limitations should also be noted: in particular, the small sample sizes and high attrition rates must be taken into account when interpreting the results. Furthermore, it is important to consider that the low baseline values shown by the participants limit the capacity to detect meaningful change. The resulting floor effect could partly explain null or unstable findings. In this sense, the generalizability of the results obtained to women with moderate or severe symptoms may be compromised. Finally, end-user perspectives have not been incorporated during development and refinement of the DDI, which may have affected the lower effect found for this condition and the low retention rate of that cohort. Similarly, the digital platform used in this study was at the initial stages of its development, which may have affected the usability and contributed to dropouts in the DDI group. These limitations necessarily require caution when interpreting the conclusions drawn regarding the effect of the intervention on the DDI group. Despite its limitations, the study has notable strengths. The main one is that the sample focuses on a typically understudied group—namely, pregnant women with concurrent mental health and substance use symptoms—and simultaneously addresses both issues, in line with recommendations for integrated treatment approaches. The second main strength lies in its analysis of technology for delivering psychotherapeutic interventions. The integration of digital and telephone modalities represents a necessary approach, especially in contexts where accessibility and flexibility are essential. By leveraging technology, the study contributes to the growing body of research exploring alternative formats of psychological care beyond traditional face-to-face therapy. It is important to note that the primary objective of the study was not to establish definitive efficacy, but rather to examine the feasibility of implementing technology-based psychotherapeutic interventions and explore preliminary trends in treatment effects. The results support the integration of technology-assisted psychotherapy into clinical settings. The results also suggest that telephone-based interventions may yield greater benefits compared to purely digital formats. This pattern highlights the potential added value of direct contact with the therapist, even when conducted remotely. Another key strength of this study is its focus on attrition, which is known to be particularly high in technology-based interventions, especially among populations with mild to moderate symptoms. The high attrition rates observed reinforce the need to better understand participants’ motivations, expectations, and barriers to continued participation. Individualization of content and delivery format emerges as a crucial consideration for improving engagement, retention, and ultimately, intervention effectiveness in this population. Together, these strengths position the study as a meaningful contribution to the development of technology-assisted psychotherapeutic models, Conclusions Our exploratory findings contribute to the limited existing literature on integrated treatments for co-occurring mental health and substance use symptoms in pregnant women and provide valuable insights into the feasibility and direction of short-term effects of different treatment approaches for this population. By using e-health, the WOMAP study addressed the need for accessible treatments for pregnant women with co-occurring mental health and substance use symptoms, which has been identified as a gap in previous research. More research is needed to clarify whether the digital approach is useful on its own, or whether it is more useful in combination with a traditional face-to-face approach. Supplementary Information Supplementary Material 1. Abbreviations Woman Mental Health and Addictions on Pregnancy Adult Screen for Co-Occurring Disorders Treatment as usual Patient Health Questionnaire − 9 General Anxiety Disorder seven-item screener Post-Traumatic Stress Disorder Checklist Post-Traumatic Stress Disorder Alcohol Use Disorders Identification Test Drug Abuse Screening Test Telephone-delivered intervention Digitally delivered intervention Human immunodeficiency virus Sexually transmitted disease Cognitive-behavioural therapy Short Message Service Secure Socket Layer/Transport Layer Security Authors' contributions Rodrigo Carmona Camacho: Conceptualization, Data curation, Project administration, Supervision, Validation, Writing – original draft; Writing – review & editing. Nayara López Carpintero: Conceptualization, Data curation, Validation, Writing – original draft; Writing – review & editing. Lucia Albarracín García: Conceptualization, Data curation, Supervision, Validation, Writing – review & editing. María Luisa Barrigón: Formal analysis, Methodology, Writing – original draft, Writing – review & editing. Igor Barahona: Formal analysis, Writing – review & editing, Enrique Baca-García: Conceptualization, Data curation, Funding acquisition, Resources, Supervision, Validation, Writing – review & editing. Cristina Ruiz Nogales: Investigation, Writing – review & editing. Inés Menéndez: Investigation, Writing – review & editing. Montserrat Sánchez Alonso : Investigation, Writing – review & editing. Carmen Vidal Mariño : Investigation, Resources, Writing – review & editing. Lucía Rodríguez Blanco: Investigation, Resources, Writing – review & editing. Silvia María Campos Soler: Investigation, Writing – review & editing. Montserrat Díaz Rosell: Investigation, Writing – review & editing. Benjamin Le Cook: Conceptualization, Investigation, Methodology, Writing – review & editing. Margarita Alegría: Conceptualization, Data curation, Funding acquisition, Resources, Supervision, Validation, Writing – review & editing, Irene Caro Cañizares: Investigation, Methodology, Resources, Supervision, Validation, Writing – original draft; Writing – review & editing. Funding This study was funded by the Spanish Government Delegation for the National Plan on Drugs (20151073) and the Spanish Ministry of Education and Universities (CRECER/GROW Project, PID2024-156704OA-I00). Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study was carried out in compliance with the Declaration of Helsinki and approved by the Jiménez Díaz Foundation Ethics Committee for Clinical Investigation (Ref. 2015/43). After a complete description of the study, all screened participants gave informed consent. To ensure data protection, access to the clinician user interface was restricted to researchers. Only the principal investigator, clinicians and researchers, using usernames and passwords, had access to identifiable information. The data provided by the clinician were encrypted via Secure Socket Layer/Transport Layer Security (SSL/TLS) between the investigator’s computer and the server. Data were stored in an external server created for research purposes, and were encrypted using the industry-standard AES-256 algorithm. Furthermore, an external auditor guaranteed that the security measures met the Organic Law for Data Protection (Spanish Government, Ley Orgánica 3/2018 de Protección de Datos Personales y garantía de los derechos digitales, 2018) standards at a high protection level. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.