Eligibility criteria include (1) participants older than 18 years of age; (2) English-speaking; (3) report cannabis use at least twice weekly over the past 60 days; (4) pregnant with one child with a gestational age no more than 28 weeks (to allow time for scheduling of baseline); (5) not currently in substance use treatment; (6) no current (past month) mania, psychosis, or active suicidal ideation; (7) no substance use other than cannabis, tobacco, or alcohol; and (8) own a personal smartphone to accommodate the data collection app. For revisions to eligibility criteria to address initial recruitment challenges, see Results section.

Participants were recruited through social media, community flyers (posted in locations such as playgrounds, Women, Infants, and Children offices [supplemental nutrition program for women, infants, and children], and obstetrics/gynecology clinics), and via BuildClinical, a platform that identifies potential research participants using online advertising campaigns and conducts initial eligibility screening. Respondents to flyers or online advertisements accessed an online screener hosted on Qualtrics via a clickable link or the QR code. This screener included questions aligned with the study’s inclusion and exclusion criteria. Those meeting initial eligibility criteria were prompted to complete a contact form specifying their preferred method and time for follow-up. Study staff then contacted these individuals to confirm eligibility through a telephone screener. Alternatively, prospective participants could reach out directly via call or text to schedule screening. Eligible participants were invited to an in-person baseline assessment and instructed to abstain from using cannabis for at least 15 hours prior to the visit.

After verifying identity and age via government-issued ID (eg, driver’s license), participants complete the informed consent process, which was presented via PowerPoint (Microsoft Corp) to facilitate comprehension. After consent is obtained, participants provide biochemical verification of eligibility. Biochemical verification includes a commercially available pregnancy test to confirm pregnancy and a 5-panel iCup urine drug screen which screens for amphetamines, cocaine, methamphetamine, opiates, and tetrahydrocannabinol, as well as validity and specimen integrity checks, including oxidants, pH level, and specific gravity, to confirm cannabis use and no other drug use. Tetrahydrocannabinol is detected at 50 ng/mL. If participants tested negative for tetrahydrocannabinol at 50 ng/mL, a tetrahydrocannabinol dipstick with a greater sensitivity of 5 ng/mL was administered. A saliva test is also conducted to ensure that the participant has not used cannabis in the past 15 hours, with a cut-off threshold of 40 ng/mL of tetrahydrocannabinol. This test was conducted to ensure no intoxication was experienced at study sessions.

Following bioverification of eligibility, the Mini-International Neuropsychiatric Interview (MINI [38]) is administered to rule out past month mania, psychosis, or active suicidal ideation, and the Timeline Followback (TLFB) is administered to rule in twice weekly cannabis use and rule out substance use other than alcohol and nicotine. The MINI is administered by a PhD-level study staff member or a student currently enrolled in an MPH program; both are supervised by a licensed clinical psychologist. Participants deemed ineligible are compensated US $15 and provided psychoeducational resources on cannabis and mental health. These resources are provided to all who were screened and enrolled in the study.

The baseline assessment lasts approximately 3-4 hours. Participants complete baseline surveys on substance use patterns and psychosocial factors that may influence maternal health. Finally, participants view a PowerPoint presentation and receive hands-on training to familiarize themselves with completing EMA surveys. This training ensured they were comfortable using MetricWire, the EMA mobile app, on their smartphones to report real-time data on daily substance use, mood, and other contextual factors. Participants are subsequently enrolled in the EMA burst procedures, which begin the day following the baseline assessment and last for 2 weeks.

Each trimester follow-up assesses recent (30-day) substance use, as well as current depression and anxiety symptoms.

A 5- to 10-minute phone call is conducted 0-14 days post delivery to gather basic information about the newborn. Participants are asked to text research staff upon the delivery of their baby. If staff do not receive a text, they refer to the participants’ estimated delivery date and send text and email reminders to let study staff know when they deliver.

A 30-minute virtual visit occurs at approximately 6 weeks post partum. Participants self-report on substance use, mother-infant bonding, birth experience, infant development, and parental substance use. Participants were refamiliarized with the EMA procedures and are enrolled in EMA burst 2, which begins the day after the Time 2 assessment.

The participant and their infant attend a 90-minute laboratory assessment at approximately 6 months post partum. The birthing person reports on their substance use, depression, and anxiety. Dyads also complete behavioral tasks, which are video-recorded for later behavioral coding. Following the mother-infant interaction tasks, study staff debrief with the parent to review their experience in the paradigm and ensure mother and baby are comfortable. Participants are refamiliarized with the EMA procedures and are subsequently enrolled in the third and final burst, which begins the day following the 6-month assessment.

EMA data are collected via MetricWire, enabling individualized data collection via participants’ smartphones. At baseline, participants are trained to use the mobile app, including specific instruction on responding to prompts and measurement of cannabis quantity. Participants are instructed to self-initiate reports just before using cannabis (begin cannabis report), which assesses for the formulation of cannabis being used, current reasons for using cannabis, and current mood. After participants complete a begin cannabis report, they are instructed to complete an end use report, and follow-up reports are triggered at 30-minute intervals to assess for the end of the cannabis use event if not self-initiated. Once the end use report is completed, cannabis use event details are assessed, including context, any other substance use, as well as impact on mood and infant presence (postpartum bursts only). Participants also complete random prompts throughout the day (sent randomly within three 3-hour blocks each day) to capture mood during noncannabis use events and cannabis use that is not self-initiated. Participants also complete morning reports, which assess overall sleep quality, mood symptoms, intentions to use cannabis that day, and summarize the prior day’s substance use. Bedtime reports capture mother-child interactions from the day (postpartum bursts only) and any additional use since the last report. Assessments are designed to be less than 2 minutes. Similar methods have been used by Gunn et al [39].

Participants could earn up to US $595, dispersed on reloadable debit cards. The compensation schedule is outlined in Tables 1 and 2. Payments are provided following each assessment. Compliance is maximized in the EMA study phase by compensating participants by increasing USD amounts based on compliance rates. EMA compliance rates are calculated following the completion of each burst and compliance is compensated following completion of each burst.

The MINI is a short structured diagnostic interview designed to assess major DSM-IV (Diagnostic and Statistical Manual of Mental Disorders [Fourth Edition]) and ICD-10 (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision) psychiatric disorders [38]. The MINI assesses for current (past month) mania, psychosis, and suicidal ideation, all of which were exclusion criteria in this study. In addition to these screening data obtained to inform study eligibility, the MINI includes assessments of current and past psychiatric diagnoses, including major depressive disorder, panic disorder, agoraphobia, social anxiety disorder, obsessive compulsive disorder, posttraumatic stress disorder, alcohol and substance use disorder, anorexia nervosa, bulimia nervosa, binge eating disorder, generalized anxiety disorder, and antisocial personality disorder.

Participants self-report their substance use with the TLFB, a structured, calendar-based interview tool used to collect detailed retrospective data on substance use [40,41]. Participants are guided to recall daily activities and contextual cues to enhance the accuracy of their reports. The TLFB is included as part of the in-person screening to identify use of cannabis (eligibility), alcohol, tobacco, as well as to capture the use of other substances (ineligibility). Participants also report mode of cannabis use. Duration covered by the TLFB varies by assessment, such that the reporting period was either past 90-day (Time 1) or past 30-day (trimester follow-ups, Time 2, and Time 3) use.

Participants report on their demographics (eg, age, biological sex at birth, gender identity, race, ethnicity, income, government assistance), as well as contextual information about the home environment, including the number of children and relationship status. Date of last menstrual period and trimester of pregnancy are queried to determine gestational age. Additional questions probe prenatal care, including the number of prenatal appointments attended and how pregnancy has been confirmed.

Participants report on their use of nicotine or tobacco products (cigarettes, e-Cigarettes, or nicotine vaporizers, other tobacco or nicotine products) and alcohol use, as well as the frequency and quantity of use for each trimester. Participants report on parallel information about their partner’s substance use (alcohol, cannabis, cigarettes, other tobacco products, and other drugs) since the participant found out they were pregnant.

Participants are asked if they have a medical card, if they use the card to purchase cannabis, and for what conditions they have been provided a medical card. Moreover, they report on their current level of cannabis use (ie, if they use the same, less, or more than when they found out they were pregnant), if they have quit or temporarily stopped using cannabis since finding out they were pregnant, and if so, why.

For each trimester, participants report number of days of cannabis use per week, average times per day of cannabis use, average number of hours high per cannabis use day (presented on a visual analog scale from 0 to 24 hours), how high they felt when using cannabis (visual analog scale ranging from 0=not at all high to 100=as high as possible). Amount of tetrahydrocannabinol versus cannabidiol in cannabis used was assessed (mostly or entirely cannabidiol/no tetrahydrocannabinol, roughly equal parts cannabidiol or tetrahydrocannabinol; mostly or entirely tetrahydrocannabinol or no cannabidiol, or unsure). Finally, participants indicate the number of days (0-7) that they used each of the following: join, hand pipe, bong, blunt, concentrates or dabs, edibles, tetrahydrocannabinol-infused alcoholic beverages, tinctures, and cannabidiol-only products (eg, lotions) [42].

Participants are provided a list of 10 physical symptoms common in pregnancy (ie, stomach or digestion problems, physical pain, neurological symptoms, sleep, fatigue, bleeding, Braxton-Hicks contractions, dyspareunia, hot flashes, and appetite) and 13 emotions (ie, sad or depressed, happy, nervous or anxious, excited, decreased pleasure or motivation, angry, optimistic, lack of energy, fidgety or restless, scared, irritable, and boredom stress). Participants indicated if they did or did not use cannabis for managing each symptom or emotion, and if they reported using cannabis to manage the symptom, they were asked how helpful cannabis was at managing that symptom (response options were did not help, helped somewhat, helped a lot). This measure was developed for this study.

The Brief Marijuana Consequences Questionnaire [43] assesses the negative consequences, grouped into 5 categories: interpersonal, health, academic or occupational, legal, and psychological, associated with cannabis use since the participant found out they are pregnant. Respondents rate how frequently they have experienced each consequence on a 5-point Likert scale (ranging from 0=never to 4=often).

Participants report on how harmful they think their cannabis use is to their own health, to the health of their baby during pregnancy, and to the health of their baby after they give birth. An additional question probes how harmful they think it is to breastfeed while using cannabis. All questions are presented with separate visual analog scales ranging from 0 (not at all harmful) to 100 (very harmful). In addition, participants are asked the extent to which they agreed or disagreed with statements about the perceived harms of alcohol use, nicotine or tobacco use, and use of other drugs to the developing child. Items were adapted from Nordeng et al [44].

Participants are asked where they sought information about substance use and who provided them with information, including unsolicited advice. Response options included physician, midwife, nurse, other health care provider, internet, pregnancy groups, family or friends, dispensaries, or other. A follow-up, open-ended question assessed the type of information received from these sources. Participants also reported if they disclosed their cannabis use to their provider (yes or no), how comfortable they are in discussing their cannabis use with their provider (4-point Likert scale from not at all comfortable to very comfortable), and how honest they have been with their provider about their cannabis use. This measure was developed for this study.

Participants are also asked how much they agree or disagree on a 5-point Likert scale, including a “not sure” option, with the following statements: “there is negative stigma around using cannabis” and “there is negative stigma around using cannabis during pregnancy.” Additionally, participants reported how acceptable or unacceptable they find both medical or therapeutic and recreational cannabis use.

The Protective Behavioral Strategies for Marijuana Scale is a self-report tool designed to assess how often individuals use strategies to minimize potential harms from cannabis use [44,45]. The scale consists of 17 items (eg, “Only use once during a day/night”), each measuring the frequency of specific protective behaviors (response options are never, rarely, occasionally, sometimes, usually, and always). To our knowledge, this is the first application of this measure in a pregnant sample. Together with the developers of the measure, we made 3 modifications to the scale. First, the question stem specifies the use of protective behavioral strategies during pregnancy. Second, replaced “marijuana” with the individual’s preferred term. Finally, we included the following open-ended question to probe for protective behavioral strategies that were not captured in the initial measure: “What other things do you do to help you avoid some of the not so good things about using [preferred cannabis term] while you are pregnant?”

The Edinburgh Postnatal Depression Scale [46] was used to screen symptoms of depression and has been validated in pregnant and postpartum samples [47]. The Edinburgh Postnatal Depression Scale consists of 10 items (eg, “I have felt sad or miserable”) scored on a 4-point Likert scale (0-3), reflecting the frequency or severity of symptoms over the past week. Scores range from 0 to 30, with higher scores indicating greater severity of depressive symptoms.

The Perinatal Anxiety Screening Scale (PASS [48]) is designed to screen for a broad range of anxiety symptoms in pregnant and postpartum individuals. The PASS consists of 31 items scored on a 4-point Likert scale ranging from 0 (not at all) to 3 (almost always). The PASS includes four subscales, including excessive worry and specific fears (eg, “Fear that harm will come to the baby”), social anxiety (eg, “Fear that others will judge me negatively”), perfectionism and control (eg, “Needing to be in control of things”), and acute anxiety and trauma. Total scores can range from 0 to 93, with higher scores indicating more severe anxiety symptoms. Anxiety thresholds are defined as low (0-20), moderate (21-41), or high (42+).

Two questions query social support in pregnancy; the first asked if participants have been adequately supported in their pregnancy (1=never to 5=always), and if they have or would be willing to attend a support group for pregnant individuals. Items were developed for this study.

Participants report their intended primary method of feeding their infant during the first 6 months (options: breastfeeding or chestfeeding, pumping, formula-feeding, milk sharing or milk bank, or other). They also complete a check-all-that-apply item listing all feeding methods they planned to use during this period. Participants indicated whether they believe cannabis can pass through breast or chest milk (response options: yes, no, or I don’t know). Items were developed for this study in collaboration with an International Board Certified Lactation Consultant.

The Maternal-Fetal Attachment Scale is a self-report questionnaire designed to measure the emotional bond that a mother develops with their unborn child during pregnancy [49]. The Maternal-Fetal Attachment Scale assesses the extent to which a pregnant person engages in behaviors that reflect attachment to their fetus. The scale consists of 24 items with four response options (definitely yes, yes, no, and definitely no). The Maternal-Fetal Attachment Scale includes 5 subscales: differentiation of self from fetus, interaction with fetus, attributing characteristics to fetus, giving of self, and role taking. Higher scores indicate stronger maternal-fetal attachment.

Demographic information collected included the number of children and adults in the home, as well as adult substance use within the household. Participants were presented with a table reflecting the previously reported number of adults in the home and were asked to indicate if any adult had used cannabis, alcohol, cigarettes, other nicotine products, or other drugs since returning home after delivery.

The Childbirth Experience Questionnaire [50] provides a comprehensive assessment of a birthing person’s perceptions and feelings about their labor and birth experience. The measure includes 19 items rated on a 4-point Likert scale ranging from 1 (totally agree) to 4 (totally disagree). The measure consists of 4 key domains: own capacity (eg, I felt strong during labor and childbirth), professional support (eg, my provider understood my needs), perceived safety (eg, I felt scared during labor and birth), and participation (eg, I felt I could have a say in deciding my birthing position).

Participants reported on whether or not the family has had Child Protective Services involvement since the birth of the baby.

Participants were asked if their baby spent any time in the neonatal intensive care unit, and if so, the number of days.

The Parental Stress Scale [51] is an 18-item questionnaire assessing parents’ feelings about their parenting role, exploring both positive (eg, emotional benefits and personal development) and negative aspects of parenthood (eg, demands on resources and feelings of stress). Respondents agree or disagree on a 5-point Likert Scale in terms of their typical relationship with their child or children. Example items include “I am happy in my role as a parent” and “I sometimes worry whether I am doing enough for my child(ren).”

The very short form of the Infant Behavior Questionnaire [52] is a widely used parent-report tool designed to assess temperament in infants aged 3 to 12 months. The measure includes 37 items with response options of 1 (never) to 7 (always), including a not applicable option. Multiple dimensions of temperament are assessed, including activity level, distress to limitations, fear, smiling and laughter, duration of orienting, soothability, and falling reactivity or rate of recovery.

Participants completed developmentally appropriate Ages and Stages Questionnaires [53] at 6 weeks and 6 months post partum. This measure is a widely used developmental screening tool designed to assess young children’s developmental progress and identify potential delays or concerns. The 2-month questionnaire (which covers babies ages 1 month 0 days to 2 months 30 days) assesses a child’s development across 5 domains: communication (eg, ability to respond to sounds and voices), gross motor skills (eg, large muscle movements), fine motor skills (eg, hand-eye coordination and small muscle control), problem solving (eg, early cognitive skills), and personal-social skills (eg, social interactions and self-regulatory behaviors). Each domain includes 6 questions that parents answer based on their own observations of their child. Similarly, the 6-month questionnaire assesses a child’s development across the same 5 domains as the 2-month version, but with milestones appropriate for a 6-month-old infant.

Parents completed 2 questionnaires to identify potential issues in bonding and early signs of difficulties, such as maternal distress or postpartum mood disorders. The Postpartum Bonding Questionnaire [54] includes 25 items that are designed to assess the quality of the emotional bond between a parent, typically the mother, and their newborn. The measure includes 25 items, rated on a 6-point Likert scale from “always” to “never.” Four subscales assess: impaired bonding (eg, “I feel distant from my baby”), rejection and pathological anger (eg, “I feel irritated by my baby”), anxiety about care (eg, “I feel anxious when my baby is with me”), and risk of abuse (eg, “I feel like I might hurt my baby”). The Mother Infant Bonding Scale [55] includes 8 items that reflect aspects of maternal feelings toward the infant (eg, “I feel loving toward my baby”) on a 4-point Likert scale ranging from “Very Much” (0) to “Not at All” (3).

Mental health, substance use, parenting stress, bonding, and child development are reassessed at 6 months. Additionally, participants complete 2 behavioral tasks that are video recorded using a 3-camera mounted recording system. This system includes one camera focused on the mother, one on the infant, and one capturing a wide-angle view of the entire room. Video signals from the two cameras are transmitted through a digital mixer to generate a split-screen recording of the mother and infant. The recording process is managed by a research assistant stationed outside the laboratory.

Still-Face Paradigm [56] is a laboratory-based assessment in which a caregiver’s unemotional face is used to evoke pronounced behavioral reactions in the infant. Mothers are seated 18-36 inches from their infants, at eye level with their infants, who are placed in an infant seat by the mother. The paradigm involves three phases of mother-child interaction: (1) Baseline: in the first interaction, mothers are instructed to interact normally with their infants. (2) Still-face: after the mother returns from a break, they are told to keep eye contact but maintain a neutral expression, remain still, and do not touch or interact with their infant. (3) Reunion: after the second break, the mother is instructed to respond to their baby in any way they feel is appropriate. Each episode lasts 120 seconds, and the onset of each episode is cued by a research assistant who is seated in the room, out of sight of the infant.

Mother and infant behaviors are coded using the Infant and Caregiver Engagement Phases framework [57], which includes a set of mutually exclusive phases of interactive engagement and additional regulatory codes based on facial expressions, direction of gaze, and vocalization. The Infant and Caregiver Engagement Phases coding scheme has been applied to the paradigm conducted with substance-exposed infants [57] and infants of mothers experiencing psychiatric symptoms and stress [58]. For infants, engagement codes include passive-withdrawn, protest, object-environment interaction, social monitoring, and social positive engagement. Caregiver engagement codes consist of hostile-intrusive, withdrawn, social monitoring without vocalizing, social monitoring with positive vocalizing, and social positive engagement. Additional infant behaviors coded involve oral self-comforting (eg, mouthing), self-clasping, distancing or turning away, and autonomic stress indicators such as hiccups or spitting up. For caregivers, additional codes capture rough touches and deviations from still-face instructions, including speaking to or touching the infant during the designated still-face period. Infant and mother phases will be coded in two separate viewings of the videotape, with a third viewing dedicated to coding of infant regulatory actions, autonomic stress indicators, mother intrusions, and procedural violations.

Mothers are instructed to play with a standardized set of developmentally appropriate toys as they normally would at home for 5 minutes. The research assistant is seated outside the room. The free play interaction will be coded using the Emotional Availability Scales [59], based on global judgements from the observers, and designed to assess the emotional connection between parent and child.

For all assessments, cannabis use is assessed by asking participants whether they used flower, concentrate, or edible (with “select all that apply”), followed by the quantity of cannabis used based on formulation. (ie, flower=grams of flower, concentrate=number of hits, and edible=milligrams of tetrahydrocannabinol). Additionally, participants report the potency of cannabis consumed and the level of intoxication or “high” (“rate how high you feel from 0=not at all high to 10=as high as possible”).

Participants selected all that apply: “to be social,” “to feel less anxious,” “to feel less depressed,” “to relieve physical pain,” “to sleep better,” “to relax,” “to enjoy the effects,” “to increase appetite,” and “other.”

Participants are prompted to report the context in which they consume cannabis, indicating the number of people they were with, the specific people they were with (eg, friend, partner, and family), and whether or not those individuals were also consuming cannabis at the time.

Participants are prompted to report any other substance use (ie, “Did you/are you using any other substance?”), reporting these substances via a select all option for alcohol, cigarettes, nicotine vaporizers, other tobacco or nicotine products, simulants, opiates, sedatives, hallucinogens, and other, if applicable.

Participants report on their current mood with an adapted version of the NIH Patient Reported Outcomes Measurement Information System adult depression short [60] form on a 5 point scale: “Rate how much you currently feel this way” from 1=not at all to 5=very much, for depressed, hopelessness, helpless, worthless, anxious, irritable, happy, excited, and relaxed.

For Time 3 only (ie, post partum), surveys assessed for the following mother-child questions, all reported on a 10 point scale from (not at all to as much as possible): “How bonded did you feel to your child today?” “How emotionally connected did you feel with your baby today?” “How sensitive did you feel toward your baby today?” “When your baby needed something today, how responsive were you to their needs?” “Did you provide the comfort and care that your baby needed today?”

The University institutional review board completed and approved human subjects’ ethics review (approval number 2022003468). Research procedures have been conducted in accordance with the Declaration of Helsinki. All participants provided informed consent, during which a research assistant reviewed all aspects of the study protocol, and knowledge checks were conducted before each participant gave consent. Participants are provided with the opportunity to opt out. Participants have received monetary compensation, outlined below, which was deemed appropriate for the time required to complete study procedures.

This project is supported by Brown University’s Center for Addiction Disease Risk Exacerbation (CADRE), funded through NIH’s Centers of Biomedical Research Excellence. All projects supported by CADRE benefit from the resources of three cores: the Clinical Laboratory Core, the Recruitment, Engagement, and Community Health Core, and the Administrative Core. The Clinical Laboratory Core provides essential research infrastructure and resources to advance CADRE-funded research. A key objective of this core is to create a database of factors (biological, environmental, social, and behavioral) associated with the development and progression of substance use disorders and chronic disease. In alignment with this objective, the current project incorporated the CADRE Core measures into its data collection process. A complete list of these measures is provided in Multimedia Appendix 1. These measures were administered at the baseline session after the study-specific measures listed below. For the purposes of this report, we focus on the research activities and measures specifically relevant to the aims of the current research.