Smoking, nicotine and pregnancy 2 (SNAP2) trial: protocol for a randomised controlled trial of an intervention to improve adherence to nicotine replacement therapy during pregnancy

Primary objective

To determine whether, when added to the usual NHS cessation support, the BMN intervention increases smoking abstinence during pregnancy, as measured in late pregnancy or at childbirth, with exhaled CO and/or saliva samples used to validate self-reported abstinence.

Secondary objectives

In all participants, to compare between the intervention and usual care groups:

1. Reported smoking abstinence at 28 days after the quit date (QD).

2. Reported smoking abstinence at both 28 days after a QD, and in late pregnancy/childbirth with and without validation in late pregnancy.

3. The number of days of NRT use in the first 28 days after the QD.

4. Mean daily nicotine dose in the first 7 days after the QD (‘intensity’ of the NRT use).

5. Adverse pregnancy outcome rates.

In the intervention group participants:

1. To assess engagement with BMN intervention components.

Economic

1. To investigate the cost-effectiveness of the BMN intervention.

Pilot phase objectives

In all pilot phase participants, the intervention and usual care groups were compared:

1. Urges to smoke and tobacco withdrawal symptoms on Day 7 after the QD.

2. NRT concerns and necessity beliefs on Day 28.

In the pilot phase, participants and practitioners in the intervention group only:

1. To assess the fidelity of the BMN intervention delivery.

Other objectives (pilot phase only)

To compare biochemically measured nicotine exposure before and after exposure to BMN intervention

Inclusion criteria

People are eligible if aged ≥16 years; pregnant and <25 weeks of gestation; they smoked ≥5 daily cigarettes before pregnancy (currently smoking ≥1 daily cigarette) and are referred for or receiving antenatal care. Participants must have sufficient understanding of English to give informed consent; agree to try quitting smoking with NRT within 14 days, receive and send short message/messaging service text messages, and instal the trial’s data collection app on their smartphone.

Exclusion criteria

They are ineligible if already using NRT or are enrolled in a smoking cessation study, NHS stop smoking support or a cessation-orientated text message service, or they intend to continue using e-cigarettes or have contraindications to NRT.

Recruitment

Participants will be identified from:

1. NHS clinical settings, by poster, direct contact from researchers (face-to-face or distanced) and online, with adverts in NHS digital spaces.

2. Online, outside of NHS settings.

NHS settings

These can be hospital antenatal care, general practice, community midwifery or stop-smoking service settings. Researchers may identify potential participants from medical records, contacting them by letter, telephone, email or text before appointments, and including QR code/links to or paper copies of a Patient Information Sheet (PIS) (see online supplemental file). They may also approach pregnant people attending clinics asking them to complete an eligibility screening questionnaire or give them a summary leaflet that contains links/QR codes leading to the PIS. Depending on the setting, researchers may consent to those who are interested and eligible, or they may pass contact details to the trial team (see online supplemental file) to enable consent to be received by them.

Posters describing the trial will be displayed in clinical areas or appropriate NHS digital spaces; these will include QR codes or links leading to the PIS and to an online version of the screening questionnaire, following which eligible and interested people will be invited to leave contact details in a secure RedCap database hosted by the University of Nottingham. The trial team can then access contact details and contact potential participants directly.

Online, outside of NHS settings

Google or social media (eg, Facebook, Instagram,) adverts will be targeted at those who smoke and are pregnant. Embedded links will lead to a study information webpage and an eligibility questionnaire, and those who are potentially eligible will be asked to enter contact details, as above.

Interested, eligible potential participants will be given at least 24 hours to consider the PIS before discussion with a researcher and informed consent is received. Discussion and documentation of consent could be by face-to-face (using ‘wet ink’) or ‘distanced’ using either an online form or by telephone (see online supplemental file). If the online form is used, a link is sent to eligible potential participants and then they will fill out the form during a telephone conversation with a researcher. Consent via telephone is similar, but in this case, the consent form will be generated by the researcher from the research database, following a strict protocol, and then signed copies will be shared with the participant. For all consent methods, a letter will be sent to the participant’s GP informing them of the enrolment.

Control

The usual care for smoking cessation, following the National Centre for Smoking Cessation Training (NCSCT) standard treatment programme,28 comprises:

1. Helping set a QD.

2. Conducting up to six telephone or video call counselling sessions.

3. Offering NRT as a patch, short-acting NRT or combined (‘dual NRT’).9 10

Intervention

The BMN intervention is offered alongside UC (described above) and integrated into an identical schedule. BMN is described in detail elsewhere26 and comprises tailored behavioural support designed to encourage adherence to NRT and increase quit rates during pregnancy. The main components are as follows:

Staff delivering interventions

All SSPs delivering support are trained to the recognised NCSCT standard required for delivering UC stop smoking support in the UK NHS.²⁶ All intervention group participants are counselled by specially trained SSPs working within the research team, who deliver the BMN components integrated into UC. Control group counselling is provided by either a separate group of research team SSPs or by NHS providers responsible for providing locally available UC stop-smoking support.

Baseline

We will ask participants about demographics, gestation, estimated date of delivery, partner smoking status, whether they smoked in previous pregnancies, nicotine dependence,30 current and pre-pregnancy smoking behaviours, experience using NRT, smoking beliefs, urges to smoke (cravings) and tobacco withdrawal symptoms.31 Where possible, those recruited in person will also provide an expired air CO reading at baseline. Researchers will also help participants to instal the ‘NicUse’ smartphone app from Google Play and Apple Store,32 a bespoke app developed for SNAP2 on which applicants record daily smoking, e-cigarette and NRT use (see below) during the 28-day intervention period. We will seek contact details for participants’ GPs and check with them if there is any reason why a participant should not be enrolled.

Follow-up

After baseline, apart from data collection by app and routine sources, the primary data collection mode is by online questionnaires, with links texted or emailed to participants. If there is no response, a reminder text and/or email will be sent first, then participants will be called and hard copy questionnaires with reply paid envelopes will be posted as a final option.

Daily NicUse recording of behaviours

NicUse works on Android or Apple smartphones. Participants will be asked to use the app daily to record smoking behaviour, NRT and/or e-cigarette use. If participants miss reporting for 1 or 2 days, they can record these data retrospectively. Participants receive text messages to prompt completion of the app. The survey on NRT will ask participants to record patch and short-acting NRT use, and the number of units of short-acting NRT consumed, which allow us to calculate their daily nicotine dose.33 Compared with questionnaires, NicUse facilitates a more complete collection of NRT adherence data, with more robust face validity, and is less likely to overestimate adherence than questionnaires.33

Day 7 after the QD (pilot phase only)

We were asked if a quit attempt was made; if any NRT had been used since the QD and, if so, how many days, which types; and if short-acting NRT is used, how many lozenges, cartridges or sprays were used. We also asked about current smoking, e-cigarette use, urges to smoke (cravings) and tobacco withdrawal symptoms. As soon as possible after Day 7, saliva samples (see above) were collected from participants. After the pilot, we discontinued data collection at this time point, as this information was only required for pilot phase outcomes.

Day 28 after the QD

We will use the same questions as in the pilot phase-only Day 7 follow-up but will ask about the previous 28 (not 7) days, and we will ask the intervention group about BMN components (eg, texts and website). Participants who do not provide information on smoking status or NRT use by app or questionnaire methods will be sent a text about NRT use since the QD and asked to reply directly.

36 weeks of gestation or delivery (if earlier)

We will ask participants about smoking, adherence to NRT since 28-day follow-up and use of NHS stop smoking support. We will send non-responders a direct-reply text message asking about smoking in the previous 7 days.

Routine data collected from medical records

NHS hospitals collect smoking status and exhaled CO from every woman from week 36 of gestation onwards, and we will collect these data from NHS records. We will also extract maternal and fetal pregnancy outcome data from medical records. In NHS hospitals, research staff will collect these data; otherwise, the research team contact relevant NHS staff to ask for this information. Where birth outcome data cannot be obtained from records, using methods outlined above, we will ask participants to provide birth weight, gestation at delivery, whether they underwent caesarean section or the baby was admitted to special care or had any congenital abnormality and whether they were a smoker or non-smoker at delivery.

Validation of smoking abstinence

For participants who report 7-day smoking abstinence at 36 weeks or later, we will collect saliva samples as they attend hospital, at home visits or postally (see above). Before giving samples, participants will be asked about any recent smoking, vaping or NRT use.

Financial incentives encouraging data return

To recognise time taken for study participation, participants will receive up to £50 in ‘Love to Shop’ gift cards, which cannot be redeemed for tobacco or alcohol. To receive maximum remuneration, participants will need to submit all adherence app data; they will receive 50p for each daily app report, and an additional £1.50 for supplying a continuous full week’s reports, plus an additional £5 if they report for all 28 days after their QD (maximum total £25). Further gift cards will be given to women if they provide requested questionnaire data and validation saliva samples.

Fidelity assessment (pilot phase only)

We will audio-record all initial intervention group consultations, selecting a random sample for further scrutiny. Two researchers will listen to the selected audio recordings, independently rating the completeness of intervention delivery against a fidelity checklist that lists key components of the BMN intervention, and inter-rater reliability between researchers will be determined. We will store recordings on a secure University of Nottingham server for a maximum of 7 years.

Data management

Each participant will be assigned a unique study identification number allocated at consent to identify their data and biological samples. Personal identifiers (name, email address and phone number) will be stored in a password-protected computer database accessible only by the researchers. Data will be entered into a REDCap database where possible, but paper case report forms may be used as source data and entered by researchers into the database. Information submitted by participants via the NicUse app is stored as pseudonymised data on the Amazon cloud.

All electronic data will be securely stored at the University of Nottingham for 15 years after which it will be destroyed. Data management will be led by the York Trials Unit, with the support of the Trial Manager at the University of Nottingham as detailed in the Data Management Plan.

The Trial Management Group is responsible for the day-to-day running of the trial, meeting regularly and is supported and reportedby the Trial Steering Committee (TSC) (see online supplemental file).

Saliva samples

Saliva samples will be collected by researchers or by participants and sent by post directly to ACM labs for storage and analysis at the end of the trial. The laboratory will quantify salivary cotinine concentrations and the presence or absence anabasine using a quantitative enzyme immunoassay technique. Once the analysis has been completed, the saliva samples will be destroyed in accordance with the Human Tissue Act 2004.

Design changes

Originally, we planned SNAP2 as an RCT to test the extent to which BMN did or did not increase adherence to NRT; adherence to NRT was intended as the primary outcome and smoking outcomes as secondary. If study findings gave a sufficiently positive ‘signal’ for an effect on NRT adherence, we planned a second RCT to test whether BMN had positive effects on smoking cessation. However, due to the COVID-19 pandemic and NHS service provision changes, the funder agreed that BMN efficacy for smoking abstinence would be better investigated instead of simply using a measure of smoking behaviour as the trial primary outcome and increasing the SNAP2 sample size. This was dependent on a successful pilot phase of the trial in which (a) BMN demonstrated a sufficiently large ‘signal’ that impacts the adherence to NRT and (b) explanatory trial outcome data were collected before being discontinued in the more pragmatic ‘full’ trial. Below we detail how, in the SNAP2 pilot phase, the impact of BNM on adherence to NRT was assessed, and progression decided on.

Assessment of BMN’s potential effect on NRT adherence’

For the original SNAP2 RCT, we defined a clinically important effect as BMN increasing NRT use by 21%. Data from a previous study indicated that if the control group was offered NRT for 28 days, they were likely to use this on a mean of 7 days,34 so a 21% increase represented an extra 1–2 days of NRT use.

We used a one-sided CI approach35 to assess whether or not the pilot phase SNAP2 trial ‘signal’ for the impact of BMN on NRT adherence was consistent with assumed effects. Using a CI approach, we calculated that for the checkpoint analysis to produce an upper limit of a one-sided 80% or 90% CI that excludes the estimate of effect, assuming the treatment estimate from the checkpoint assessment was zero or less would require 34 or 54 in the analysis. A trial statistician estimated BMN efficacy using data from 49 participants followed up to the primary outcome point. The point estimate and the upper confidence limit (whether 80% or 90%) were greater than the pre-specified clinically relevant effect size of 1.21. The funder deemed this sufficient demonstration of potential efficacy for the pilot to progress to a full trial.

Full trial sample size estimate

To determine the sample size for the full trial, we assume a control group quit rate of 10%, consistent with similar UK studies,34 36 37 and seek to detect an absolute increase in the risk of abstinence of 6% (corresponding to a risk ratio (RR) of 1.6). We think this is a reasonable effect size to seek as all previous trials of NRT used by pregnant women tested the use of only one NRT product and these show a relatively weak, imprecise effect (RR 1.37, 95% CI 1.08 to 1.74),13 with smaller point estimates seen in the least biased studies (RR 1.21, 95% CI 0.95 to 1.55).13 However, in non-pregnant smokers where adherence to NRT is much stronger, depending on the type of NRT used, Cochrane review point estimates for NRT efficacy range between 1.49 (for gum) and 2.02 (inhaler),38 and dual NRT (patch+fast acting) is more effective than single product use, with an OR for abstinence with NRT of 1.25 (dual vs single product NRT use).12 We aim to recruit 1320 participants, providing a minimum of 90% power for a two-sided test of size 5%, assuming the control group’s quit rate is at least 10%.

Assessment of efficacy

The primary efficacy endpoint is reported abstinence in late pregnancy or around childbirth, validated by appropriate biochemical measures (binary—abstinent or non-abstinent). The primary analysis model will include all randomised participants as part of the groups to which they were allocated, with any missing primary endpoints imputed as being non-abstinent. We will compare abstinence between treatment groups using logistic regression, with fixed effects of the treatment group, gestation at baseline (the stratification factor) and other predictive baseline covariates (pre-specified in the SAP). In addition to the point and interval estimate of the OR for allocation, we will use the fitted model to obtain estimates of treatment effects on both the relative risk and risk difference scales. Binary secondary outcomes (eg, other abstinence outcomes) will be analysed in a broadly similar manner to the analysis of the primary outcome. Secondary outcomes including adherence to NRT, NRT adherence intensity, cotinine (smoking exposure), necessities/concerns and birth outcomes will be compared between arms using appropriate generalised linear models. All other outcomes will be summarised descriptively by a randomised arm. If the relevant data are sufficiently complete, we will perform exploratory analyses to decompose the total effect of allocation on the primary outcome into indirect effects (ie, those mediated by improved NRT adherence/usage) and direct effects (ie, those not mediated by improved NRT adherence/usage).

Procedures for missing, unused and spurious data

We will assume that participants with missing smoking status are smoking, in line with the Russell standard, meaning people who do not provide data are assumed to be smoking.39 Similarly, participants not providing a NRT usage response(s) will be assumed to be not using NRT in the corresponding period. We will investigate the sensitivity of results from the main analyses to departures from these strict missing not at random (MNAR) assumptions, by imputing these missing data under a range of missing at random (MAR) and MNAR scenarios.

Safety

The study tests a behavioural intervention aimed at optimising NRT use.26 As this is a standard NHS treatment, we do not anticipate any harm being caused, and there is no adverse event monitoring.

Economic analysis

The economic analysis will determine the cost-effectiveness using a lifetime time horizon. To estimate long-term benefits, costs and cost-effectiveness, and potential longer-term cost savings, we will use the Economics of Smoking in Pregnancy (ESIP) model, a bespoke, dynamic economic model designed specifically for valuing smoking cessation in pregnancy in economic terms.40 41 Smoking cessation rates from the trial and costs of intervention delivery and usual care will be used as ESIP inputs. As SNAP 2 has a short follow-up period, we will not use measures of participants’ quality of life (QoL) as one would not expect changes in response to the SNAP2 intervention within the short study timeframe, and any QoL changes resulting from smoking cessation would likely be variation in QoL reflecting physiological changes in pregnancy.