Table 2 MOHMQuit data collection and analysis

 Biochemically verified 7-day point prevalence abstinence at the end of pregnancy among pregnant smokers and recent quitters, confirmed by salivary cotinine testing.

(Aim 1)

Structured telephone interview^a and salivary cotinine testing

Postpartum women

Within 2 weeks after giving birth

• The primary analysis will compare the odds of quitting smoking (cotinine-confirmed 7-day point prevalence abstinence) after the adoption of the intervention relative to the baseline rate.

• To achieve this, logistic regression will be used within a generalised linear mixed model framework.

• In addition to the fixed effect for intervention, a fixed effect for time will be included to adjust for any secular trends in the outcome. Random effects for site and time within site will be included to account for clustering by site and repeated measures over time.

• Women who do not complete the biochemical verification will be classified as continuing smokers, as recommended in analyses of smoking cessation trials (e.g., see [56]).

 Self-reported 7-day point prevalence abstinence at the end of pregnancy among pregnant smokers and recent quitters.

(Aim 1)

Structured telephone interview^a

Postpartum women

Within 2 weeks after giving birth

Self-reported 7-day point prevalence abstinence will be analysed in the same way as the primary outcome.

 Documented provision of smoking cessation support (SCS) at booking visit and any subsequent visits. Offer of:

  • referral to a smoking cessation service including NSW Quitline

  • behavioural support

  • nicotine replacement therapy

(Aim 2)

Deidentified data extraction of specified variables for all births to eligible women at intervention sites

eMaternity^b

Throughout the 32-month collection period

• eMaternity SCS scores will be calculated based on the documented provision of SCS in eMaternity for each woman at each visit.

• Scores will be modelled using a linear mixed model.

Fixed effects will be included for time, intervention, and the interaction of these.

•Random effects for site and participant will be included to account for clustering by site, and repeated measures on participants.

•The interaction term will directly compare the rate of change in SCS scores over antenatal visits between pre and post-intervention.

 Self-reported:

  • Provision of Advice to quit smoking

  • Provision of Assistance to quit smoking

  • Provision of referral to a quit smoking service (e.g. NSW Quitline or local service).

(Aim 3)

Study specific questionnaire

Clinicians

Before training and 6 months post-training

• The outcome measure will be a composite score of self-reported provision of SCS, calculated from the questionnaires completed by clinicians prior to training and at 6 months post-training.

• Post-intervention clinician SCS scores will be modelled using a linear mixed model. Fixed effects will be included in the model for the pre-intervention SCS score and time. A random effect will be included for site to account for randomisation by cluster.

 Women’s reported receipt of:

  • Advice on quitting

  • Referral to a quitting program

  • Written self-help materials

(Aim 5)

Structured telephone interview^a

Postpartum women

Within 2 weeks after giving birth

• Receipt of SCS reported by women will be modelled using logistic regression in a generalised linear mixed model framework. Fixed effects for time and intervention will be included in the model and a random effect for site will be used to account for randomisation by cluster.

 • Capability

 • Personal priority

 • Tracking systems

 • Work environment

 • Effectiveness of Quitline referral

 • Client relationship

 • Lifestyle choice

 • Personal discomfort

(Aim 4)

Study-specific questionnaire

Clinicians

Before training, immediately post-training and 6 months post-training

• Factor analysis of the initial questionnaire used to identify barriers/enablers identified several factors that predicted provision of SCS (e.g. capability, personal priority). These are included in the current questionnaire with clinicians.

• Additional items to measure constructs/barriers identified in our qualitative research, but not in the original questionnaire have been added to enable assessment of whether these barriers have been addressed.

• Initially, confirmatory factor analysis will be undertaken to confirm the factor structure.

• A composite score across all items within each of the factors will be the outcome measure.

• Any items not loading onto a factor will be modelled separately.

• To assess changes in factor scores the total score on each factor will be modelled using a linear mixed model where one parameter will estimate the score immediately post-intervention and another parameter will estimate the score 6 months after intervention. The model will adjust for pre-intervention total score and time. A random effect will be included for site to account for randomisation by cluster.

• Note - work environment will only be assessed immediately pre-training and 6 months post-training, as this is unlikely to have changed during the training.

 Leadership

(Aim 8)

Study specific questionnaire includes the Implementation Leadership Scale [57]

All leaders involved in training

Three months after leader training

• The Implementation Leadership Scale (supervisor version) is comprised of twelve items that make four subscales: the proactive subscale, the knowledgeable subscale, the supportive subscale, and the perseverant subscale.

• For each subscale, a mean score for each set of items that load onto the relevant subscale will be computed. Additionally, a mean of the scale scores will be computed to generate the total score for the Implementation Leadership Scale [57]. Scores will be computed for each clinician and an aggregated clinical level score will be calculated.

Study specific questionnaire includes the Leadership Engagement Scale [58]

Clinicians

6 months post-training

The Leadership Engagement Scale is comprised of four items. A mean score for these items will be computed for each clinician. An aggregated clinic level score will also be calculated.

Interviews

Selected leaders

6 months post-training

Analysis of qualitative data is described below under process evaluation.

 Implementation climate

(Aim 8)

Study-specific questionnaire includes the Implementation Climate Scale [58]

Clinicians

6 months post-training

The Implementation Climate Scale is comprised of four items. A mean score for these items will be computed for each clinician. An aggregated clinic level score will also be calculated.

 Acceptability

(Aim 7)

Interviews

Selected leaders

6 months post-training

Qualitative data will be analysed thematically [59] to understand perceptions of acceptability, variation in this, and factors impacting it.

 Adoption

(Aim 7)

Study specific questionnaire

Clinicians

6 months post-training

At each site the proportion of respondents reporting utilisation of each of the key resources will be calculated.

Study specific questionnaire

Leaders

3 months post-training

For each site, the reported completion of each of the leader components will be calculated.

Interviews

Selected leaders

6 months post-training

Qualitative data will be analysed thematically [59] to understand the extent of implementation undertaken and factors impacting this.

 Appropriateness

(Aim 7)

Interviews

Selected leaders

6 months post-training

Qualitative data will be analysed thematically [59] to understand perceptions of the appropriateness of the intervention, and any adaptations made.

 Feasibility

(Aim 7)

Interviews

Selected leaders

6 months post-training

Qualitative data will be analysed thematically [59] to understand perceptions of the feasibility of implementing MOHMQuit and the extent to which it has been integrated into usual care.

 Fidelity (of delivery of the MOHMQuit intervention to the services)

(Aim 7)

Training logs

Researcher notes

Research team

Throughout

Data will be analysed to determine the proportion of eligible clinicians and leaders participating in training; the extent to which the training and other strategies were delivered as intended.

 Context

Multiple

Internet, Partner Investigators and Leaders

Throughout

Information on service size, capacity, structure will be collated. Information from Partner Investigators and leaders regarding factors impacting delivery will be documented throughout the study period.

 Cost-effectiveness—the incremental cost per additional quitter from a service-provider perspective

(Aim 6)

Within-trial data collection

Data will be sourced from study and site records as well as the existing literature.

Data collection throughout the study, with analysis at the end.

Economic evaluation conducted from the perspective of the Australian health care system with 8-month time horizon for cost-effectiveness and lifetime time horizon for cost utility. Probabilistic sensitivity analysis and non-parametric bootstrapping will be used to assess robustness of the model.

 Cost-utility—lifetime analysis of the incremental cost per Quality adjusted life year (QALY) gained

(Aim 6)

Modelled economic evaluation

Data will be sourced from study and site records as well as the existing literature.

Data collection throughout the study, with analysis at the end.