Table 1 Pragmatic and explanatory approaches in the ACCESSATION study

Participants

Eligibility criteria

· Adult smokers (aged 18–75 years)· Prospective subjects were motivated to make a quit attempt within 14 days· Subjects were recruited from community-based sites across Canada· Comorbid patients included· Past compliance not considered· No risk stratification

· Smokers excluded if they have existing SCT; female subjects of childbearing potential could be included if not pregnant or nursing, and practising effective contraception (not eligible for within-label pharmacotherapy due to safety reasons)

· Smokers excluded if they had a quit attempt in the past 30 days, had unstable comorbid conditions, possessed SCTs, lived with another study participant or were currently participating in another study· Participants smoked ≥10 cigarettes per day· Eligibility had to be determined by a central service

· Mostly pragmatic

Intervention and expertise

Experimental intervention – flexibility

· Physicians and patients controlled the type of smoking cessation interventions according to usual practice Any on-label combination of non- and pharmacological intervention was permitted· Participants could switch medications· Type or number of co-interventions used was not limited· Drug reimbursement card allowed any pharmacist to fill the prescription, rather than the doctor’s office· Side effects were to be managed at the discretion of the provider· Fewer contact minutes than efficacy trials with allowance for as-needed visits to assist the smoker

· Following randomization, a participant could not be re-randomized to a more favourable arm

· Provision of the card by the practitioner is not how insurance coverage cards are provided to patients· x`More contact than usual practice with scheduled rather than as-needed visits

· Mostly pragmatic

Experimental intervention – practitioner expertise

· All physicians were licensed to practice in Canada, and included GPs/FPs (86%) and specialists in cardiovascular medicine, surgery, cardiology, pulmonology and psychiatry· Most investigators (80%) had no prior clinical research experience· Primarily treatment- seeking patients from physician practice (93%)· None of the authors were associated with sites of recruitment

· Some investigators (20%) had prior clinical research experience

· Mostly pragmatic

Comparison intervention – flexibility

· Participants received counselling about smoking cessation options and chose their SCT prior to randomization. This reflects usual practice

· Participants were given a drug reimbursement card, which included $5 to partially offset pharmacist dispensing fees, to enable tracking of SCT use if obtained from a pharmacy

· Mostly pragmatic

Comparison intervention – practitioner expertise

· Randomization was at the patient level to mimic the real world, so that within the same practice some smokers received reimbursement while others did not. Practitioner expertise was therefore the same for both arms at each site, reflecting the real-world variation in expertise in smoking cessation· Given the central randomization process, re-randomization to a more favourable arm (i.e., SCT coverage) was not possible

· Some training of smoking cessation occurred at the investigator meeting to ensure all practitioners had knowledge of evidence-based interventions for smoking cessation

· Mostly pragmatic

Follow-up and outcomes

Follow-up intensity

· Participants did not attend the clinic on a weekly basis· At clinic visits/telephone contact, outcomes were measured, and AEs and method(s) used to quit (if any) were recorded· No exhaled CO levels were measured to reflect real-world practice

· Randomization visit involved more procedures than would usually be included at the first visit to discuss quitting· Although participants attended fewer contact visits than a traditional explanatory trial, they received more frequent contact than usual in clinical practice to balance the needs of data collection with the clinical reality of access to care· Strong focus on measuring outcomes and AEs from SCT use and/or quitting· A suicide severity scale was included because of recent reports of possible correlation with the use of SCTs

· Favours explanatory

Primary trial outcome

· Primary endpoint: comparison of the self-reported seven-day point prevalence of abstinence between the full and no reimbursement groups at Week 26 as opposed to 4 weeks continuous abstinence in efficacy trials in smoking cessation from the quit date. This endpoint is used in observational studies in telephone quitline and NRT distribution studies· Outcome assessed locally and no central adjudication of outcome

· Scheduled time for assessment of outcome at 26 weeks

· Balances both explanatory and pragmatic elements

Secondary endpoint

· A key secondary endpoint was self-reported 7-day point prevalence of abstinence at Week 26, confirmed by urine cotinine analysis. Urine collection is common in practice and the least intrusive method to confirm self-reports· Exhaled CO was not used, since this is less real-world practice

· Urine cotinine confirmation was used as a secondary endpoint to independently confirm self-reported abstinence. However, results were not shared with the practice

· Balances both explanatory and pragmatic elements

Compliance/adherence

Participant compliance with intervention

· Once randomized, participants could choose to use (or not) any currently available SCT· SCT use patterns were passively measured by drug reimbursement card use across both groups· Data regarding adherence were collected but not shared with prescribers

· All or a portion of the participants were compensated for costs incurred to visit the clinic. They could have been compensated between $25 and $50 for each on-site visit (four on-site visits at V1, V4, V7 and V10). There was variation between some sites with the total allocation between $100 and $200 to account for local factors

· Mostly pragmatic

Practitioner adherence to study protocol

· There was no measure of how practitioners provided SCT per se or developed their source documentation for the study· There were some deviations from protocol, including prescribing off-label, no smoking cessation counselling before unblinding, and enrolling participants who were not motivated to quit

· Sites and physicians were visited by the site monitor twice during the study but kept to a minimum· Randomization was centrally controlled to prevent gaming. However, participants could use any available SCT· Standard study auditing processes discovered one site was not compliant with GCP (inadequate source documentation)· Study monitors would bring off-label use of medication to the attention of the investigator

· Balances both pragmatic versus explanatory trials

Analysis

Analysis of primary outcome

· The main analysis will be the ITT population post-randomization regardless of use of the drug reimbursement card or making a quit attempt, with missing data being counted as being a smoker

· Mostly pragmatic