Brazier 2022.

Study characteristics
Methods	Study design: randomised controlled trial Study grouping: parallel group
Participants	Baseline characteristics Dear Doctor Age (mean ± SD): NR Sex (N (% female)): NR Sample size: 74 Years of experience (mean ± SD): NR Control (no intervention) Age (mean ± SD): NR Sex (N (% female)): NR Sample size: 79 Years of experience (mean ± SD): NR Overall Age (mean ± SD): NR Sex (N (% female)): NR Sample size: 153 Years of experience (mean ± SD): NR Included criteria: trainees were invited to participate if they were registered as being in Core Training Year 2 (CT2), or Speciality Training Years 3 or 4 (ST3 or ST4) of the UK training programme at the time of recruitment. Excluded criteria: no exclusion criteria were applied. Pretreatment: randomisation was well‐balanced in the final sample: participants in the two trial groups did not differ in terms of profile or covariates. Compliance rate: 1 of the 139 opted out (99%) Response rate: 18% Type of healthcare worker: exclusively trainee anaesthetists
Interventions	Intervention characteristics Dear Doctor Type of the intervention: Intervention type 4 ‐ Combination of two or more Description of the intervention: The intervention consisted of 22 text messages, including an introduction message with details of how to opt out of receiving the messages, sent fortnightly for approximately 10 months. The programme was named‘Dear Doctor’ and the messages appeared to come from this name on recipients’ mobile phones. The intervention content was informed by factors associated with burnout in the existing literature (summarised in the online Supporting Information Appendix S1, Table S1) and by factors identified by UK trainee anaesthetists in a preceding interview study (online Supporting Information Appendix S1, Table S2). As contributors to burnout may vary across medical specialities, the interview study ensured that the intervention content was tailored to the target population. The intervention combined this tailored approach with evidenced practices from the behavioural science and well‐being literature. Messages drew on 11 evidence‐based themes, with some messages drawing on more than one theme. The 11 themes could be grouped into six higher level categories: gratitude; self‐efficacy; connection to purpose; social support; support resources (including mindfulness and self‐compassion); and planning prompts. The number of sessions: 22 text messages Duration of each session on average: NA Duration of the entire intervention: 10 months Duration of the entire intervention short vs long: Long Intervention deliverer: NA Intervention form: Individual text messages Control (no intervention) Type of the intervention: NA Description of the intervention: NA The number of sessions: NA Duration of each session on average: NA Duration of the entire intervention: NA Duration of the entire intervention short vs long: NA Intervention deliverer: NA Intervention form: NA
Outcomes	Copenhagen Burnout Inventory ‐ work‐related subscale Outcome type: ContinuousOutcome
Identification	Sponsorship source: The study was funded by the National Institute for Health Research Imperial Patient Safety Translational Research Centre and registered (ISRCTN11418903) Country: UK Setting: Royal College of Anaesthetists Comments: NR Authors name: A Brazier Institution: National Institute for Health Research Imperial Patient Safety Translational Research Centre,  Faculty of Medicine, Imperial College London, London, UK Email: a.brazier19@imperial.ac.uk Address: NR Time period: 2019‐2020
Notes	CBI included in analysis 4.1
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were randomly assigned to intervention or control conditions with a 1:1 allocation ratio using a random number generator (Stata â, StataCorp, College Station, TX, USA). Randomisation was stratified by training year and across five broad training regions. "
Allocation concealment (selection bias)	Unclear risk	Insufficient information to understand whether intervention allocations could have been foreseen in advance of, during, enrolment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "Participants were not blinded to their condition assignment: intervention group participants received the intervention; control participants did not." Participants were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Participants were not blinded whereas outcomes are self‐reported.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "Despite the high attrition, there were no significant differences in participant characteristics (v 2 s < 3.80, ps > 0.15) or baseline outcomes measures (ts < 0.81, ps > 0.42) between those who did and did not complete the final survey (see also online Supporting Information Table S4)." High attrition, reasons other than did not complete survey not given. Lost to follow‐up appears to be at random.
Selective reporting (reporting bias)	Unclear risk	Quote: "The following exploratory findings should be regarded as indicative as they were not pre‐specified in the trial registry or protocol (the relevant outcomes were added after the start of the trial)." Trial registration number not provided, Difficult to assess whether there is selective outcome reporting.
Other bias	Unclear risk	Quote: "The RCoA identified and invited 1549 trainees to participate via email. Of the 647 (response rate 42%) who completed the baseline survey, 274 trainees (18% of the original cohort) consented to participate and remained in the trial throughout the trial period." Quote: "Female participants (v 2 = 8.07, p = 0.018), those with lower burnout (t = ‐2.42, p = 0.016), higher ‘meaningful’ score (t = 3.29, P = 0.001) and ‘valued’ score (t = 2.97, p = 0.003) were more likely to sign up for the trial after completing the baseline survey." Low selective response rate.