Kesselheim 2020.

Study characteristics
Methods	Study design: cluster‐randomised controlled trial Study grouping: parallel group
Participants	Baseline characteristics Novel training (intervention group) Age (N (%)) 26‐30, 31‐35, 36‐40, 41‐50: 17 (29%), 38 (64%), 4 (7%), 0 Sex (N (% female)): 45 (76%) Sample size: 59 Years of experience (mean ± SD): NR Usual training (control group) Age (N (%)) 26‐30, 31‐35, 36‐40, 41‐50: 12 (29%), 27 (66%), 1 (2%), 1 (2%) Sex (N (% female)): 31 (76%) Sample size: 41 Years of experience (mean ± SD): NR Overall Age (N (%)) 26‐30, 31‐35, 36‐40, 41‐50: NR Sex (N (% female)): NR Sample size: 100 Years of experience (mean ± SD): NR Included criteria: NR Excluded criteria: NR Pretreatment: there were no significant differences between the UT and intervention arms with respect to age, gender, or additional professional degrees. At baseline, pretest data reveal that UT and intervention groups did not differ significantly in their scores on the PHOSAH, MBI, PPOS, or Empowerment at Work Scale (Table 2). Mean scores on the PHOSAH, the primary outcome measure, were 7.4 (SD 4.2) for fellows in the UT group and 8.2 (SD 3.3) for the intervention group (P = 0.35). However, baseline performance on the PHOSAH was somewhat lower than previously published, with mean score of 9 (SD 3.4) .15 Fellows in both groups had similar levels of satisfaction with their fellowship training in several domains of humanism and professionalism (Table 3). Compliance rate: NR Response rate: NR Type of healthcare worker: paediatric haematology‐oncology fellows
Interventions	Intervention characteristics Novel training (intervention group) Type of the intervention: Intervention type 1 ‐ to focus one’s attention on the experience of stress Description of the intervention: The intervention arm of this study futilises a novel, 4‐module, case‐based curriculum which aims to foster PHO fellows’ reflection on the feelings, challenges, and conflicts arising in the care of children and families affected by cancer or blood disorders. The number of sessions: NR Duration of each session on average: NR Duration of the entire intervention: NR Duration of the entire intervention short vs long: short Intervention deliverer: faculty facilitator Intervention form: group Usual training (control group) Type of the intervention: NR Description of the intervention: NR The number of sessions: NR Duration of each session on average: NR Duration of the entire intervention: NR Duration of the entire intervention short vs long: NR Intervention deliverer: NR Intervention form: NR
Outcomes	Maslach Burnout Inventory ‐ Emotional Exhaustion Outcome type: ContinuousOutcome Maslach Burnout Inventory ‐ Depersonalisation Outcome type: ContinuousOutcome Maslach Burnout Inventory ‐ Personal accomplishment (lack of) Outcome type: ContinuousOutcome
Identification	Sponsorship source: NR Country: USA Setting: Hospital Comments: NR Authors name: Jennifer Kesselheim Institution: Department of Pediatric Oncology, Dana‐Farber/Boston Children’s Cancer and Blood Disorders Center, Boston, Massachusettes Email: Jennifer_kesselheim@dfci.harvard.edu Address: Jennifer Kesselheim, Dana‐Farber/Boston Children’s Cancer and Blood Disorders Center, Boston, Massachusettes 450 Brookline Avenue, D1107, Boston, MA02215 Time period: 2016‐2017
Notes	MBI‐EE included in analysis 1.1
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: Program directors committed to study participation sought local approval from their Institutional Review Board after which their program was randomized to either the intervention or UT arm of the study.  Sequence generation process insufficiently described.
Allocation concealment (selection bias)	Unclear risk	Unable to judge whether participants and/or investigators could possible foresee assignment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	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	Low risk	89 of the 100 randomised participants included in the analysis (89%). Reasons not described nor whether missing was at random. However loss to follow‐up is below our pre‐defined cut‐off point.
Selective reporting (reporting bias)	Unclear risk	No trial registration or no study protocol reported, nor did we fine one online
Other bias	Unclear risk	Unit of analysis error (i.e. when a study ignored the clustering of the data in their analysis). Compliance not reported.