Abstract Study question What are fertility staff and patients’ evaluation of staff performance during sharing bad news (SBN) conversations? Summary answer In positive SBN conversations, staff and patient ratings of staff performance are similarly good, but in negative SBN conversations patients’ ratings are lower than staff’s. What is known already Bad news, defined as news that negatively alters patients’ view of their future, happens at all stages of fertility care. Fertility staff perceive SBN among the biggest challenges of their job. 47% of patients express staff could improve bad news delivery. SPIKES is an evidence-based protocol that has shown promise in guiding SBN in fertility care via breaking the task in six steps: Setting the scene, evaluating patients’ Perception, getting patients Invitation, sharing Knowledge, addressing Emotions, and Summarizing and planning future care. Knowledge about how staff perform during SBN conversations and which factors are associated with performance is lacking. Study design, size, duration Two cross-sectional mixed-methods online surveys were distributed to staff (via ESHRE) and patients (via charities, social media). Inclusion criteria for staff were working at a fertility clinic and SBN at least once a month and for patients being 18 years of age and having had a SBN conversation with staff in past two months. 334 staff and 345 patients clicked survey links, of which 217 staff and 222 patients completed it (65% and 64% completion). Participants/materials, setting, methods Dependent variable was staff SBN performance across SPIKES steps, measured with CARE (empathy) and the Breaking Bad News Assessment Schedule (other steps). Independent variables were actor (staff, patient) and overall evaluation of the SBN conversation (positive, negative). Factors assessed related to staff and patients’ background (e.g., gender, age), fertility care context (e.g., patients: parental status, treatment stage, staff: role, burnout, time pressure, SBN training), and SBN conversation (e.g., patients: person vs remotely, alone vs accompanied). Main results and the role of chance Crosstabs of actor and evaluation showed a higher proportion of staff (85%) than patients (67%) evaluated the SBN conversation as positive (Χ2(1) = 38.863,P<.001). MANOVA investigating differences in staff’s SBN performance according to actor and overall evaluation revealed significant effects of actor (Hotelling’s Trace T = .236,F(388,6) = 15.279,p<.001,η2p = .191), overall evaluation (T = .344,F(388,6) = 22.233,p<.001,η2p = .256), and their interaction (T = .127,F(388,6) = 8.243,p<.001,η2p = .113). Follow-up ANOVAs showed the same pattern for all SPIKES steps: in positive SBN staff and patient ratings of staff performance were similar (∼4 from 1 = poor/not at all to 5 = excellent/completely), but in negative SBN patients’ ratings were lower than staff’s (∼2.5 vs 4, η2p ranged .160[Setting] to .289[Perception]). Among staff, physicians and nurses tended to rate their SBN performance better than embryologists across all SPIKES steps except Knowledge and Emotions (β ranged .166[Invitation] to .312[Perception]). Reporting ≥1 burnout symptom was associated with worse performance in Emotions (β = -.168,P = .049), time pressure was associated with worse performance in Perception (β = -.185,P = .019), and having attended SBN training with better performance in Setting (β = .295,P<.001). Patients who were alone during the SBN conversation rated staff SBN performance worse than patients who were accompanied in all SPIKES steps (β from -.154[Summary] to -.403[Setting]). Patients with children rated staff performance in Setting (β = .156,P = .015), Emotions (β = .158,p = .031), and Summary (β = .142,p = .048) better than childless patients. Limitations, reasons for caution Online data collection attracted mostly UK patients while staff were from all over the world. SBN performance was reported and not based on behavioral observation, though based on sound questionnaires that evaluate different aspects of performance. Wider implications of the findings Staff and patients agree about what good SBN conversations are, but patients are more critical of bad SBN conversations. Staff SBN performance seems more a function of the context in which news are shared than of actors’ background. SBN training can support staff recognizing bad SBN conversations and improving performance. Trial registration number not applicable