In a multicentre randomised controlled trial (DIRECT), we evaluate whether an intervention of providing direct letters from healthcare professionals to at-risk relatives (ARRs) affects the proportion of ARRs contacting a cancer genetics clinic, compared with patient-mediated disclosure alone (control). With the aim to explore how the patients included in the trial perceived and performed risk communication with their ARRs we analysed 17 semi-structured interviews with reflexive thematic analysis. All patients described that they disclosed risk information to all close relatives themselves. No integrity-related issues were reported by patients offered the intervention, and all of them accepted direct letters to all their ARRs. Patients’ approaches to informing distant relatives were unpredictable and varied from contacting all distant ARRs, sharing the burden with the family, utilising the offer of sending direct letters, vaguely relying on others to inform, or postponing disclosure. Most patients limited their responsibility to the disclosure, although others wanted relatives to get genetic counselling or felt a need to provide additional information to the ARRs before ending their mission. We also identified confusion about the implication of test results, who needed risk information, and who was responsible for informing ARRs. These misunderstandings possibly also affected risk disclosure. This study revealed that despite accepting the direct letters to be sent to all relatives, the patients also contributed to risk disclosure in other ways. It was only in some situations to distant relatives that the healthcare-assisted letter was the only means of communication to the ARRs.