Human Voice vs AI-Cloned Voice Audio Added to Written Consent Before Ureteroscopy or Retrograde Intrarenal Surgery (AUDIO-CONSENT) (AUDIO-CONSENT)

Standard care: the unified written consent form plus the physician's usual verbal explanation. No audio recording.

The institution's unified written informed consent form for ureteroscopy and retrograde intrarenal surgery, read and signed by the patient, together with the physician's usual verbal explanation. Delivered identically in all three arms.

Standard care plus an audio recording of the entire consent text read by a urologist (the same recording at all centers), presented after the standard consent process.

The institution's unified written informed consent form for ureteroscopy and retrograde intrarenal surgery, read and signed by the patient, together with the physician's usual verbal explanation. Delivered identically in all three arms.

An approximately 12-minute audio recording of the complete consent text, read word for word by one urologist (the same recording at all centers; not the patient's own surgeon). Presented after the standard written consent and physician explanation, on a tablet with headphones in the patient's room without companions; the patient may replay it and the number of replays is recorded. Loudness-normalized to -24 LUFS. Introduced to the patient as "a recording of the entire consent form read by a urologist".

Standard care plus an audio recording of the same text generated with AI from a voice clone of the same urologist, presented after the standard consent process; patients are told that the voice was generated with AI.

The institution's unified written informed consent form for ureteroscopy and retrograde intrarenal surgery, read and signed by the patient, together with the physician's usual verbal explanation. Delivered identically in all three arms.

The same consent text, generated once with Resemble.ai from a voice clone of the same urologist (reference sample: a separate recording of at least 5 minutes of a different text), frozen before enrollment with no live AI calls. Duration approximately 12 minutes; loudness-matched (-24 LUFS), codec and delivery identical to the human-voice recording, presented after the standard written consent and physician explanation, on a tablet with headphones; replay allowed and counted. Patients are informed that the voice was generated with AI from a urologist's voice.

Score on a 10-item multiple-choice test derived from the consent text (range 0-10; higher scores indicate better knowledge of the procedure, its risks, alternatives and patient responsibilities). Each item is mapped to a specific paragraph of the consent form. Analyzed as a continuous variable.

State-Trait Anxiety Inventory, State form (Turkish version, Öner & Le Compte), range 20-80; higher scores indicate greater anxiety. Measured immediately after the consent process and analyzed adjusted for the pre-consent STAI-S score (measured immediately before the consent process) and STAI-T.

Three study-specific 5-point items (perceived clarity of the information, adequacy of the information for decision-making, overall satisfaction with the consent process); each scored 1-5, higher scores indicate better outcomes. Reported item by item; total score (3-15) secondary. Not validated.

Two study-specific 5-point items: naturalness of the voice (1 = not at all natural to 5 = completely natural) and usefulness of receiving the information as audio (1 = not at all to 5 = very much). Higher scores indicate greater perceived naturalness/usefulness. Asked in both audio arms.

Single study-specific 5-point item: "Did it bother you that the voice was generated by artificial intelligence?" (1 = not at all to 5 = very much; higher scores indicate lower acceptability). Administered on a separate sheet after the naturalness items have been collected.

Number of times the patient chose to replay the audio recording.

Single item with four categories: read all of it / read most of it / skimmed it / did not read it.

Number and percentage of patients who decline or postpone the planned surgery after the consent process, and of patients who stop the audio recording before completion. Reported descriptively by arm; not interpreted as an adverse outcome.