Questions older people and caregivers ask about kidney therapies: A computer-assisted analysis

Older adults with advanced chronic kidney disease (CKD) face complex and often overwhelming decisions regarding kidney therapy (KT), yet little is known about the specific information they and their caregivers seek during these discussions. Understanding informational needs is critical for improving shared decision-making.

To identify and compare questions asked by patients and caregivers during KT decision-making visits.

A dataset of 36 transcribed kidney therapy (KT) decision-making focused outpatient palliative care encounters-conducted by a single clinician in patient homes and/or a single clinic-was analyzed using computational linguistic techniques supplemented by manual verification. Participants included 25 patients aged ≥ 75 years with an estimated glomerular filtration rate ≤ 25 ml/min/1.73 m², identified by nephrologists as needing KT decision-making support. Caregivers (n = 15) were included when present. All patients received a decision aid and question prompt list. Visits were led by a nephrologist board-certified in palliative care.

Across 36 visits, 640 questions were recorded, averaging 17.78 per visit. The presence of a caregiver did not significantly affect the number of questions patients asked (10.42 vs. 10.00). Patients were more likely to ask conversational cooperation questions (37 % vs. caregivers' 19 %), while caregivers focused more on dialysis (38 % vs. patients' 25 %). Both groups asked a similar amount of questions about chronic kidney disease (patients 11 %, caregivers 10 %) and conservative kidney management (patients 4 %, caregivers 5 %). End-of-life inquiries were minimal, with patients asking 4 % of questions and caregivers none.

Patients and caregivers have different information needs during KT decision-making, highlighting the importance of addressing educational, decisional, and EOL planning gaps.

Targeted educational interventions that address patients' and caregivers' unique preferences and information gaps are essential, including question prompt lists and decision aids with coaching, which AI-driven approaches could further optimize.