Manz ML nudges (long-term)

Long-term Effect of Machine Learning-Triggered Behavioral Nudges on Serious Illness Conversations and End-of-Life Outcomes Among Patients With Cancer

Patient / Population Intervention / Exposition Comparaison Critère de jugement

In 20,506 patients with cancer (41,021 encounters) at 9 medical oncology clinics in a large academic health system, machine-learning-triggered behavioural nudges to clinicians over 40 weeks increased serious illness conversations and reduced end-of-life systemic therapy compared with usual care, without changing hospice, inpatient-death or ICU outcomes.

N
20 506
Schéma
Stepped-wedge cluster RCT, prespecified 40-week analysis, 9 oncology clinics
Critère
Serious illness conversation rates for all and high-risk patient encounters; secondary end-of-life outcomes among decedents
Pertinence
2
RésultatSICs in high-risk encounters 13.5% vs 3.4% (unadjusted); all patients aOR 2.09 (95% CI 1.53–2.87; P<0.001). End-of-life systemic therapy 7.5% vs 10.4% (aOR 0.25, 95% CI 0.11–0.57; P=0.001); no effect on hospice enrolment or length of stay, inpatient death or end-of-life ICU use.
Manz CR, et al. JAMA Oncol. 2023;9(3):414-418. 10.1001/jamaoncol.2022.6303
Discussion et critique

Extends the 2020 stepped-wedge trial (Manz ML nudges (SIC)) to 40 weeks and to what matters at the end of life: the conversation gain persisted and systemic therapy near death fell — a patient-level end-of-life outcome, not just a process measure. Hospice, ICU and inpatient-death outcomes were unchanged, only 1,417 patients (6.9%) had died by the end of follow-up, and the single-system setting limits generalisability.