Electronic health record-facilitated symptom surveillance and collaborative care intervention in oncology (E2C2): a cluster-randomised, population-level, stepped-wedge, pragmatic trial
Mené aux Etats-Unis sur 50 207 patients atteints d'un cancer, cet essai évalue l'intérêt du dossier médical électronique pour la surveillance des symptômes et la coordination des soins en oncologie et hématologie
Background: Patient-reported outcome measure (PROM) surveillance and collaborative care improve cancer symptom control. However, human resource requirements constrain their implementation and reach. Electronic health record (EHR) facilitation reduces resource needs and might allow population-level scaling. We aimed to assess the effect of EHR facilitation of PROM-directed collaborative care on clinical and health services outcomes.
Methods: E2C2 was a cohort cluster-randomised, unblinded, stepped-wedge, pragmatic trial, in which we randomly assigned 15 clusters of medical oncology and haematology clinics in the USA sharing a common EHR, Epic, to five sequences to compare an intervention of remotely delivered electronic PROM (ePROM) symptom surveillance and EHR-facilitated collaborative care (ECC) management with a usual care (UC) control of ePROM surveillance alone. Sequences transitioned from the UC control to the ECC intervention state at 8-month intervals. All adult (aged ≥18 years) patients who received medical oncology or haematology care in a US multi-state health system were enrolled. All cancer stages, cancer types, and treatment phases were included, except for patients enrolled in hospice or with acute leukaemia. SPPADE symptoms (sleep interference, pain, impaired physical function, anxiety, depression, and energy deficit or fatigue) were assessed with 0–10-point numerical rating scales linked to clinical encounters. The prespecified co-primary outcomes were all post-baseline SPPADE scores, and clinically actionable scores (≥4/10), among participants who completed at least two ePROMs. The outcomes were assessed using multivariate regression of cluster-period mean SPPADE symptom scores against intervention exposure, baseline SPPADE scores, fixed-cluster, and secular time effects. The trial is registered at ClinicalTrials.gov (NCT03892967) and is now closed to recruitment.
Findings: From March 28, 2019, to Jan 31, 2023, 50 207 patients were enrolled and administered ePROMs in association with oncology or haematology visits. In the analytical cohort of 24 874 participants, 10 390 (42%) were assigned ePROMS in both the ECC and UC periods. Of the 19 084 [77%] in the ECC group, 11 138 (58%) were female, 7946 (42%) were male, and 18189 (95%) were White; and of the 16 180 (65%) in the UC group, 9621 (60%) were female, 6559 (40%) were male, and 15468 (96%) were White. 21 153 (85%) participants reported one or more clinically actionable symptoms (defined as SPPADE score ≥4/10). In multivariate analyses, mean population joint SPPADE symptom burden favoured ECC periods (p=0·0055) with adjusted mean differences of –0·12 (95% CI –0·19 to –0·05) for anxiety, –0·08 (–0·15 to –0·01) for depression, –0·06 (–0·16 to 0·03) for fatigue, –0·04 (–0·14 to 0·07) for pain, 0·03 (–0·07 to 0·14) for physical function, and –0·07 (–0·16 to 0·02) for sleep. ECC benefit was also noted following actionable scores (p<0·0001) with adjusted mean differences of –0·10 (–0·17 to –0·03) for anxiety, –0·09 (–0·16 to –0·02) for depression, –0·09 (–0·18 to –0·01) for fatigue, 0·04 (–0·05 to 0·12) for pain, 0·07 (–0·03 to 0·17) for physical function, and –0·02 (–0·10 to 0·07) for sleep.
Interpretation: Centralised EHR-facilitated, symptom surveillance and collaborative care management are more beneficial than symptom surveillance alone in reducing the population burden of SPPADE symptoms in oncology patients.
The Lancet Oncology , résumé, 2025