Implementation outcomes of a symptom management intervention in ambulatory oncology practices evaluated using a cluster randomized stepped-wedge trial design

Objective: To test a package of clinician- and system-level implementation strategies on the adoption and reach of an electronic health record (EHR)-integrated cancer symptom assessment and management program, called cPRO, within a large academic healthcare system.

Methods: This hybrid type 2 effectiveness-implementation study used a cluster randomized stepped-wedge trial design to test a package of strategies targeting system operations, clinician practices, and patient experience to support implementation of cPRO. Six clusters, comprised by 26 oncology clinic sites, were randomly allocated to one of six sequences which dictated the time at which each cluster underwent a 6-month implementation preparation period followed by a transition to the post-implementation phase in which 46 discrete implementation strategies were deployed. The primary implementation outcome was patient adoption of cPRO, measured by the proportion of patients completing cPRO assessments. Secondary outcomes included the reach of patient enrollment in the cPRO system and clinician adoption of referrals using an EHR “dot phrase” (snippets of text that can be quickly inserted into patient charts for referrals, orders, etc.) triggered by elevated cPRO scores. Data were analyzed using a cluster-period level analysis (generalized least squares linear regression with fixed cluster effects and adjustment for calendar time).

Results: The study included 34,643 unique outpatients receiving cancer treatment at 26 clinics between October 2020 and March 2024. The primary analysis showed no significant difference between the pre- and post-implementation periods on the mean difference in the proportion of patients who complete the assessments (25% vs. 40%). Secondary outcomes indicated that the implementation strategy package did not significantly improve the reach of cPRO enrollment among patients (RR = 1.00, CI: 0.78 to 1.27). Clinician adoption of referrals in response to elevated cPRO symptom scores showed a marginal positive, alebeit non-statistically significant association with the implementation strategy package (RR = 1.66, CI: 0.79 to 3.48), although this varied over time.

Conclusions: The implementation strategies tested did not significantly alter patient adoption rates of cPRO when comparing pre- and post-implementation periods, but might improve clinician adoption of the EHR dot phrase function. Future studies should explore strategies to enhance the integration of digital symptom management systems into routine cancer care to improve patient outcomes.

Trial registration: ClinicalTrials.gov NCT03988543; registered 8 May 2019 https://clinicaltrials.gov/study/NCT03988543?term=NCT03988543&rank=1.