Becker, Robert A.; Bonnet, Kemberlee; Shah, Mauli V.; Dang, Elizabeth; Ancker, Jessica S.; Malone, Daniel C.; Trinkley, Katy E.; Gomez-Lumbreras, Ainhoa; Del Fiol, Guilherme; Kawamoto, Kensaku; Tawfik, Abdelrahman G.; Cornia, Ryan; Jones, Aubrey E.; Mitchell, James; Reese, Thomas J. (2026).听.听BMJ Open, 16(3).听
This study explores what helps or hinders the use of a digital tool designed to support听shared decision-making鈥攁 process where patients and clinicians work together to choose treatments鈥攕pecifically for managing听drug interactions involving anticoagulants听(blood-thinning medications that can increase bleeding risk when combined with certain other drugs). Researchers conducted interviews with 36 participants, including both clinicians and patients, across three U.S. medical centers before the tool was introduced into routine care.
Participants identified several challenges, such as limited time during medical visits, difficulty fitting the tool into existing clinical workflows, restrictions related to healthcare roles, and differences in patients鈥� comfort with digital technology (digital literacy). At the same time, they noted helpful features, including clear visual displays of bleeding risk, access to supporting medical evidence, and a user-friendly design. Some factors could act as either barriers or benefits depending on the situation鈥攆or example, the tool might save time in some cases but add complexity in others.
Overall, the findings emphasize that for such tools to be successfully adopted, they must fit smoothly into clinical routines, align with the responsibilities of healthcare providers, and be easy for both clinicians and patients to use. Addressing these issues early in the design and rollout process could improve adoption and effectiveness in real-world healthcare settings.
Figure 1
DDInteract user interface layouts for visualising anticoagulant drug interaction risks. (A) Design option 1. The interface guides users through three steps: (1) patient-specific clinical data are imported from the electronic health record to identify relevant risk factors for gastrointestinal (GI) bleeding; (2) clinicians select among pain management and antidepressant options, with associated costs and estimated GI bleed risks displayed and (3) estimated GI bleed risk per 100鈥塸atients is visualised using an icon array. Red icons represent expected bleeding events, peach indicates bleeds prevented relative to ibuprofen, and grey indicates patients not expected to experience a bleed. Additional buttons provide definitions of key terms and access to the underlying evidence model. Feedback and summary print options are available in the top right. (B) Design option 2. This layout presents the same clinical information and interaction elements as option 1 but combines patient profile data (step 1) and treatment selection (step 2) into a single panel for a more streamlined experience. The estimated risk display and additional information resources remain consistent with option 1.