Background: Secure text messaging is a primary mode of clinician-to-clinician communication in large acute care hospitals. However, the accessibility and ease of secure messaging, and a lack of best practice recommendations, raise the risk of overuse. Multitasking demands from secure messaging challenge clinicians as they engage in focus-intensive patient care tasks. To address this, we created a multi-pronged educational intervention to enhance messaging practices, aiming to decrease overall message interruptions. We then assessed the impact of the intervention on both the volume of messages received and the clinician experience with secure messaging.
Methods: To develop the intervention, a workgroup of hospitalist, nurse (RN), and advance practice provider (APP) leaders was formed. The workgroup created and refined two educational secure messaging best practice documents. The first was a bulleted list of tips on messaging etiquette aimed to reduce unnecessary messages. The second was an infographic adopted from Mendel et al illustrating preferred communication methods for specific clinical scenarios (see Figure 1). In our quaternary care academic medical center, two independently staffed 30-bed medical units were selected to serve as intervention and control sites. Over a 10-week trial period on the intervention unit, the materials were 1) reviewed weekly at unit meetings; 2) included as attachments in weekly unit emails; 3) displayed on unit computer workstations; and 4) posted in the unit APP workroom. The number of messages received by APPs was tracked over this 10-week period (APP messaging data was selected as a focus based on initial work showing higher messaging burdens among APPs compared to RNs and hospitalists). The control unit received no education. Results were compared to messaging data collected for both units during a pre-intervention 10-week period. Additionally, pre- and post- intervention surveys were sent to each member of the care teams to gauge the effect of the educational intervention on attitudes towards the interruptive nature of secure messaging.
Results: An interrupted time series analysis revealed a non-significant decrease in the message volume slope post-intervention on the intervention unit, with a difference of -60.7 messages per week (95% CI: -190.0, 68.7, p=0.335). Similarly, the difference-in-difference analysis showed a non-significant decrease of 360.4 messages between control and intervention (95% CI: -895.1, 174.3, p=0.180. See Figure 2). 56 pre-intervention survey responses were completed, and 49 post-intervention surveys were completed by the unit teams. Interestingly, APPs on the intervention floor showed an unexpected increase in agreement that “Epic chats can become distracting from patient care,” contrary to the hypothesized outcome.
Conclusions: Our educational interventions fell short of achieving the desired improvements in message volumes and attitudes towards secure messaging. Effecting change in the utilization of clinical secure messaging may necessitate more than passive education. Surprising post-intervention survey responses from APPs on the intervention unit indicated heightened frustration with messaging interruptions, suggesting that our interventions may have served only to highlight a simmering issue. To foster actual change, future efforts could explore using clinician quality champions already participating in the secure messaging clinical environment to assist in real-time education within message chains.
