Automated real-time text messaging as a means for rapidly identifying acute stroke patients for clinical trials

Abstract

Background: Recruiting stroke patients into acute treatment trials is challenging because of the urgency of clinical diagnosis, treatment, and trial inclusion. Automated alerts that identify emergency patients promptly may improve trial performance. The main purposes of this project were to develop an automated real-time text messaging system to immediately inform physicians of patients with suspected stroke and to test its feasibility in the emergency setting.

Methods: An electronic standardized stroke algorithm (SSA) was implemented in the clinical information system (CIS) and linked to a remote data capture system. Within 10 minutes following the documentation and storage of basic information to CIS, a text message was triggered for patients with suspected stroke and sent to a dedicated trial physician. Each text message provided anonymized information on the exact department and unit, date and time of admission, age, sex, and National Institute of Health Stroke Scale (NIHSS) of the patient. All necessary information needed to generate a text message was already available - routine processes in the emergency department were not affected by the automated real-time text messaging system. The system was tested for three 4-week periods. Feasibility was analyzed based on the number of patients correctly identified by the SSA and the door-to-message time.

Results: In total, 513 text messages were generated for patients with suspected stroke (median age 74 years (19-106); 50.3% female; median NIHSS 4 (0-41)), representing 96.6% of all cases. For 48.3% of these text messages, basic documentation was completed within less than 1 hour and a text message was sent within 60 minutes after patient admission.

Conclusions: The system proved to be stable in generating text messages using IT-based CIS to identify acute stroke trial patients. The system operated on information which is documented routinely and did not result in a higher workload. Delays between patient admission and the text message were caused by delayed completion of basic documentation. To use the automated real-time text messaging system to immediately identify emergency patients suitable for acute stroke trials, further development needs to focus on eliminating delays in documentation for the SSA in the emergency department.

Figure 1

Overview of the Standardized Stroke Algorithm (SSA) and the resulting text message. (1) Physicians at the different EDs document routine clinical data in the electronic Standard Stroke Sheet (SSS) at the clinical workstations. (2) Data is saved and made available to interconnect by the clinical information system (CIS). (3) The CIS-server is polled by the e-Gate (interconnection gateway) every 10 minutes and data is retrieved from the CIS. (4) The filter server provides anonymous data and is controlled by a service desktop to adjust the filter settings. (5) The external provider sends the anonymized data to the (4a) pre-assigned recipients. (6) Screenshot of exemplary text messages received on mobile phone on January 8^th, 2013, with information about hospital site and department as well as patient’s sex, age, date and time of admission, and NIHSS score. While the SSA operates on personalized patient data within the Charité intranet, no personal data is transmitted via internet.