StreaMD multilingual orthopaedic SMS chatbot
Purpose #
To coach total joint arthroplasty (TJA) patients through the perioperative period in their own language via automated text messages, delivering perioperative education and remote monitoring to patients — including those with limited English proficiency (LEP) — who might otherwise face language barriers to perioperative instructions. The service (StreaMD, Chicago, IL) is available in English and 19 additional languages.
"The multilingual SMS chatbot service used in this study was provided by StreaMD (Chicago, IL). Their orthopaedic chatbot service coaches patients before and after their surgery through automated text messages." (Rainey, 2023, p. S79)
"This chatbot is available in English and 19 additional languages and has been shown to improve patient engagement and clinical outcomes in the early postoperative period following TJA [17]." (Rainey, 2023, p. S79)
Mechanism #
An SMS (short message service) chatbot pushes scheduled, procedure- and physician-specific messages that include preparation/recovery information, short instructional therapy videos, and personalized video messages from the patient's own surgeon. It is interactive: patients can text questions or topics (e.g., pain, showering) and receive further automated responses. Because it runs over plain SMS, the only technology a patient needs is an SMS-capable cell phone. Engagement is quantified as the total number of patient-generated text messages sent to the chatbot.
"The messaging content includes information about preparing and recovering from surgery, short instructional therapy videos, and personalized video messages from their surgeons. In addition to receiving scheduled messages, patients are able to send text messages to the chatbot regarding questions or topics (such as pain and showering) and receive further automated responses to these topics." (Rainey, 2023, p. S79)
"The advantage of the aforementioned chatbot is that the only technology required by patients is a cell phone capable of receiving SMS-based text messages, with current estimates showing that 97% of the US population owns an SMS-capable cellphone [29]." (Rainey, 2023, p. S81)
Example(s) of usage #
Deployed at a single academic center (University of Utah) for TJA patients from 2020-2022; 1,282 TJA patients enrolled, of whom 47 received their messages in a language other than English (a total of 17 languages represented across the LEP cohorts, Spanish most common). Figure 1 shows a typical automated bot response (English and Polish) to patient questions about pain and dressing removal.
"Figure 1 demonstrates a typical automated bot response to patients in response to questions regarding pain and dressing removal." (Rainey, 2023, p. S79)
Other Notes #
The chatbot is the concrete system whose "How?" the outcome EVDs from this paper reference. The plugin's relation grammar does not yet define EVD/CLM→ART edges, so links to this artifact are wikilinks for now. A prior study (Campbell et al. 2019, ref [17]) established the underlying automated text-messaging system's efficacy in a general (not specifically LEP) TJA population.