This was a prospective, block randomized trial of adult patients (aged 18 years or older) undergoing elective orthopedic procedures (ClinicalTrials.gov NCT03532256). The study took place over 6 weeks between July 1, 2019, and August 12, 2019. The primary objective was to determine if the method of patient engagement would impact response and completion rates. The primary outcome was response rate and total completion of PRO questionnaires.

Eligible patients included adult patients undergoing an elective outpatient orthopedic surgery (including knee, hip, shoulder, and elbow) and prescribed an acute opioid for postoperative pain. Acute opioid pain medications included oxycodone, oxycodone/acetaminophen, hydromorphone, or hydrocodone. Exclusion criteria included no access to an SMS-capable device or no opioid prescription. These inclusion and exclusion criteria were consistent with prior published research protocols using an established automated postoperative messaging program [24,25]. All patients were recruited from the University of Pennsylvania Health System Department of Orthopedics, which performs approximately 13,000 surgeries per year. In addition to general procedures (~3500 annual surgeries), the department consists of the following divisions: (1) Sports Medicine (~1700 annual surgeries), (2) Hand (~1900), (3) Joints (~4500), and (4) Trauma (~1500).

The research team had previously worked with the institution’s legal, privacy, and patient safety departments to obtain remote, SMS text message–based consent for data collection [25]. This was an intentional design of a larger institutional program aimed at improving acute opioid prescribing. The approach allows for remote consent in an attempt to reduce clinical providers’ workload and improve the scale of engaging eligible patients.

Following an elective outpatient surgery within the Department of Orthopedics, patients who underwent surgery received an initial SMS text message on the second day after the procedure. This message informed the patient of safe SMS text messaging data practices, provided links to further information regarding the follow-up research study, and offered the ability to opt in or opt out of further messaging. Patient were asked to consent via a simple SMS text message response of “yes.” Consenting patients were then sent SMS text messages on postoperative days number 4, 7, 14, 21, and 28. Patients were block randomized in groups of 2 or 4 to receive either (1) a hyperlink to a web-based mobile survey or (2) automated bidirectional SMS text messaging in a conversational format. Mobile survey questions and conversational questions were identical and included PROs related to self-reported pain intensity, ability to manage pain, use of prescribed medications, and ability to control pain. Block randomization was used to achieve balance in the two groups across demographics, procedure date, and dates of mobile engagement.

Individuals who either did not respond, opted out, or self-reported no further planned use of acute opioid medications were not subsequently messaged. For example, if a participant completed the mobile survey or replied to the bidirectional SMS text messaging on postoperative day 4 and indicated they were no longer planning on using their opioid prescription, no further messaging was sent on day 7, 14, 21, or 28. Participants who did not respond to any messaging would receive a reminder message within 30 minutes; if participants did not respond, then no further messaging was sent. The follow-up intervals were determined with input from a key clinician and surgeon and were in line with prior studies that evaluated PROs at 1 or 2 weeks postoperation.

Descriptive summary statistics were used to characterize the study population, using mean and standard deviation for age, and frequencies and percentages for categorical variables including gender, type of surgery, opioid tablet quantity, comorbidities, and invasiveness of the procedure (open surgical approach vs laparoscopic). To determine differences in the primary outcome of response rate at day 4 between the messaging arms and categorical demographic variables, Fisher exact test was used. To determine the difference in response at day 4 by age, Student t test was used. To determine if there were differences in secondary outcomes such as opioid use over time, Kaplan-Meier product-limit method with a log-rank test was used. A 2-sided α of .025 was considered statistically significant. All opioid prescription types were converted to equivalent doses of 5 mg oxycodone tablets [26]. All analyses were performed using SAS statistical software (version 9.4; SAS Institute).

The Institutional Review Board of the University of Pennsylvania approved this study (number 827461).

This study has two key findings within the context of mobile patient engagement and data capture. First, immediate postoperative patient engagement and research consent are feasible using digital methods. Second, there were no significant differences in the overall response rates between the modalities of bidirectional SMS text messaging and mobile surveys. We deployed a direct-to-patient approach in obtaining mobile consent to prospectively capture PROs for postoperative pain and pain management. Prior studies aiming to understand patients’ pain and use of prescription opioids following surgeries have been limited by retrospective design, telephone or paper surveys, and recall bias [19,22,27]. We worked collaboratively within the health system to develop an approach that offloads clinical providers from obtaining written consent during preoperative visits and reaches out to patients following their procedure through SMS text messaging. This mobile consent process may be further studied to decrease in-person time, allow for researchers to link important research and other patient information, and offer patients the ability to opt out.

Second, the method and approach used to capture patient-reported data may not significantly impact initial response rates and overall completion rates. The ways in which patients use digital technology and mobile apps continue to change [28]. This study begins to analyze direct-to-patient methods that capture patient-reported data [29]. These early results indicate mobile methods can be used to engage postoperative patients and may provide a scalable approach for engaging larger patient populations. Traditional paper-based surveys rely on mail services and may introduce time delays and recall bias, whereas digital methods allow patients to respond in the moment [30]. Though not statistically significant, we found that patients undergoing more invasive procedures (ie, open surgeries and nonlaparoscopic surgeries) or any procedure on the hip were more likely to respond and remain engaged. This suggests an opportunity to provide tailored content and messaging as more procedures become outpatient and more recovery is based in the home. Ultimately, we describe similar and consistent completion rates as the weeks passed following patients’ surgeries.

This study has limitations. First, the study and data collection were performed at a single academic medical center and thus the study is not generalizable. Second, selection bias may be present as patients needed to have a mobile device and opt to answer questions through either the mobile survey or the bidirectional SMS text messaging system. Those patients receiving a mobile link to the survey must have a smartphone and may represent a select population. Third, patients were undergoing elective outpatient procedures and may represent a population that is not generalizable to all procedures. This study compares two techniques that have been rapidly adopted; it does not compare these directly to paper methods, which have more traditionally been used.

This study demonstrates the early feasibility of PRO capture using two methods of patient engagement following orthopedic surgery. The findings suggest no major differences between bidirectional SMS text messaging and mobile surveys to collect PRO data. In an increasingly digital era, clinicians and researchers may employ digital surveys as a tool for rapid collection of PROs, which can inform research and a learning health system. Future studies will need to investigate larger scale programs and generalizability outside of surgical settings and for broader populations.