Abstract
Objective
To evaluate patient satisfaction with the experience of using an autonomous wheelchair to transport patients in a large outpatient clinical environment.
Methods
The autonomous wheelchair pilot was approved as a feasibility pilot by the institutional committees and deemed a quality improvement project by the Institutional Review Board (IRB). A total of 409 adult patients using an autonomous wheelchair at a large academic medical center who volunteered to complete a paper survey were included. The survey was administered immediately after autonomous wheelchair use, using a cross-sectional, anonymous survey, between 15 Oct 2025 and 14 Jan 2026. Of 409 completed surveys, six were excluded because participants did not identify their endpoint for stratification purposes. Descriptive analysis included frequencies and percentages of responses.
Results
No collisions or adverse events were observed during the pilot, and the system operated reliably within the predefined routes. Most survey respondents were first-time users (335/402 [83.3%]). A majority reported they would use the autonomous wheelchair again (341/395 [86.3%]) and would recommend it to others (364/397 [91.7%]). Overall, the experience was rated better than expected by 293 of 393 participants (74.6%). When given a choice, 271 of 379 respondents (71.5%) preferred the autonomous wheelchair over a staff-operated wheelchair.
Conclusion
These findings suggest that autonomous wheelchairs are feasible and acceptable to patients in a controlled outpatient setting and support continued piloting and prospective evaluation.
Keywords
Introduction
Over the past 10 years, the use of automation and artificial intelligence (AI) has grown in ambulatory and hospital medical settings.1,2 Although automation and AI adoption have grown recently in the healthcare sector, implementation has lagged behind other sectors, such as the automotive industry, finance and information services. 2 The prevailing model of patient mobility within healthcare centers has historically relied on staff-assisted transport, most commonly using manual wheelchairs. In this framework, patient movement is coordinated and executed by trained personnel, typically in response to a patient’s service request. Current staff-assisted transport models may be limited by staffing constraints, delays, and scalability challenges, particularly in large and expanding healthcare environments. These limitations can lead to inefficiencies in patient flow, increased wait times, and variability in patient experience. As healthcare systems face ongoing workforce constraints and rising patient volumes, there is a need to explore alternative models of patient transport that are scalable, efficient, and patient-centered. These challenges provide a more specific rationale for evaluating autonomous mobility solutions in clinical settings.
A recent 2025 survey across various sectors of the economy showed that AI use in healthcare was 8.9%, which was lower than in all other areas surveyed, such as finance and insurance (11.6%), education (15.1%), technical services (19.2%), and information services (23.2%). 1 Early examples of implementation in healthcare have largely focused on administrative automation within the electronic medical record (EMR), including automated scheduling, inbox management, prior authorizations, and billing assistance, with the goal to reducing clinical burden. 3 Wider implementation of automation and AI in the US healthcare sector has been slowed by several factors, including concerns over data protection, patient privacy, and safety, as well as the need for large studies to gain regulatory approval from authorities such as the FDA. 2
Large-scale, successful implementation of clinical automation and AI in healthcare has centered on robotic-assisted surgery, AI-assisted radiology diagnostics, and clinical decision-making (CDM). 4 Autonomous AI-enabled robots have been used in healthcare settings, serving roles such as rehabilitation assistance, food service delivery, and custodial services. Rehabilitation robots can assist in physical therapy and stroke recovery.5-7 These robotic rehabilitation devices can assist with gait training, extremity rehabilitation, and individualized therapy programs. 8 In addition, robotics has been used for the transportation of supplies and to enhance patient mobility in the hospital setting.8,9
To our knowledge, no attempts have been made in the United States to date to transport patients in hospital settings using automated mobile devices. With recent innovations and technological advancements in autonomous mobility wheelchairs and automobiles, they are now widely deployed across industries beyond healthcare. Autonomous mobile wheelchairs have been used to transport people to airports around the world. A small pilot study of the use of AI-enabled wheelchairs in a Japanese hospital has recently been published. 10 The primary goal was to evaluate the safety and general feasibility of deploying autonomous wheelchairs in the hospital. A total of 51 patients with musculoskeletal disorders reported high satisfaction (86%), but over 50% reported feeling unsafe at times. Despite these feelings of being unsafe, there were no reported safety incidents during the autonomous rides. 10 Besides this single-center pilot, there is a paucity of other examples of the deployment of autonomous wheelchair transport for patients.
This manuscript reports patient-reported outcomes and implementation insights from a real-world pilot of autonomous wheelchair transport at a large medical center. Patients who used autonomous wheelchairs were surveyed, with the hypothesis that they would report high levels of satisfaction with their experience.
Methods
This study was conducted at a large tertiary academic medical center in the Midwest USA, within a high-volume outpatient clinical environment. This project was undertaken as a quality improvement initiative and operational implementation initiative and did not constitute human subjects research as defined by federal regulations because the surveys were anonymous, voluntary, and collected solely for service evaluation purposes. Therefore, a formal Institutional Review Board review was not required under institutional policy (Figures 1 and 2). Image of autonomous wheelchairs waiting to be used Schematics of the autonomous wheelchair used in this transportation pilot
Implementation of Autonomous Wheelchairs
The autonomous wheelchair program was initiated to address increasing demand for efficient, patient-centered mobility solutions within a large tertiary medical center. Due to significant disruption to patient flow at the facility, alternative modes of transportation were evaluated. Early conceptual development began with discussions focused on identifying operational gaps in patient transport and on opportunities to automate to enhance independence, efficiency, and experience. Initial engagement with an external mobility vendor (WHILL Mobility Inc) occurred through structured meetings in early 2025, followed by on-site demonstrations of the autonomous wheelchair (Figures 1 and 2) and feasibility discussions to assess alignment with institutional goals and patient needs.
A multidisciplinary stakeholder group was convened to guide implementation, including representatives from facilities and support services, robotics and automation leadership, patient safety, clinical operations, and the Office of Patient Experience (OPX). This stakeholder group provided oversight throughout planning, deployment, and evaluation, ensuring that operational, safety, and patient-centered considerations were addressed concurrently. The term ” self-driving scooter” was selected to be included in the survey. This was based on feedback from our OPX, to be easier to read and understand, at a basic level. In addition, the term ‘self-driving scooter’ is more patient-centric, emphasizing mobility and autonomy rather than disability, which is often associated with the term wheelchair.
A comprehensive site evaluation was conducted prior to deployment to determine routes for autonomous operation. Routes were selected based on pedestrian traffic patterns, environmental predictability, accessibility, and relevance to patient transport needs. Selected routes connected to high-volume clinical and parking locations and were intentionally limited to controlled environments without elevator use, highly congested areas, or dynamic construction obstacles during the initial pilot phase. The route used in the current pilot was approximately 800 feet long and connected the patient parking garage to the main outpatient clinic building (Figure 3). Facility mapping was performed by the vendor and the operational user teams. This process included digital scanning of corridors, identification of fixed landmarks, establishment of geofenced pathways, and definition of speed zones to optimize safety and ride smoothness. Mapping was refined during testing to accommodate real-world operational conditions and ensure reliable autonomous navigation. Map of the route chosen for the autonomous wheelchair pilot, which ran 800 feet, connecting the parking garage and outpatient building
Prior to launch, the program underwent formal review through institutional governance pathways, including robotics oversight, facilities safety (with specific attention to lithium battery management), and patient safety endorsement. Safety validation included hardware performance testing, emergency stop verification, sensor functionality checks, and live-environment trial runs. No adverse events or collisions occurred during testing or pilot operations.
Operational readiness included structured training for operational staff for daily monitoring and oversight. Training covered device availability checks, battery management, basic troubleshooting, passenger assistance, and escalation pathways for technical or safety concerns. Passenger simulation exercises were conducted prior to launch to validate workflows and identify opportunities for refinement. This training also offered reassurance about job stability for this group of employees, as they could see how the autonomous wheelchairs filled a need, but the reliance on manual wheelchairs will remain high.
A centralized fleet management platform was implemented to enable real-time monitoring of device location, status, and diagnostics. This platform supports dispatch coordination, auto-return functionality, and data capture for subsequent evaluation of utilization patterns and system performance. The platform was one of the main safety features to ensure oversight of all riders.
The six autonomous wheelchairs used in the pilot were monitored at all times. If a patient pushed the pause button at any time during the ride, a notice was sent to the portal, and a staff member could walk to the location and offer assistance. There were staff at each of the beginning and end locations of the route that gave instructions, answered questions, cleaned the equipment between each use, and monitored the user experience. The pause button was not used for any emergency situation during the pilot. There are retail spaces along the route where patients stopped the wheelchair and proceeded into the shop. The frequency of this was minimal. The program followed a phased deployment strategy, including mapping, testing, and staff training. Once these phases were completed, a limited release to allow for real-time observation and adjustment under live conditions was conducted in September 2025. Following patient safety endorsement and operational validation, the program transitioned to an official go-live pilot on October 15, 2025 (ending on January 14, 2026). During the pilot period, autonomous wheelchairs operated along predefined routes with operational staff available to assist users (patients) at boarding locations. The implementation emphasized patient autonomy while maintaining visible support to promote comfort and confidence, particularly for first-time users.
Study Population
During the pilot period, sequential adult patients who used the autonomous wheelchair were invited to participate in an anonymous paper survey. Participation was voluntary, and no incentives were provided. There were no formal exclusion criteria beyond the inability or unwillingness to complete the survey. The number of patients approached was not tracked, and therefore, a response rate could not be calculated. There were two routes where patients could use the autonomous wheelchair, both of which connected via a subterranean walkway. The routes were from the main patient parking ramp to the main outpatient building and from the main outpatient building to the main patient parking ramp, totaling 800 feet (Figure 3).
Survey Development
A survey was adapted from a previously validated patient satisfaction instrument.11-13 However, it was modified for the operational context of this pilot and was not formally validated for autonomous wheelchair use. The survey was deliberately designed to be brief and low burden for participants to complete in a real-world operational setting. The survey included 7 quantitative items with binary and categorical response options to assess experience, preferences, and willingness to reuse, along with 2 open-ended comments soliciting additional input on their experience. It was anonymous (i.e.,no patient identifiers), and all surveys were administered and completed on paper. The survey included: prior wheelchair use, overall experience, willingness to reuse the wheelchair, recommendations to others, preference compared with staff-operated wheelchairs, and open-ended comments on benefits and areas for improvement. As noted previously, based on feedback from OPX, the term “self-driving scooter” was chosen because it was easier for patients with varying levels of knowledge to comprehend.
Data Collection
A non-clinical operational staff member collected the anonymous surveys at the conclusion of each route and forwarded them for data entry. Subsequently, all returned surveys were digitized using the institution-hosted electronic data capture (REDCap) platform.14,15 REDCap is a secure, web-based platform for researchers to collect study data. It features a user-friendly interface, detailed audit trails, easy export to statistical software, and supports integration with external systems.14,15
Data Analysis
Quantitative data were summarized using descriptive statistics, including frequencies and percentages. Survey questions were summarized by the direction patients were heading (towards the parking ramp or the main outpatient clinic building), and the frequencies of their responses were reported. Stratification by direction of travel was performed by destination context (arrival vs. departure) because we assumed a priori that an anonymous sample of those completing the return included many patients who also used the autonomous wheelchair on arrival. This sample may be biased toward those who viewed the wheelchair favorably on arrival. An exploratory analysis to compare the transportation mode preference between repeated users and first-time users was completed using a Chi-square test. First-time users were also compared using a Chi-square test to identify potential differences in preference by direction of travel. SAS statistical software was used for all analyses (SAS version 9.4, SAS Institute Inc.).
Qualitative responses were analyzed using content analysis. Two members of the statistical study team independently reviewed responses and grouped them into thematic categories; discrepancies were resolved through consensus. Themes were developed inductively from the data. Comments on improvements for the autonomous wheelchair were grouped into thematic categories and displayed on a bar graph. The comments about the patient’s preference were graphed based on whether the patient preferred the autonomous wheelchair or the standard wheelchair.
Results
Of the total 409 who completed the survey, six patients were excluded from the analysis because they did not identify their endpoint for stratification purposes. No demographic or clinical characteristics were collected; therefore, the study population cannot be further described.
Satisfaction Survey
Results From Survey by Endpoint
(A) Themes of preference comments by which method of transport they preferred. (B) Themes of improvement comments in the current autonomous wheelchair pilot
Exploratory analyses were conducted to examine potential factors influencing preference for the autonomous wheelchair versus the standard wheelchair. Repeated users showed no significant difference in preference when compared to those who were first-time users (70.8% vs 71.6% preferred autonomous wheelchair, p=0.90). Comparisons between endpoints were restricted to first-time users, and no significant difference in preference was found (p=0.26).
Survey Comments
Comments about why a patient preferred or did not prefer the autonomous wheelchair were also categorized. Those who did not prefer the autonomous wheelchair had comments that fell into four categories: they had braking problems with the autonomous wheelchair (“Humans are better at anticipating when to slow or stop. Not as jerky when pushing the wheelchair.”), they wanted the autonomous chair to move at a faster speed (“Regular wheelchair is faster but it’s a great option for long hallways”), they had other problems besides speed or braking with the autonomous wheelchair (“Unreliable, my coat blocked forward sensor”), or they wanted the human component of a standard wheelchair (“Enjoy interaction with escort staff, they are encouraging when you are having a tough day.”). For those who preferred the autonomous wheelchair, the common themes were that it was convenient (“I find this to be a more relaxing experience.”), allowed them to be autonomous (“I like how it eases the burden for the caregiver and gives the patient freedom to do things themselves.”), or was a fun and novel opportunity (“It’s comfortable, careful, slow, and fun!”). Miscellaneous comments from those who preferred the autonomous wheelchair were also reported (“I can see pluses and minuses to both the scooter and wheelchair.”).
The improvement comments were categorized into the following themes: faster/speed improvements, additional features to the automated wheelchair, brakes/stops, addition of other destinations, and no improvement needed. The theme with the most comments (n=75) was the speed category. Most of the comments in this category called for increased speed. Additional features spanned from adding small features, such as cupholders, to adding a flag or other markers to make it more visible. Another common comment on improvement was that the brakes were jerky or that the autonomous chair would suddenly stop when an object or person was near its path.
Implementation Findings
Key operational considerations included structured facility mapping, predefined routes, and controlled environments without elevators or high congestion. Staff presence supported first-time users and enhanced confidence.
Discussion
To our knowledge, this is the first study to describe the implementation of an autonomous wheelchair program within a tertiary medical center in the United States. In this study, patients who used autonomous wheelchairs reported high satisfaction and strong acceptance of this mobility technology. Most users were first-time riders, yet the majority expressed willingness to reuse and recommend the system, with many reporting a better-than-expected experience. These observations provide early insight into how autonomous mobility technologies may be received in a controlled outpatient setting. Overall, most patients were happy with the autonomous wheelchair and accepted it well. The steps involved in establishing the program are also comprehensively described, creating a roadmap that may support implementation at similar centers.
Importantly, this pilot was not designed as a formal safety study, and patient safety perceptions were not assessed using validated safety instruments or predefined clinical safety endpoints. A review of the literature found one previous example of an autonomous wheelchair program in a hospital setting, reported from a Japanese center. In that study, 51 patients with musculoskeletal disorders traveled about 100 meters between clinical locations. Most patients were satisfied, with 86% rating their experience at 3 or higher on a 5-point scale. These results are similar to ours, showing high patient satisfaction and willingness to use the device again. In the Japanese study, 26 of 51 patients reported feeling unsafe at some point during the ride, but there were no actual accidents or collisions. Within the program, no collisions or adverse events were observed. Although specific questions regarding safety were not included, some participants in this study, as well as in the Japanese study, reported abrupt stops or jerky movements that led to brief feelings of instability or concern about falling. In the Japanese study, it was especially evident in those recovering from surgery or those who were more vulnerable. 10 Still, several respondents in our study said they felt safe overall and found the device especially helpful for people with physical disabilities, especially when no escort was available. In addition, key operational considerations in our study included the need for structured facility mapping, predefined routes, and controlled environments without elevators or high congestion. Staff presence at boarding locations supported first-time users and enhanced confidence in the system. These results show that no adverse events were observed; however, safety was not formally measured. Perceived safety remains an important consideration for future studies.
In addition to patient satisfaction, autonomous wheelchairs may offer operational benefits. Healthcare systems are encountering significant staffing shortages and workforce constraints. While AI technologies, including autonomous wheelchairs, are designed to support, not replace, humans, autonomous mobility solutions can help address gaps in patient transport during staffing shortages. By reducing reliance on staff for routine transport, these systems may lower wait times, improve patient flow, and increase overall patient satisfaction. It was clear in our study that patients were very satisfied with the autonomous wheelchairs on the current routes, with the majority preferring them to regular wheelchairs. This implementation was in the context of our institution, which is undergoing expansion and construction, resulting in significant disruptions to parking. In these situations of disruption, autonomous wheelchairs could provide a scalable solution to changing infrastructure needs.
Limitations
It should be noted that the paper surveys used in this study were administered by non-clinical operational staff; as a result, no tracking of the number of people approached was conducted, which prevents calculation of a response rate and introduces potential selection bias. Other limitations within this project included that it was a single-center study conducted along a defined route. In addition, because demographic and clinical variables were not collected, the current study cannot determine which patient populations may benefit most or least from autonomous wheelchair transport, limiting interpretation of generalizability and subgroup applicability. Furthermore, this limitation restricts the interpretation of which patient populations may derive the greatest benefit and limit external applicability. This also limits the ability to tailor autonomous mobility solutions to specific patient needs and to identify individuals most likely to benefit from this technology. The route did not include elevators, highly congested areas, or dynamic construction-related obstacles, and therefore may not fully represent real-world variability. In addition, although some participants had prior experience with an autonomous wheelchair, repeat users were not tracked longitudinally, and safety perceptions were not assessed using a standardized instrument. Another limitation is that participation in the survey was voluntary, which may introduce response bias, as individuals with particularly positive or negative experiences may have been more likely to respond. In addition, surveys were administered by non-clinical operational staff involved in the pilot workflow, which may have influenced participant responses and introduced additional response bias. The survey instrument was adapted but not formally validated, and safety perceptions were not systematically measured. Although exploratory subgroup comparisons were conducted (e.g., first-time vs. repeat users), these analyses were limited in scope and lacked power to draw definitive subgroup conclusions. The high proportion of first-time users also raises the possibility of a novelty effect, in which initial exposure to new technology may positively influence satisfaction. Although early receptivity is encouraging, the current study cannot determine whether these perceptions would remain stable with repeated use. Importantly, implementation required a highly controlled environment, including predefined routes, the absence of elevators, and limited congestion. The system also depended on vendor-led facility mapping and geofencing. These constraints may limit scalability and highlight the need for further evaluation in more complex environments. Longitudinal assessment of repeat users will be important to distinguish sustained acceptability from initial enthusiasm. Future studies should evaluate performance with expanded routes, multiple devices, and increasingly complex navigation scenarios. Although no safety issues were detected in this pilot, investigation into patient-reported safety, device speed optimization, smoother braking algorithms, and integration into more extensive hospital transport workflows will be essential to inform next-generation implementation. Furthermore, future studies should also incorporate demographic and clinical variables to evaluate whether satisfaction and usability differ across patient populations.
Conclusion
In summary, these findings suggest that autonomous wheelchairs may be feasible and acceptable in a controlled outpatient setting, with generally favorable patient-reported experiences and willingness to reuse. However, these findings should be interpreted as exploratory given the descriptive design and operational nature of the pilot.
Supplemental Material
Supplemental material - Patient Perspectives on an Autonomous Wheelchair Transport Pilot in a Tertiary Medical Center: A Cross-Sectional Survey
Supplemental material for Patient Perspectives on an Autonomous Wheelchair Transport Pilot in a Tertiary Medical Center: A Cross-Sectional Survey by Stephanie Decker Hurt, Jithinraj Edakkanambeth Varayil, Jon E. Bowman, Karen Fischer, Darrell R. Schroeder, and Ivana T. Croghan in Journal of Primary Care & Community Health.
Supplemental Material
Supplemental material - Patient Perspectives on an Autonomous Wheelchair Transport Pilot in a Tertiary Medical Center: A Cross-Sectional Survey
Supplemental material for Patient Perspectives on an Autonomous Wheelchair Transport Pilot in a Tertiary Medical Center: A Cross-Sectional Survey by Stephanie Decker Hurt, Jithinraj Edakkanambeth Varayil, Jon E. Bowman, Karen Fischer, Darrell R. Schroeder, and Ivana T. Croghan in Journal of Primary Care & Community Health.
Footnotes
Acknowledgements
We would like to thank Val Kleinhans, Amanda Yang, Alyse Schroeder, Doug Holtan, the General Service management team, and the dedicated patient movement team members for their ongoing support of this pilot. We thank our partners in Facilities, Security, and the Office of Patient Experience for their guidance in making this pilot possible.
Ethical Considerations
Therefore, it was not submitted for Institutional Review Board review. All data were handled in accordance with institutional privacy and confidentiality policies.
Consent to Participate
This project was undertaken as a quality improvement initiative and did not constitute human subjects research as defined by federal regulations.
Author Contributions
All the authors participated in the study concept and design, the development of the survey, identification of the survey cohort, data collection, analysis and interpretation of data, drafting and revising the paper, and have seen and approved the final version of the manuscript.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported in part by Mayo Clinic, Rochester, MN. The data entry system used was REDCap, supported in part by the Center for Clinical and Translational Science award (UL1 TR000135) from the National Center for Advancing Translational Sciences (NCATS) funding for the project was Mayo Clinic and the award was only for the REDCap system used for data entry.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: All authors declare no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; and no other relationships or activities that could appear to have influenced the submitted work.
Data Availability Statement
All data supporting the study findings are contained within this manuscript.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
