A Systematic Review of the Use of Google Glass in Graduate Medical Education Open Access

Graduate medical education (GME) has emphasized the importance of assessing resident competencies and milestones so that residency programs throughout the United States will produce physicians demonstrating proficiency across multiple domains.^1,2  In order to comply with new requirements, there is renewed emphasis on resident assessment methods. Studies have documented clear associations between level of supervision and educational and patient-related outcomes,^3–13  resident autonomy,^5,6,9,13  satisfaction,^5,9,12  and clinical competence and preparedness after graduation.^{3,5,9,14–16}  Moreover, trainees have identified lack of supervision as a major source of suboptimal patient outcomes.¹⁶ 

Due to scheduling limitations, supervising physicians are not always available to provide in-person oversight, which may be ideal for direct observation and supervision. This is most challenging during time-sensitive emergency consultations and after-hours rotations, such as night float.⁶  Lefrak and colleagues⁸  found night float residents encountered lower resident supervision compared with their day rotation colleagues. One strategy to ensure patient safety and quality outcomes while reinforcing resident autonomy and clinical education is adding in-house faculty, which is a costly and resource-limited solution.^6,10,11,17 

Telemedicine offers a potential medium for enhancing remote supervision, education, and evaluation of residents without necessitating increased physical presence.^17–30  A variety of specialties have implemented conventional telemedicine in GME by assessing its feasibility for educating trainees via recorded videos^18,27,31  and live video-teleconferencing (VTC) with good feedback.^{17,19–21,23–30}  However, stationary telemedicine end points are not always practical for supervising residents performing hands-on physical examinations or procedures in multiple clinical environments.

Wearable technology has emerged as an alternative to stationary telemedicine and is of recent interest because of its applicability in the clinical setting. Google Glass (GG) in particular is a popular hands-free wearable device with telemedicine capability (figure 1). Multiple reviews have demonstrated GG's feasibility, usability, and acceptability in surgical settings^32–34  and have called for increased research into GG's role in clinical education.

In this review, we specifically investigate the use of GG in the clinical learning environment and evaluate its use in both surgical and nonsurgical settings. We appraise the literature for experiences with GG used in GME for resident supervision and education, which distinguishes it from previous reviews of clinical uses for GG. We also review the clinical utility and technical limitations of GG, again with a focus on feasibility in the GME clinical learning environment.

Our systematic review was conducted according to the guidelines in the 2009 PRISMA statement for reporting systematic reviews.³⁵ 

We established a search strategy after consulting 2 medical librarians. In August 2018, 3 databases (PubMed, MEDLINE, and Web of Science) were independently searched for articles that referenced the use of GG in clinical practice by physicians or trainees for the purpose of GME. To complete the search, we queried articles published in the last 5 years using the key word “Google Glass.” Two reviewers (C.C.W. and W.C.) independently screened titles, abstracts, and full-text articles, selecting only articles pertaining to use of GG in the GME setting. Only articles defined as original research or case reports that had full text available were included for analysis. The findings of original research articles, and particularly those with objective outcome measures, were weighted more heavily in the analysis. Conference proceedings or abstracts with no accompanying full text, editorials, commentaries, or online/news reports were not included. Duplicate records, non-English, and nonhuman studies were also omitted. Table 1 details our full eligibility requirements. Our review focuses on the use of GG due to the paucity of available literature on other wearable devices.³⁶ 

Each article satisfying our search criteria underwent extraction of the following data points: author names, publication year, specific medical specialty described (if any), outcomes assessed, and a summary of findings. We specifically aimed to highlight clinical uses and how GG, as a novel device, might serve as a substitute for conventional telemedicine to augment education and supervision.

The literature search yielded a total of 349 publications. After excluding duplicates, we performed an initial screening process of scanning the titles (removing 107 articles), then reading the abstracts (removing 25 articles), finally producing 48 unique publications. After reviewing the full texts, we removed 8 additional articles not pertaining to GME, and 3 review articles. A total of 37 articles were included in the final review (figure 2). The 37 publications included 32 original research articles and 5 case reports. The number of publications for each year is as follows: 2013 (n = 1), 2014 (n = 6), 2015 (n = 11), 2016 (n = 13), 2017 (n = 6), and 2018 (n = 0). The included articles were mostly from surgical or subsurgical specialties (n = 23, 62%), and the remainder were from medical specialties (n = 14, 38%). The most common uses for GG in these articles was surgical/procedural skills teaching or supervision (n = 23, 62%), imaging/study interpretation (n = 10, 27%), and simulation-based exercises (n = 4, 11%). A summary of included articles is presented in table 2.

Studies of GG in GME are in the early stages, with most studies (n = 26, 70%) describing feasibility, general use, and feedback on the device's utility. A small subset of these studies (n = 3, 8%) compared GG to other devices with regard to performance in the clinical learning environment on the basis of video quality, technical limitations, and ease of use. Few studies (n = 4, 11%) had objective measures of surgical/procedural performance comparing GG wearers to nonwearers, including time to completion, number of attempts, and need for repositioning/redirection. Only 2 studies (5%) specifically evaluated differences between GG-based assessment and conventional assessment methods (in-person supervision or third-person video assessment). Articles felt to be of particularly high quality are identified in table 2.

Currently available uses of GG broadly include VTC, photo and video capture, and custom prism displays (what can be viewed by the device wearer; figure 1).

GG's ability to conduct VTC has been demonstrated for surgery,^31,36–46  dermatology,^42,47  neurology,⁴⁸  pediatrics,⁴⁹  cardiology,⁵⁰  and toxicology.^51,52  With VTC, GG also doubles as a telementoring device, allowing trainees to broadcast their point of view to supervising physicians^{43,44,50,51,53}  and vice versa.³⁶  Vallurupalli et al⁵⁰  found GG could assist trainees in simulated clinical scenarios and allow for improved patient communication. Skolnik et al⁵¹  showed high agreeability between in-person and remote supervisors when toxicology fellows wore GG for live-streaming physical examinations and transmitting still photos of electrocardiograms (ECGs). Chai et al⁵²  reported that residents seeing toxicology consults in the emergency room using GG to aid their consultation (VTC with remote supervisors) resulted in a change in medical management in approximately one-half of cases. Thus, trainees were able to gain clinical experience and direct supervisor feedback even in the absence of direct physical oversight.

Beyond person-to-person VTC, existing literature highlights a range of GG uses, including live-streaming vital signs during simulated surgical scenarios,⁵⁴  recording video for resident assessment purposes,^{36,37,40,55,56}  building a video library to log improvements,²⁶  and capturing point-of-view procedures performed by senior physicians.^57,58  Sahyouni et al³⁷  reported that reviewing video clips recorded via GG enhanced neurological surgery residents' technical understanding of procedures, ultimately leading to increased confidence and level of comfort. Multiple studies reported improved trainee techniques following review of recorded surgical videos, with Chimenti and Mitten³⁸  noting that for percutaneous pinning of hand fractures in a cadaveric lab, use of GG led to decreased time to pin fracture, a decrease in the number of pin attempts, and a decrease in the number of fluoroscopic images obtained.³⁸  Lastly, photo capture has been used to obtain intraoperative photos,^40,42  obtain photos of electrocardiograms,⁵⁹  document airway placement in the operating room,⁶⁰  and photos of chest x-rays.⁶¹  Schaer et al⁶²  simulated 6 differing ECG rhythms thought to require urgent attention, and trainees using GG images for interpretation demonstrated no significant difference from live ECG readings on a laptop.

Custom-designed GG applications provide an avenue to enhance basic GG functionality. Custom-designed applications have been developed for pediatric anesthesia⁶³  and cardiac surgery⁶²  to continuously monitor patient parameters. Additionally, custom-designed GG applications have been utilized in image-guided procedures, such as ultrasound-guided central venous access,⁶⁴  intracranial tumor resection,⁶⁵  and urologic prosthetics.⁶⁶ 

In 30% (n = 11) of studies reviewed, the most commonly reported hardware criticism of GG in GME relates to battery life.^{27,28,37,38,45,57,58,65,67,68,71}  Additionally, GG use was affected by incompatibility with select surgical equipment,^{31,38,57,66,67}  inability to follow gaze and correct to the line of sight,^{31,36,57,66,67}  distractibility,^{38,40,65,67,72}  and propensity to overheat.^55,64  Use of GG may be limited by room lighting,^{36–38,57,66,67}  a need for additional glasses, or surgical magnification loupes.^31,36,37  However, at least 1 study⁵⁷  found that GG fit with magnifying loupes for scleral buckling surgery, and others utilized strong adhesive tape for loupe attachment.^65,68  GG software obstacles include a lack of zoom capability,^57,58,61,67  and connectivity issues that hamper VTC and recorded multimedia audiovisual quality.^{31,36,44,50,52,53,64,67,69}  To address zooming, Stetler et al⁷⁰  disclosed an increased accuracy of remote ECG interpretations on GG with the use of third-party software that enabled hands-free zoom and pan capability, addressing a drawback previously reported by Jeroudi et al.⁵⁹ 

Wearer distractibility is also a major concern in the GME setting. One study likened GG voice control features to hands-free communication while driving.⁶⁷  Another had reservations about inattention blindness,³⁸  but this may be reduced by increasing the emphasis on layout, space, colors, and timing design of the GG software. An additional study found that residents adopted a “listening attitude” during live VTC that negatively impacted performance.⁴⁹ 

As with all forms of telemedicine, patient privacy and data security represent a potential concern among medical GG users.^{36,38,39,44,47,48,55,57,67}  Under the Health Insurance Portability and Accountability Act (HIPAA) and the Health Information Technology for Economic and Clinical Health Act (HITECH),⁷³  many commercial wearable devices with GG-like features do not comply with the extensive information technology standards.⁵²  Muensterer et al⁶⁷  noticed GG automatically syncs with Google servers when charged and connected to a Wi-Fi network, leading to concerns about sensitive information being transmitted outside the medical center firewall.⁶⁷  To address concerns regarding privacy, many users retain the services of certified partners of GG in order to operationalize them for use in the health care setting.^47,51,52  From patient perspectives, use of GG appears to be acceptable.^63,67  Patients evaluated with use of GG in teledermatology consults felt comfortable when physicians used third-party HIPAA-compliant GGs.⁴⁷ 

A majority of studies expressed striking acceptability among physicians^{37,40,43,45,46,49,54,56,58,66,67,70}  and patients.^51,52,56,67  One study found that younger trainees rated GG as more educationally useful and less distracting in the operating room compared with their faculty counterparts.⁶⁶  Despite results that did not fully support the use of GG in GME, most researchers remained optimistic about future iterations and the potential to improve provision of clinical care and medical education.^{27,49,53,63,70} 

Our review highlights that much of the literature investigating use of GG in GME has been focused on the feasibility and general use of GG within the clinical learning environment. The literature suggests that the ability to execute VTC, particularly using the wearer's point of view, holds promise as a tool for procedural skills acquisition, remote supervision, and assessment.

Interestingly, we did not encounter any reports of patients wearing GG for point-of-view–recorded encounters with a resident as a tool for self-assessment. This use has been described in both nursing and dental training settings.^74,75  As did many of the articles in our review, much of the existing literature regarding use of GG within the medical realm is focused on reporting feasibility. Unfortunately, the literature to date does not include much quantitative or qualitative data reporting the utility of GG from the resident's perspective. Nor are there significant data specifically investigating if this enhanced supervision led to improvement in patient outcomes or resident performance.

Limitations of our review include the relative paucity of literature specific to GG applications in GME, and the published articles primarily include single institution studies or small sample sizes. We chose to exclusively study GG due to its early penetration among wearable devices into the clinical learning environment, although there are other wearable devices commercially available that may offer alternative characteristics for future application. Among the literature describing GG and its acceptance as an educational tool, few objective data exist, and most experience remains observational. Unsurprisingly, much of the current published literature focuses on GG incorporation in surgical settings. More studies of its utility and feasibility in nonsurgical settings will allow for a more complete picture of its future incorporation into medical education and training. Lastly, with the rapid pace of technologic innovation, results may not reflect current advances in wearable mobile health technology.

The body of literature evaluating GG in GME would be bolstered significantly by additional studies more specifically investigating benefits to resident learning and performance as opposed to simply the feasibility of use in the clinical learning environment. A qualitative assessment of the resident perspective when using GG in their education could provide greater insight into trainee experience beyond acceptability, which is what has been published most extensively. Evaluation of patients' attitudes to residents' use of GG in their care would also be informative.

Despite the promise and early experience with GG as a health care and GME tool, further investigation aimed at demonstrating specific educational benefits and feasibility in more diverse settings is needed before it can be fully integrated into the clinical learning environment. This review made evident potentially solvable issues with usability, durability, and acceptance currently seen across specialties. As mobile health technology continues to evolve, GG and other mobile, hands-free devices may serve as effective media for remote supervision and evaluation of medical trainees in the clinical learning environment.