RESEARCH/Original Article

Plastic surgery telehealth consultation expedites Emergency Department treatment

Journal of Telemedicine and Telecare 2017, Vol. 23(2) 321–327 ! The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1357633X16639459 jtt.sagepub.com

Angie M Paik, MD1, Mark S Granick, MD2 and Sandra Scott, MD3

Abstract Plastic surgery is a field that is particularly amenable to a telehealth milieu, as visual exam and radiographs guide proper diagnosis and management. The goals of this study were to evaluate telehealth feedback executed through an iPad app for plastic surgeryrelated consultations. A Quality Assurance/Quality Improvement (QA/QI) study was conducted over a 1-month period during which patients with hand injuries, facial injuries, or acute wounds presenting to the Emergency Department (ED) of a level-one trauma centre and university hospital were monitored. The study utilized a commercial iPad application through which up to four images and a brief history could be sent to a remote Plastic Surgery Educator (PSE) for evaluation. The PSE would respond with best practice information, references and videos to assist ED point-of-care providers. During the 1-month period of this study, there were 42 ED consultations for plastic surgical conditions. There was a highly significant difference in overall mean response time between consultants and PSEs (48.3 minutes vs. 8.9 minutes respectively, p < 0.001). The agreement between PSEs and consultants regarding patient assessment and care was 85.7% for in-person consultations and 100% for phone consultations. In four cases of telephone consultations, the ED providers placed splints incorrectly on hand-injured patients. Our results show that telehealth consultations to a remote plastic surgeon based on digital images and a brief history were able to produce timely and accurate responses in an emergency care facility. This design may have significant impact in rural areas, underserved populations, or regions abroad. Keywords Telemedicine, Telehealth, Emergency Consultation, Quality Assurance/Quality Improvement (QA/QI), plastic surgery Date received: 15 June 2013; Date accepted: 15 April 2014

Introduction Plastic surgery is a field that relies primarily on visual information for proper diagnosis. Photographic documentation remains a critical component of initial patient work-up and is also used to track the course of treatment. These characteristics make plastic surgery particularly amenable to telemedical/telehealth services, since digital images can be readily transmitted to remote plastic surgeons who can use this material to make accurate assessments. Sending photographs to colleagues to diagnose and monitor patients is already in popular practice.1–3 Similarly, a plastic surgeon can use digital images not only to evaluate a patient’s condition, but also to select appropriate web-based information, references, and videos that discuss the patient’s evaluation, treatment, and triage. There is a growing body of literature on the use of telehealth implementation in the real-time practice of plastic surgery.4–7 Murphy et al. showed that photographs were highly representative of the wound with the exception of depth assessment.6 Trovato et al. conducted a study where ‘‘medium’’ and ‘‘low’’-quality images were

captured and evaluated using the store and forward approach of telemedicine.4 This study reported that remote plastic surgeon evaluation had a negative predictive value of 96.38 when compared with the on-site physician, showing the ability to accurately rule out diagnoses. Diver et al. conducted a study where 95% of trauma referrals from distant hospitals were appropriately diagnosed with photographs when compared with face-to-face evaluation.5 It was also found that in 25% of these patients, photographic assessment alone could have precluded patient transfer altogether. 1

Brown University, Alpert Medical School, Providence, RI; Department of Plastic and Reconstructive Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA 2 Department of Surgery, Division of Plastic Surgery, Rutgers New Jersey Medical School, Newark, NJ, USA 3 Department of Emergency Medicine, Rutgers New Jersey Medical School, Newark, NJ, USA Corresponding author: Mark S Granick, MD, 140 Bergen Street, E1620, Newark, NJ 07103, USA. Email: [email protected]

322 These previous studies on the use of photographic telemedicine have shown promising results where remote plastic surgeons were able to make accurate assessment of a wound based on electronic images. They also suggest that its use in a clinical setting has the potential to triage injuries and recommend initial management. In order to better test this conclusion, a commercially available iPad mini (AppleÕ Inc., Cupertino, California) application (‘‘app’’) was utilized to send digital images obtained by the device as well as a brief history to a remote American Board of Plastic Surgery-certified plastic surgeon – the Plastic Surgery Educator (PSE) (ReconSTAT, LLC, Piscataway, New Jersey). The app did not require FDA approval and is proprietary, on loan from ReconSTAT, LLC. Since no patient identifiers are transmitted and the transmissions are encrypted, the device is HIPAA compliant. The PSE would then respond via the iPad, providing educational information including a best practice commentary, references and instructional videos critical to the evaluation, initial treatment and triage of the patient. The goal of this study was to compare the time to response and the congruency between the PSE and the hospital onsite consultants for plastic surgery-related cases presenting to the Emergency Department (ED).

Methods This study was conducted as a Quality Assurance (QA) and Quality Improvement (QI) initiative in order to understand the plastic surgery consultation process and evaluate an option for improved efficiency. It was conducted over a 1-month period during which the Plastic Surgery Research Fellow (AMP) monitored patients who presented to the ED at a level-one trauma centre and university hospital (the University Hospital, Newark, NJ). There is a fully accredited Emergency Medicine residency programme on-site. Active study time was restricted to Monday–Friday 8AM until 6PM. During that time all of the hospital-based consultants were physically located at the hospital. All patients who visit the ED are initially triaged depending on the urgency of their chief complaint to various areas that are staffed accordingly with ED physicians and nurse practitioners. After examination and work-up, these healthcare providers determine whether there is a need to consult a specialist service. Patients who presented with complaints that fell under a plastic surgeon’s area of expertise were followed for the purpose of this study to determine whether these cases required consultation from the ED staff. At this institution, Plastic Surgery and Orthopaedic Surgery alternate to cover the ‘‘Hand’’ consult service on a daily basis. Plastic Surgery, Otolaryngology, and Oral Maxillofacial Surgery rotate to cover the ‘‘Facial Injury’’ consult service. In addition, Plastic Surgery is available continuously for any acute wounds that necessitate consultation. All cases of hand injuries, facial injuries, and acute wounds that were deemed by ED clinicians (physician extender,

Journal of Telemedicine and Telecare 23(2) resident, or attending physician) to require a consultant were included in this study regardless of the consultant’s specialty. The consultants that evaluated patients were resident physicians in their respective fields. An iPad mini was used to obtain, send, and receive all relevant information and images, although no patient identifiers were transmitted. An iPad app was designed with an encrypted platform through which an abbreviated patient history, physical exam findings, and imaging could be securely conveyed to a PSE. Up to four images were obtained and sent to the PSE using the built-in 5 megapixel camera on the iPad mini, without compression. Photos were taken of the patient’s injury as well as of pertinent imaging (radiographs, commuted tomography, and MRI) from hospital computers equipped with imaging software. One inquiry was submitted for every patient case included in this study. Three PSEs were available during the study period to respond to inquiries on a first-come, first-served basis. PSEs could review submitted inquiries on either their personal computer or an iPad or iPhone with the downloaded app. PSEs were instructed to use the photos and brief history of the patient to select a package of educational materials for the ED clinicians to utilize in making treatment and triage decisions. The time to in-house consultant management was deemed the time between the initial call, by pager or directly to a cell phone, to the consultant and the consultant’s response with a discrete plan for the patient. The time to PSE response was the time between inquiry submission and inquiry response. The responses submitted by the PSE were compared with the recommendation of the consulted specialist. The Plastic Surgery Research Fellow additionally observed the management administered by the ED clinicians after telephone consultations. All data were collected in May 2013. As a QA/QI initiative an institutional review board (IRB) review was not required to initiate the project. An IRB submission was subsequently approved prior to publication. Nonparametric statistical analysis was conducted using the Fisher’s exact test for comparison of categorical data and the two-sample t-test and Mann–Whitney U test for evaluation of continuous variables. Statistical significance was set at p < 0.05.

Results During the 1-month period of this study, there were 42 visits to the ED at our institution that required consultation for a hand injury, facial injury, or acute wound. The iPad app was used for PSE evaluation for every consultation during the study. The majority of cases were hand injuries (73.8%, n ¼ 31) followed by facial injuries (21.4%, n ¼ 9) and acute wounds (4.8%, n ¼ 2). Of the hand injuries, 48.4% (n ¼ 15) were seen by plastic surgery and 51.6% (n ¼ 16) were seen by orthopaedics. Of the facial injuries, 55.6% (n ¼ 5) were seen by oromaxillofacial surgery, 22.2% (n ¼ 2) were seen by plastic surgery, 11.1%

Paik et al. (n ¼ 1) were seen by otolaryngology, and 11.1% (n ¼ 1) were seen by oculoplastic surgery. The data were analysed in two separate cohorts (in person and phone) with respect to time of response. However, the data were pooled for all consultants vs. PSEs when concurrence or management was analysed. Of the 28 consultations that were responded to in person, the mean response time was 68.5 minutes (standard deviation ¼ 41.4) and the median response time was 56 minutes (Figure 1). The median consultant response time in person was significantly higher than corresponding median PSE response time of 4.5 minutes (Mann– Whitney U test ¼ 51.5, p < 0.001) (Table 1). Consultants that provided a phone consultation had access to the patient’s pertinent history as given by ED clinicians as well as imaging through the hospital’s imaging software. Of the 14 consultations that were responded to by phone, the mean response time was 7.8 minutes (standard deviation ¼ 6.6) and the median response time was 6 minutes (Figure 2). There was no significant difference between the median consultant response time by phone and corresponding median PSE response time of 3 minutes (Mann–Whitney U test ¼ 63, p ¼ 0.11). When the data from these two cohorts were pooled, the response time for consultats (phone and in person) was a mean of 48.3 minutes. The mean response time for PSE responses, when pooled, was 8.9 minutes. This difference in response time was highly significant (two-sample t-test, p < 0.001). In order to determine whether there was agreement of patient assessment and initial management between the consultant and the PSE, each case was examined to

323 compare the consultant response with the corresponding PSE recommendations. The overall agreement rate between consultant and PSE was 90.5% (n ¼ 38). Of those consultations that were responded to in person, there was an 85.7% (n ¼ 24) agreement rate with the PSE (Figure 3). Of those consultations that were responded to by phone, there was a 100% (n ¼ 14) agreement rate with the PSE. There was no significant relationship with consultant response modality (in person vs. phone response) and agreement with PSE (Fisher’s Exact Test, p ¼ 0.37). The four cases showing variance between the PSE and the consultant were further evaluated. Three cases involved a treatment variance. The first was a difference in responses regarding the safe period of time for wound closure after injury: 12 hours vs. 24 hours. The second and third involved whether or not to obtain an X-ray after fracture reduction in the ED. One of these cases was a Bennett’s fracture which is typically treated operatively (Figure 4). The orthopaedic intern consultant reduced the fracture in the ED, while the PSE recommended a thumb spica splint and follow-up with a hand surgeon

Table 1. Distribution of cases receiving an in-person consultation, phone consultation, and the pooled responses.

In-person consultation Phone consultation Pooled responses

Number of cases

Number of corresponding PSE responses

28 14 42

28 14 42

Figure 1. Time to response comparing PSE and in-person consultant. PSE response time range was 1–50 minutes; median, 4.5 minutes. Inperson consultant response time range was 2–150 minutes; median, 56 minutes. *p < 0.001.

324

Journal of Telemedicine and Telecare 23(2)

Figure 2. Time to response comparing PSE and phone consultant. PSE response time range was 2–14 minutes; median, 3 minutes. In-person consultant response time range was 1–25 minutes; median, 6 minutes. p ¼ 0.11.

Figure 3. Number of cases that showed congruence in patient assessment between consultant and PSE. The red column shows the number of ED errors in implementing phone consultants’ recommendations.

within a week for this condition. The other, a Salter– Harris Type II fracture, was reduced by the orthopaedic intern consultant, while the PSE response suggested stabilization with a thumb spica splint and follow-up with a hand surgeon within a week. The remaining variance was a case where the PSE interpreted a red mass in the palm as a possible abscess, while the consultant deemed it to be a sebaceous cyst (Figure 5). The Plastic Surgery Research Fellow noted that in four instances the ED clinicians applied splints improperly on

patients with hand injuries. Although the ED clinicians reviewed the PSE responses, they did not look at the app’s attached instructional videos on splint application, which has information about optimal splint construction and positioning.

Discussion A QA initiative was undertaken to better understand the process of consultation for plastic surgery issues at a busy

Paik et al.

Figure 4. Paediatric patient presents to the ED with X-ray consistent with a Bennett’s fracture. Consultant recommended immediate reduction in the ED whereas the PSE recommended a thumb spica splint and follow-up with a hand surgeon within a week for operative management.

Figure 5. Patient presented to the ED with an erythematous mass of the palm. Consultant diagnosed this condition as a sebaceous cyst whereas the PSE interpreted this image as an abscess.

urban ED. There is a concern that the consultation process is unduly prolonged, causing patient retention in the ED for an unnecessarily long time. The QI portion of the evaluation assessed a telehealth system that could potentially be used to facilitate the plastic surgery consultation process. Our study reviewed 42 cases that presented to an

325 inner city hospital ED. The hours of the study were limited to daily working hours so that all of the consulting physicians would be physically present in the hospital during the time of the study. The majority of cases were hand or mid-face injuries where plastic surgery and other related sub-specialty consultants were called. Consultants could respond in two modalities to direct ED clinicians on proper initial management – in person or by phone. Our study showed a highly significant five-fold difference in overall response time between hospital’s on-site consultants and our PSE when the data was pooled for the two cohorts (48.3 vs. 8.9 minutes). It is not unreasonable to presume that during off-hours, holidays, and weekends, the response times by consultants would be even greater. When separating consultant responses by response modality, we noticed that in-person response times varied widely, with a maximum response time of 150 minutes. One must keep in mind that when a specialist is consulted, the patient must remain in an ED bed until the specialist evaluates the case and offers a plan. We found that patients would often be waiting for over 2 hours before a specialist was able to see them, even though all of the consultants were working within the hospital at the time of the consultation. This wait time in addition to the total time spent in the ED is taxing for patients and often leads to frustration, conflicts in scheduling, and overall inefficient delivery of care. As a QA finding, the process was found to be inefficient at best. There was a highly significant difference of 12-fold in median response time between in-person consultants and our PSE (56 vs. 4.5 minutes). It is clear that having a PSE respond conveniently from their personal electronic devices allowed for a quick and efficient response that ED clinicians could use to formulate a treatment plan for their patients. Although this response time differential was also evident for phone consultations compared with the associated PSE response in half of the time, it was not statistically significant as the size of this cohort was too small to generate statistically relevant analysis. The second component of this study was to determine how closely consultants’ plans for initial management aligned with those of our PSE. We report a 90.5% rate of congruency, which demonstrates that a PSE is able to successfully assess a patient based on digital images, pertinent radiographs, and a brief history. This is slightly better than the expected congruence, since the literature suggests that there is typically a 70–80% agreement between different physicians, even when directly examining patients.4 As noted in the Trovato et al. study, variance between on-site and off-site evaluators is greatest in patients with erythematous and edematous conditions, as was seen in the study case (Figure 5). With regard to the Bennett’s fracture and the Salter–Harris Type II fractures, delayed treatment after immediate stabilization is perfectly appropriate for an ED provider, particularly if the hand specialist is not immediately available to see the patient in the ED. As plastic surgery is a field that relies heavily on visual exam, we found that this telehealth

326 intervention could be used to adequately provide materials to facilitate triage and initial management of patients presenting to the ED. The QI findings are of considerable interest. The telehealth intervention provided significantly improved efficiency and a high degree of reliability. In addition, the telephone consults were found to result in errors of treatment. Even though the consultants suggested placing splints and triaging the patients to the subsequent clinics, the ED physicians occasionally placed the splints improperly. There were no cases of improperly placed splints during in-person consultations. This suggests that additional training for the ED providers might be necessary. It also suggests that the ED staff could use the videos, supplied as part of the educational materials by the telehealth service, showing how to apply different splints. In no cases did the ED staff take advantage of the availability of this instructional material. The greatest strength of this study is the novelty of the subject matter. While the use of telehealth in plastic surgery consultations is not currently practised in the United States, this study introduces a technology that could be feasibly integrated into practice for higher-quality care in the ED. One of the limitations of the study is the sample size. This study was conducted for a 1-month period of time due to the limited research personnel involved in data collection. Another limitation of this study is the potential for inherent differences between those residents that respond in person and those that respond by phone. This could create a bias in time to respond by these consultants. Furthermore, there was an unanticipated error rate in initiating therapy discussed over the phone. As a result, all residents are now required to perform in-person consultations to minimize treatment errors. This process, however, further lengthens consultation time. Based on this study, we see promise in implementing a similar telehealth design in real practice. In a survey of primary care physicians, 60% of respondents felt comfortable with telemedical consultations for management of problem wounds.8 Furthermore, 93% of respondents stated that they would continue to use telemedical consultation in their practice. Despite all of the positive feedback received, there is no consensus on the optimal venue for effective use of a telehealth resource in plastic surgery. In a review of the literature, while 96% of manuscripts reported benefits of access to specialist care and diminished costs, only 17% reported a benefit from the gold standard of face-to-face consultation in plastic surgery.9 This study reaffirms the value of face-to-face consultation where a specialist is able to both diagnose and offer the initial management for the patient. Our study reveals, however, that a telehealth service, like the one employed, may be the next best option over phone consultations. The iPad app provides a package of best practice educational and instructional materials that can be quickly reviewed by the ED provider on the iPad app and utilized to enhance their diagnostic and initial treatment capabilities.

Journal of Telemedicine and Telecare 23(2) Proper use of these educational materials would prevent technical errors in initial management, as seen in our study with splinting. This telehealth design has the promise of providing needed specialty input to rural areas, underserved populations, or regions where there is no immediate access to a specialist. In these situations, primary care physicians, ED physicians and related health providers could use telehealth responses to both assist them in managing their patients while providing them with worthwhile best practice educational materials. The app has demonstrated in this study to be a valuable QA/QI tool. As the QI study demonstrates, even in an academic training programme there is room for improvement in providing optimal care for patients with highly specialized medical conditions. In an academic setting, the use of the app could enhance resident training opportunities as well. For a resident, it is analogous to carrying a reference guide; however, the PSE educational responses are focused to the specific patient condition.

Conclusions Our study shows that telehealth consultations to a remote plastic surgeon based on store and forward digital images and a brief history were able to produce timely and accurate responses in an emergency care facility. The universal goal of current medical practice is to optimize the quality of patient care, while maximizing efficiency. Time and cost savings are critical for patient satisfaction and hospital financing. The use of a telehealth system in emergency care settings could dramatically decrease the overall time patients spend in the ED, yet accurately facilitate triage and management decisions. Furthermore, a telehealth system could be used in the training of ED providers and as an effective QA/QI tool. Acknowledgement This work will be presented at the 2014 American Telemedicine Association meeting in Baltimore, MD.

Declaration of Conflicting Interests The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AM Paik and S Scott declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. MS Granick is a founder and chairman of ReconSTAT, LLC.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

References 1. Knobloch K, Gohritz A and Vogt PM. Cell phone-based multimedia messaging service in reconstructive microsurgery: A novel telemedicine application. Plast Reconstruct Surg 2009; 123: 220e–222e.

Paik et al. 2. Pap SA, Lach E and Upton J. Telemedicine in plastic surgery: E-consult the attending surgeon. Plast Reconstruct Surg 2002; 110: 452–456. 3. Varkey P, Tan NC, Girotto R, et al. A picture speaks a thousand words: The use of digital photography and the Internet as a cost-effective tool in monitoring free flaps. Ann Plast Surg 2008; 60: 45–48. 4. Trovato MJ, Scholer AJ, Vallejo E, et al. eConsultation in plastic and reconstructive surgery. Eplasty 2011; 11: e48. 5. Diver AJ, Lewis H and Gordon DJ. Telemedicine and trauma referrals – a plastic surgery pilot project. Ulster Med J 2009; 78: 113–114. 6. Murphy RX Jr., Bain MA, Wasser TE, et al. The reliability of digital imaging in the remote assessment of wounds: Defining a standard. Ann Plast Surg 2006; 56: 431–436.

327 7. Wallace DL, Jones SM, Milroy C, et al. Telemedicine for acute plastic surgical trauma and burns. J Plast Reconstruct Aesthet Surg 2008; 61: 31–36. 8. Dobke MK, Bhavsar D and Herrera F. Do telemedicine wound care specialist consults meet the needs of the referring physician? A survey of primary care providers. Int J Telemed Appl 2011; 2011: 321376. 9. Gardiner S and Hartzell TL. Telemedicine and plastic surgery: A review of its applications, limitations and legal pitfalls. J Plast Reconstruct Aesthet Surg 2012; 65: e47–e53.