This study was performed at the Seoul National University Bundang Hospital (SNUBH), located in the Seoul metropolitan area of South Korea. SNUBH is a tertiary national university hospital founded in May 2003 as a fully digital hospital. It was accredited with a Stage 7 of Electronic Medical Record (EMR) adoption model from the Healthcare Information and Management Systems Society (HIMSS) Analytics on October 2010 [16]. The hospital has 910 beds and was visited by an average of approximately 4,100 outpatients daily during 2011.

SNUBH launched VDI system in November 2011 after the system was implemented using VMware 5.0 over approximately 4 months, beginning in July 2011. The system was implemented so that the use of all hospital information systems, including clinical and administrative applications such as EMR, Picture Archiving Communication System (PACS) and groupware, was possible on all devices, including desktops, iOS-based tablet PCs, Android-based tablet PCs, and laptops. The system was concurrently accessible by 400 users.

Figure 1 shows the overall system architecture of VDI, consisting of virtual private networks (VPN), virtualization network, VDI servers, management and access servers, and storage area network (SAN) switches. The system architecture was established to achieve high availability and stability of the VDI service through dual implementation of all servers such as a management server, an access server, an active directory (AD) server, switch servers, and VPN. And, VDI server was clustered so that the virtualization service can function even with a 20% hardware failure.

In addition, in order to implement a mobile hospital environment an iPad was provided to every medical professor so that they can access VDI at anytime and anywhere.

As a case study of using VDI for patient care, we surveyed the possible use of the system in physician's round. We examined the relationship between patient satisfaction with the point-of-care mobile consultation service, which used hospital information systems on a virtual desktop with an iPad, and 3 variables representing the influence of their experience on their likelihood of revisiting the hospital (labeled the revisit), on their likelihood of recommending the hospital to other people (labeled recommendation), and on whether they believed the service would help to improve the brand image of the hospital (labeled brand image) (Figure 2).

A survey was conducted on the satisfaction of 35 inpatients and their next-of-kin admitted to SNUBH during the 3-month period from December 2, 2011 to February 13, 2012 in which VDI system was used during physician rounds. Thirty-five participants were selected from each of 7 specialties (pediatrics neurology, pediatrics pulmonology, cardiology, neurology, otorhinolaryngology, rehabilitation medicine, and urology). Eligibility criteria for the survey were: age over 20 years, inpatients or their next-of-kin who can express their opinions without evidence of delirium, confusion or clouded consciousness and had radiology exams during their hospitalization period.

Seven physicians from 7 specialties participated in the survey.

This study was approved by SNUBH's institutional review board (IRB).

In the survey, physicians used a mobile environment in which VDI system was accessed through an iPad. The physicians accessed hospital information system while making their rounds through the wards and also used the iPad to directly show patients their progress and results. After these rounds, the physicians who had used the VDI were interviewed about the usefulness of its use, and the patients were also surveyed with a questionnaire.

The survey used in this study was composed by a researcher after a literature review, and it consisted of 9 questions in total [17-21]. Of these questions, 6 concerned patient satisfaction, and the remaining 3 questions concerned the revisit, recommendation, and brand image factors (see Appendix A for the questionnaire items in detail). The patient satisfaction consists of the following variables: sufficiency of explanation, ease to understand, trustworthy, impression, necessity, and satisfaction. By performing factor analysis on the six variables, a single patient satisfaction factor was extracted and measured on the average of the variables.

The survey results, measured on a 5-point Likert scale, were statistically analyzed using SAS ver. 9.2 (SAS Inc., Cary, NC, USA).

During the implementation of the mobile hospital environment using VDI, several concerns arose regarding the design of fault-tolerant system architecture, the privacy and security of patient healthcare information, a mobile device management, the speed and seamlessness of the internal wireless network, the support of screen resolution of legacy clinical applications on tablet PCs, and user education. The followings described the major issues that were derived and resolved in more detail.

Regarding the design of fault-tolerant system architecture, high system availability and stability should be considered to enable the organization-wide adoption of the VDI system. The clustering architecture and dual implementation of servers were considered in our system.

Regarding the privacy and security for the protection of patient healthcare information, secure communication methods and institutional privacy policies should be developed. Technically, the system security was reinforced by employing transmission protocols such as PC-over-Internet protocol (PCoIP) and AD based authorization. We classified end-user groups into private and public users and restricted unauthorized user access to the VDI system with AD-based authorization technique. Demands for the use of the VDI system outside the hospital were issued and supported by allowing access through the VPN and the firewall. The authority was granted only to healthcare faculty in our hospital.

In addition, security concerns created mobile device management challenges. For efficient management, it was necessary to adopt a mobile device management (MDM) system that gave IT staff the ability to view and control all mobile devices.

The speed and seamlessness of the internal wireless network were frequently cited as requirements by the end users who accessed the VDI system using their mobile devices. Thus, the internal wireless network infrastructure should be thoroughly tested before launching a VDI system. Additionally, as the use of the VDI system grew more frequent, it was necessary to expand the access circuit to maintain the network speed.

Regarding the support of screen resolution of legacy clinical applications on tablet PC, during our implementation process, the initial VDI client program did not support the 1,024 × 768 resolution of our clinical application on the iPad and Galaxy Tab. Verification problems were discovered and resolved before launching the system. Finally, education was necessary for users to become familiar with the user interface of VDI client application, especially when using the system with iPad or Galaxy Tab. Efficient education methods and materials are likely to be a key success factor in system adoption. We utilized online and offline educational methods such as video materials and offline courses.

Seventeen patients and 18 patients' next-of-kin participated in the survey. Twelve (34.3%) survey participants were men and 23 (65.7%) were women. Eleven (31.4%) participants were aged 30-39 years, 5 (14.3%) were 40-49 years old, 8 (22.9%) were 50-59 years old, and 11 (31.4%) were 60 years or older. The participants were from the departments of pediatrics, cardiology, neurology, otolaryngology, rehabilitation medicine, pulmonology, and urology. Table 1 shows the general demographic characters of the participants.

Six questions about sufficiency of explanation, easy to understand, trustworthy, necessity, and satisfaction to mobile consultation service were classified as a single averaged variable labeled Patient Satisfaction by factor analysis (Cronbach's alpha, 0.8448).

Table 2 shows the results of the survey, which displayed a "satisfied-level" degree of satisfaction, with average scores over 4 points across all categories. Analysis showed that patient satisfaction did not differ by gender, age, or specialty. A statistically significant difference in patient satisfaction was found between respondents of patients (mean, 4.2) and their next-of-kin (mean, 4.5; p = 0.0369).

Table 3 shows the results of a cross tabulation between patient satisfaction and the 3 influence variables. The results were statistically significant for all 3 items of influence at a level of significance of p < 0.05. The results revealed a significant positive relationship with the patients' decisions on revisiting the hospital and whether the use of VDI improved the brand image of the hospital (p < 0.05).

According to the interviews with the 7 physicians who provided the mobile consultation service, the physicians found it useful to be able to explain the patient's progress and results while presenting visual information and, in particular, to be able to instantly check the patient's charts and answer the patient's questions.

However, the degree of physician's satisfaction with the VDI system seemed to differ greatly depending on the user's proficiency with the VDI applications and with iPad operation. In fact, some physicians expressed difficulties in using VDI applications with an iPad, for examples, double-clicking (for 3 physicians), setting up the network (for 4 physicians), and using the automatic screen rotation/hold functions (for 3 physicians). Additionally, concerns arose about the insufficient battery life of iPad and potential extended time of consultation.