Training Digital Divide Seniors to use a Telehealth System: A Remote Training Approach

Abstract

As the use of health information technologies continues to proliferate amongst seniors, many of whom lack computer experience, there is a need to develop effective training approaches to foster basic competencies. This paper describes the REmote Patient Education in a Telemedicine Environment (REPETE) system, a component of the IDEATel telemedicine architecture. The REPETE architecture supports simultaneous visual and audio teaching modes over low bandwidth connections. This paper presents an in-depth qualitative analysis of two patients being trained to use the IDEATel patient web portal. The results indicate that this method of instruction was useful in facilitating patients’ use of the web application. However, the observations suggest that there is learning curve for the trainer to use the resources effectively to establish common ground and foster competencies in the patient.

Introduction

Chronic illness affects over 100 million individuals in the United States. Patient self-management is increasingly seen as an integral component of the healthcare process. The IDEATel project is a randomized controlled study of the efficacy of a home telemedicine system for diabetes care of the underserved rural and inner-city residents.¹ Central to this project is the home telemedicine unit (HTU), which provides the following functions: synchronous videoconferencing, electronic transmission of finger-stick glucose and blood pressure readings, secure messaging, web-based review of one’s clinical data, and access to web-based educational materials.

Despite best efforts to make the HTUs easy to use, field usability studies have shown that many older adults had difficulty in developing mastery of the system.² These studies documented that many of these patients experienced difficulty in using the mouse and accessing web applications. Certain patients required multiple training sessions in their homes. However, in geographically distributed telemedicine projects such as IDEATel, it is costly and time consuming to provide as much in-person training as is sometimes necessary. A potentially viable solution to this problem would be to develop a method to remotely educate patients to use their HTUs.

A common solution to the remote training problem is for the user to place a telephone call to a support center. A significant problem with telephone support is the difficulty in orienting users to specific objects on the computer screen. Phone communication lacks many of cues normally available in a face-to-face interaction. In addition, older adults frequently have minimal experience with computers and lack the appropriate vocabulary for speaking the language of graphical user interfaces (e.g., referring to widgets such as buttons, menus, and scrollbars.). This is compounded by problems associated with older adults’ lack of visual acuity and limits in their ability to selectively attend to relevant screen features. Providing additional visual cues to draw their attention may serve to offset some of these limitations.³

There is evidence to indicate that the use of a mutually visible pointer (i.e. a telepointer) enhances computer-mediated instruction in distance learning.⁴ On this basis, we thought a good strategy might be one in which the trainer is able to visually monitor the trainee’s progress as well as have the ability to take over a user’s session. This form of remote training has benefits of co-reference, the ability to orient attention, as well as provide cues normally available in a face-to-face session such as gestures in a remote support situation that would not otherwise be possible.

We developed the REmote Patient Education in a Telemedicine Environment (REPETE) architecture to provide a timely and effective training method. Similar forms of this type of remote assistance have been used for technical support with computer savvy individuals in large corporations over local area networks, and are usually used in conjunction with or simultaneously with telephone support. This interaction has not been used in the telemedicine environment or with a novice computer population. In addition, few formal evaluations of this type of remote training have been performed.

Architecture

The REPETE architecture (as shown in Figure 1) leverages the existing H.323 VOIP audio chat infrastructure used by many home telehealth devices. In addition to using the H.323 capabilities, one can support remote training by adding remote control through the use of a remote control protocol (RCP). This architecture allows for simultaneous voice conferencing and remote control over a single telephone line. Little research has been done on the use of RCPs simultaneously with H.323 audio chat over narrow-band connections.

Figure 1.

REPETE Architecture

In this architecture, the trainer and the patient can do voice communication over the VOIP audio chat infrastructure. In addition to voice, they have a shared workspace of the HTU’s screen. The trainer and the patient have shared control over the workspace. A major component of this interaction is the mouse pointer. The trainer can use his/her control over the mouse pointer as a telepointer to make gestures at items on the screen. The architecture has been implemented on all second-generation IDEATel telemedicine units and installed in over 400 patient homes.

Methods

The testbed for this study consisted of a provider station and an HTU equipped with the REPETE software. The provider station and HTU were located in different rooms and connected by a telephone line. We evaluated the architecture by attempting to train IDEATel patients to perform the various tasks supported by the IDEATel data review website, myIDEATel, over a telephone line.

We developed a training benchmark based on a cognitive task analysis.² A cognitive task analysis (CTA) for using myIDEATel, on the HTU was performed. The tasks, goals, and actions generated from the CTA were used as a list of steps that a subject would need to perform. This list of steps was used as the optimal approach to completing tasks and is used as the basis of comparison to steps the patient performed. The web skills competency instrument includes five tasks: 1) logging into the myIDEATel website; 2) reviewing monitoring data; 3) entering pedometer data; 4) sending messages to a provider; and 5) reviewing messages from a provider. These tasks comprise the core activities an IDEATel patient would perform when interacting with the myIDEATel website. In order to limit subject fatigue, training sessions were limited to 1 hour. Therefore, we focused on tasks 2–4. In addition to this training benchmark, we interviewed the patients concerning their experience with training using the REPETE architecture.

The subjects were two IDEATel patients who were familiar with the IDEATel HTU. The training sessions coincided with the introduction of a new website design, which provided several new functions. The subjects had not received prior training on the new website. Subjects were assessed pre- and post-training using the web skills competency instrument. All interactions of the patient using the HTU were videotaped. The study was approved by the Columbia IRB.

First, the patients were asked to perform tasks 2–4 Then, they were trained using the remote training method in a laboratory setting. A picture of a user interacting with the HTU can be seen in Figure 2. The training mirrored that of a face-to-face one-on-one training session. The patients were trained using a scaffolded instruction approach that adjusted to their level of skill and autonomy. The patients were first asked to perform each task with minimal assistance. If the patients were unable to perform the requisite sequence of actions, the trainer provided additional assistance by guiding and prompting them through the myIDEATel system. After training, the patients were asked to repeat tasks 2–4 without assistance. Lastly, they were asked questions concerning user satisfaction, user confidence, and their opinion of the training methodology in order to evaluate the acceptability of the remote training methodology. We reviewed the videotapes and analyzed how well the user completed each task. The pre- and post-training evaluations of their web skills competency were compared as a measure of the efficacy of remote training.

Figure 2.

Patient using the calendar

In addition to the outcome measures pertaining to change in task performance, we also completed a detailed analysis of the training process using goal-action coding.² This analysis captures user’s intentions (goals), actions and the system responses. The coding provides a basis for characterizing progress as well as for diagnosing user problems. It also enables us to understand how the training mediates changes in performance.

Results

Prior to training, neither of the subjects was able to complete any of the tasks. After training, subject 1 completed all tasks and subject 2 completed one task without assistance. Subject 2 completed two other tasks with minimal assistance. Detailed analysis of the videotapes revealed a number of other important issues with this type of remote training.

Subject 1 was a 70-year-old man with a college education and basic skills in computing, but had not been using the HTU for the past several months. The patient had a good grasp of related concepts (and spoke the language of computers) and this greatly facilitated the interaction.

The first task was to find previous glucose and blood pressure results in a particular date range. After some prompting from the trainer, the subject was able to navigate to the correct web page to select a date range. However, after he was presented with on-screen calendars, the subject was able to correctly choose months, but neglected to choose specific dates, resulting in an error. The trainer needed to explain to the subject that specific dates were required. After being shown how to choose a valid date range, the subject was then asked to repeat the task using a different date range. He was then able to successfully complete the task. A goal-action coding of the patient’s actions are shown in Figure 3. While the patient was unsuccessful at mastering the task after receiving training, it was enough training to enable him to use the system on his own. In the pre-test, the patient was unable to even begin the task. With training, the goal is not necessarily to produce people who are competent at using the system, but those who are able to take the skills learned and apply them to learn on their own.

Figure 3.

Goal-action coding of patient trying to view blood pressure readings in a date range.

Subject 1’s experiences with the messaging task is also informative (Figure 4). Prior to training, the user was unfamiliar with this task and was unable to compose a message. After training, the subject was able to compose a message to his provider without mistakes.

Figure 4.

In this excerpt, the trainer is teaching the patient how to read a message from his provider.

The interaction illustrates various aspects of the remote training scenario. For example, when the trainer was attempting to demonstrate the messaging component, a secure email system to communicate with healthcare providers, of myIDEATel, both the trainer and patient attempted to control the system simultaneously, producing a temporary clash of control. The turn taking was then negotiated through dialog between the trainer and patient. In addition, when the trainer was describing the red flag denoting a new message, but neglected to use the telepointer to gesture and to provide additional visual cues to the patient. This remote training method requires the trainer to learn a number of techniques in order for the interaction to be optimal. At the end of the task, the trainer and patient had another tug-of-war over control of the system.

Subject 2 was a 76-year-old woman, with a high school education. Like the first subject, she was also familiar with the use of the computer and the jargon concerning GUI widget elements. She also had good mastery of the mouse and keyboard. Despite her familiarity with the use of the computer, she was unable to use the pedometer module without an initial training. However, after receiving training, she was able to correctly enter the number of steps from her pedometer. For her to be able to do this correctly, she required some repetitive training, but was ultimately successful in entering of a pedometer measurement without mistakes. A goal-action coding of the patient’s actions in her two attempts at entering pedometer results is shown in Figure 5.

Figure 5.

Goal-action coding of patient trying to enter a pedometer reading into myIDEATel.

We asked the patient to enter a pedometer result into the system and to choose a date. During attempt #1, after the pedometer screen comes up, she says “I’ll put in 16,” stating that she’s interested in selecting the date shown in bold. However, the date in bold was a holiday. She incorrectly assumed that bold meant that it was the date selected for the pedometer result. The actual system response to selection of a date is a red box around the number. After she submitted the result, the trainer and patient realized that the patient had not performed the action correctly, and the trainer asked her to repeat the exercise. During the second attempt, the trainer asked the patient to “make sure it comes up on top correctly” referring to the text box above denoting the date selected in textual form. This time, the patient again chose the 16th, and after she touched the touch screen, a red box appeared around the number. She then said, “I see… a red square,” indicating that she now notices that was the system response to look for when selecting a date. Because the trainer was not face to face with the patient, the trainer was unaware that the patient had merely pointed at the date and had not actually tapped on the screen in the first attempt.

In the follow up interviews, both patients felt that the training was appropriate for their skill level. In addition, both patients felt that the training helped them learn and that they would use the skills they had learned. The first patient stated that he felt that the training was worthwhile and that he was leaving “a little smarter” than when he came. The second patient expressed the thought that “it would be fantastic if more training was available” using this architecture.

Discussion

This study provides a proof-of-concept for the REPETE architecture. Patients not only reported that the training was beneficial, but also showed measurable changes in skills. There is negligible literature on remote training over narrowband connections. Surprisingly, few, if any, computer trainings studies with novice users have assessed changes in task performance. Most studies have focused only on changes in user attitudes.

To our knowledge, this is the first documented integration of H.323 and remote control protocols for remote training and also the first use of using remote control training in conjunction with a touch screen computer. This combination provides a set of unique challenges to this paradigm. In a mouse-based system, the trainer sees the cursor move before the trainee selects a target. However, when the trainee is reaching toward a touch screen, no such feedback is given to the trainer; the only feedback is given at the time the screen is touched. This requires the trainer to be cognizant of these issues and to continually ask for status verbally. For this reason, the parallel audio channel in the REPETE architecture becomes even more critical.

The results also demonstrate that in-person training approaches must be modified for the remote training domain. One example was pointer “tug-of-war.” We observed situations in which both the trainer and the patient were trying to control the pointer at the same time. This tug-of-war was the source of some frustration and confusion. In addition, the trainer frequently uses the pointer to gesture to different areas of the screen. With experience, both trainer and trainee soon learned to recognize these situations. Technology solutions, such as visually indicating who had cursor control might also prove beneficial.

The study has some obvious limitations. Data collection is ongoing; only two patients have been evaluated to date. However, as proof-of-concept, even this small number is adequate to demonstrate the functionality of the architecture. Although the two subjects presented in this paper were relatively more literate and experienced computer users compared to the overall IDEATel population, they were significantly less computer literate than subjects in many computer training studies. Therefore, we believe the results are likely to be generalizable to the typical elderly population.

In summary, this study demonstrates a new architecture for remote technology training over narrowband connections. The architecture has been successfully demonstrated with a small number of elderly subjects.