This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on http://www.researchprotocols.org, as well as this copyright and license information must be included.
Innovative interventions that empower patients in diabetes self-management (DSM) are needed to provide accessible, sustainable, cost-effective patient education and support that surpass current noninteractive interventions. Skills acquired in digital virtual environments (VEs) affect behaviors in the physical world. Some VEs are programmed as real-time three-dimensional representations of various settings via the Internet. For this research, a theoretically grounded VE that facilitates DSM was developed and pilot tested. It offered weekly synchronous DSM education classes, group meetings, and social networking in a community in which participants practiced real world skills such as grocery shopping, exercising, and dining out, allowing for interactive knowledge application. The VE was available 24/7 on the Internet, minimizing access barriers.
The objective of this study was to evaluate the feasibility and efficacy of participation in a VE for DSM education and support.
This study utilized a single group, pre-mid-post measure design. At 0, 3, and 6 months, we assessed participants’ perceived VE usability and usefulness, self-efficacy, diabetes self-management behaviors, perceived social support, and diabetes knowledge using validated survey measures; and we recorded metabolic indicators (HbA1c, BP, BMI). Process data were continuously collected in the VE (log-ins, voice recordings, locations visited, objects interacted with, and movement). Data analysis included descriptive statistics,
We enrolled 20 participants (13/20, 65% white, 7/20, 35% black), with an age range of 39-72 years (mean age, 54 years) and diabetes duration from 3 months to 25 years. At baseline, 95% (18/19) and 79% (15/19) of participants rated usefulness and ease of use as high on validated surveys with no significant changes at 3 or 6 months. Participants logged into the site a mean of 2.5 hours/week over the course of 6 months. High DSM class attendance was reflected by the largest percentage of time spent in the classroom (48.6%). Self-efficacy, social support, and foot care showed significant improvement (
Because relatively little is known about usability, acceptability, and efficacy of health interventions in VEs, this study constitutes an important, innovative first step in exploring the potential of VEs for facilitating DSM. The preliminary data suggest that VEs provide a feasible and useful platform for patients and educators that affects self-management and related mediators. Flexible access to both synchronous and asynchronous diabetes education, skill building activities, and support from a home computer remove barriers to attending clinic-based meetings. This program has potential for improving DSM in an easily disseminated alternative model.
“The medium is the message,” a phrase coined over 40 years ago, signifies not only that the content of the message, but also the characteristics of the medium itself affect perception of the message [
Diabetes affects 23.6 million US adults, most of whom have type 2 diabetes (T2D) [
Currently, patients and providers with T2D face barriers to accessing even minimal self-management support or short-term diabetes self-management training (DSMT), including a lack of referral sources for DSMT, fear of losing patients to DSMT if offered outside their clinic, a belief that DSMT was not needed, inability to fit it into their schedule, cost, and transportation [
Digital VEs have the potential to solve the contextual disconnect between the clinic or other education-center based educational programs and the challenges of patients’ daily lives [
For this project, a Second Life [
Social cognitive theory [
This feasibility study involved a one-arm, pre-mid-post measure design. A convenience sample of adult participants with T2D was enrolled in the VE intervention with access to the site for 6 months. Pre-intervention measures at baseline, mid-intervention measures at 3 months, and post-intervention measures at 6 months provided data addressing the primary and secondary aims of the study.
Approval from the Duke University Institutional Review Board was obtained prior to study initiation. Endocrinologists and the study coordinator recruited a convenience sample within the endocrinology clinic at the Duke University Medical Center. Potential participants were identified by endocrinology providers either during scheduled visits or through medical record review. If recipients were interested in learning more or participating, letters were sent to them introducing the study and giving them the study coordinators’ contact information. Eligible participants were those with a diagnosis of T2D who (1) were between 21 and 75 years old, (2) were able to speak and read English, (3) were computer literate (have used a computer for at least 6 months), (4) understood how to use the Internet (have accessed the Internet on at least 6 occasions), (5) had access to a computer with a non-dial-up Internet connection in a private location, (6) were mentally capable of informed consent, (7) were reachable by telephone, (8) had no comorbidities or severe diabetes-related complications that would interfere with study participation or measures (eg, renal failure, stage III hypertension, severe orthopedic conditions or joint replacement scheduled within 6 months, paralysis, bleeding disorders, or cancer), and (9) were able to travel to the clinic for follow-up appointments. Subjects at various stages of T2D with treatment regimens including both oral and injectable medications were included if the above criteria were met.
The SLIDES community was developed using an iterative usability design, informed by the programmer, researchers, study team, and patients with diabetes [
The SLIDES site was designed to provide DSMT and support based on social cognitive theory, effective diabetes interventions, and American Diabetes Association and American Association of Diabetes Educator standards for diabetes care and education [
Weekly 1-hour classes (using the American Diabetes Association/American Association of Diabetes Educators self-management training curriculum facilitated by nurse practitioners, diabetes educator or health professional guest presenters) were held [
Each participant had access to the SLIDES community site in Second Life 24 hours a day, 7 days per week. Access to SLIDES allowed participants to utilize resources and weblinks such as those in the bookstore or grocery store and to interact with other participants at their convenience. Participants were advised to consult their health care provider regarding any medication regimen changes or side effects, symptoms, or health status changes. Medical management remained outside the domain of this intervention.
Class session in SLIDES.
Demographic data included age, marital status, ethnicity, race, cohabitation, family history of diabetes, level of education, employment status, and household income range. We also assessed duration of diabetes, diabetes-related medications, attendance at prior diabetes education, and prior attendance at a diabetes support group. Finally, we assessed Internet use in general by hours per week, and then specifically hours per week using social networking sites, emailing, playing games, searching the Internet, and reading newspapers or magazines.
To meet our primary aim of developing a virtual diabetes community (SLIDES) and assessing its feasibility and acceptability, we tracked participation rates (number of log-ins, time spent in SLIDES) and use (locations visited in SLIDES, objects manipulated, voice and text tracking), perceived usefulness, perceived ease of use, and attitudes toward use.
Process data were measured via the number of log-ins into SLIDES, duration of each log-in, locations visited in the site, and objects touched such as books, food items, etc. Additionally, all verbal and text communication during the duration of the study were recorded continuously (24/7). Process data were analyzed via average number of log-ins per month, average amount of time spent in the site, average amount of time spent in each location within the site, average number of objects touched, and a list of objects touched. Data points such as participant avatar location were collected every 15 seconds, or any time an object was “touched,” to ensure that object interactions were captured in the database.
Perceived usefulness was assessed using a verbal 7-point Likert scale based on work by Davis [
Perceived ease of use was assessed based on work by Davis [
The purpose of the focus group was program evaluation. Questions asked during the focus group included: What are your perceptions of the SLIDES site? What did you get out of the SLIDES site? What were positive and negative aspects of the SLIDES program? What changes would you make to the SLIDES program?
To meet our second aim of determining the preliminary effects of participation in the SLIDES intervention, we collected data on metabolic outcomes and potential psychosocial mediating variables as outlined below.
Metabolic control was measured by glycosylated hemoglobin (HbA1c), blood pressure (BP), and body mass index (BMI) obtained from medical records. HbA1c is a routine laboratory measure of metabolic control in clinical practice, obtained at 3-month intervals by diabetes care standards [
Self-efficacy was assessed using the Diabetes Empowerment Scale-Short Form (DES-SF), an 8-item Likert scale that ranged from 1 (strongly disagree) to 5 (strongly agree), measuring participants’ confidence in their ability to perform DSM behaviors and self-assess satisfaction with diabetes care. The DES-SF was created by choosing the item from the original 28 items with highest item to subscale correlation from each of the original eight conceptual domains (Cronbach alpha, 0.85) [
Diabetes knowledge was measured using a subset of the Assessment of Diabetes Knowledge [
The 11-item Summary of Diabetes Self-Care Activities [
Perceived support for diabetes management was assessed using a 12-item survey with Likert scale items ranging from 1 (strongly disagree) to 7 (strongly agree). This diabetes support scale was developed specifically for assessment of social support in a diabetes Internet intervention. It has demonstrated internal consistency reliability of 0.90-0.93, sensitivity to intervention effects, and construct validity when compared to the Interpersonal Support Evaluation List and Chronic Illness Support Survey (
Presence (sense of actually being in the site) evaluation was based on the work of Witmer and Singer [
Copresence (sense that others are actually present in the environment) [
During the baseline participant visit in a private room at our study office, the study coordinator obtained signed informed consent and administered surveys to assess demographic factors and psychosocial mediators as described above. The most recent height and weight, BMI, BP, and HbA1c were obtained from the participant’s medical record by the study coordinator. During the baseline visit, subjects were oriented to the SLIDES site in Second Life and each subject’s avatar was developed by the study staff with the participant’s input and approval. Although participants could keep their individualized avatar indefinitely, they could only access the SLIDES site during the study period. Participants were given headphones with a microphone for home use to allow for synchronous voice communication. Online tutorials that were developed in Second Life helped to facilitate participant learning and integration into the VE. Written instructions for accessing the site were given to each participant along with their username and password during the initial, baseline visit. The initial baseline orientation with the study coordinator was approximately 60-90 minutes long. Most participants required at least 1 follow-up phone conversation with the coordinator to ensure that the application was set up and that they were able to use the basic functions on their home computer, which took about 30 minutes. Although periodic technical challenges required follow-up support, most were either related to Second Life server issues beyond our control or limitations of participant computers (processing capabilities, etc).
Individuals were told that they could visit the SLIDES site as frequently as they wanted, but were instructed to sign in to the SLIDES intervention at least twice a week for the first 4 weeks, after which they could log in to the SLIDES site at any time for the remaining 5 months. The rationale for the initial required log-ins were to encourage use and increase familiarity with the site, resources in the site, and to meet others and attend weekly diabetes education meetings. This was only “required” and participants received reminders from the study coordinator to log in if they did not meet this requirement for the first month so that their use of the site after that point could be observed for the remainder of the study as a part of feasibility testing. Three days after the baseline appointment, participants were telephoned to address any questions about Second Life or study procedures. A re-orientation was provided to those who continued to have difficulty with how to use the site, which was necessary for approximately 4 participants.
Weight, BP, and HbA1c at 3 and 6 months were obtained from medical records (with a 2-week window before or after the study follow-up time point). Subjects were sent a link via email to the follow-up surveys (as outlined above), which were administered online at 3 and 6 months through REDCap [
All data, including the process data recorded in Second Life, were stored on a secure research server at the Duke University Office of Instructional Technology. Statistical data analysis was conducted in SAS software version 9.3 [
Addressing our secondary aim, the preliminary effects of our VE program on measures of metabolic control (HbA1c levels, BP, and BMI) and potential psychosocial mediating variables (perceived support, self-efficacy, diabetes knowledge, diabetes self-management behaviors, presence, and copresence) were assessed. Means and standard deviations were used to describe these parameters at baseline. Paired
Of the 42 patients contacted, 20 (48%) agreed to participate in this feasibility study.
Type and amount of Internet use.
Sample characteristics of participants in SLIDES (N=20).
Attribute | n (%) | |
|
|
|
|
Female | 19 (95) |
|
|
1 (5) |
|
|
|
|
<45 | 3 (15) |
|
45-54 | 7 (35) |
|
55-64 | 6 (30) |
|
65-74 | 3 (15) |
|
Missing | 1 (5) |
|
|
|
|
White | 13 (65) |
|
African American/black | 7 (35) |
|
|
|
|
Single | 7 (35) |
|
Married | 11 (55) |
|
Divorced | 2 (10) |
|
|
|
|
Spouse/Partner | 11 (55) |
|
Children | 4 (20) |
|
Other relatives | 2 (10) |
|
Other | 2 (10) |
|
None of the above | 5 (25) |
|
|
|
|
Technical/trade school | 2 (10) |
|
Some college | 2 (10) |
|
Associates degree | 3 (15) |
|
Bachelor’s degree | 4 (20) |
|
Master’s degree | 9 (45) |
|
|
|
|
Full-time | 7 (35) |
|
Part-time | 4 (20) |
|
Retired | 4 (20) |
|
Not employed | 5 (25) |
|
|
|
|
25,000-34,999 | 4 (20) |
|
35,000-49,999 | 2 (10) |
|
≥50,000 | 14 (70) |
|
|
|
|
Mother | 9 (45) |
|
Father | 12 (60) |
|
Sister(s) | 5 (25) |
|
Brother(s) | 6 (30) |
|
Children | 2 (10) |
|
|
|
|
On oral medications | 14 (70) |
|
On insulin | 10 (50) |
|
On other medications | 2 (10) |
|
14 (70) |
Participants logged into SLIDES a total of 766 times with a mean of 38 times, a median of 35 times, and a range of 1-113 times per participant over the 6 months they were in the study. Seventy-five percent of the log-ins occurred in the first 3 months with the majority in the first month (
The participants in this study visited every location in the VE. However, they spent the majority of their time in the classroom (48.6%) because DSMT classes were held there twice weekly. The second most frequently visited area was the “outdoors” area on the site. As shown in the heat map (
A total of 297 of 394 (75%) objects such as food items, books, menus, websites, videos, and pharmacy items were “handled” by 19 of the participants while in SLIDES. Participants interacted with these 297 objects a total of 1180 times. Participant interaction with specific types of objects within these locations is delineated in
Number of log-ins into the SLIDES site.
Heat map of locations visited.
Overall, participants anticipated SLIDES to be useful at baseline (mean 2.14 SD 0.73) and this continued at 3 months (mean 2.02 SD 1.21) and at 6 months toward extremely likely (mean 1.68, SD 0.79). Although there was a trend toward extremely useful, there was not a statistically significant improvement (
Summary of all object interactions by participants.
Location of objects | Object category (items in store/items examined) | Object subcategory (number of interactions, %) | No. of participants who interacted with objects in location | Total no. of times object was interacted with by all participants | Number of interactions with objects per participant, mean (SD) |
Grocery store | All food items (146/211) |
|
19 |
408 | 22.66 (29.79) |
|
|
Meat, poultry, fish, nuts, beans (90, 22%) |
|
|
|
|
|
Beverages (82, 20%)a |
|
|
|
|
|
Fats, oils, sweets (61, 15%)a |
|
|
|
|
|
Cereal/breads (33, 8%)a |
|
|
|
|
|
Dairy (33, 8%)a |
|
|
|
|
|
Vegetables (33, 8%)a |
|
|
|
|
|
Frozen foods (33, 8%)a |
|
|
|
|
|
Snack foods (20, 5%)a |
|
|
|
|
|
Fruit (16, 4%)a |
|
|
|
|
|
Rice/pasta (8, 2%)a |
|
|
|
Bookstore | Books (20/22) |
|
15 |
63 |
4.2 (3.74) |
|
|
Nutrition (56, 89%)a |
|
|
|
|
|
Diabetes management (4, 6%)a |
|
|
|
|
|
Medication/treatment (3, 5%)a |
|
|
|
|
Websites (self-care) (11/18) |
|
15 | 105 | 7.0 (6.15) |
Restaurant | Menus (56/64) |
|
16 | 192 | 12 (18.46) |
|
|
Chain eat-in restaurants (106, 55%)a |
|
|
|
|
|
Chain drive-through restaurants (86, 45%)a |
|
|
|
Pharmacy | All items (49/63) |
|
15 | 104 | 6.93 (5.67) |
|
|
Blood glucose monitors (42, 40%)a |
|
|
|
|
|
Dental, foot, skin care (25, 24%)a |
|
|
|
|
|
Injection supplies (15, 14%)a |
|
|
|
|
|
Mobility products, scales, diabetes specialty supplements (23, 22%)a |
|
|
|
|
Medications (Rx) (10/11)a |
|
|
|
|
|
|
Lipid lowering (21, 20%) |
|
|
|
|
|
Oral antihyperglycemic (66, 63%) |
|
|
|
|
|
Insulin (18, 17%) |
|
|
|
Gym | Exercise videos (cardio, yoga, strength training) (3/3) |
|
17 | 134 | 7.88 (9.31) |
Community center | Recorded classes (11/12) |
|
16 |
73 |
6.63 (4.05) |
|
|
Intro to diabetes (15, 21%)a |
|
|
|
|
|
Healthy eating (10, 14%)a |
|
|
|
|
|
Exercise (10, 14%)a |
|
|
|
|
Forum |
|
12 | 101 | 8.41 (10.1) |
aItems reviewed most often.
Eight of the participants attended the focus group. Six of these focus group participants were active users of the SLIDES VE and 2 were much less active participants. Four categories were discussed based on the focus group questions: expectations of the program, positive aspects of SLIDES, negative aspects of SLIDES, and suggestions for changes. The themes noted in the data coding were consistent with the theoretical framework (environmental factors, personal factors, and behavior) for the study approach, and are summarized in
The focus group confirmed existing problems that we were aware of with the site and offered new insights for improvement of the site in future versions. With regard to the user expectations, all users thought that site provided resources that improved their knowledge of diabetes. The majority of the comments regarding the positive aspects of the site revolved around the informational resources and social interaction. The participants thought that the site provided a source of new information not only in the various locations in the site such as the grocery store, restaurant, and bookstore, but they also liked the interactive nature and informational aspects of the weekly classes. One participant stated that she learned something new every week. The majority of the positive comments on social interaction addressed their ability to interact with others synchronously and hearing about the experiences of others and learning through other participants “stories” about their disease. The majority of the comments about the negative aspects of the site were categorized as usability problems. For example, participants found the background noises made by other participants chewing food or just general home sounds, such as people talking or dogs barking, very annoying. Because the majority of conversations were synchronous, if the users did not mute their microphones, all of the sounds within their respective homes were heard by all participants and thus distracting. This required the moderators of social events such as the classes to ask participants to mute their individual microphones until they were ready to speak. Other problems identified by the participants had to do with the complex functionality of Second Life. These problems included navigational issues such as moving their avatar from one location to another, having difficulty with sound, and general computer problems that did not meet the specifications required by Second Life. Additionally, 1 participant who thought that there would be more participants in the site was referring to an age diversity issue. She was hoping for more people in her age group. Finally, the suggestions for future changes provided us with ideas on how to improve the site in the future. The majority of these suggestions were about how to improve social interaction such as including tasks to complete together as group such as homework, scavenger hunts, and in general group bonding activities. The participants thought that these types of activities would have helped them to bond earlier during their time in the site.
Focus group results.
Categories | Themes (n, %) | Examples |
|
|
|
|
Informational resources (3, 30%) | “The educational material was a plus too” |
Met expectations (2, 20%) | “Met expectations in terms of discussion” | |
Diabetes knowledge (2, 20%) | “Learn a few things” | |
No expectations (2, 20%) | “I had no expectations and was delighted” | |
Social interactions (1, 10%) | “Intrigued by being with other people with diabetes” | |
|
|
|
|
Informational resources (17, 31%) | "Liked the comments on the items in the grocery store" |
|
Social interaction (16, 29%) | “I did enjoy interacting with others” |
|
Diabetes knowledge (6, 11%) | “have had diabetes for 25 years, but learning new things” |
|
Usability (6, 11%) | “I am a click and find person—like things immediately at my disposal” |
|
Community resources (5, 9%) | “Liked the gym, helped me to exercise” |
|
Physical realism (3, 5%) | “Liked the seasons changing in the site” |
|
Self-management behavior change (2, 4%) | “Literally changed my life in terms of treatment with insulin—rarely now takes insulin during the day” |
|
|
|
|
Usability (26, 76%) | “background noises from others – home sounds, chewing” |
|
Social interactions (3, 9%) | “Expected larger group of participants” |
|
Informational resources (3, 9%) | “Need clarity on nutrition information by serving” |
|
Behavioral realism (2, 6%) | “Avoided gym ‘just like in real life’” |
|
|
|
|
Social interaction (13, 42%) | “Would be good to have group exercises” |
|
Informational resources (9, 29%) | “More variety in the grocery store” |
|
Usability (5, 16%) | “Would like to hear (bots)and read feedback” |
|
Diabetes knowledge (3, 10%) | “Would like nutritionist or other specialists at classes (podiatry, pharmacist) |
|
Community resources (1, 3%) | “Would like a walking path” |
There were no changes noted from baseline to 6 months in systolic or diastolic blood pressure; however, overall the participants had good blood pressure control (
Physiological and psychosocial outcomes.
Variable |
|
Baseline (N=18) |
3 Months (N=14) |
Change at 3 months | 6 Months (N=13) |
Change at 6 months |
HbA1c (%) | 7.51 (1.15) | 7.14 (1.24) |
|
6.92 (1.37) |
|
|
Weight (lb) | 217.5 (45.2) | 215.7 (45.8) |
|
208.4 (43.9) |
|
|
BMI (kg/m2) | 37.4 (7.9) | 37.2 (8.3) |
|
36.2 (8.5) |
|
|
Systolic blood pressure (mm Hg) | 131.3 (13.0) | 129.6 (14.5) |
|
130.1 (10.5) |
|
|
Diastolic blood pressure (mm Hg) | 74.8 (10.8) | 74.7 (11.2) |
|
78.1 (9.4) |
|
|
Self-efficacy (score scale 1-5) | 3.89 (0.81) | 4.45 (0.67) |
|
4.64 (0.39) |
|
|
Social support (score scale 1-7) | 4.61 (1.25) | 5.45 (1.07) |
|
6.35 (0.44) |
|
|
Diabetes knowledge (% score) | 89.1 (4.04) | 93.9 (5.25) |
|
88.2 (15.0) |
|
|
|
|
|
|
|||
|
Dietary | 4.13 (1.42) | 4.5 (1.67) |
|
4.75 (1.45) |
|
|
Exercise | 3.07 (2.03) | 2.43 (1.74) |
|
2.79 (2.26) |
|
|
Blood sugar testing | 5.15 (2.04) |
4.79 (2.08) |
|
4.83 (2.28) |
|
|
Foot care | 3.68 (2.08) | 4.61 (2.19) |
|
6.17 (1.54) |
|
aSignificant difference at
Participant perception of presence [
Copresence (the sense of being with other virtual humans) changed slightly from 3 months to 6 months. At 3 months, the mean score on a 7-point Likert scale for copresence was (slightly agree) 5.25 (SD 1.42) and at 6 months, the score was not significantly different (toward agree) at 5.78 (SD 1.35). Although a higher score indicates a greater level of copresence, and the participants average score did increase from 3 months to 6 months, there was not a statistically significance change. Embarrassment and likability are two other constructs to measure social response to others in a VE [
This study examined the feasibility and acceptability of an eHealth program utilizing a digital VE platform that provided diabetes education and self-management support. In addition, we examined physiological outcomes (HbA1c levels, BP, BMI) and psychosocial outcomes or mediators (diabetes knowledge, self-management behaviors, self-efficacy, perceived support, presence, and copresence) to determine preliminary effects of participation in the VE intervention. To our knowledge, this is the first study to explore the feasibility of VEs as a medium to improve T2D self-management in patients. This study showed that not only was the VE easy to use even for our elderly participants (age >65), but the participants overall found the environment and the synchronous interaction with peers and educators useful.
Participants were very engaged in the VE particularly within the first 3 months or the total time of one diabetes class series (12 weeks). Once they completed the class series, they continued to come into the site, but not as frequently. There were two possible reasons for this: (1) the class series was completed, and (2) they had reviewed all informational resources on the site. A subgroup of 6 participants who had formed a supportive social group continued to return to the site to attend weekly group sessions. This ongoing group demonstrated that the classes and resources were important, but social support was very significant in their disease management, as we know from the literature to date [
Participants spent the majority of their time in the classroom (48.6%), followed by the outside areas of the VE (16.8%), and the Social Center (9.0%). It was expected that the majority of participant time would be spent in the classroom because that is where they attended class once or twice a week. However, it was surprising to find that a significant amount of time was spent outside. The time spent moving from one location in the VE to another (see path around site on heat map
Participants also spent a significant amount of time interacting with objects. The top objects interacted with were food items in the grocery store. This is not surprising as people with diabetes find dietary issues and guidelines to be the most challenging [
Statistically significant improvements were found in 3 behavioral and psychosocial outcomes at 6 months: social support, self-efficacy, and foot care. This was not expected given the small pilot sample size. However, we interpret these with caution given the sample size and heterogeneity in terms of baseline demographics and metabolic control. Although we did not find statistical significance in the physiological indicators, we did find some weight loss, and associated decrease in BMI. The mean weight loss of 9.1 lb is clinically relevant and we will need to explore case-based analysis further regarding improvements among subgroups by participation and behavioral outcomes characteristics. Interestingly, this weight loss occurred in the absence of significant dietary and physical activity changes, which could also indicate inaccuracy or variation in weight measurement in the clinic settings and needs to be interpreted with caution. Attaining higher quality and consistently collected measures are being studied in our current larger trial. HbA1c decreased from baseline to 6 months post-intervention, reaching a level of less than 7% at 6 months, although this is interpreted in the context that metabolic control was relatively good at baseline. It is thought that interactive health interventions exert their effects by a combination of enhanced self-efficacy and knowledge, enabling patients to change their health behaviors, leading in turn to changes in clinical outcomes [
Our findings are limited by the small pilot study sample and lack of comparison or control group, and will need to be tested in our future larger randomized controlled trial. This study primarily focused on feasibility and usability of VEs in the context of DSM and support. Although the study sample was diverse in terms of age and number of years with diabetes, it was primarily well-educated women with at least moderate income levels. We need to demonstrate the efficacy in a larger, more diverse sample. Broadening the locations and demographics of targeted participant recruitment in our larger trial, and ensuring that a VE is developed that appeals across age, race/ethnicity, and gender groups is a primary focus. We will also need to address the issue of enrolling a sample of activated participants who start with fairly good metabolic control in our future efficacy studies, and determine the ability to engage and improve outcomes among those with poor metabolic control who are less likely to participate in DSME in general. In terms of results, the reliance on clinical measures (weight, BP, HbA1c) from medical records due to the financial constraints of a pilot study resulted in some missing data and possible inconsistency in collection of these measures within the clinic setting. Therefore, the clinical outcomes must be interpreted with caution. Finally, participation within the site dropped after 12 weeks. We believe this was problematic because we did not keep the site content dynamic [
e-Health applications such as SLIDES enable patients with chronic diseases such as diabetes to become more engaged with the self-management of their disease. Evidence shows that people who use e-Health resources have better social support for their disease, increased knowledge, and gains in self-efficacy [
body mass index
blood pressure
diabetes empowerment scale-short form
diabetes self-management
diabetes self-management training
glycosylated hemoglobin
second life impacts diabetes education & self-management
type 2 diabetes
virtual environment
This study was funded by the National Library of Medicine (1R21LM010727-01).
None declared.