Published on in Vol 21, No 9 (2019): September

Preprints (earlier versions) of this paper are available at, first published .
Using Relational Agents to Promote Family Communication Around Type 1 Diabetes Self-Management in the Diabetes Family Teamwork Online Intervention: Longitudinal Pilot Study

Using Relational Agents to Promote Family Communication Around Type 1 Diabetes Self-Management in the Diabetes Family Teamwork Online Intervention: Longitudinal Pilot Study

Using Relational Agents to Promote Family Communication Around Type 1 Diabetes Self-Management in the Diabetes Family Teamwork Online Intervention: Longitudinal Pilot Study

Original Paper

1United States Department of Agriculture/Agricultural Research Service Children's Nutrition Research Center, Baylor College of Medicine, Houston, TX, United States

2Department of Pediatrics, Baylor College of Medicine, Houston, TX, United States

3Diabetes and Endocrinology Section, Department of Pediatrics, Baylor College of Medicine, Houston, TX, United States

Corresponding Author:

Debbe Thompson, PhD

United States Department of Agriculture/Agricultural Research Service Children's Nutrition Research Center

Baylor College of Medicine

1100 Bates St

Houston, TX, 77030

United States

Phone: 1 713 798 7076

Fax:1 713 798 7098


Background: Family conflict can reduce adolescent adherence to type 1 diabetes management tasks. The Family Teamwork in-person intervention was shown to be efficacious in reducing conflict and low adherence to diabetes-related tasks. Its reach and potential impact, however, were limited by the need to deliver the intervention sessions in person. Relational agents (ie, computerized versions of humans) have been shown to appeal to diverse audiences and may be an acceptable replacement for a human in technology-based behavior change interventions.

Objective: The purpose of this paper is to present the results of a pilot study assessing feasibility and acceptability of Diabetes Family Teamwork Online, an adapted version of the Family Teamwork intervention, delivered over the internet and guided by a relational agent.

Methods: Parent-adolescent dyads were recruited through a diabetes care clinic at a large tertiary care hospital in the southwestern United States. A one-group design, with assessments at baseline, immediate postintervention, and 3 months later, was used to assess feasibility. A priori feasibility criteria included an assessment of recruitment, completion, attrition, program satisfaction, therapeutic alliance, attitudes toward the relational agent, and data collection. The institutional review board at Baylor College of Medicine approved the protocol (H-37245).

Results: Twenty-seven adolescents aged 10 to 15 years with type 1 diabetes and their parents were enrolled. Criteria used to assess feasibility were (1) recruitment goals were met (n=20), (2) families completed ≥75% of the modules, (3) attrition rate was ≤10%, (4) program satisfaction was high (≥80% of families), (5) therapeutic alliance was high (average score of ≥60/84), (6) families expressed positive attitudes toward the relational agent (average item score of ≥5 on ≥4 items), (7) ≥80% of data were collected at post 1 and post 2, and (8) few technical issues (≤10%) occurred during intervention delivery. All feasibility criteria were met. Qualitative data confirmed that adolescents and parents had positive reactions to both the content and approach.

Conclusions: The Diabetes Family Teamwork Online intervention proved to be a feasible and acceptable method for enhancing communication around diabetes management tasks in families with an adolescent who has type 1 diabetes.

International Registered Report Identifier (IRRID): RR2-10.2196/resprot.5817

J Med Internet Res 2019;21(9):e15318



According to the most up-to-date statistics from the SEARCH for Diabetes in Youth Study Group [1,2], about 154,000 youth are living with type 1 diabetes (T1D), and each year approximately 15,000 youth aged 20 years and younger are diagnosed with T1D, which is second only to asthma as the most common chronic disease of childhood in the United States [3]. Among youth, from 2001 to 2015, 27,000 new cases were diagnosed annually, with children aged 10 to 14 years having the greatest incidence of T1D [4].

Results from the Diabetes Control and Complications Trial (DCCT) increased awareness of how important maintaining near-normal blood glucose levels is to the management of T1D, particularly for delaying and/or preventing T1D complications [5]. However, translation of the rigorous DCCT regimen has been limited by low patient adherence to treatment plans [6]. Adolescents are particularly at risk for low adherence to the constant demands associated with T1D disease management. An efficacy trial of continuous glucose monitoring systems (CGMSs) reported that the adult cohort, but not the adolescent cohort, improved blood glucose control using CGMSs. This may be partially explained by the significantly lower CGMS adherence in the adolescent cohort [7], emphasizing that technology alone, without impacting the environments in which adolescents live and manage T1D, may not produce improved blood glucose control.

It is well documented that the family is important in adolescent adherence [8]. Family conflict and negative communication around diabetes management, especially blood glucose monitoring, are common barriers to adolescent adherence to the diabetes regimen [9]. A meta-analysis of T1D interventions promoting adherence in youth concluded that behavioral interventions without a focus on the personal and interpersonal aspects of the disease were less likely to affect blood glucose control [10]. Therefore, interventions focusing on adolescent adherence and disease management should not ignore the importance of the family, particularly family communication around T1D management tasks, in this process [8].

The Family Teamwork (FT) intervention is an 8-session, in-person intervention delivered by a trained research assistant during routine clinic visits to youth aged 10 to 14 years with T1D and their parents. It was designed to improve hemoglobin A1c (HbA1c) by promoting adherence and blood glucose control in adolescents with T1D [11]. Sessions focused on helping parents and adolescents work together as a team to increase positive parent involvement and reduce family conflict around T1D management. Its efficacy has been demonstrated in two randomized controlled trials, with significant improvements observed in blood glucose monitoring adherence, HbA1c [12,13], and self-reported quality of life [9] in adolescents who participated in FT as compared with those who received standard care only. There was no increase in diabetes-related family conflict, and parents maintained or increased involvement in diabetes management tasks [9]. Reach, however, was limited by the costs associated with having a trained research assistant available to deliver the intervention, the number of families the intervention could be delivered to, and the need for families to travel to the clinic to participate in the intervention.

Communication technologies offer an opportunity to strengthen the reach and effectiveness [14] of interventions while reducing delivery cost. A review found effective eHealth technologies extended the reach of adult diabetes management interventions [15]. However, no studies of online self-management interventions for youth with T1D that met methodological criteria were identified [16], and most studies found lacked details on intervention components and how interventions were tailored for individual patients. An online program to meet the social and informational needs of older adolescents and young adults with T1D was found to be feasible, but only with multiple reminders from the clinic team [17]. Several small online interventions with youth with T1D improved self-management behavior [18,19]. One reported that translating an in-person coping skills intervention for adolescents with T1D that improved glycemic control to the internet was feasible [20].

Self-management is essential to successful treatment and control of chronic diseases like T1D [5,21]. A meta-analysis of pediatric eHealth interventions suggested that interventions with behavioral methods like self-monitoring, goal-setting, and problem-solving were much more effective than those that were solely educational [22], complementing the findings of the meta-analysis cited previously [10]. Both support that adherence and blood glucose control would be improved by intervening with the parent/adolescent dyad to address family barriers around adolescent adherence, particularly when supported by behavioral self-management (goal-setting, problem-solving, self-monitoring). The gold standard for health behavior change is face-to-face interaction with a health care provider [23,24]; however, limitations such as reach [24], time [23], cost, and fidelity can reduce effectiveness [23].

Relational agents (Ragts) may address these limitations. Ragts are computer representations of health care providers that simulate face-to-face interaction with an individual in the real world [23]. They mimic characteristics of face-to-face interactions, including verbal and nonverbal behaviors associated with trust-, rapport-, and relationship-building. Programs delivered by Ragts are convenient, readily accessible, and relatively low in cost, particularly when delivered online [23]. Ragts have been found to be feasible and acceptable with a variety of age groups and health behaviors [23,25-30]. The purpose of this paper is to report the feasibility of an intervention, Diabetes Family Teamwork Online (FTO), an updated version of FT adapted for online delivery by a Ragt.

Adaptation of Family Teamwork Into Diabetes Family Teamwork Online

Using an approach informed by the authors’ previous work in the development of technology-based interventions, computers as persuasive technologies [31], and self determination theory (SDT) [32], FT was systematically adapted for online delivery. Informed by social cognitive theory [33], FT contained 8 informational sessions and 1 review session focused on healthy family communication around key T1D management tasks and common T1D-related issues encountered by families, such as avoiding diabetes burnout (Table 1). A trained research assistant delivered the sessions to parent/adolescent dyads during routine visits to the adolescent’s diabetes care provider, meaning that an extended period of time likely elapsed between clinic visits (eg, 3 to 4 months). Therefore, by necessity, there was a fair amount of repetition included in the sessions. After reviewing FT session content, the research team determined that the sessions could be collapsed into 4 online modules delivered to families every 2 weeks without loss of fidelity to the original program.

To facilitate translation and ensure a balance between didactic components and interactivity, a flow diagram was developed to guide the online program (Figure 1). Using feedback obtained from interviews with families during development [34], a team consisting of a key member of the FT development team (BJA), a pediatric endocrinologist (MJR), and a child psychologist (AB) collapsed the 8 sessions into 4 online modules (Figure 2). Using the flow diagram as a guide, scripts were written for each module. Consistent with SDT, scripts emphasized satisfaction of the basic psychological needs of autonomy (choice and control), competence (knowledge, skill), and relatedness (connection to important others). This was done by providing families with choices as they navigated the modules, offering guidance regarding responses to common situations families encounter regarding T1D management, and emphasizing how to work together to manage blood glucose.

Table 1. Family Teamwork session topics.
  • Diabetes and the family
  • Challenges of diabetes
  • Tools
  • Blood sugar monitoring
  • Hemoglobin A1c
  • Checking blood sugars
  • Talking about blood sugars
  • Avoiding blame
  • Sharing the burden
  • Identifying blood sugar patterns
  • Individualizing care
  • Flexibility in meal planning
  • Carbohydrate counting
  • Exercise
  • Reassessment of the burden
  • Preventing burnout
  • Achieving flexibility
  • Miscarried helping
  • Interdependence versus independence
  • Reducing conflict
  • Review
  • Research and technology update
  • Advances in monitoring devices
Figure 1. Family Teamwork Online module flow.
View this figure
Figure 2. Family Teamwork Online module topics.
View this figure

As part of the development work [34], parents and adolescents were interviewed separately to elicit their opinions about what the Ragt should look like, including appearance, sex, age, and clothing. They were also asked for suggestions regarding names for the Ragt. Using this information as a guide, a female Ragt named Ashley, with a professional, yet approachable, appearance was created (Figure 3). She was animated and designed to mimic in-person delivery by a human diabetes care provider. Prerecorded vocal segments by a professional voice actor were synced to Ashley’s character, enabling her to voice the scripted segments and convey emotion, concern, and respect for the families through vocalizations, facial expressions, and body movements. Consistent with past feedback, Ashley was designed to be “exotic” (ie, she was designed so that she appeared to be of no particular race or ethnicity), thus enabling families to assign their preferred race/ethnicity to her. In addition to Ashley, there were eight other characters—a family, consisting of a father, mother, son, and daughter, and four adult characters, consisting of a male and female physician, a teacher, and a coach (Figure 4).

Each module lasted approximately 20 minutes and focused on a different topic relevant to family communication around diabetes management (Figure 2). Guided by the flow diagram (Figure 1), modules were delivered in a set format that included topic introduction by Ashley; typical family scenarios common to a particular problem or issue relevant to the topic of a particular module; animatics, where typical family reactions (ranging from ideal to less than ideal) to common parent/adolescent situations arising around T1D were portrayed; and interactive components, where Ashley posed questions and the parent/adolescent dyad selected responses from a preselected list of responses, followed by a summary delivered by Ashley regarding the question or scenario. Each module ended with a joint parent/adolescent goal-setting task related to the module topic and family communication around diabetes management. Families were provided with several goals from which they could select. They could then print their goal and a tip sheet. At the beginning of the next module, the family reported goal attainment and received feedback from Ashley. Following completion of the module, they also received an email with the goal they selected and the tip sheet to minimize potential problems associated with not having ready access to a printer.

The program was designed to be viewed over a high-speed internet connection on a variety of devices (eg, desktop computer, laptop, mobile device). It was hosted over a secure, password-protected website. Parents and adolescents received separate passwords with which to log on and view the module. For the initial viewing of module, parents and adolescents were required to log on together. After completing the initial viewing, parents and adolescents could log on separately to view previously completed modules unlimited times. Families were eligible to view a new module and set a new goal 2 weeks after completing the previous module. If a family logged on before they were eligible to view the next module, they had access to previously completed modules only. Families were given approximately 3 months to complete the program.

A secure, password-protected database associated with FTO recorded each family’s information as they navigated the program, including log-ons, responses, and module completion. The researcher accessed the database through a secure dashboard that enabled them to view each family’s progress through the program including log-ons, responses, and goals. After completion of baseline data collection, the intervention coordinator entered the family into the database and assigned passwords. To protect confidentiality, each family received a unique study identifier, and this identifier, rather than their actual names, was entered into the database. A master list of actual names, identifiers, and passwords was stored on a secure, password-protected server available only to the study team.

Figure 3. Image of Ashley, the relational agent in Diabetes Family Teamwork Online.
View this figure
Figure 4. Cast of characters.
View this figure


Families were recruited from a diabetes clinic at a large, tertiary care hospital for children. The study coordinator identified eligible families using the upcoming clinic schedule and the electronic medical record. Inclusionary criteria for the adolescents included being a current patient, aged 10 to 15 years old, diagnosed with T1D for at least 1 year, with English fluency, having high speed internet access, staying in the area for the duration of the study, and having a parent or legal guardian willing to participate in the study. Exclusionary criteria for the adolescents included having an average HbA1c over the past year greater than 12% or less than 7%, inability to attend regular clinic visits, unwillingness to have interviews audio recorded, or having a physical or mental disease or condition that would conflict with study activities.

Inclusionary criteria for the parents included being the primary diabetes caregiver of the adolescent participating in the study, willingness to participate in study activities, with English fluency, having access to high speed internet, and staying in the area for the duration of the study. Exclusionary criteria included inability to attend regular clinic visits, unwillingness to have interviews audio recorded, or having a physical or mental disease or condition that would conflict with study activities.

Eligible families were sent a letter and flyer in the mail notifying parents that a study coordinator would meet the family at their upcoming clinic visit to speak about the study. Prior to the clinic visit, a study coordinator contacted the parent, informed them about the study, and if the parent was interested, screened for eligibility. If the parent was not available by phone, a study coordinator met the family at their clinic visit, described the study, and if the family was interested, screened for eligibility. Parents provided written informed parental consent and each child assented prior to study participation. The institutional review board at Baylor College of Medicine approved the protocol (H-37245).

Intervention Procedures

After completing baseline data collection, the intervention coordinator sent the parent and adolescent separate emails with the website link, username, and private password. Families were provided with a program guide with instructions on how to use the website, including how to log on, how to view the modules, how to click the Back and Next buttons, and how to print the goal and tip sheets.

After the family completed a module, the intervention coordinator sent a thank you email to the parent’s email address with the available date for the next module and a reminder that the family could log on between modules and view the previously completed modules. The goal (set by the family) and tip sheets for the module were also attached to the thank you email. The intervention coordinator kept track of the families logging on to the modules. If families were not logging on regularly to watch the next eligible module, the coordinator sent reminder emails to both the parent and adolescent and followed up with a phone call after the initial reminder email.

Data Collection Procedures

Several types of data were collected: self-report data at baseline, post 1 (immediately after completing the intervention), and post 2 (3 months after post 1); interviews (post 1, post 2); intervention (ie, log-ons, in-module responses); and staff logs (Table 2). Parents and adolescents each received monetary incentives for completing surveys and/or interviews at each data collection time point ($50 baseline, $70 post 1, $80 post 2). No incentives were provided for completing intervention modules.

Table 2. Data sources.
TypeWhoMethodRecruitmentBaselineInterventionPost 1Post 2
Self-reportParent; adolescentOnline

InterviewsParent; adolescentTelephone

ModuleFamilyBackend database

LogsStaffTracking system

Feasibility Criteria

A priori feasibility criteria were established to guide the determination of whether FTO was feasible and acceptable [34]. Minor adjustments were made to the criteria to account for differences in measures or scoring approaches. Final criteria used to assess feasibility were (1) recruitment goals were met (n=20), (2) families completed ≥75% of the modules, (3) attrition rate was ≤10%, (4) program satisfaction was high (≥80% of families), (5) therapeutic alliance with the Ragt was high (average score of ≥60/84), (6) families expressed positive attitudes toward the Ragt (average item score of ≥5 on ≥4 items), (7) ≥80% of data were collected at post 1 and post 2, and (8) few technical issues (<10%) occurred during intervention delivery.


Recruitment was assessed with logs maintained by the research staff (eg, number of eligible families with clinic visits in the recruitment period, number of eligible families who enrolled and reasons, number of eligible families who did not enroll and reasons). The recruitment goal was 20 families. This criterion was assessed by comparing the number of eligible families who enrolled to the criterion.

Module completion was defined as parent and adolescent logging on to the program website at the same time to complete the module activities. This information was automatically collected when families logged on and interacted with the program. Module completion was calculated by dividing the total number of modules completed by the family by the total number of available modules.

Attrition rate was defined as the number of families who did not complete the intervention (ie, all 4 modules) and all data collection activities (ie, baseline, post 1, post 2). It was calculated by dividing the total number of families who did not complete all these activities by the total number of families who enrolled in the study.

Program satisfaction was assessed among parents and adolescents at post 1 with a 10-item measure of satisfaction used in previous studies [30,35,36]. Items were rated using a 3-point Likert scale (yes, not sure, no). Responses were summed to determine overall program satisfaction. Scores could range from 10-30, with lower scores representing higher satisfaction. Separate scores were calculated for parents and adolescents.

Therapeutic alliance was assessed among parents and adolescents at post 1 with the Bond subscale of the Working Alliance Inventory, a 12-item measure assessing trust and belief the participant can work with their provider (ie, Ragt) to achieve the desired outcomes [37]. Items were rated on a 7-point Likert scale, ranging from always hard to always easy. For 3 items, the desired response was never instead of always; therefore, these items were reverse coded, with never being the desired response. Responses were summed to create a total scale score. The total score could range from 12 to 84, with higher scores representing greater therapeutic alliance. Separate scores were computed for parents and adolescents.

Attitudes toward the Ragt were assessed among parents and adolescents at post 1 with a 6-item survey adapted from Bickmore et al [27]. Attitudes were assessed with items addressing different dimensions of the Ragt, including usability, continuance, relationship, preference, adherence, and satisfaction [27]. Items were rated on a 7-point scale, ranging from less desirable to more desirable attitudes (Table 3). Response scales differed for items; therefore, responses were scored and reported individually. Higher scores represented more positive attitudes.

Data completeness was defined as the percentage of families with complete self-report data at all assessment points (baseline, post 1, post 2).

Technical issues were defined as the number of incidents reported by families that limited access to the program and were determined from a review of staff logs.

Table 3. Attitude scale (adapted from Bickmore et al [27]).
AttitudeQuestionsResponse anchors
UsabilityHow easy was it to “talk” with Ashley?1 (always hard) to 7 (always easy)
ContinuanceHow much would you like to continue working with Ashley?1 (not at all) to 7 (very much would like)
RelationshipHow would you describe your relationship with Ashley?1 (complete stranger) to 7 (always a friend)
PreferenceWould you have rather talked with your doctor or nurse rather than Ashley?1 (definitely prefer doctor or nurse) to 7 (definitely prefer Ashley)
AdherenceHow likely is it that you will follow Ashley’s advice?1 (very unlikely) to 7 (very likely)
SatisfactionHow satisfied were you with Ashley?1 (very dissatisfied) to 7 (very satisfied)

To obtain an understanding of parent and adolescent perceptions toward the intervention, telephone interviews were conducted following completion of post 1 and post 2 data collection surveys. The purpose of the post 1 interview was to identify perceptions related to the intervention, while the post 2 interview explored perceptions related to maintenance. Only post 1 interviews will be reported here to provide a deeper understanding of program feasibility and acceptability from the perspective of the adolescents and parents who participated in the study. Interviews were scripted and conducted by trained interviewers. Probes and prompts were used as needed to expand, clarify, and understand responses.

Statistical Analyses

Descriptive statistics were calculated on survey data. Separate analyses were calculated for parents and adolescents. Results were compared to the criteria to make a determination of feasibility. Audio recordings of interviews were reviewed and key point summaries created to capture important thoughts that emerged from adolescent and parent interviews to enhance understanding of results [38]. Verbatim quotes were used to provide voice to adolescent and parent perceptions and experiences.


Parents were mostly female (25/26, 96%), white (21/26, 81%), non-Hispanic (22/26, 85%), and married or living with significant other (22/26, 85%). Families had college-level education or higher (20/26, 77%) and reported household incomes of greater than $61,000 (20/26, 80%; Table 4).Adolescents were mostly female (19/26, 73%), white (22/26, 85%), and non-Hispanic (21/26, 81%). There were slightly more adolescents aged 13 to 15 years in the study (14/26, 54%). Disease duration ranged from 2 (4/26, 15%) to 13 years (1/26, 4%; Table 5).

Table 4. Descriptive statistics of parents who completed the Diabetes Family Teamwork Online study (n=26).
CharacteristicValue, n (%)

Male1 (4)

Female25 (96)

American Indian/Alaskan Native1 (4)

Black3 (12)

White21 (81)

Other1 (4)

Hispanic4 (15)

Non-Hispanic22 (85)
Marital status

Married/living with significant other22 (85)

Single, never married2 (3)

Divorced, separated, or widowed1 (4)

Other1 (4)
Highest household education

High school graduate or equivalent3 (12)

Technical school1 (4)

Some college2 (8)

College graduate11 (42)

Postgraduate study9 (35)
Household incomea

$21,000-$41,0003 (12)

$42,000-$61,0002 (8)

Greater than $61,00020 (80)

aOne parent did not report household income.

Table 5. Descriptive statistics of adolescents who completed the Diabetes Family Teamwork Online study (n=26).
CharacteristicValue, n (%)

Male7 (27)

Female19 (73)

American Indian/Alaskan Native1 (4)a

Black3 (12)

White22 (85)

Hispanic5 (19)

Non-Hispanic21 (81)
Age (years)

10-1212 (46)

13-1514 (54)
Disease duration (years)

24 (15)

38 (31)

42 (8)

53 (12)

63 (12)

71 (4)

82 (8)

101 (4)

121 (4)

131 (4)

aParticipant also identifies as white.

Feasibility Outcomes

Feasibility criteria and status are summarized in Table 6. Feasibility of each criterion is briefly described below.

Recruitment began in October 2016 and ended in December 2016. A review of clinic records indicated that 48 families met general eligibility criteria and were notified of the study and their eligibility to participate. Of these, 5 declined to participate, 9 could not be contacted, 3 were enrolled in another study that made them ineligible for our study, and 4 were not included for reasons ranging from clinic visit scheduled after recruitment closed (2), not meeting inclusionary criteria (1), and not returning recruitment packet (1). Of the eligible families, 27 enrolled, completed baseline data collection, and were allocated to the intervention (Figure 5). Our recruitment goal was 20 families; therefore, this criterion was exceeded.

FTO consisted of 4 modules. All 4 modules were completed by 26 out of 27 families enrolled in this study. The feasibility criterion was >75%; therefore, this criterion was met.

Of the 27 families who enrolled in the study, 26 completed the intervention and all data collection (baseline, post 1, post 2) and intervention activities. The feasibility criterion was ≤10%; therefore, this criterion was met.

The mean program satisfaction scores were 11.08 (SD 1.32) for adolescents and 10.46 (SD 0.65) for parents, indicating that over 90% of adolescents and parents reported high program satisfaction. The feasibility criterion was ≥80% of adolescents and parents reported high program satisfaction. Therefore, this criterion was met, with both groups reporting high program satisfaction.

The mean therapeutic alliance score for adolescents was 64.58 (SD 14.43) and 64.65 (SD 14.33) for parents. The feasibility criterion was an average total scale score of ≥60/84. Therefore, this criterion was met for both adolescents and parents.

Table 6. Feasibility criteria.
Recruitment20 familiesExceeded
Module completion≥75% modulesMet
Attrition rate≤10%Met
Program satisfaction≥80%Met
Therapeutic allianceAverage score of ≥60/84Met
Attitudes toward RagtScore of ≥5 on ≥4 itemsMet
Data completeness≥80%Met
Technical issues≤10%Met
Figure 5. Family Teamwork Online CONSORT (Consolidated Standards of Reporting Trials) diagram.
View this figure

Both adolescents and parents reported attitude scores toward Ashley of at least 5 on 5 of the 6 attitude items, respectively: usability (5.85 [SD 1.43], 5.50 [SD 1.39]), continuance (5.42 [SD 1.17], 5.23 [SD 1.03]), relationship (5.42 [SD 1.55], 5.08 [SD 1.52]), preference (2.73 [SD 1.59], 3.54 [SD 1.49]), adherence (5.73 [SD 1.12], 6.00 [SD 0.69]), and satisfaction (6.0 [SD 1.02], 5.96 [SD 0.72]). Therefore, this criterion was met.

Twenty-six of the 27 families had complete self-report data at all assessment points. The feasibility criterion was ≥80%; therefore, this criterion was met.

No technical issues were reported by the families. The feasibility criterion was “few technical issues”, defined as <10%. Therefore, this criterion was met.

The qualitative findings generally supported the survey results. Most adolescents and parents liked the program. They generally thought the content was appropriate, easy to understand, and realistic.

Well, I thought it was useful in what we deal with every day—high blood sugars or the low blood sugars or what we need to be doing better when it comes to managing [daughter’s name] diabetes. The examples were pretty common to what we deal with, so it helped to see it being played out for us and how families responded and if our response was appropriate for that situation. I thought it was helpful in helping us understand how to react with the high blood sugar or low blood sugar or just certain situations we deal with every day.

They thought that the module number and length were appropriate.

I think they were just about right. I mean, I liked them as they were. I wouldn’t have liked them to be longer or shorter.

This sentiment seemed to be echoed by parents.

I liked the time. I didn’t think they were too long or too short ‘cause it was so detailed with every session, but it didn’t take a lot of my time or my child’s.

A few families wanted longer or more modules on certain topics. For example, a parent suggested adding information on tricky topics.

I think four is okay. I think for some of the more tricky topics, maybe adding a follow-up for that that goes into more detail might be helpful.

A adolescent expressed also wanting more information.

I think it was too little...6 or 7...cause I wanted to learn more about what they were talking about.

Parents and adolescents also seemed to like the family focus taken by the program.

I liked it. It was pretty fun experiencing it with my mom and how we got to like work out my diabetes better. It was pretty fun.

Overall, parents shared positive feedback on the Ragt, including her appearance, responses, and voice.

I thought she was very helpful. She was very helpful in answering things, and like for something quick, like a quick question or whatever, I mean her response to things was better than having a doctor right there in your face. She was professional. She was child friendly and adult friendly.

Adolescents had similar reactions.

I think she was helpful, and I think that—I don’t know if it’s just me—since she’s a girl, it’s easier to listen to someone the same as me.

A few adolescents expressed that although they liked Ashley, she had limitations.

I think she was fine, but I would rather have an actual person...because Ashley didn’t really give that much detail.

Families were also asked what information in the program was most helpful or useful. Parents shared a variety of information that was most helpful for them, including the information presented on parent reactions to high and low blood sugars, not referring to blood sugars as good and bad, communication between the parent and adolescent, setting goals, and diabetes burnout. Two parent quotes eloquently summarize parent thoughts regarding what they perceived as most helpful:

Well for me it would be the reminder to let the teen take charge and let the parent sort of step back and say what would you do differently instead of managing it myself.
That [diabetes burnout] was probably the best one for us because we really didn’t know a lot about that.

Adolescents had similar reactions and identified a variety of topics that were helpful. They seemed to particularly focus on issues related to family communication around diabetes management.

How to control like no blaming and everyone’s part of the team.

The focus on using different language was also viewed as important.

The most helpful activity was the one where you had to use different words instead of it was like a bad or good blood sugar, it was a high or a low.

When asked what was least helpful in the program, few parents or adolescents had many suggestions to offer. Most thought the program was helpful.

I don’t really think there was just anything so unhelpful that I would leave it out.

This sentiment was echoed by adolescents.

I don’t know. I don’t think anything was bad about the program. I liked it myself, so I don’t think anything was least helpful. Everything was pretty helpful to me.

A few suggestions were offered, however.

For me, it was all a good reminder, but the probably least helpful was looking at the numbers as good or bad. I always try to look at them as information...that was probably the least helpful because I kind of already look at it that way.
Taking the snack and checking your blood sugars and stuff like that when you’re exercising.

Principal Findings

This paper reports on the feasibility of delivering a program to enhance communication around blood glucose management tasks in families with an adolescent who has T1D. Feasibility criteria were met or exceeded. Therefore, we concluded this program was a feasible method for promoting enhanced communication around diabetes management tasks.

Recruitment goals for this study were exceeded in a relatively short period of time. Conflict around diabetes management tasks in families with adolescents who have T1D and its association with poorer diabetes outcomes have been well documented in the literature [9,39,40]. There is also clear evidence that both parents and adolescents perceive there is conflict around diabetes management [41,42]. The short time period within which recruitment goals were exceeded suggests that parents and adolescents are interested in learning to manage or reduce this conflict.

Participation by families was high, as indicated by the low attrition rate and high program completion rates. This suggests that families found the program helpful and a beneficial use of their time. This is not surprising, however, given the success of FT, the program on which FTO was based [9,12,13]. This finding also adds to the body of literature supporting communication technologies as a potentially effective way to extend the reach of diabetes management programs for adolescents [16]. Future research is needed to assess the efficacy of this approach.

Therapeutic alliance is the emotional dimension associated with working with a provider and their perceived ability to help the patient achieve a desired outcome [37]. Although parents and adolescents were aware that Ashley was a digital image and not an actual person, therapeutic alliance was established. This was similar to results reported by Bickmore et al [43] when a Ragt was used to encourage exercise in a healthy population and deliver hospital discharge instructions to patients with depressive symptoms [27]. Given the differences between initiating a healthy behavior such as exercise and receiving instructions for posthospital care versus potential consequences of poor glycemic control [44], this suggests that Ragts may be an acceptable way to extend care for a variety of diverse behaviors outside of traditional settings. In the management of a chronic condition, such as T1D, patient-provider communication is an essential component of care [45]. Therefore, our goal was not to supplant this relationship but to extend care outside the clinical setting. Although this study suggests Ragts are a feasible and acceptable method of achieving this goal, future research is needed to examine ways in which Ragts may further extend the patient-provider relationship outside the clinical setting. One way to achieve this may be to design Ragts that respond emphatically in real time, based on the participant’s expressed emotions [46]. Further, as suggested by the qualitative results, it may be that therapeutic alliance takes on a different form with a Ragt. Future research is needed to more fully investigate this possibility.

Both parents and adolescents reported positive attitudes regarding ease of use, desire to continue working with Ashley, their relationship with Ashley, the likelihood of adhering to her advice, and satisfaction with her. The exception was their stronger preference for talking with their doctor or nurse, rather than Ashley, about T1D. As discussed above, given that the patient-provider relationship is essential to the care of T1D, particularly in adolescents [45], this finding is not surprising. The goal when working with chronic disease management should not be to supplant or interfere with the patient-provider relationship but to augment and support it and extend care beyond the clinical setting. Given that this is the first study to use Ragts to enhance communication around blood glucose management tasks in families with an adolescent who has T1D, it adds important information to the literature regarding expectations around chronic disease management and ways in which to augment and extend care between visits with the health care provider. Future research with providers may provide insight into ways Ragts could be used to further extend care into the home environment.

Although missing data are a common problem in research that can present challenges in the analytic stage [47], we had complete data on all but one family. Consistent with guidelines developed to minimize missing data [48], data collection was online, making participation convenient, and the families received notifications when they were eligible to participate in data collection activities. We also ensured contact information for each family was current, study coordinators were experienced and well trained in study procedures, participation in data collection activities was monitored and tracked, and incentives were provided to the families for completing data collection at each assessment point. All of these activities likely contributed to the high data collection rates.

Finally, families reported no technical issues accessing or using the program to the intervention team. This is likely due to working with a highly skilled local Web design company with whom we have a history of successful collaboration. Further, experience has taught us the importance of knowing in advance how the individual pieces of the intervention fit together, as well as the overall program flow. A template was created for the modules, and this template guided each module. This consistency may have contributed to fewer technical glitches. A program guide was also created and shared with families prior to their viewing of the first module, which may have further contributed to fewer technical issues. Because technical issues can be annoying and may disrupt program participation, particularly when they occur multiple times or at critical points in program delivery, the lack of technical issues may have also contributed to our high program participation and satisfaction ratings.


Although this study has strengths, including being an adaptation of an intervention previously found to be efficacious and its theoretical grounding, it is not without limitations. The small sample size and one-group design limits its ability to find statistical significance in key outcome variables, such as family conflict and communication around blood glucose. However, this was designed to be a feasibility study, with the goals of assessing the viability of this approach, needed changes, and whether it should move forward for an assessment of efficacy with a larger, more fully powered randomized controlled trial [49]. Parents and adolescents in the study were predominantly female and the Ragt was female, which may have influenced the appeal of this approach to families who participated in the study. However, the appearance, sex, and age of the Ragt designed for this study was informed by formative research with parents and adolescents other than those who participated in the feasibility study; this suggests the Ragt may have broad appeal regardless of sex. This study was also conducted in a limited geographic region with a sample of families who were predominately white, of higher socioeconomic status, and well educated, limiting the ability to draw conclusions about the generalizability of this approach with other populations. Finally, because this study uses technology, it may not be appropriate for use in locations that may have limitations regarding internet access or speed, such as some rural areas, and it may have broader appeal among families with adolescents who use technology to facilitate blood glucose management (eg, continuous glucose monitoring, insulin pumps).


An online intervention to enhance communication around blood glucose management task in families with an adolescent who has diabetes was found to be feasible and acceptable. The contribution of this research to the literature is that it demonstrates that Ragts are a feasible adjunct to traditional care and may be a way to extend care outside the clinical setting. Future research is needed to identify the effects on family communication, quality of life, blood glucose, and diabetes management in a larger, more fully powered sample.


This work was supported by the US Department of Agriculture/Agricultural Research Service (USDA/ARS) and the Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine (Houston, Texas), funded in part with federal funds from the USDA/ARS under Cooperative Agreement No 58-3092-5-001. This project was also supported by National Institute of Diabetes and Digestive and Kidney Diseases, grant #DK106505 (to BJA and DT). The contents of this paper do not necessarily reflect the views or policies of the USDA, and the mention of trade names, commercial products, or organizations does not imply endorsement from the US Government. We would like to thank Jerald Reichstein, Jackson Lau, and the team at Bouncing Pixel for their assistance in developing the program, as well as the parents and adolescents who participated in this study.

Conflicts of Interest

None declared.

  1. Maahs DM, West NA, Lawrence JM, Mayer-Davis EJ. Epidemiology of type 1 diabetes. Endocrinol Metab Clin North Am 2010 Sep;39(3):481-497 [FREE Full text] [CrossRef] [Medline]
  2. Halfon N, Newacheck PW. Evolving notions of childhood chronic illness. JAMA 2010 Feb 17;303(7):665-666. [CrossRef] [Medline]
  3. Miller GF, Coffield E, Leroy Z, Wallin R. Prevalence and costs of five chronic conditions in children. J Sch Nurs 2016 Oct;32(5):357-364 [FREE Full text] [CrossRef] [Medline]
  4. Rogers MAM, Kim C, Banerjee T, Lee JM. Fluctuations in the incidence of type 1 diabetes in the United States from 2001 to 2015: a longitudinal study. BMC Med 2017 Dec 08;15(1):199 [FREE Full text] [CrossRef] [Medline]
  5. Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993 Sep 30;329(14):977-986. [CrossRef] [Medline]
  6. Narayan KM, Gregg EW, Engelgau MM, Moore B, Thompson TJ, Williamson DF, et al. Translation research for chronic disease: the case of diabetes. Diabetes Care 2000 Dec;23(12):1794-1748 [FREE Full text] [CrossRef] [Medline]
  7. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group, Beck RW, Buckingham B, Miller K, Wolpert H, Xing D, et al. Factors predictive of use and of benefit from continuous glucose monitoring in type 1 diabetes. Diabetes Care 2009 Nov;32(11):1947-1953 [FREE Full text] [CrossRef] [Medline]
  8. Lansing AH, Crochiere R, Cueto C, Wiebe DJ, Berg CA. Mother, father, and adolescent self-control and adherence in adolescents with Type 1 diabetes. J Fam Psychol 2017 Jun;31(4):495-503 [FREE Full text] [CrossRef] [Medline]
  9. Anderson BJ, Vangsness L, Connell A, Butler D, Goebel-Fabbri A, Laffel LMB. Family conflict, adherence, and glycaemic control in youth with short duration Type 1 diabetes. Diabet Med 2002 Aug;19(8):635-642. [Medline]
  10. Hood KK, Rohan JM, Peterson CM, Drotar D. Interventions with adherence-promoting components in pediatric type 1 diabetes: meta-analysis of their impact on glycemic control. Diabetes Care 2010 Jul;33(7):1658-1664 [FREE Full text] [CrossRef] [Medline]
  11. Anderson BJ, Brackett J, Ho J, Laffel LM. An office-based intervention to maintain parent-adolescent teamwork in diabetes management. Impact on parent involvement, family conflict, and subsequent glycemic control. Diabetes Care 1999 May;22(5):713-721 [FREE Full text] [Medline]
  12. Laffel LMB, Connell A, Vangsness L, Goebel-Fabbri A, Mansfield A, Anderson BJ. General quality of life in youth with type 1 diabetes: relationship to patient management and diabetes-specific family conflict. Diabetes Care 2003 Nov;26(11):3067-3073. [Medline]
  13. Laffel LMB, Vangsness L, Connell A, Goebel-Fabbri A, Butler D, Anderson BJ. Impact of ambulatory, family-focused teamwork intervention on glycemic control in youth with type 1 diabetes. J Pediatr 2003 Apr;142(4):409-416. [CrossRef] [Medline]
  14. King DK, Glasgow RE, Leeman-Castillo B. Reaiming RE-AIM: using the model to plan, implement, and evaluate the effects of environmental change approaches to enhancing population health. Am J Public Health 2010 Nov;100(11):2076-2084. [CrossRef] [Medline]
  15. Lorig K, Ritter PL, Laurent DD, Plant K, Green M, Jernigan VBB, et al. Online diabetes self-management program: a randomized study. Diabetes Care 2010 Jun;33(6):1275-1281 [FREE Full text] [CrossRef] [Medline]
  16. Stinson J, Wilson R, Gill N, Yamada J, Holt J. A systematic review of internet-based self-management interventions for youth with health conditions. J Pediatr Psychol 2009 Jun;34(5):495-510 [FREE Full text] [CrossRef] [Medline]
  17. Gerber BS, Solomon MC, Shaffer TL, Quinn MT, Lipton RB. Evaluation of an internet diabetes self-management training program for adolescents and young adults. Diabetes Technol Ther 2007 Feb;9(1):60-67. [CrossRef] [Medline]
  18. Mulvaney SA, Rothman RL, Osborn CY, Lybarger C, Dietrich MS, Wallston KA. Self-management problem solving for adolescents with type 1 diabetes: intervention processes associated with an Internet program. Patient Educ Couns 2011 Nov;85(2):140-142 [FREE Full text] [CrossRef] [Medline]
  19. Mulvaney SA, Rothman RL, Wallston KA, Lybarger C, Dietrich MS. An internet-based program to improve self-management in adolescents with type 1 diabetes. Diabetes Care 2010 Mar;33(3):602-604 [FREE Full text] [CrossRef] [Medline]
  20. Whittemore R, Grey M, Lindemann E, Ambrosino J, Jaser S. Development of an Internet coping skills training program for teenagers with type 1 diabetes. Comput Inform Nurs 2010;28(2):103-111 [FREE Full text] [CrossRef] [Medline]
  21. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998 Sep 12;352(9131):837-853. [Medline]
  22. Cushing CC, Steele RG. A meta-analytic review of eHealth interventions for pediatric health promoting and maintaining behaviors. J Pediatr Psychol 2010 Oct;35(9):937-949 [FREE Full text] [CrossRef] [Medline]
  23. Bickmore T. Relational agents for chronic disease self management. In: Hayes BM, Aspray W, editors. Health Informatics: A Patient-Centered Approach to Diabetes. Cambridge: MIT Press; 2010:181-204.
  24. Bickmore T, Giorgino T. Some novel aspects of health communication from a dialogue systems perspective. 2004 Presented at: Proceedings of the AAAI Fall Symposium on Dialogue Systems for Health Communication; 2004; Arlington.
  25. Bickmore TW, Schulman D, Sidner CL. A reusable framework for health counseling dialogue systems based on a behavioral medicine ontology. J Biomed Inform 2011 Apr;44(2):183-197 [FREE Full text] [CrossRef] [Medline]
  26. Bickmore T, Gruber A. Relational agents in clinical psychiatry. Harv Rev Psychiatry 2010;18(2):119-130. [CrossRef] [Medline]
  27. Bickmore TW, Mitchell SE, Jack BW, Paasche-Orlow MK, Pfeifer LM, Odonnell J. Response to a relational agent by hospital patients with depressive symptoms. Interact Comput 2010 Jul 1;22(4):289-298 [FREE Full text] [CrossRef] [Medline]
  28. Puskar K, Schlenk EA, Callan J, Bickmore T, Sereika S. Relational agents as an adjunct in schizophrenia treatment. J Psychosoc Nurs Ment Health Serv 2011 Aug;49(8):22-29. [CrossRef] [Medline]
  29. Watson A, Bickmore T, Cange A, Kulshreshtha A, Kvedar J. An internet-based virtual coach to promote physical activity adherence in overweight adults: randomized controlled trial. J Med Internet Res 2012;14(1):e1 [FREE Full text] [CrossRef] [Medline]
  30. Thompson D, Bhatt R, Lazarus M, Cullen K, Baranowski J, Baranowski T. A serious video game to increase fruit and vegetable consumption among elementary aged youth (Squire's Quest! II): rationale, design, and methods. JMIR Res Protoc 2012;1(2):e19 [FREE Full text] [CrossRef] [Medline]
  31. Fogg B. Persuasive Technology: Using Computers to Change What We Think and Do. 1st Edition. San Francisco: Morgan Kaufmann Publishers; 2002.
  32. Ryan RM, Deci EL. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. Am Psychol 2000 Jan;55(1):68-78. [Medline]
  33. Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs: Prentice Hall; 1986.
  34. Thompson D, Cullen KW, Redondo MJ, Anderson B. Use of relational agents to improve family communication in type 1 diabetes: methods. JMIR Res Protoc 2016;5(3):e151 [FREE Full text] [CrossRef] [Medline]
  35. Thompson D, Mahabir R, Bhatt R, Boutte C, Cantu D, Vazquez I, et al. Butterfly Girls; promoting healthy diet and physical activity to young African American girls online: rationale and design. BMC Public Health 2013;13:709 [FREE Full text] [CrossRef] [Medline]
  36. Cullen KW, Thompson D, Boushey C, Konzelmann K, Chen T. Evaluation of a web-based program promoting healthy eating and physical activity for adolescents: teen choice: food and fitness. Health Educ Res 2013 Aug;28(4):704-714 [FREE Full text] [CrossRef] [Medline]
  37. Horvath AO, Greenberg LS. Development and validation of the working alliance inventory. J Couns Psychol 1989;36(2):223-233. [CrossRef]
  38. Schneider M, Hall WJ, Hernandez AE, Hindes K, Montez G, Pham T, HEALTHY Study Group. Rationale, design and methods for process evaluation in the HEALTHY study. Int J Obes (Lond) 2009 Aug;33 Suppl 4:S60-S67 [FREE Full text] [CrossRef] [Medline]
  39. Hilliard ME, Holmes CS, Chen R, Maher K, Robinson E, Streisand R. Disentangling the roles of parental monitoring and family conflict in adolescents' management of type 1 diabetes. Health Psychol 2013 Apr;32(4):388-396 [FREE Full text] [CrossRef] [Medline]
  40. Williams LB, Laffel LMB, Hood KK. Diabetes-specific family conflict and psychological distress in paediatric Type 1 diabetes. Diabet Med 2009 Sep;26(9):908-914 [FREE Full text] [CrossRef] [Medline]
  41. Ye CY, Jeppson TC, Kleinmaus EM, Kliems HM, Schopp JM, Cox ED. Outcomes that matter to teens with type 1 diabetes. Diabetes Educ 2017 Dec;43(3):251-259 [FREE Full text] [CrossRef] [Medline]
  42. Weissberg-Benchell J, Nansel T, Holmbeck G, Chen R, Anderson B, Wysocki T, Steering Committee of the Family Management of Diabetes Study. Generic and diabetes-specific parent-child behaviors and quality of life among youth with type 1 diabetes. J Pediatr Psychol 2009 Oct;34(9):977-988 [FREE Full text] [CrossRef] [Medline]
  43. Bickmore T, Gruber A, Picard R. Establishing the computer-patient working alliance in automated health behavior change interventions. Patient Educ Couns 2005 Oct;59(1):21-30. [CrossRef] [Medline]
  44. Rausch JR, Hood KK, Delamater A, Shroff PJ, Rohan JM, Reeves G, et al. Changes in treatment adherence and glycemic control during the transition to adolescence in type 1 diabetes. Diabetes Care 2012 Jun;35(6):1219-1224 [FREE Full text] [CrossRef] [Medline]
  45. Patel NJ, Datye KA, Jaser SS. Importance of patient-provider communication to adherence in adolescents with type 1 diabetes. Healthcare (Basel) 2018 Mar 28;6(2):30 [FREE Full text] [CrossRef] [Medline]
  46. Scholten MR, Kelders SM, Van Gemert-Pijnen JE. Self-guided web-based interventions: scoping review on user needs and the potential of embodied conversational agents to address them. J Med Internet Res 2017 Nov 16;19(11):e383 [FREE Full text] [CrossRef] [Medline]
  47. Kang H. The prevention and handling of the missing data. Korean J Anesthesiol 2013 May;64(5):402-406 [FREE Full text] [CrossRef] [Medline]
  48. Little RJ, D'Agostino R, Cohen ML, Dickersin K, Emerson SS, Farrar JT, et al. The prevention and treatment of missing data in clinical trials. N Engl J Med 2012 Oct 4;367(14):1355-1360 [FREE Full text] [CrossRef] [Medline]
  49. Bowen DJ, Kreuter M, Spring B, Cofta-Woerpel L, Linnan L, Weiner D, et al. How we design feasibility studies. Am J Prev Med 2009 May;36(5):452-457 [FREE Full text] [CrossRef] [Medline]

CGMS: continuous glucose monitoring system
DCCT: Diabetes Control and Complications Trial
FT: Family Teamwork
FTO: Diabetes Family Teamwork Online
HbA1c: hemoglobin A1c
Ragt: relational agent
SDT: self-determination theory
T1D: type 1 diabetes
USDA/ARS: US Department of Agriculture/Agricultural Research Service

Edited by G Eysenbach; submitted 01.07.19; peer-reviewed by S Brigitte, T Oser; comments to author 24.07.19; revised version received 30.07.19; accepted 13.08.19; published 13.09.19


©Debbe Thompson, Chishinga Callender, Caroline Gonynor, Karen W Cullen, Maria J Redondo, Ashley Butler, Barbara J Anderson. Originally published in the Journal of Medical Internet Research (, 13.09.2019.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on, as well as this copyright and license information must be included.