You are on page 1of 8

C l i n ic al C a r e / E d u c a t i o n / N u t r i t i o n / P s y c h o s o c i a l Rese arch

O R I G I N A L ARTICL E

Internet Psycho-Education Programs


Improve Outcomes in Youth With
Type 1 Diabetes
MARGARET GREY, DRPH, RN, FAAN1 MELISSA S. FAULKNER, DSN, RN, FAAN3 As youth transition to adolescence
ROBIN WHITTEMORE, PHD, APRN, FAAN1 ALAN DELAMATER, PHD4 and take on greater responsibility for their
SANGCHOON JEON, PHD1 FOR THE TEENCOPE STUDY GROUP* diabetes self-management, educational
KATHRYN MURPHY, PHD, RN2 needs are higher. Standards of care for
youth with type 1 diabetes identify the
importance of education; yet, the pro-
OBJECTIVEdThe purpose of this study was to determine the efficacy of two Internet-based
vision of education in the clinical setting
psycho-educational programs designed to improve outcomes for youth with type 1 diabetes
transitioning to adolescence. is mostly informal and inconsistent. Con-
siderable evidence indicates that in-
RESEARCH DESIGN AND METHODSdThe study was a multisite clinical trial of 320 person psycho-educational interventions,
youth (aged 11–14 years; 37% minority; 55% female) randomized to one of two Internet-based such as Coping Skills Training (CST),
interventions: TeenCope or Managing Diabetes. Primary outcomes were HbA1c and quality of life improve metabolic control of type 1 di-
(QOL). Secondary outcomes included coping, self-efficacy, social competence, self-management, abetes as well as psychosocial adjustment
and family conflict. Data were collected at baseline and after 3, 6, and 12 months online. Youth
and QOL in youth (6–8). However, im-
were invited to cross over to the other program after 12 months, and follow-up data were
collected at 18 months. Analyses were based on mixed models using intent-to-treat and per- plementing these evidence-based pro-
protocol procedures. grams in clinical care is challenging
because of provider and organizational bar-
RESULTSdYouth in both groups had stable QOL and minimal increases in HbA1c levels over riers, such as lack of time, resources, and
12 months, but there were no significant differences between the groups in primary outcomes. expertise (9). Rapid advances in technology
After 18 months, youth who completed both programs had lower HbA1c (P = 0.04); higher QOL and access to the Internet have made it not
(P = 0.02), social acceptance (P = 0.01), and self-efficacy (P = 0.03) and lower perceived stress (P = only a viable mode for the delivery of psy-
0.02) and diabetes family conflict (P = 0.02) compared with those who completed only one
program.
cho-educational interventions but also
a platform that can be widely dissemi-
CONCLUSIONSdInternet interventions for youth with type 1 diabetes transitioning to nated and implemented. Internet inter-
adolescence result in improved outcomes, but completion of both programs was better than ventions allow for standardization of
only one, suggesting that these youth need both diabetes management education and program content, can be targeted to spe-
behavioral interventions. Delivering these programs via the Internet represents an efficient cific ages and developmental phases, al-
way to reach youth and improve outcomes. low for social interaction, and can be

Y
easily updated. Access to the Internet is
Diabetes Care 36:2475–2482, 2013
increasingly available nationwide and has
risen to its highest level ever, with 93% of
outh with type 1 diabetes in transi- particularly challenging for these youth. youth using the Internet regularly for
tion to adolescence are a vulnerable Having diabetes involves fear of hypogly- school assignments, hobbies or special
population. They exhibit deteriorating cemia, fear of future complications, feel- interests, entertainment, and connection
metabolic control (1,2), poorer self- ings of guilt for possible wrongdoing, with others (10,11). The Internet, there-
management, and increased social stres- and feelings of stress associated with fore, represents an efficient way to deliver
sors and psychosocial distress, as well as challenging self-management tasks (3). psycho-educational interventions to
lower quality of life (QOL) compared In addition, these stressors occur within youth with type 1 diabetes.
with children with type 1 diabetes at the broader context of increased expect- Psycho-educational interventions
other ages (2). Resolving independence/ ations for adolescents to maintain ex- delivered via the Internet have demon-
dependence issues and acquiring positive cellent metabolic control (4), despite strated efficacy in improving symptoms
acceptance by peers within the context of the insulin resistance associated with and health behaviors in youth of different
a complex treatment regimen are puberty (5). ages and illness experiences (12–14). In a
pilot study, youth with type 1 diabetes
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c
who completed an Internet program
From the 1Yale University School of Nursing, New Haven, Connecticut; the 2Department of Pediatrics, with a focus on problem solving and so-
Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania; the 3School of Nursing, University of
Arizona, Phoenix, Arizona; and the 4Department of Pediatrics, University of Miami, Miami, Florida. cial networking demonstrated improved
Corresponding author: Margaret Grey, margaret.grey@yale.edu. self-management and problem solving
Received 26 October 2012 and accepted 11 February 2013. compared with a control group (15).
DOI: 10.2337/dc12-2199. Clinical trial reg. no. NCT00684658, clinicaltrials.gov. Internet-based interventions that can
*A complete list of the TeenCope Study Group can be found in the APPENDIX.
© 2013 by the American Diabetes Association. Readers may use this article as long as the work is properly
reach large numbers of youth with diabe-
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ tes have the potential to result in signifi-
licenses/by-nc-nd/3.0/ for details. cant improvements in long-term health,

care.diabetesjournals.org DIABETES CARE, VOLUME 36, SEPTEMBER 2013 2475


Internet psycho-education for youth

as well as reductions in the costs of care positive self-talk, and conflict resolution program. A unique password was pro-
for diabetes-related complications. They (16,19). A monitored discussion board vided to each participant, and they were
also have the potential to improve access allowed TeenCope participants to com- instructed to change this password the
for diverse youth with type 1 diabetes. municate with youth from the other first time they logged onto the program.
While there is ongoing evidence of the participating clinical sites. Managing Each program had five sessions that were
digital divide, this has been decreasing Diabetes, the control condition, was released weekly and took ~30 min to
over time, particularly with English- designed as a diabetes education and complete. The average time to complete
speaking youth in the U.S. (10). problem-solving program to be delivered all sessions was 40 6 23.2 days. Research
TeenCope, a new Internet-based ver- via the Internet. It used visuals and an in- staff at each site sent e-reminders if youth
sion of CST, was developed by our group. teractive interface that allowed youth to did not complete a session within 2
It is based on social cognitive theory and learn about healthy eating, physical activ- weeks. After completion of 12-month
posits that improving coping skills will ity, glucose control, sick days, and diabe- data, participants were invited to com-
lead to improved self-efficacy and self- tes technology. Interactivity consisted plete the alternate program.
management of diabetes that result in of active links to more detailed infor- A framework developed by our group
better outcomes, as has been demon- mation, polling about diabetes care is- guided our measurement design (20).
strated in studies of CST delivered in a sues, and problem-solving exercises Data were collected at baseline and 3, 6,
group-based in-person format (16). Man- with tailored feedback to participant re- 12, and 18 months. The primary out-
aging Diabetes was developed to serve as sponses. Content of Managing Diabetes comes were HbA1c levels and QOL. Sec-
the control condition and was a diabetes was based on standards of care for diabe- ondary outcomes included stress, coping,
education and problem-solving program. tes management in youth (8), with an em- self-efficacy, self-management, social
Thus, the purpose of this multisite phasis on decision making for optimal competence, and family conflict. Data
randomized clinical trial was to compare outcomes. were collected by online survey with the
the efficacy of two Internet-based pro- A convenience sample was recruited exception of HbA1c and other clinical data
grams on the primary outcomes of HbA1c from four university-affiliated clinical that were collected by chart review and
and QOL and on the secondary out- sites (Children’s Hospital of Philadelphia, demographic data that were provided
comes of stress, coping, self-efficacy, University of Arizona, University of Mi- at consent by parents/guardians. All
self-management, social competence, ami, and Yale University). Inclusion crite- measures were thoroughly evaluated for
and family conflict at 12 months. At 12- ria were as follows: diagnosis with type 1 reliability and validity.
month follow-up, youth were invited to diabetes for at least 6 months, age 11–14
participate in the alternate program, al- years, no other significant medical prob- Primary outcomes
lowing us to explore the effect of partici- lem, school grade appropriate to age HbA1c levels were determined using the
pating in two programs compared with within 1 year, ability to speak and write DCA2000 (Bayer, Tarrytown, NY) at each
participating in only one. We hypothe- English, and access to high-speed Internet of the sites. Very few (3%) of the results
sized that youth who participated in Teen- at home or school or in the community. were done by outside laboratories, and
Cope would have lower HbA1c levels and The sample size was determined by a these results were not significantly differ-
better QOL after 12 months than those power analysis of the primary outcomes ent from those from the DCA, so these
who participated in Managing Diabetes. as well as a mediator analysis that will be data were combined. QOL was measured
reported in a future article. For the pri- by the Pediatric Quality of Life Inventory
RESEARCH DESIGN AND mary analyses, the sample of 320 youth (PedsQL) (teen version)-Core (21), a 23-
METHODSdThe study was a multi- yielded a power of 90% for HbA1c and item measure of global QOL. Higher
site, randomized, parallel-group trial de- 82% for QOL. Institutional review boards scores reflect better QOL. High reliability
signed to evaluate the comparative at all clinical sites reviewed and approved and validity have been established in clin-
efficacy and combined effect of two In- the study. Trained research personnel ap- ical and community samples. Cronbach a
ternet psycho-educational programs for proached youth and parents/guardians in for our sample was 0.87.
youth with type 1 diabetes transitioning the clinic setting and obtained informed
to adolescence (17). The two programs consent/assent. Demographic data were Secondary outcomes
were an Internet CST program (Teen- supplied by parents or guardians at en- Stress/coping was measured with the
Cope) and an Internet diabetes health ed- rollment, and e-mail communication Perceived Stress Scale, a 14-item scale
ucation program (Managing Diabetes) was subsequently established with youth. that measures the degree to which situa-
(18). Each program consisted of five ses- If online data collection was not com- tions in one’s life are appraised as stressful
sions with content tailored to transition- pleted within 3 months of enrollment, (22). Items assess feelings of stress, has-
ing adolescents with type 1 diabetes that youth were considered to be passively re- sles, and coping during the past month.
were released once per week for 5 weeks. fusing study participation. Escalating in- Respondents rate items on a 5-point
TeenCope used a cast of ethnically diverse centives (20–50 USD) were provided for Likert scale ranging from 0 (never) to 4
characters with type 1 diabetes and a completion of each round of data collec- (very often), with higher scores indicative
graphic novel video format to model com- tion. No incentives were provided for of greater perceived stress and less effec-
mon problematic social situations (i.e., completing the programs. tive coping. The reliability estimate in our
parent conflict) and different coping skills After completion of baseline data data was 0.80.
to solve the problems. Content of CST collection, an automated e-mail was sent Coping style in response to diabetes-
was based on our previous studies and to youth and their parents/guardians to related stresses was assessed with the
included communication skills, social identify their group assignment and Responses to Stress Questionnaire (23).
problem solving, stress management, provide a link to the randomly assigned The first 10 items of the measure list

2476 DIABETES CARE, VOLUME 36, SEPTEMBER 2013 care.diabetesjournals.org


Grey and Associates

stressors specific to adolescents with di- order, and sex of child with diabetes) at black/Hispanic/other. Approximately
abetes (24), followed by 57 items asking baseline from the consenting parent/ 50% of families had incomes $80,000
how the individual responds to these guardian. Pubertal status was assessed USD. Nearly 59% of the youth used
stressors. There are three coping factors: with the Pubertal Development Scale pump therapy, and 53% began the study
primary engagement coping (problem (30) to control for the level of pubertal with HbA1c .8%. At baseline, 97 (30%)
solving, emotional modulation, and emo- development, which has been shown to of the subjects had not yet entered pu-
tional expression), secondary control en- correlate well with clinical observations. berty. Figure 1 shows the CONSORT
gagement coping (positive thinking, Diabetes clinical variables, such as length flow diagram.
cognitive restructuring, acceptance, and of time since diagnosis and treatment type The two groups were comparable at
distraction), and disengagement coping (injections or pump), were collected by baseline, with the exception of years of
(avoidance, denial, and wishful think- research staff from the medical record. parental education, with those in Manag-
ing). Proportion scores were used in the Satisfaction was evaluated by youth ing Diabetes having 0.7 years more edu-
analyses to control for response bias with a 6-item survey on how helpful, en- cation. There were, however, differences
and base rates of item endorsement. joyable, easy to use, and worthwhile the among the four sites in race/ethnicity,
Cronbach a ranged from 0.77 to 0.87 in program was. Items were rated on a income, therapy type (pump or injec-
our sample. 5-point Likert-type scale from not at all tions), parent education, and HbA1c, and
The Self-Efficacy for Diabetes Scale to very satisfied, with higher scores indic- these were controlled in the analyses.
measures self-perceptions or expectations ative of higher satisfaction. Scale reliabil- Participation in sessions was high,
held by people with diabetes about their ity was 0.73 in our sample. with 78% of all youth completing four
personal competence, power, and re- of five sessions, and 90% of youth com-
sourcefulness for successfully managing Data analyses pleted at least one session. TeenCope
their diabetes (25). The diabetes-specific The sample and each of the variables were participants completed 82% of sessions
self-efficacy subscale (24 items) was used described using frequency distributions and Managing Diabetes participants com-
in this study, with lower scores indicative and appropriate summary statistics. pleted 74% of sessions, and these differ-
of higher self-efficacy. The reliability co- Group differences at baseline were tested ences were not statistically different. More
efficient was 0.88 in our sample. with t tests or x2. The main hypotheses than one-half (52%) of TeenCope partic-
Self-management was assessed with Self- tested were that youth who participated ipants participated in the discussion
Management of Diabetes - Adolescents, a in the TeenCope program would demon- boards. Satisfaction was high with both
new self-report measure of self-management strate better HbA1c and QOL than those programs, with no significant difference
for adolescents with type 1 diabetes (26) who participated in the Managing Di- between groups. The mean satisfaction
that was developed to encompass a view abetes program. For testing of these score was 3.97 6 0.71 (median = 4) for
of self-management that moves beyond hypotheses, a series of mixed-effects TeenCope and was 3.89 6 0.56 (median =
adherence to treatment regimens. There models (repeated-measures linear regres- 4) for Managing Diabetes. Both groups
are five subscales (Collaboration with Pa- sion with arbitrary within-subject corre- had slight increases in HbA1c levels (P =
rents, Diabetes Care Activities, Diabetes lation structures) in the SAS procedure 0.05) and improved QOL (P , 0.001)
Problem Solving, Diabetes Goals, and Di- MIXED was conducted using an intent- (Table 1) over time, but there were no
abetes Communication) with reliability to-treat approach and a per-protocol significant differences between the two
estimates in our sample ranging from analysis (completion of $4 lessons), con- groups on either of these primary out-
0.62 to 0.80. trolling for sex, age, race/ethnicity, dura- comes over 12 months of follow-up in
Social competence was measured tion, income, therapy type, and site. The the intent-to-treat analyses (Table 2).
with the five-item social acceptance sub- moderation effect of puberty was exam- Thus, the primary hypotheses were not
scale of the Self-Perception Profile for ined by testing the interaction between supported. Mean HbA1c levels increased
Adolescents (27). Statements are scored time and puberty level. Our second and slightly (mean 0.12%) (TeenCope 8.43 6
on a 4-point rating scale, such that high exploratory hypothesis was that youth 1.5% [68.6 mmol/mol] vs. Managing Di-
scores reflect greater perceived compe- who participated in both programs abetes 8.25 6 1.3% [66.7 mmol/mol])
tence. Cronbach a for this sample was would demonstrate better outcomes and stayed .8% in both groups. Those
0.75. compared with youth who participated with baseline HbA1c ,8% at baseline had
The revised Diabetes Family Conflict in only one program. For this analysis, worsened control at 12 months of follow-
Scale was used to evaluate diabetes- youth who completed both programs up (mean change = 0.06%), whereas
related treatment conflict (28). The scale were compared with youth who com- those who had HbA1c $8% at baseline
is a 19-item questionnaire adapted from pleted one program in an intent-to-treat improved over 12 months by 0.5%. This
the Diabetes Responsibility and Conflict analysis and on a per-protocol analysis difference was significant but not affected
Checklist (29) and is used to measure the ($4 lessons for initial program) in a series by group assignment. Pubertal level did
degree of conflict between family mem- of mixed-effects models, controlling for not moderate HbA1c levels.
bers on diabetes management activities. the same variables. There were few differences in second-
Diabetes conflict is rated on a 3-point ary outcomes after 12 months (Table 3).
scale with higher scores indicative of RESULTSdThe final sample of 320 Group-by-time analyses showed that
greater conflict. Cronbach a for this youth had a mean age of 12.3 6 1.1 years youth in Managing Diabetes had signifi-
sample was 0.87. with diabetes duration of 6.1 6 3.5 years. cantly less (P = 0.02) diabetes family con-
Data were also collected on sociode- Mean HbA1c was 8.46 (69.0 mmol/mol) flict than those in TeenCope. There were
mographic data (i.e., ethnicity, socioeco- (61.42%). The sample was 55% female, no other differences in secondary out-
nomic status, number of children, birth 62.2% non-Hispanic white, and 37.8% comes in the group-by-time analyses,

care.diabetesjournals.org DIABETES CARE, VOLUME 36, SEPTEMBER 2013 2477


Internet psycho-education for youth

Figure 1dCONSORT flow diagram. x-over, crossover.

although there were a number of time ef- completed at least four sessions of the re- After 12 months, youth were invited
fects in each group that indicated im- spective program. The results of this per- to cross over and do the other program.
provement over time. We then analyzed protocol analysis were similar to those of To examine whether participating in both
these outcomes using only those who the intent-to-treat analysis. programs was significantly better than

2478 DIABETES CARE, VOLUME 36, SEPTEMBER 2013 care.diabetesjournals.org


Grey and Associates

Table 1dMeans (SD) of primary and secondary outcomes over 18 months: intent-to-treat in examining the relative impact of estab-
analyses, n = 320 lished interventions. Further, with regard
to HbA1c levels, the mean HbA1c for the
Baseline 3 months 6 months 12 months 18 months sample at baseline was just over 8% (63.9
mmol/mol), creating the potential for a
HbA1c floor effect.
TeenCope 8.29 (1.50) 8.32 (1.66) 8.18 (1.65) 8.43 (1.47) 8.59 (1.83) The results after 18 months were
Managing Diabetes 8.15 (1.33) 8.18 (1.19) 8.20 (1.29) 8.25 (1.31) 8.40 (1.31) more in line with our primary hypotheses
QOL than those at 12 months. One intriguing
TeenCope 79.95 (11.01) 80.20 (12.27) 81.68 (12.06) 82.03 (13.51) 83.36 (12.04) possibility is that youth who actually
Managing Diabetes 82.91 (10.32) 85.35 (11.22) 86.31 (9.96) 85.65 (10.02) 85.59 (10.59) completed the second program were dif-
ferent in some way from those who only
did one program, even though analyses of
participating in one program, we com- CONCLUSIONSdWhile both study clinical and demographic data showed no
pared the 128 youth who completed both groups experienced improvement over differences between these groups. Youth
programs with those who completed only the 12 months of follow-up in QOL as who did the second program may have
their initially assigned program (n = 122). well as slightly higher HbA1c levels, the been more motivated and engaged in the
Those who did two programs were not hypothesis that TeenCope would yield process of taking responsibility for self-
different in demographics or clinical superior outcomes after 12 months was management because the timing was right
data than those who did only one pro- not supported. This was unexpected and for them. In addition, it may be that just
gram. With baseline scores controlled may be the result of several factors. First, longer contact, regardless of content, led
for, significant improvements in HbA1c in this comparative effectiveness trial, we to the improved longer-term outcomes.
levels were found for those who com- compared two new relatively sophisti- These results may also suggest that these
pleted both programs compared with cated Internet programs aimed at differ- youth need ongoing support and encour-
those who completed only one (P = ent needs of youth transitioning to agement from nonparents and nonpro-
0.04; one program 8.74 6 1.8% [72.0 adolescence. TeenCope, built on a highly viders to support this transition. Further
mmol/mol] vs. both programs 8.32 6 successful group-based model designed studies comparing such programs are
1.4% [67.3 mmol/mol]) and QOL (P = for a wider range of adolescent ages, fo- necessary.
0.02) (Table 4). Social acceptance (P = cused on teaching a series of coping skills Nonetheless, both interventions re-
0.01), diabetes family conflict (P = 0.02), shown previously to improve both QOL sulted in minimal deterioration in meta-
self-efficacy (P = 0.03), and perceived and HbA1c (16). Managing Diabetes was bolic control over months during the
stress (P = 0.02) were also improved in designed as a program to teach advanced transition period to adolescence. The
those who did the two programs com- diabetes problem solving and healthy life- onset of puberty is associated with de-
pared with those who did only one. After styles using an interactive format. The re- terioration in metabolic control, which is
controlling for race, therapy type, in- sults of the 18-month analyses suggest usually associated with increases in levels
come, and site in addition to baseline that youth in this age-group require of growth hormone associated with sex-
score, participation in both programs both sets of skills to transition success- ual development (5). Thus, in this group
resulted in significantly improved QOL fully to adolescence without the risk of of youth ages 11–14 years, who were en-
(P = 0.045), social acceptance (P = poorer outcomes. Secondly, most previ- tering (30% of the sample) or progressing
0.023), and diabetes family conflict (P = ous studies of behavioral interventions for through puberty (70%), it would be ex-
0.044) and trends in HbA1c (P = 0.16; youth with type 1 diabetes compared the pected that metabolic control would
mean difference 0.3%), self-efficacy (P = new intervention with usual care or a worsen over the 12 months of follow-up
0.07), and perceived stress (P = 0.08) minimally active control condition. In coupled with an increase in family con-
compared with participating in only this comparative effectiveness trial, the flict (31). Both programs resulted in min-
one. Results using intent-to-treat proce- control condition was a very active con- imal worsening of HbA1c levels over 12
dures were similar. dition. Such work is extremely important months as well as improvement in
HbA1c, better QOL, and less family con-
flict with participation in both programs,
Table 2dMixed-effects model of HbA1c and QOL controlling for covariates at 12 months: suggesting that both programs are useful
intent-to-treat analysis, n = 320 during this period of transition for youth.
Nonetheless, HbA1c levels increased by
Numerator Denominator ~0.3% over the course of the study in
df df F P both groups. It is not possible to know
with the design of this study what would
HbA1c
be the likely increase over the same period
Months at visit from first session 1 839 3.91 0.05
without the interventions, but in several
Intervention (TeenCope vs. Managing Diabetes) 1 286 1.18 0.28
previous studies it was reported that in-
Months at visit 3 intervention (TeenCope) 1 839 0.21 0.65
creases of 2% in HbA1c are common dur-
QOL (PedsQL total)
ing adolescence (2,32).
Months at visit from first session 1 645 31.92 ,0.01
Previous reports have suggested that
Intervention (TeenCope vs. Managing Diabetes) 1 291 1.52 0.22
youth from minority and underserved
Months at visit 3 intervention (TeenCope) 1 645 0.66 0.42
groups may participate less in diabetes

care.diabetesjournals.org DIABETES CARE, VOLUME 36, SEPTEMBER 2013 2479


Internet psycho-education for youth

Table 3dCoefficients of change in secondary outcomes at 12 months, controlling for covariates: intent-to-treat analysis, n = 320

TeenCope Managing Diabetes Time effects: Time effects:


Variable Coefficient SE Coefficient SE TeenCope P Managing Diabetes P Group-by-time P
Social acceptance 0.220 0.232 0.713 0.220 ,0.01 ,0.01 0.12
Family conflict 0.801 0.488 21.551 0.500 0.04 0.03 0.02
Self-efficacy 25.606 1.120 23.230 1.080 ,0.01 ,0.01 0.13
Perceived stress 21.010 0.699 20.861 0.201 0.04 0.04 0.81
Coping
Primary control 0.010 0.003 0.006 0.003 ,0.01 0.02 0.30
Secondary control 0.011 0.004 0.007 0.004 ,0.01 ,0.01 0.44
Disengagement 0.007 0.003 0.003 0.002 ,0.01 ,0.01 0.23
Stress reactivity 20.010 0.003 20.009 0.003 ,0.01 ,0.01 0.81
Self-management
Collaboration 22.587 0.466 22.311 0.469 ,0.01 ,0.01 0.68
Activities 23.359 0.402 25.990 0.405 0.37 0.14 0.68
Problem solving 0.625 0.287 0.927 0.290 0.03 ,0.01 0.46
Communication 0.867 0.427 0.212 0.663 0.04 0.62 0.28
Goals 0.744 0.252 0.570 0.253 ,0.01 0.03 0.63

care and behavioral interventions, which that lower-income youth were less likely program alone, further analyses of medi-
has been partially explained by percep- to participate (34). Race/ethnicity and in- ation will allow for examination of model
tions of greater perceived risks for short- come did not moderate outcomes in this testing. It may be that different forms of
term versus long-term complications and study, however. coping are facilitated by the two programs
greater risks to others with diabetes than An important contribution of this that would support the similar outcomes.
to self (33). Sites for this study were se- study is the use of a well-developed For example, Managing Diabetes may
lected to purposefully oversample minor- conceptual framework that guided inter- help with primary engagement coping,
ity youth. In a recent article, our group vention development and measurement. which includes problem solving, but
reported that there was no difference in While secondary outcomes do not appear TeenCope enhances secondary control
participation by sex or race/ethnicity but to be changed by participation in one engagement coping (positive thinking,

Table 4dMeans (SD) of outcomes in youth who participated in one program versus both programs (n = 250)

Baseline 3 months 6 months 12 months 18 months


HbA1c
1 program 8.21 (1.39) 8.31 (1.38) 8.22 (1.44) 8.24 (1.23) 8.74 (1.85)
Both programs 8.23 (1.45) 8.21 (1.51) 8.17 (1.50) 8.40 (1.48) 8.32 (1.37)
QOL (range 0–100)
1 program 81.50 (10.81) 81.99 (12.16) 83.97 (12.00) 82.92 (10.71) 82.11 (11.57)
Both programs 81.28 (10.75) 83.47 (11.88) 84.11 (10.75) 84.35 (12.86) 85.97 (11.07)
Social competence (range 5–20)
1 program 16.35 (3.11) 16.51 (3.37) 16.93 (3.19) 16.66 (3.53) 16.34 (3.38)
Both programs 16.58 (3.24) 16.86 (3.08) 17.22 (3.10) 17.08 (3.17) 17.41 (2.89)
Self-management: activity (range 0–45)
1 program 32.44 (5.37) 31.68 (6.10) 31.65 (5.65) 32.09 (6.35) 30.93 (6.51)
Both programs 33.12 (5.52) 32.48 (5.72) 32.56 (5.56) 32.20 (5.98) 32.16 (6.17)
Self-management: problem solving (range 0–21)
1 program 14.49 (4.17) 14.77 (4.39) 15.06 (3.72) 15.58 (4.03) 15.27 (3.61)
Both programs 14.97 (3.59) 14.82 (3.45) 15.18 (3.45) 15.75 (3.59) 16.19 (3.14)
Diabetes family conflict (range 19–57)
1 program 25.64 (4.90) 25.83 (5.88) 24.72 (5.39) 26.12 (6.52) 26.14 (6.10)
Both programs 25.69 (5.44) 24.39 (4.09) 24.90 (5.81) 24.84 (4.54) 24.27 (4.48)
Diabetes self-efficacy (range 4–96)*
1 program 44.68 (11.22) 42.48 (13.99) 40.64 (14.91) 41.45 (14.43) 40.83 (10.62)
Both programs 44.38 (11.55) 41.99 (12.40) 40.98 (12.33) 39.78 (12.60) 37.15 (11.43)
Perceived stress scale (range 0–56)
1 program 21.19 (7.29) 21.45 (7.03) 20.53 (7.68) 19.16 (7.73) 23.04 (8.41)
Both programs 20.36 (8.29) 20.09 (8.00) 19.19 (8.13) 19.80 (8.53) 19.83 (7.75)

2480 DIABETES CARE, VOLUME 36, SEPTEMBER 2013 care.diabetesjournals.org


Grey and Associates

emotional modulation, etc.), and both investigator. S.J. researched data and reviewed 11. Nielsen J. Usability of websites for teen-
contribute to improved outcomes in ado- and edited the manuscript. K.M., M.S.F., and agers (Jakob Nielsen’s Alertbox) [article
lescents transitioning to independent di- A.D. reviewed and edited the manuscript and online], 2005. Available from http://www
abetes self-management. contributed to the discussion. M.G. is the .useit.com/alertbox/20050131.html. Ac-
guarantor of this work and, as such, had full cessed 8 August 2009
As with any study, there are limita-
access to all the data in the study and takes 12. Siemer CP, Fogel J, Van Voorhees BW.
tions. Despite efforts to recruit a highly responsibility for the integrity of the data and Telemental health and web-based appli-
diverse sample, low-income youth were the accuracy of the data analysis. cations in children and adolescents. Child
more likely to passively refuse to partic- Parts of this study were presented in abstract Adolesc Psychiatr Clin N Am 2011;20:
ipate after consent than those from form at the 37th Annual Meeting of the In- 135–153
higher-income families, probably as a re- ternational Society for Pediatric and Ado- 13. Stinson J, Wilson R, Gill N, Yamada J,
sult of more limited access to the Internet. lescent Diabetes, Miami Beach, Florida, 19–22 Holt J. A systematic review of internet-
Since previous studies led to the conclu- October 2011. based self-management interventions for
sion that low-income youth have poorer youth with health conditions. J Pediatr
metabolic control (35), our sample is bi- Psychol 2009;34:495–510
ased toward youth with better metabolic References 14. Ritterband LM, Tate DF. The science of
1. Amiel SA, Caprio S, Sherwin RS, Plewe G, internet interventions. Introduction. Ann
control. Thus, our findings cannot be
Haymond MW, Tamborlane WV. Insulin Behav Med 2009;38:1–3
generalized to all youth with type 1 dia- resistence of puberty: A defect restricted 15. Mulvaney SA, Rothman RL, Wallston KA,
betes who are transitioning to adoles- to peripheral glucose metabolism. J Clin Lybarger C, Dietrich MS. An internet-
cence. Attrition over 12 months was Endocrinol Metab 1991;72:277–282 based program to improve self-manage-
28%, even with multiple approaches 2. Insabella G, Grey M, Knafl GK, ment in adolescents with type 1 diabetes.
taken to retain subjects. Nonetheless, Tamborlane WV. Transition from ado- Diabetes Care 2010;33:602–604
intent-to-treat procedures were used for lescence to young adulthood in youth 16. Grey M, Boland EA, Davidson M, Li J,
analyses, which mostly likely resulted in with T1D. Pediatr Diabetes 2007;8:228– Tamborlane WV. Coping skills training
conservative results. In addition, the sam- 234 for youth with diabetes mellitus has long-
ple was limited in age range. 3. Karlsson A, Arman M, Wikblad K. Teen- lasting effects on metabolic control and
agers with type 1 diabetesda phenome- quality of life. J Pediatr 2000;137:107–
In summary, the results of this study
nological study of the transition towards 113
indicate that youth with type 1 diabetes autonomy in self-management. Int J Nurs 17. Grey M, Whittemore R, Liberti L,
transitioning to adolescence will partici- Stud 2008;45:562–570 Delamater A, Murphy K, Faulkner MS. A
pate in Internet programs and were highly 4. Schilling LS, Knafl KA, Grey M. Changing comparison of two internet programs for
satisfied. Such programs result in rela- patterns of self-management in youth adolescents with type 1 diabetes: design
tively stable HbA1c levels and improve- with type I diabetes. J Pediatr Nurs 2006; and methods. Contemp Clin Trials 2012;
ment in QOL and perception of social 21:412–424 33:769–776
acceptance, along with a decrease in fam- 5. Amiel SA, Sherwin RS, Simonson DC, 18. Whittemore R, Grey M, Lindemann E,
ily conflictdcritical outcomes at this de- Lauritano AA, Tamborlane WV. Impaired Ambrosino J, Jaser S. Development of
velopmental phase. Research is needed on insulin action in puberty. A contributing an Internet coping skills training pro-
factor to poor glycemic control in ado- gram for teenagers with type 1 diabetes.
how to use such programs in the routine lescents with diabetes. N Engl J Med 1986; Comput Inform Nurs 2010;28:103–
care of youth with type 1 diabetes. 315:215–219 111
6. Grey M. Interventions for children and 19. Grey M, Whittemore R, Jaser S, et al. Ef-
APPENDIXdThe following individu- adolescents with diabetes. Annu Rev Nurs fects of coping skills training in school-age
als and institutions constitute the Teen- Res 2000;22:149–170 children with type 1 diabetes. Res Nurs
7. Anderson B, Loughlin C, Goldberg E, Health 2009;32:405–418
Cope Study Group (1R01NR004009) Laffel L. Comprehensive, family-focused 20. Whittemore R, Jaser SS, Guo J, Grey M.
(*principal investigator or director): Yale outpatient care for very young children A conceptual model of childhood adap-
University School of Nursing, M.G.*, living with chronic disease: Lessons tation to type 1 diabetes. Nurs Outlook
R.W.*, W. Tamborlane, L. Liberti, S. Jaser, from a program in pediatric diabetes. 2010;58:242–251
N. Hunter, S.J., and T. Ma; Children’s Child Serv (Mahwah NJ) 2001;4:235– 21. Varni JW, Seid M, Rode CA. The PedsQL:
Hospital of Philadelphia, K. Murphy* and 240 measurement model for the pediatric
S. Dumser; University of Arizona College 8. Laffel LM, Vangsness L, Connell A, quality of life inventory. Med Care 1999;
of Nursing, M.S.F.*, S. Michaliszyn, and Goebel-Fabbri A, Butler D, Anderson BJ. 37:126–139
E. Crockett; and University of Miami De- Impact of ambulatory, family-focused 22. Cohen S, Kamarck T, Mermelstein R.
teamwork intervention on glycemic con- A global measure of perceived stress.
partment of Pediatrics, A.D.*, J. Hernandez,
trol in youth with type 1 diabetes. J Pediatr J Health Soc Behav 1983;24:385–396
and D. Wile. 2003;142:409–416 23. Connor-Smith JK, Compas BE,
9. Silverstein J, Klingensmith G, Copeland Wadsworth ME, Thomsen AH, Saltzman
K, et al.; American Diabetes Association. H. Responses to stress in adolescence:
AcknowledgmentsdThis research was sup- Care of children and adolescents with type measurement of coping and involuntary
ported by grant R01NR04009 from the Na- 1 diabetes: a statement of the American stress responses. J Consult Clin Psychol
tional Institute of Nursing Research. Diabetes Association. Diabetes Care 2005; 2000;68:976–992
No potential conflicts of interest relevant to 28:186–212 24. Davidson M, Penney ED, Muller B, Grey
this study were reported. 10. Pew Research Center. Trend data for teens M. Stressors and self-care challenges faced
M.G. wrote the manuscript, researched [article online], 2009. Available from by adolescents living with type 1 diabetes.
data, and was a principal investigator. R.W. http://www.pewinternet.org/static-pages/ Appl Nurs Res 2004;17:72–80
reviewed and edited the manuscript, contrib- trend-data-for-teens/whos-online.aspx. 25. Grossman HY, Brink S, Hauser ST. Self-
uted to the discussion, and was a principal Accessed 3 June 2011 efficacy in adolescent girls and boys with

care.diabetesjournals.org DIABETES CARE, VOLUME 36, SEPTEMBER 2013 2481


Internet psycho-education for youth

insulin-dependent diabetes mellitus. Di- in diabetes care. Diabetes 1989;38(Suppl. 33. Patino AM, Sanchez J, Eidson M,
abetes Care 1987;10:324–329 1):28 Delamater AM. Health beliefs and regimen
26. Schilling LS, Dixon JK, Knafl KA, et al. A new 30. Petersen AC, Crockett L, Richards M, adherence in minority adolescents with
self-report measure of self-management of Boxer A. A self-report measure of pubertal type 1 diabetes. J Pediatr Psychol 2005;
type 1 diabetes for adolescents. Nurs Res status: Reliability, validity and initial norms. 30:503–512
2009;58:228–236 J Youth Adolesc 1988;17:117–133 34. Whittemore R, Jaser S, Faulkner M,
27. Harter S. Manual for the Self-Perception 31. Johnson SB, Pollak RT, Silverstein JH, et al. Murphy K, Delamater A, Grey M. Re-
Profile for Adolescents. Denver, CO, Cognitive and behavioral knowledge about cruitment, participation, and satisfaction
Uni- versity of Denver, 1988 insulin-dependent diabetes among children with an eHealth psycho-educational pro-
28. Hood KK, Butler DA, Anderson BJ, Laffel and parents. Pediatrics 1982;69:708–713 gram for youth with type 1 diabetes. J Med
LM. Updated and revised Diabetes Family 32. Bryden KS, Peveler RC, Stein A, Neil A, Mayou Internet Res 2013;15:15–23
Conflict Scale. Diabetes Care 2007;30: RA, Dunger DB. Clinical and psychological 35. Delamater AM, Shaw KH, Applegate EB,
1764–1769 course of diabetes from adolescence to young et al. Risk for metabolic control problems
29. Rubin R, Young-Hyman D, Peyrot M. adulthood: a longitudinal cohort study. Di- in minority youth with diabetes. Diabetes
Parent-child responsibility and conflict abetes Care 2001;24:1536–1540 Care 1999;22:700–705

2482 DIABETES CARE, VOLUME 36, SEPTEMBER 2013 care.diabetesjournals.org