DIABETICMedicine

DOI: 10.1111/dme.14078

Short Report: Educational and Psychological Aspects

A pilot randomized controlled trial of a gratitude
intervention for adolescents with Type 1 diabetes

K. R. Schache1, P. L. Hofman2 and A. S. Serlachius1

1
Department of Psychological Medicine, and 2Liggins Institute, University of Auckland, Auckland, New Zealand

Accepted 13 July 2019

Abstract
Aim Cost-effective psychosocial interventions that can feasibly be implemented into busy clinical settings are needed to
improve psychological and physical health outcomes in adolescents with Type 1 diabetes. We examined the efficacy of a
gratitude journalling intervention to improve psychological well-being and glycaemic control in adolescents aged 10–16
years with Type 1 diabetes.
Methods Eighty adolescents were randomized to the 8-week gratitude intervention (N = 40) or standard care
(N = 40). Self-reported measures of stress, quality of life, self-care, depression and gratitude were assessed at baseline
and 8 weeks after baseline. Glycaemic control (HbA1c) was assessed at baseline and 12 weeks after baseline. A per-
protocol analysis was conducted with the adolescents who completed all questionnaires (N = 60). Analysis of covariance
(ANCOVA) was used to examine differences between treatment arms at follow-up adjusting for baseline scores.
Results There was no evidence of any between-group differences in the psychological or behavioural measures at follow-up
(all P-values > 0.05). Glycaemic control slightly increased in the control group while remaining stable in the gratitude group,
with a between-group difference of 6.1 mmol/mol [95% confidence interval (CI) 2.6 to 14.7; 0.6%, 95% CI 0.2 to 1.3] at
12 weeks after baseline. After adjusting for baseline HbA1c, this between-group difference was significant (P = 0.048).
Conclusions This is the first randomized trial of a gratitude journalling intervention for adolescents with Type 1
diabetes. Gratitude journalling interventions represent a clinically usable approach. If and how it helps to stabilise
glycaemic control in adolescents with Type 1 diabetes remains to be confirmed in future research.
Diabet. Med. 00, 1–5 (2019)

on their efficacy for influencing physiological health outcomes

Introduction
Adolescents with Type 1 diabetes are at a high risk for
psychiatric disorders and suboptimal glycaemic control as they
navigate a new developmental period while managing a chronic
illness [1,2]. Reviews and meta-analyses of psychosocial
interventions for adolescents with Type 1 diabetes have
demonstrated efficacy in improving psychological well-being,
but improvements are often harder to achieve for glycaemic
control [2,3]. There is also a lack of studies that are cost-
effective and can realistically be integrated into clinical care
[4,5]. A promising approach may be gratitude interventions.
Observational studies in healthy populations have found a
consistent association between highgratitude and lower rates of
depression and stress [6,7]. Gratitude interventions have
demonstrated improvements in psychological well-being in
healthy adolescent populations [8,9] and research is emerging
Correspondence to: Anna Serlachius. E-mail: a.serlachius@auckland.ac.nz
[10]. Gratitude interventions come in different formats with the
most common being a gratitude journal [8], where participants
are asked to write a list of things for which they are thankful; the
format also used in the current study. Our objective was to
examine the efficacy of a gratitude journalling intervention for
adolescents with Type 1 diabetes. We hypothesized that the
gratitude intervention would improve psychological well-being
and glycaemic control in comparison with standard care.
Research design and methods
A pilot randomized controlled trial (RCT) of an 8-week
gratitude journalling intervention (treatment group) was
compared with standard care (control group) in 80 adolescents
with Type 1 diabetes between March and June 2017. Ethical
approval was granted by the Health and Disability Ethics
Committee and all participants provided written informed
consent. CONSORT guidelines were followed and the trial

ª 2019 Diabetes UK 1
DIABETICMedicine Gratitude intervention in adolescents with Type 1 diabetes  K. R. Schache et al.
What’s new?
• There is a growing need for clinically usable interven-
tions to improve psychological well-being in adoles-
cents with Type 1 diabetes.
• Gratitude journalling interventions are a novel
approach for improving psychological and physical
health outcomes in adolescents.
• We found no evidence to suggest that the gratitude
intervention improved psychological outcomes in ado-
lescents with Type 1 diabetes.
• The gratitude journalling intervention may help to
stabilize glycaemic control in adolescents with Type 1
diabetes compared with standard care.
• Further research is needed to clarify whether gratitude
journalling interventions are a useful approach to
improve health outcomes in this population.
was registered with the Australian New Zealand Clinical
Trials Registry (ANZCTR) (ACTRN12617000248369).
Participants and randomization
Participants were recruited from three paediatric and ado-
lescent diabetes clinics in Auckland, New Zealand. Inclusion
criteria included being 10–16 years old and diagnosed with
Type 1 diabetes for > 6 months. Exclusion criteria included
non-English-speaking adolescents and adolescents diagnosed
with serious developmental or psychiatric disorders requiring
ongoing treatment. During recruitment 150 adolescents were
approached, of which 80 consented to participate and were
randomized. The most common reason for non-participation
was lack of time to participate in research. Treatment
allocations were computer generated by a biostatistician
independent of the intervention delivery and participants
were randomized using sequentially numbered sealed opaque
envelopes.
Measures and intervention
At baseline, participants completed the following self-report
questionnaires: Diabetes Stress Questionnaire for Youths
(higher scores indicate higher stress) [11], the Centre for
Epidemiological Studies Depression Scale for Children
(higher scores indicate higher levels of depressive symptoms)
[12], the ‘Impact of Diabetes’ subscale from the Diabetes
Quality of Life for Youths scale (higher scores indicate worse
quality of life) [13], the Self-Care Inventory-Revised Version
(higher scores indicate more optimal self-care behaviours)
[14] and the Gratitude Questionnaire-6 (higher scores
indicate greater levels of gratitude) [15]. Glycaemic control
(HbA1c) was also measured as part of their routine clinic visit
2 ª 2019 Diabetes UK
and at their next clinic visit 12 weeks after baseline (window
of 12–16 weeks after baseline). Other covariates which were
assessed included sex, age, ethnicity, comorbidities, insulin
regimen and date of diagnosis.
After completing baseline questionnaires, participants
were randomized into the treatment group (gratitude inter-
vention) or the control group (standard care). The treatment
group was provided with a journal and based on previous
gratitude interventions [8] were asked to write down three
things they were grateful for every day for 8 weeks. The
heading on each page of the journal read ‘Today I am
thankful for:’ and had three lines underneath for adolescents
to write down anything they wanted. Additionally, the pages
were labelled with each day of the week, serving as a cue to
action. Participants were also told that the journals were
confidential and they would not be collected at the end of the
study. During the 8-week trial, no additional reminders were
provided.
Eight weeks after baseline, both groups were asked to
complete the same set of questionnaires. At the follow-up
assessment, an additional question was added to the treat-
ment group’s questionnaires to assess adherence to the
journal (daily/almost every day/two or three times a week/
once a week or less).
Statistical analyses
The sample size was based on a hypothesized difference in
stress levels (primary outcome) at follow-up between the two
groups using an independent samples t-test (one-tailed) with
80% power and a significance level of 0.05. For the power
calculation, we used an effect size from a stress management
intervention for youth [16], which found a large effect in
stress levels between groups (Cohen’s d =0.6). A minimum of
72 participants were required, which was increased to 80 to
allow for attrition.
Eleven participants from the control group and nine
from the treatment group were lost to follow-up and did
not complete their post-intervention questionnaires.
Because of the high attrition, a per-protocol analysis was
conducted, in which the 20 participants who were lost to
follow-up were not included (N = 60). For HbA1c, there
were two missing values at follow-up, which we treated as
missing data. Data were tested for violations of statistical
assumptions. Analysis of covariance (ANCOVA) was used
to examine differences between groups at follow-up
adjusting for baseline scores. Means, standard deviations
(SD) and 95% confidence intervals (CI) are presented with
the analyses.
Results
The baseline characteristics of the two treatment arms
(N = 60) as well as the group lost to follow-up (N = 20)
are presented in Table S1.
 Short Report DIABETICMedicine

Per-protocol Between-group differences at follow-up (per protocol)

The baseline demographic and clinical characteristics were
similar between the treatment and control groups. The
mean age for the total sample (N = 60) was 12.2 years (SD
1.8) and there were 32 girls (53.3%). The majority of the
sample identified as New Zealand European (N = 39,
65%), with eight participants (13.3%) identifying as Maori
(indigenous New Zealanders), four as Pasifika (6.7%) and
nine (15%) as other ethnic groups. The baseline HbA1c for
the total sample was 69 mmol/mol (SD 12.2) (8.4%, SD 1.1).
Thirty-three participants (55%) were on multiple daily
injections and 27 participants (45%) were on pump
therapy.
Lost to follow-up
The adolescents who were lost to follow-up had a higher
baseline HbA1c (78 mmol/mol, SD 18.4; 9.3%, SD 1.7), higher
levels of stress, worse quality of life, and reported lower self-
care behaviours than the participants who completed the
intervention (see Table S1).
Self-reported adherence to the gratitude intervention
Forty-five per cent of adolescents in the treatment group
reported writing in their gratitude journal every day or
almost every day. An additional 26% reported using it two
or three times per week and 29% once a week or less during
the 8-week trial period.
Table 1 depicts the mean differences in outcome measures
between the groups at baseline and 8 weeks after baseline (12
weeks after baseline for HbA1c). As reflected in the mean
differences and 95% CIs, there was no evidence of any group
differences in stress, quality of life, self-care behaviours,
depression or gratitude. For HbA1c there was a between-
group difference of 6.1 mmol/mol (95% CI 2.6 to 14.7;
0.6%, 95% CI 0.2 to 1.3) at 12 weeks after baseline.
Lastly, ANCOVAs were also conducted to examine group
differences at follow-up while adjusting for baseline scores.
The results of the ANCOVA mostly supported the above
findings. The effect of the gratitude journalling intervention
on HbA1c was statistically significant after adjusting for
baseline scores, F(1, 56) = 4.08, P = 0.048. Figure 1
demonstrates that while HbA1c remained stable during the
12 weeks in the gratitude group, HbA1c increased in the
control group.
Supplementary analyses
Although underpowered, subgroup analyses were carried out
to explore whether the lack of between-group differences in
the psychological measures at follow-up might be due to a
ceiling effect (e.g. the GQ-6 has a range between 7 and 42
and both groups scored > 30 at baseline) or due to non-
adherence to the intervention. No between-group differences
were found at follow-up between the participants in the
intervention group who had lower/suboptimal scores on the
psychosocial outcomes at baseline (e.g. using the cut-off for

Table 1 Mean differences in outcome measures at baseline and follow-up (per protocol analysis)

Measures Gratitude N = 31) Control N = 29) Mean difference 95% CI) P-value*

Stress
Baseline 32.8 (8.8) 34.3 (7.8) 1.5 ( 2.8 to 5.8)
8 weeks 31.2 (8.5) 32.5 (8.2) 1.3 ( 3.0 to 5.6) 0.999
Impact of diabetes
Baseline 47.0 (12.5) 48.7 (10.6) 1.6 ( 4.4 to 7.6)
8 weeks 48.2 (13.1) 48.5 (10.2) 0.3 ( 5.8 to 6.4) 0.585
Self-care
Baseline 58.4 (7.6) 57.8 (8.2) 0.6 ( 4.7 to 3.5)
8 weeks 59.3 (5.7) 58.2 (8.2) 1.1 ( 4.7 to 2.6) 0.591
Depression
Baseline 14.9 (12.6) 14.1 (10.7) 0.7 ( 6.8 to 5.3)
8 weeks 13.9 (12.7) 12.2 (7.9) 1.7 ( 7.2 to 3.8) 0.600
Gratitude
Baseline 35.5 (6.6) 34.1 (5.1) 1.4 ( 4.5 to 1.7)
8 weeks 36.3 (6.7) 36.0 (4.4) 0.3 ( 3.2 to 2.7) 0.530
HbA1c
Baseline (mmol/mol) 69 (10.9) 68 (13.7) 0.3 ( 6.6 to 6.1)
Baseline (%) 8.4 (1.0) 8.4 (1.3) 0.02 ( 0.6 to 0.6)
12 weeks (mmol/mol) 68 (10.7) 74 (21.0) 6.1 ( 2.6 to 14.7)
12 weeks (%) 8.3 (1.0) 8.9 (1.9) 0.6 ( 0.2 to 1.3) 0.048*

P-values refer to analysis of covariance (ANCOVA) for between-group comparisons at follow-up. Data are unadjusted means (SD).
*P < 0.05. ANCOVAs adjusting for baseline scores.

ª 2019 Diabetes UK 3
DIABETICMedicine Gratitude intervention in adolescents with Type 1 diabetes  K. R. Schache et al.
90
HbA1c (mmol/mol)
80
70
60
50
Baseline 12 weeks post-intervention
FIGURE 1 Change in HbA1c from baseline to 12 weeks after baseline
for the per protocol analysis (N = 60). Means at 12 weeks are adjusted
for baseline HbA1c. Error bars represent 95% confidence intervals.
Solid line, gratitude group; dashed line, control group.
the depression scale or dichotomizing gratitude) and the
control group (all P > 0.05). Similarly, no differences were
found at follow-up when we compared participants who
adhered to the gratitude journal (wrote in the journal daily/
almost daily, N = 14) and the control group (all P > 0.05).
Lastly, we also compared the adherent participants (N = 14)
with those who wrote in the gratitude journal less regularly
(N = 16) in the intervention group only. Although no between-
group differences were found in the psychological measures at
baseline or 3 months, a group difference was observed for
HbA1c at follow-up after adjusting for baseline scores, F(1,
27) = 4.24, P = 0.049, with the more adherent group demon-
strating lower HbA1c. Although underpowered, these findings
suggest that those participants who adhered to the gratitude
intervention had a greater improvement in HbA1c compared
with those who were less adherent, with a between-group
difference of 9.2 mmol/mol (95% CI 16.6 to 1.9; 0.8%,
95% CI 1.5 to 0.2) at 12 weeks after baseline.
Conclusions
This is the first RCT of a gratitude journaling intervention for
adolescents with Type 1 diabetes. The results did not support
our hypotheses that a gratitude intervention would improve
psychological well-being in adolescents with Type 1 diabetes
compared with standard care; however, the findings did
suggest that a gratitude intervention may have the potential
to stabilize glycaemic control in adolescents with Type 1
diabetes. Gratitude interventions may influence health out-
comes via an increase in healthy behaviours (e.g. improved
diet) [17] or through biological functioning (e.g. inflammation,
blood pressure) [18,19]. Changes in HbA1c due to a psycho-
logical intervention alone are difficult to achieve [3] and so this
pilot study provides a basis for future research into this area.
Unexpectedly we did not see any changes in the psycholog-
ical or behavioural measures, including self-care behaviours or
gratitude, which makes it difficult to ascertain that the
between-group differences in HbA1c were due to the gratitude
4 ª 2019 Diabetes UK
intervention. There are a few possible reasons for why self-
reported gratitude did not improve in the treatment group,
including that gratitude scores at baseline were relatively high
in both groups, indicating that a ceiling effect may have
occurred, although our supplementary analyses did not sup-
port this theory. Another likelihood for the lack of improve-
ments in psychological outcomes include both the high
attrition (N = 20) and relatively high non-adherence to the
gratitude intervention (> 50% of adolescents allocated to the
treatment group did not write in their gratitude journal every
day), therefore diluting the effects of the intervention and
decreasing power. Notably, in our supplementary analyses,
the more adherent adolescents demonstrated significantly
lower HbA1c at follow-up compared with those who were
less adherent, but there were no group differences in the
psychological measures. This provides further evidence that
the gratitude journaling intervention may help to stabilize
HbA1c, but the mechanisms remain unknown. Finally, the
adolescents who were lost to follow-up had higher HbA1c
and reported higher stress and worse quality of life, indicat-
ing that they may have had the most scope to improve. It is
well known that adolescents are a notably hard-to-reach
group and adolescents with Type 1 diabetes are already
overburdened with diabetes-related tasks. Therefore, devel-
oping psychosocial interventions that are feasible and appeal
to this population remains an important goal for future
research.
Strengths of this study include its clinical utility. Many
psychological interventions for people living with chronic
disease are not translated well or translatable at all into a
clinical setting [5]. Limitations of this study include its short-
term follow-up, high levels of attrition and relatively high
levels of non-adherence to the journalling intervention. We
also did not have an active control group. We plan to follow-
up this cohort of adolescents to examine if these results are
maintained. Future research is required to confirm these
findings in a larger sample with an active control group and
to examine the underlying pathways through which gratitude
may be influencing glycaemic control.
Funding sources
None.
Competing interest
None declared.
Ethical approval
The study was approved by the Health and Disability Ethics
Committee [Auckland, New Zealand], and informed consent
was obtained from all participants. This research study was
conducted in accordance with the guidelines of the Declara-
tion of Helsinki
 Short Report
Acknowledgements
Sincere thanks to the adolescents and their families who
participated in this study.
Author contributions
K.S., P.H. and A.S. came up with the conception and design.
K.S. recruited the participants and acquired the data. K.S.
and A.S. conducted the analyses and wrote the manuscript.
P.H. reviewed and edited the manuscript.
References
1 Rausch JR, Hood KK, Delamater AM, Pendley JS, 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; 35: 1219–1224.
2 Delamater AM, de Wit M, McDarby V, Malik J, Acerini CL.
ISPAD Clinical Practice Consensus Guidelines 2014. Psychological
care of children and adolescents with type 1 diabetes. Pediatr
Diabetes 2014; 15(Suppl 20): 232–244.
3 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; 33: 1658–1664.
4 Hampson SE, Skinner TS, Hart JO, Storey L, Gage H, Foxcroft D
et al. Behavioral interventions for adolescents with type 1 diabetes:
how effective are they? Diabetes Care 2000; 23: 1416–1422.
5 Leventhal H, Weinman J, Leventhal EA, Phillips LA. Health
psychology: the search for pathways between behavior and health.
Ann Rev Psychol 2008; 59: 477–505.
6 Wood A, Maltby J, Gillett R, Linley PA, Joseph S. The role of
gratitude in the development of social support, stress, and
depression: two longitudinal studies. J Res Pers 2008; 42: 854–871.
7 Emmons RA, Mishra A. Why gratitude enhances well-being: what we
know, what we need to know. In: Sheldon KM, Kashdan TB, Steger MF,
editors. Designing Positive Psychology: Taking Stock and Moving
Forward. New York: Oxford University Press, 2011: 248–262.
8 Wood AM, Froh JJ, Geraghty AW. Gratitude and well-being: a
review and theoretical integration. Clin Psychol Rev 2010; 30:
890–905.
ª 2019 Diabetes UK
DIABETICMedicine
9 Froh JJ, Sefick WJ, Emmons RA. Counting blessings in early
adolescents: an experimental study of gratitude and subjective well-
being. J School Psychol 2008; 46: 213–233.
10 Schache K, Consedine N, Hofman P, Serlachius A. Gratitude –
more than just a platitude? The science behind gratitude and
health. Br J Health Psychol 2019; 24: 1–9.
11 Delamater AM, Smith JA, Lankester L et al. (1988). Stress and
metabolic control in diabetic adolescents. Paper presented at the
9th Annual Meeting of Behavioral Medicine, Boston, MA.
12 Weissman MM, Orvaschel H, Padian N. Children’s symptom and
social functioning self-report scales comparison of mothers’ and
children’s reports. J Nerv Ment Dis 1980; 168: 736–740.
13 Ingersoll GM, Marrero DG. A modified quality-of-life measure for
youths: psychometric properties. Diabetes Educ 1991; 17: 114–
118.
14 Lewin AB, LaGreca AM, Geffken GR, Williams LB, Duke DC,
Storch EA et al. Validity and reliability of an adolescent and
parent rating scale of type 1 diabetes adherence behaviors: the
Self-Care Inventory (SCI). J Pediatr Psychol 2009; 34: 999–
1007.
15 McCullough ME, Emmons RA, Tsang J-A. The grateful disposi-
tion: A conceptual and empirical topography. J Pers Soc Psychol
2002; 82: 112–127.
16 Boardway RH, Delamater AM, Tomakowsky J, Gutai JP. Stress
management training for adolescents with diabetes. J Pediatr
Psychol 1993; 18: 29–45.
17 Fritz MM, Armenta CN, Walsh LC, Lyubomirsky S. Gratitude
facilitates healthy eating behavior in adolescents and young adults.
J Exp Soc Psychol 2019; 81: 4–14.
18 Redwine LS, Henry BL, Pung MA, Wilson K, Chinh K, Knight B
et al. Pilot randomized study of a gratitude l intervention on heart
rate variability and inflammatory biomarkers in patients with stage
B heart failure. Psychosom Med 2016; 78: 667–676.
19 Jackowska M, Brown J, Ronaldson A, Steptoe A. The impact of a
brief gratitude intervention on subjective well-being, biology and
sleep. J Health Psychol 2016; 21: 2207–2217.
Supporting Information
Additional supporting information may be found online in
the Supporting Information section at the end of the article.
Table S1. Baseline characteristics of the two treatment arms
(N = 60) and the participants lost to follow-up (N = 20).
5