Abstract
Individuals with type 1 diabetes (T1D) must engage in a variety of complex and burdensome self-management behaviors daily to maintain near normal blood glucose levels and prevent complications. There is a need for interventions to improve use of sophisticated diabetes technologies, such as insulin pumps, during adolescence - a very high-risk developmental period for individuals with T1D. All diabetes devices, including insulin pumps, store large amounts of behavioral data that can be downloaded and analyzed to evaluate adherence to recommended T1D self-management behaviors. The overall objective of the present study, Pump it Up!, was to use objectively downloaded insulin pump data to inform and test two interventions to optimize insulin pump use in adolescents with T1D and their caregivers. Multiphase Optimization Strategy (MOST) was used to achieve the overall goal of this study – to separately test the main effect of the Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and the main effect of Pump It Up! Problem-Solving Skills intervention to improve T1D self-management behaviors using a 2 × 2 factorial design. The purpose of this paper is to describe the Pump It Up! study design and rationale, and participant baseline characteristics. Longitudinal data analyses will be conducted, and moderating effects of psychosocial factors will be examined in relation to primary (insulin pump self-management behaviors) and secondary (A1C) outcomes.
1. Background and rationale
Type 1 diabetes mellitus (T1D) is an autoimmune disorder that occurs when the body destroys the insulin-producing beta cells in the pancreas, resulting in little to no insulin production. Without insulin, glucose is unable to be transferred into cells and used for energy. Instead, glucose remains in the blood stream and the affected individual develops hyperglycemia (i.e., abnormally high blood glucose levels). The goals of T1D self-management are to: 1) achieve blood glucose levels between 70 and 180 mg/dL on a daily basis; 2) achieve A1C <7% (i.e., an indicator of glycemic control providing an average of blood glucose during the previous 2–3 months); and 3) reduce risk for acute- (e.g., severe hypo- and hyperglycemia) and long-term complications (e. g., retinopathy, neuropathy, organ damage). To achieve these goals, individuals with T1D must engage in a variety of complex and burdensome self-management behaviors including monitoring glucose, counting carbohydrates, administering insulin via injection or insulin pump, and engaging in exercise. All diabetes devices, including blood glucose and continuous glucose monitors and insulin pumps store large amounts of behavioral data that can be downloaded and analyzed. However, use of these objective data to examine individual T1D self-management behaviors and the complicated relationship among them is limited [1–3] despite common weaknesses associated with self-report of these behaviors [4,5].
1.1. Conceptual framework: T1D self-management behaviors and adherence
Adherence is conceptualized as the extent to which an individual’s behaviors are consistent with medical recommendations. Historically, A1C has been used as an indicator of adherence, which is problematic because an individual with T1D can be “very adherent,” but still have suboptimal A1C because of puberty, stress, illness, and insulin resistance. Therefore, using A1C as an indicator of adherence is not currently recommended because T1D self-management behaviors (or lack thereof) are the actual precursors to decreases or increases in A1C [6].
As diabetes technology advances, some may argue that the impact of human behavior on T1D health outcomes will be substantially reduced or even eliminated especially as hybrid closed loop systems evolve into the artificial pancreas. This argument is flawed because: 1) it fails to take into account that there will always be people who do not adopt new technologies; 2) some individuals experience significant barriers resulting in discontinuation of technologies; and 3) even the most advanced technology can only be as effective as the person who uses it. Currently only 63% of individuals with T1D in the United States use insulin pumps and even fewer, 30%, use continuous glucose monitors [7]. In addition, engagement in self-management behaviors required for use of these sophisticated technologies is far from optimal [1–3,8,9]. Indeed, the majority of children and adolescents using insulin pumps need some intervention to optimize their use [1] and Pump It Up! is the first study to provide targeted intervention to improve specific T1D insulin pump self-management behaviors.
2. Study aims
Self-report questionnaires (e.g., Diabetes Self-Management Profile, Self-Care Inventory) are the most common method used to assess T1D self-management behaviors and adherence despite their inherent weaknesses including inaccurate recall and social desirability [4]. In contrast, objective data from diabetes devices, including insulin pumps, can be downloaded and analyzed to describe T1D self-management behaviors (e.g., dates, times, and values for blood glucoses, carbohydrates, insulin boluses). Using data downloaded from T1D devices provides the closest objective approximation to an individual’s completion of actual self-management behaviors. Studies that use objective data are typically limited to examining frequency of blood glucose monitoring [6]; however, use of more sophisticated objective data is slowly increasing [1–3,8,10,11].
The overall objective of Pump It Up! was to use objectively downloaded insulin pump data to inform and test two interventions to optimize insulin pump use in adolescents with T1D and their caregivers. The first intervention tested the hypothesis that T1D self-management behaviors would improve in participants who received a Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report compared to those in the Standard of Care group (Control) who did not receive a personalized report. The second intervention tested the hypothesis that T1D self-management behaviors would improve in participants who engaged in Pump It Up! Problem-Solving Skills sessions to identify a specific goal to improve T1D self-management behaviors compared to those in the Standard of Care group who did not engage in problem-solving. In addition, an exploratory study aim was to assess how psychosocial variables, including fear of hypoglycemia, insulin pump and T1D responsibility, depressive symptoms, and quality of life, moderated the effect of the two interventions on primary (i.e., T1D self-management behaviors) and secondary (i.e., A1C) outcomes.
3. Research design and methods
Multiphase Optimization Strategy (MOST) is a comprehensive conceptual framework for optimizing and evaluating the components of behavioral interventions [12]. The overall purpose of MOST is to identify the effective components of an intervention using a 2 × 2 factorial design. After determining whether an intervention component is effective, only the intervention(s) that work are then tested in a larger randomized clinical trial [13], which ultimately saves resources by not including an intervention that does not work. MOST was used to achieve the overall goal of this study – to test the main effect of the Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and then to separately test the main effect of Pump It Up! Problem-Solving Skills - to improve T1D self-management behaviors.
3.1. Participants
Pump It Up!, an unblinded randomized clinical trial, was conducted in pediatric endocrinology clinics in Florida and Colorado from 2012 to 2017. Adolescents were eligible to participate based on the following criteria: 1) age 10–18 years; 2) T1D diagnosis >1 year; 3) use of a Medtronic insulin pump >6 months; 4) willingness to provide caregiver written informed consent and adolescent written assent; 5) English fluency; and 6) parent(s) agreed to be present during the clinic visit. Exclusion criteria included adolescents with a developmental delay or intellectual disability. There were no exclusion criteria imposed on eligible participants regarding T1D self-management behaviors or A1C.
Of the 130 adolescent-caregiver dyads approached, 106 dyads enrolled (82% recruitment rate); n = 75 (71%) in Florida and n = 31 (29%) in Colorado. All 106 dyads were randomized, and 6 dyads withdrew resulting in 93.5% retention. Reasons for declining participation included disinterest in research and lack of time. Reasons for withdrawal included family relocation, no longer attending clinic, discontinuation of the insulin pump, and not wanting to spend extra time in clinic to complete study visits.
3.2. Randomization
Stratified block randomization scheme was computer generated by a study statistician with block sizes of 2 or 4, chosen randomly within each stratum (Florida, Colorado). Consistent with a 2 × 2 factorial design, the first randomization occurred at Visit 1; n = 52 participants to Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and n = 54 to Standard of Care. Every participant was re-randomized at Visit 3 resulting in n = 49 participants to Pump It Up! Problem Solving Skills and n = 51 to Standard of Care. It is important to note that a 2 × 2 factorial design is more efficient than a 4-arm randomized clinical trial because the former requires far fewer participants to test the main effects compared to the number of participants required to achieve adequate power to test comparisons in a 4-arm trial. Cost savings also occur when fewer participants are needed to maximize power.
As expected with a 2 × 2 factorial design, the sample size in each cell was: 1) n = 25 did not receive either intervention; 2) n = 26 received only Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report; 3) n = 25 received only Pump It Up! Problem-Solving Skills; and 4) n = 23 received Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and Pump It Up! Problem-Solving Skills (Table 1). All 6 participants that withdrew from the study did so before the second randomization (Visit 3). See Fig. 1 for the study CONSORT diagram.
Table 1.
Pump It Up! 2 × 2 factorial design.
| Pump It Up! Personalized T1D Self-Management Behaviors Feedback report | |||
|---|---|---|---|
| NO | YES | ||
| Pump It Up! Problem-Solving Skills | NO | Group 1: No intervention | Group 2: Feedback Report only |
| YES | Group 3: Problem-Solving Skills only | Group 4: Feedback Report + Problem-Solving Skills | |
Fig. 1.
CONSORT diagram.
3.3. Procedures
This study was approved by the Florida State University Institutional Review Board and the University of Colorado Multiple Institutional Review Board. This study was registered on the clinicaltrials.gov website (NCT02333539).
Adolescents with T1D using insulin pumps and their caregiver(s) participated in 5 study visits immediately after their routine T1D clinic visits, which occurred approximately every 3 months, for a total study duration of 12 months. All participants received Pump it Up! Education at Visit 1. All participants, regardless of randomization group, were exposed to the Medtronic Carelink report during their all clinic visits with the T1D medical provider. Those randomized to the Pump it Up! Personalized T1D Self-Management Feedback Report, but not Standard of Care, received feedback reports at Visits 1 and 2 during the first 6 months of the study (Table 1). All participants were then rerandomized at Visit 3 and those randomized to Pump it Up! Problem-Solving Skills, but not Standard of Care, participated in problem-solving skills sessions at Visits 3 and 4 during the second 6 months of the study (Table 1). No interventions occurred at Visit 5, which was a follow-up visit where final study measures were collected. All participants completed a battery of questionnaires at each visit. Adolescents were compensated for their participation using a graduated incentive schedule; they received $5 at baseline with an increase of $5 per study visit. Total possible compensation was $75.
3.4. Pump It Up! Education
To eliminate the possibility that weaknesses in knowledge and insulin pump skills contributed to suboptimal engagement in T1D self-management behaviors, every participant regardless of randomization group completed three assessments developed specifically for this study: Insulin Pump Knowledge Questionnaire, Carbohydrate Counting Assessment, and Blood Glucose Monitoring Skills Assessment. The assessments were developed by T1D experts (KAD and SBJ) and then reviewed by local certified diabetes educators to ensure that they accurately reflected diabetes devices and what was required of them for proper use at the time the study was completed. When incorrect answers were provided or blood glucose monitoring tasks were skipped (e.g., washing hands), education was provided to rectify any knowledge or skills weaknesses. In addition, two handouts were provided, Insulin Pump Tricks and Tips and Carbohydrate Counting Results, to reinforce correct responses.
3.5. Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report
T1D providers used Medtronic Carelink reports for routine clinical care; however, weaknesses of the Carelink report included lack of cumulative data across time and a complicated visual presentation that was not patient-friendly making it difficult to interpret data. In fact, some T1D providers are uncertain about how to interpret and use the reports in clinical care and may not be able to invest the time necessary to become proficient in interpreting the data [14,15].
Thus, the Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report was designed specifically for this study. First, the participant’s insulin pump was downloaded into the Medtronic Carelink software program. Then, the export feature within the Medtronic Carelink software program was used to download the participant’s unique comma-separated values (csv) file. Next, the csv file was uploaded into a Pump it Up! computer program created for this study and programmed using Python language. The final Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report was then created.
The Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report consisted of 2 pages of summary data and 1 page of personalized recommendations with presentation of data that was easy to understand and clearly highlighted the relationship between T1D insulin pump self-management behaviors in the context of optimal goals. Summary data from insulin pumps focused on frequency of blood glucose readings, carbohydrate entries, insulin bolusing, and or administering insulin in response to high blood glucoses (151–249 mg/dL) or very high blood glucoses (≥250 mg/dL) blood glucoses with and without food (see 3.7 Primary and Secondary Outcomes). A printed copy of the Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report was reviewed with and provided to the participant and caregiver(s).
3.6. Pump It Up! Problem-Solving Skills
The goal of engaging in two Pump It Up! Problem-Solving Skills sessions was to improve communication between adolescents and caregivers, identify a realistic and achievable goal, and develop a feasible behavior change and action plan. Specifically, adolescents and their caregivers collaborated to choose one goal which typically involved increasing the frequency of blood glucose monitoring, carbohydrate inputs, insulin bolusing, or administering insulin in response to high (150–249 mg/dL) or very high (≥250 mg/dL) blood glucoses. The interventionist then assisted families with generating multiple solutions to achieve the goal and then selecting one solution to implement. As part of Pump It Up! Problem-Solving Skills, participants were provided with a printed summary of the goal chosen, the solution generated, and what role each family member would play in achieving the goal. Problem-Solving Skills intervention is derived from Behavioral Family Systems Therapy (BFST) [16] and has been applied successfully with adolescents with T1D [17].
3.7. Measures
3.7.1. Demographic and clinical characteristics
Sociodemographic data (e.g., racial/ethnic identity, household income, insurance type) were collected using a standardized interview conducted at Visit 1. Clinical data including height, weight, T1D duration, and duration of insulin pump use were obtained from medical record review.
3.7.2. Primary outcomes
The five primary outcomes in Pump It Up! were multiple T1D insulin pump self-management behaviors assessed at 6- and 12-months including mean frequency per day: 1) blood glucose readings; 2) carbohydrate entries; 3) insulin boluses administered; and insulin boluses after: 4) high (150–249); and 5) very high (≥250) blood glucose readings. A panel of T1D physicians, nurses, and dietitians recommended ≥4 blood glucose readings per day on 90% of days, ≥3 carbohydrate entries per day on 90% of days, and ≥3 insulin boluses per day on 90% of days and correcting after high and very high blood glucoses >70% of the time. These recommendations incorporated individual lifestyles such as exercise frequency (i.e., someone who exercises to reduce blood glucose may not always bolus when blood glucose is high). These data were collected by downloading the participant’s Medtronic insulin pump at each study visit, which occurred at the time of the participant’s routine T1D clinic visit, scheduled approximately 3 months apart. At the time this study was conducted, Medtronic insulin pumps typically stored 2–3 months of data. Table 4 shows the percentage of participants that met T1D self-management behavior goals prior to the implementation of any intervention.
Table 4.
Self-report questionnaires completed by participants and their caregivers.
| Measure | Construct measured | Adolescent α | Caregiver α |
|---|---|---|---|
| Depressive Symptoms | |||
| Children’s Depression Inventory 2 [25,26] | 27-items; assesses depressive symptoms in youth ages 6–17 years. Higher scores = greater depressive symptoms. | 0.81 | – |
| Beck Depression Inventory-II [27] | 21-item; assesses depressive symptoms for youth ≥18 years. Higher scores = greater depressive symptoms. | Unable to calculate due to n = 2 | – |
| Center for Epidemiological Studies Depression Scale [28] | 20-item; assesses depressive symptoms in adults. Higher scores = greater depressive symptoms. | – | 0.88 |
| Anxiety Symptoms | |||
| Hypoglycemia Fear Survey [29,30] | 25 items; Total scores = 0–100; 3 subscales: Maintain High Blood Glucoses, Helplessness About Low Blood Glucoses, Worry About Social Consequences. Higher scores = greater fear of hypoglycemia. | Total 0.84 Maintain High BG 0.85 Helplessness 0.85 Social Consequences 0.83 | Total 0.85 Maintain High BG 0.85 Helplessness 0.90 Social Consequences 0.69 |
| Responsibility for T1D Tasks | |||
| Diabetes Family Responsibility Questionnaire [31] | 17-items; assesses general T1D responsibility. Higher scores = greater caregiver responsibility. | 0.81 | 0.85 |
| Insulin Pump Responsibility Questionnaire* | 16-items; assesses T1D responsibilities specific to monitoring blood glucoses and using an insulin pump. Higher scores = greater caregiver responsibility. | 0.88 | 0.92 |
| T1D-Specific Quality of Life | |||
| Peds QL Diabetes Module [32,33] | 28-items; assesses quality of life in 8 – 18 year-olds; yields 1 Total score and 5 subscales: Diabetes, Treatment I, Treatment II, Worry, Communication. Higher scores = better quality of life. | Total 0.91 Diabetes 0.82 Treatment I 0.76 Treatment II 0.74 Worry 0.79 Communication 0.80 | – |
| Well-being and Satisfaction of CAREgivers of Children with Diabetes Questionnaire [34] | 37-items; assesses quality of life in caregivers; yields 4 subscales: Psychosocial Well-being, Treatment Satisfaction, Acceptance of Insulin Administration, Ease of Insulin Use. Higher scores = better quality of life. | – | Total 0.90 Psychosocial Well-being 0.88 Treatment Satisfaction 0.69 Acceptance of Insulin Administration 0.62 Ease of Insulin Use 0.81 |
| Skills Assessment | |||
| Blood Glucose Meter Skills* | 17-items; assesses accurate use of blood glucose meters and if best practices are followed. Higher scores = better skills. | N/A | – |
| Carbohydrate Counting Skills* | Assesses accuracy of carbohydrate counting. Higher scores = better accuracy. | N/A | – |
| Insulin Pump Knowledge Questionnaire* | 20-items; assesses insulin pump knowledge. Higher scores = greater knowledge. | N/A | N/A |
| Child Pubertal Status | |||
| Pubertal Development Scale [35] | Self-report of pubertal status yielding estimate of Tanner stage. Higher scores = further along in puberty. | Boys 0.84 Girls 0.67 | – |
The full battery of questionnaires was completed at Visits 1, 3, and 5. Only the Insulin Pump Responsibility and Diabetes Family Responsibility Questionnaires were completed at Visits 2 and 4. The Insulin Pump Knowledge Questionnaire was completed at Visits 1, 2, and 5.
α = Cronbach’s reliability statistic.
Indicates measures created specifically for Pump it Up!.
3.7.3. Secondary outcome
One secondary outcome, Hemoglobin A1C, was collected at point of care using a Siemens Healthcare Diagnostics DCA Vantage (reference range 4.2–6.5%), which is National Glycohemoglobin Standardization Program certified as having documented traceability to the Diabetes Control and Complications Trial Reference Method. The average A1C at baseline was 8.8% with 12.3% meeting recommended goal of A1C <7.5% (the American Diabetes Assocation recommended goal for A1C at the time this study was conducted).
3.7.4. Psychosocial variables and T1D skills
Adolescents and caregivers completed a full or reduced battery of questionnaires at all study visits which took between 5 and 15 min to complete (Table 2). Parents and adolescents completed questionnaires to assess depressive symptoms, fear of hypoglycemia, T1D and insulin pump responsibility, quality of life, and insulin pump knowledge. Meter skills were assessed by having the adolescent check their blood glucose in the presence of a study team member who then scored skills based on standardized criteria. Carbohydrate counting skills were assessed using a 24-h diet recall. Reliability analysis of each questionnaire was assessed using Cronbach’s alpha; all alphas were acceptable to strong [18] (Table 5). A Post-Intervention Feasibility-Acceptability Questionnaire was completed at the final study visit.
Table 2.
Participants meeting goals for insulin pump self-management behaviors at Visit 1.
| T1D self-management insulin pump goals | Met goal n (%) |
|---|---|
| 1. ≥4 blood glucose checks/day on ≥90% days | 59 (55.7%) |
| 2. ≥3 carbohydrate entries/day on ≥90% days | 41 (39.4%) |
| 3. ≥3 boluses administered/day on ≥90% days | 59 (55.7%) |
| 4. Bolusing after high blood glucose ≥70% of the time | 25 (24.0%) |
| 5. Bolusing after very high blood glucose ≥70% of the time | 61 (61.0%) |
| Total Number of Goals Met | |
| 0 goals | 16 (16.5%) |
| 1 goal | 14 (14.4%) |
| 2 goals | 15 (15.5%) |
| 3 goals | 26 (26.8%) |
| 4 goals | 20 (20.6%) |
| 5 goals | 6 (6.2%) |
Note: percentages are valid percentages.
Table 5.
Baseline characteristics for intervention and standard of care participants at Visit 1.
| Standard of Care | Intervention | Overall | |
|---|---|---|---|
| Demographic Characteristics M(SD) or n (%); range | |||
| Adolescent age (years) | 14.0 (2.4) | 13.7 (2.2) | 13.8 (2.2); 10.2–18.6 |
| (Female) | 23 (42.6%)* | 35 (66.0%)* | 58 (54.7%) |
| Insurance type (public insurance) | 12 (22.6%)* | 23 (42.6%)* | 35 (33.0%) |
| Family income | |||
| <$20,000 | 4 (7.5%) | 4 (7.5%) | 8 (7.5%) |
| $20,000–$39,000 | 10 (18.5%) | 4 (7.5%) | 14 (13.2%) |
| $40,000–$59,000 | 6 (11.1%) | 3 (5.7%) | 9 (8.5%) |
| $60,000–$79,999 | 8 (14.8%) | 12 (22.6%) | 20 (18.9%) |
| $80,000–$99,999 | 6 (11.1%) | 9 (17.0%) | 15 (14.2%) |
| >$100,000 | 19 (35.2%) | 15 (28.3%) | 34 (32.1%) |
| Did not report | 1 (1.9%) | 6 (11.3%) | 6 (5.7%) |
| Marital status (caregivers married) | 41 (75.9%) | 42 (72.9%) | 82 (78.9%) |
| Education level of primary caregiver (education > high school diploma) | 29 (53.7%) | 31 (58.5%) | 59 (56.2%) |
| Years of education primary caregiver | 14.5 (2.5) | 14.4 (2.4) | 14.4 (2.5); 12.0–22.0 |
| Caregiver attending appointment (female) | 44 (83.0%) | 43 (79.6%) | 87 (82.1%) |
| Number of people living in home | 4.5 (1.1) | 4.1 (1.1) | 4.1 (1.1); 1.0–7.0 |
| T1D Clinical Characteristics M(SD); range | |||
| T1D duration (yrs) | 6.7 (3.4) | 6.2 (3.4) | 6.5 (3.5); 1.0–14.1 |
| Insulin pump use duration (yrs) | 4.2 (3.2) | 4.3 (3.2) | 4.3 (3.2); 0.4–13.0 |
| A1C% | 8.6 (1.3) | 8.9 (1.2) | 8.8 (1.5); 6.6 – ≥14.0 |
| Average blood glucose checks/day | 4.0 (2.2) | 4.2 (2.1) | 4.1 (2.1); 0.2–9.0 |
| Average carbohydrate entries/day | 3.6 (2.0) | 3.5 (1.5) | 3.5 (1.7); 0.1–8.5 |
| Average insulin boluses administered/aay | 5.1 (2.4) | 4.5 (1.8) | 4.8 (2.1); 0.7–12.2 |
| Insulin boluses administered following a high blood glucose (i. e., 151–249 mg/dL) | 46.3 (31.4) | 43.5 (31.0) | 44.9 (31.1); 0–100 |
| Insulin boluses administered following an extremely high blood glucose (i.e., >250 mg/dL) | 70.8 (24.0) | 71.0 (23.7) | 70.9 (23.7); 0–100 |
| Pubertal Development Scale | |||
| Boys & Girls | 3.3 (1.1) | 3.4 (0.9) | 3.3 (1.0); 1–5 |
| Boys | 2.8 (1.1) | 2.9 (0.9) | 2.8 (1.0); 1–4 |
| Girls | 3.9 (0.7) | 3.5 (0.8) | 3.7 (0.8); 2–5 |
| Adolescent Questionnaires; M(SD); range | |||
| Children’s Depression Inventory 2 Total Score | 5.8 (4.6) | 6.7 (5.9) | 6.2 (5.3); 0–23 |
| Hypoglycemia Fear Survey | |||
| Maintain High Blood Glucoses | 4.6 (3.1) | 3.9 (3.0) | 4.2 (3.1); 0–12 |
| Helplessness About Low Blood Glucoses | 10.3 (6.4)* | 13.8 (7.7)* | 12.0 (7.2); 1–32 |
| Worry About Social Consequences | 5.6 (3.8) | 6.6 (5.2) | 6.1 (4.6); 0–20 |
| Diabetes Family Responsibility Questionnaire | 32.1 (5.4) | 32.0 (5.2) | 32.0 (5.2); 17–44 |
| Insulin Pump Responsibility Questionnaire | 25.8 (6.4) | 24.9 (6.4) | 25.4 (6.4); 16–42 |
| Peds QL Diabetes Module | |||
| Total Score | 69.9 (14.8) | 67.4 (17.3) | 68.8 (15.9); 20.9–94.2 |
| Diabetes | 56.8 (15.2) | 59.4 (16.6) | 58.0 (15.9); 22.7–95.5 |
| Treatment I | 78.6 (18.9) | 74.9 (23.0) | 76.8 (21.0); 6.3–100 |
| Treatment II | 69.2 (12.4) | 67.8 (17.8) | 68.6 (15.0); 25.4–85.7 |
| Worry | 69.1 (24.1) | 64.8 (26.3) | 67.0 (25.2); 0–100 |
| Communication | 76.3 (24.5) | 73.9 (24.2) | 75.1 (24.3); 0–100 |
| Blood Glucose Meter Skills | 12.7 (2.1) | 12.8 (1.9) | 12.7 (2.0); 8–18 |
| Carbohydrate Counting Skills | 59.8 (19.0) | 64.0 (18.1) | 61.9 (18.6); 11.1–100 |
| Insulin Pump Knowledge Questionnaire | 68.8 (6.7) | 69.6 (6.7) | 68.2 (6.7); 49–83 |
| Caregiver Questionnaires M(SD); range | |||
| Center for Epidemiological Studies Depression Scale | 10.6 (9.1) | 9.2 (7.7) | 9.9 (8.5); 0–41 |
| Hypoglycemia Fear Survey | |||
| Maintain High Bkood Glucoses | 3.3 (2.5) | 3.7 (2.4) | 3.5 (2.4); 0–12 |
| Helplessness About Low Blood Glucoses | 17.1 (8.0) | 18.7 (9.0) | 18.0 (8.5); 0–40 |
| Worry About Social Consequences | 5.0 (3.8) | 4.8 (3.4) | 4.9 (3.6); 0–20 |
| Diabetes Family Responsibility Questionnaire | 34.8 (5.0) | 35.2 (5.7) | 35.0 (5.3); 17–46 |
| Insulin Pump Responsibility Questionnaire | 26.0 (7.4) | 26.4 (7.6) | 26.2 (7.5); 16–43 |
| Well-being and Satisfaction of CAREgivers of Children with Diabetes Questionnaire | 122.1 (11.8) | 121.6 (15.7) | 121.8 (13.7); 79–144 |
| Psychosocial Well-Being | 42.5 (7.8) | 43.3 (8.2) | 42.9 (8.0); 18–52 |
| Treatment Satisfaction | 25.5 (3.2) | 24.8 (3.8) | 25.2 (3.5); 16–32 |
| Acceptance of Insulin Administration | 22.5 (1.6) | 21.9 (2.9) | 22.2 (2.3); 11–24 |
| Ease of Insulin Use | 31.2 (4.3) | 30.5 (4.9) | 30.9 (4.6); 17–36 |
| Insulin Pump Knowledge Questionnaire | 75.5 (4.9) | 74.7 (5.8) | 75.1 (5.3); 60–85 |
p < 0.05.
3.8. Analysis of group differences
Means, standard deviations, and ranges for all of the variables included in Pump It Up! are provided by randomization group at Visit 1 (Table 3). There were minimal differences for any of the variables between the Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and Standard of Care groups at Visit 1; only adolescents’ sex, insurance type, and score on the Helplessness About Low Blood Glucoses subscale of the Hypoglycemia Fear Survey differed (Table 3). If these differences are associated with study outcomes, they will be controlled for in statistical analyses.
Table 3.
Schedule of measure completion.
| Visit 1 | Visit 2 | Visit 3 | Visit 4 | Visit 5 | |
|---|---|---|---|---|---|
| Insulin Pump Downloads | X | X | X | X | X |
| Medical Chart Review | X | X | X | X | X |
| Child Questionnaires | |||||
| Depressive Symptoms (CDI2/BDI-II) | X | X | X | ||
| Fear of Hypoglycemia (HFS-C) | X | X | X | ||
| Diabetes Responsibility (DFQR) | X | X | X | X | X |
| Insulin Pump Responsibility (IPR) | X | X | X | X | X |
| Quality of Life (PedsQL) | X | X | X | ||
| Carbohydrate Counting | X | X | |||
| Meter Skills | X | X | |||
| Insulin Pump Knowledge | X | X | X | ||
| Puberty | X | X | X | X | X |
| Acceptability & Feasibility | X | ||||
| Caregiver Questionnaires | |||||
| Background Information | X | ||||
| Depressive Symptoms (CES-D) | X | X | X | ||
| Fear of Hypoglycemia (HFS-P) | X | X | X | ||
| Diabetes Responsibility (DFQR) | X | X | X | X | X |
| Insulin Pump Responsibility (IPR) | X | X | X | X | X |
| Quality of Life WeCARE | X | X | X | ||
| Insulin Pump Knowledge | X | X | X | ||
| Acceptability and Feasibility | X |
3.9. Planned statistical analyses and power
It is important to note that a 2 × 2 factorial design is not a 4-arm trial in which each condition is compared in turn to a control condition. The main effect of Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report is estimated by comparing the results of Groups 1 and 3 combined against the results of Groups 2 and 4 combined. In other words, the main effect estimate is based on all 106 participants. The main effect of Pump It Up! Problem-Solving Skills is estimated by comparing the results of Groups 1 and 2 combined against the results of Groups 3 and 4 combined, again involving all participants (Table 1). Factorial designs do not require a larger number of participants than alternative designs (e.g., randomized clinical trials) when used to address suitable research questions.
Linear mixed models will be used to test the main effects of Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report and Pump It Up! Problem-Solving Skills on the 5 insulin pump self-management behaviors in parallel (primary outcomes) because we are interested in determining which specific behaviors can be improved as usually only A1C is examined as the outcome variable in T1D studies. Linear mixed models will also be used to examine the one secondary outcome of A1C. Linear mixed modeling is used to assess change across time and is similar to linear regression. However, linear mixed modeling is more sophisticated and has several advantages compared to other statistical techniques used to analyze data from repeated measures designs. Importantly, linear mixed modeling accounts for data that are intra-individually correlated as data are collected from the same participant at multiple time points in longitudinal studies. Linear mixed modeling also accounts for some missing data, an uneven number of observations between participants, and time-varying covariates and it allows modeling of individual change/trajectory across time to better understand variability of patterns of change (e.g., growth curves) within a larger population of participants [19].
4. Discussion
Pump It Up! was a 2 × 2 factorial randomized clinical trial designed to test two distinct interventions to improve insulin pump self-management behaviors in adolescents with T1D. Pump It Up! participants included youth from Colorado and Florida with sociodemographic characteristics consistent with the populations in these regions. The average A1C for study participants upon study entry was 8.8% which is representative of national registry data for individuals with T1D 10–18 years old [7,20]. The need for intervention to improve insulin pump behaviors was clearly demonstrated at study entry and prior to the start of any interventions given that only 2% of participants met all 5 insulin pump self-management goals. Differences between the treatment and control groups at Visit 1 were minimal suggesting that randomization was successful; consequently, any changes in insulin pump self-management behaviors at the study endpoint can be attributable to the intervention.
The high retention rate of 93.5% is likely the result of a combination of factors. First, adolescents received graduated incentives for participating in each study visit. All visits coincided with the adolescent’s regularly scheduled T1D visit to reduce burden on families. Study visits lasted between 10 min (Standard of Care) and approximately 45 min (Pump It Up! Personalized T1D Self-Management Behaviors Feedback Report or Pump it Up! Problem-Solving Skills), and the content was personalized to the youth and their caregivers.
Although diabetes technology has significantly changed in the years since Pump It Up! was completed, behavioral interventions such as those tested in Pump It Up! will always be applicable because diabetes devices can only be as effective as the user. Multiple behaviors, including consistent wear [21] and continual interaction with the devices, are essential to achieve optimal glycemic control. Further, not all individuals with T1D will desire or be able to adopt new technologies due to various barriers [22] including adverse skin reactions and skin infections, insurance coverage, and psychological factors such as fear of hypoglycemia, technological difficulties/failures, or mistrust in technology.
Insulin pump self-management behaviors were specifically chosen as the primary outcomes in Pump It Up! because these behaviors are fundamental to optimizing glycemic control. Without engaging in a sufficient number of specific T1D self-management behaviors consistently, individuals will not be able to change their A1C. In fact, average A1C in adolescents with T1D in the United States has increased from 9.0% to 9.3% [7,20] during a six year period despite increased adoption of technology and technological advances made concurrently. Alarmingly, 83% of youth from a 2016–2018 United States national registry cohort did not meet the recommended target of A1C <7.5% [7]. Both American Diabetes Association and the International Society for Pediatric and Adolescent Diabetes now recommend an A1C goal of <7.0% to prevent future complications [23,24]. Therefore, intervention studies such as Pump It Up! are needed to assess how behavioral interventions can support adolescents in improving T1D self-management behaviors, and thereby mitigating acute and long-term complications.
Acknowledgments
We are grateful to the adolescents and their parents who participated in this study.
Funding source
National Institute of Diabetes and Digestive and Kidney Diseases (K23DK091558) and American Diabetes Association (1-12-JF-15).
Footnotes
Declaration of Competing Interest
The authors have indicated that they have no potential conflicts of interest to disclose.
References
- [1].Driscoll KA, et al. , Insulin bolusing software: the potential to optimize health outcomes in type 1 diabetes mellitus, J. Diabetes Sci. Technol 7 (3) (2013) 646–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Clements MA, et al. , Association of HbA1c to BOLUS scores among youths with type 1 diabetes, Diabetes Technol. Ther 18 (6) (2016) 351–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Patton SR, Driscoll KA, Clements MA, Adherence to insulin pump behaviors in young children with type 1 diabetes mellitus, J. Diabetes Sci. Technol 11 (1) (2017) 87–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Driscoll KA, Young-Hyman D, Use of technology when assessing adherence to diabetes self-management behaviors, Curr. Diab. Rep 14 (9) (2014) 521. [DOI] [PubMed] [Google Scholar]
- [5].Guilfoyle SM, Crimmins NA, Hood KK, Blood glucose monitoring and glycemic control in adolescents with type 1 diabetes: meter downloads versus self-report, Pediatr. Diabetes 12 (566) (2011) 560–566. [DOI] [PubMed] [Google Scholar]
- [6].Rausch JR, et al. , Changes in treatment adherence and glycemic control during the transition to adolescence in type 1 diabetes, Diabetes Care 35 (6) (2012) 1219–1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Foster NC, et al. , State of type 1 diabetes management and outcomes from the T1D exchange in 2016–2018, Diabetes Technol. Ther 21 (2) (2019) 66–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Driscoll KA, et al. , Importance of manually entering blood glucose readings when wireless-compatible meters are not being used with an insulin pump, J. Diabetes Sci. Technol 7 (4) (2013) 898–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Patton SR, et al. , Frequency of mealtime insulin bolus as a proxy measure of adherence for children and youths with type 1 diabetes mellitus, Diabetes Technol. Ther 15 (2) (2013) 124–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Burdick J, et al. , Missed insulin meal boluses and elevated hemoglobin A1c levels in children receiving insulin pump therapy, Pediatrics 113 (3 Pt 1) (2004) e221–e224. [DOI] [PubMed] [Google Scholar]
- [11].Vigers T, et al. , cgmanalysis: an R package for descriptive analysis of continuous glucose monitor data, PLoS One 14 (10) (2019) e0216851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Collins LM, et al. , The multiphase optimization strategy for engineering effective tobacco use interventions, Ann. Behav. Med 41 (2) (2011) 208–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Collins LM, Murphy SA, Strecher V, The multiphase optimization strategy (MOST) and the sequential multiple assignment randomized trial (SMART): new methods for more potent eHealth interventions, Am. J. Prev. Med 32 (5 Suppl) (2007) S112–S118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Bergenstal RM, et al. , Recommendations for standardizing glucose reporting and analysis to optimize clinical decision making in diabetes: the Ambulatory Glucose Profile (AGP), Diabetes Technol. Ther 15 (3) (2013) 198–211. [DOI] [PubMed] [Google Scholar]
- [15].Bergenstal RM, et al. , Recommendations for standardizing glucose reporting and analysis to optimize clinical decision making in diabetes: the ambulatory glucose profile, J. Diabetes Sci. Technol 7 (2) (2013) 562–578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Robin A, Foster SL, Negotiating Parent-Adolescent Conflict: A Behavioral-Family Systems Approach, The Guilford Press, New York, NY, 1989. [Google Scholar]
- [17].Wysocki T, et al. , Randomized, controlled trial of behavior therapy for families of adolescents with insulin-dependent diabetes mellitus, J. Pediatr. Psychol 25 (1) (2000) 23–33. [DOI] [PubMed] [Google Scholar]
- [18].Taber KS, The use of Cronbach’s alpha when developing and reporting research instruments in science education, Res. Sci. Educ 48 (2018) 1273–1296. [Google Scholar]
- [19].West BT, Welch KB, Galecki AT, Linear Mixed Models: A Practical Guide Using Statistical Software, 2nd edn., CRC Press Taylor & Francis Group, Boca Raton, FL, 2015. [Google Scholar]
- [20].Miller KM, et al. , Current state of type 1 diabetes treatment in the U.S.: updated data from the T1D Exchange clinic registry, Diabetes Care 38 (6) (2015) 971–978. [DOI] [PubMed] [Google Scholar]
- [21].Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study, G, Effectiveness of continuous glucose monitoring in a clinical care environment: evidence from the Juvenile Diabetes Research Foundation continuous glucose monitoring (JDRF-CGM) trial, Diabetes Care 33 (1) (2010) 17–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Forlenza GP, et al. , Biopsychosocial factors associated with satisfaction and sustained use of artificial pancreas technology and its components: a call to the technology field, Curr. Diab. Rep 18 (11) (2018) 114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].American Diabetes A, 6. Glycemic targets: standards of medical care in diabetes-2020, Diabetes Care 43 (Suppl. 1) (2020) S66–S76. [DOI] [PubMed] [Google Scholar]
- [24].DiMeglio LA, et al. , ISPAD Clinical Practice Consensus Guidelines 2018: glycemic control targets and glucose monitoring for children, adolescents, and young adults with diabetes, Pediatr. Diabetes 19 (Suppl. 27) (2018) 105–114. [DOI] [PubMed] [Google Scholar]
- [25].Kovacs M, The children’s depression inventory (CDI), Psychopharmacol. Bull 21 (1985) 995–998. [PubMed] [Google Scholar]
- [26].Kovacs M, Children’s Depression Inventory 2nd Edition Technical Manual, Multi-Health Systems Inc., Canada, 2011. [Google Scholar]
- [27].Beck AT, Steer RA, Brown G, Manual for the Beck Depression Inventory-II, Psychological Corporation, San Antonia, TX, 1996. [Google Scholar]
- [28].Radloff LS, The CES-D scale: a self-report depression scale for research in the general population, Appl. Psychol. Meas 1 (3) (1977) 385–401. [Google Scholar]
- [29].Cox DJ, et al. , Fear of hypoglycemia: quantification, validation, and utilization, Diabetes Care 10 (5) (1987) 617–621. [DOI] [PubMed] [Google Scholar]
- [30].Shepard JA, et al. , Understanding the construct of fear of hypoglycemia in pediatric type 1 diabetes, J. Pediatr. Psychol 39 (10) (2014) 115–125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Anderson BJ, et al. , Assessing family sharing of diabetes responsibilities, J. Pediatr. Psychol 15 (4) (1990) 477–492. [DOI] [PubMed] [Google Scholar]
- [32].Varni JW, et al. , PedsQL 3.2 Diabetes module for children, adolescents, and Young adults: reliability and validity in type 1 diabetes, Diabetes Care 41 (10) (2018) 2064–2071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Varni JW, et al. , The PedsQL in type 1 and type 2 diabetes: reliability and validity of the Pediatric Quality of Life Inventory Generic Core Scales and type 1 Diabetes Module, Diabetes Care 26 (3) (2003) 631–637. [DOI] [PubMed] [Google Scholar]
- [34].Cappelleri JC, et al. , Development and validation of the WEll-being and Satisfaction of CAREgivers of Children with Diabetes Questionnaire (WE-CARE), Health Qual. Life Outcomes 6 (2008) 3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Petersen AC, et al. , A self-report measure of pubertal status: reliability, validity, and initial norms, J. Youth Adolescence 17 (2) (1988) 117–133. [DOI] [PubMed] [Google Scholar]