Design and Development of a Stepped-Care Behavioral Intervention to Support Parents of Young Children Newly Diagnosed with Type 1 Diabetes

Abstract

One of the most common chronic conditions of childhood, the prevalence of type 1 diabetes (T1D) in young children is increasing. Early childhood development complicates optimal T1D management and glycemic outcomes. Parents are at risk for elevated psychological distress, especially immediately following diagnosis. Few empirically supported interventions are available to support parents and promote optimal T1D management during this vulnerable period. This paper reports on the development and study design of First STEPS: Study of Type 1 in Early childhood and Parenting Support. The aim of this trial is to evaluate the efficacy of a stepped care behavioral intervention for parents of young children over the first year following a new T1D diagnosis. The stepped care design provides participants with up to three intensity levels, or steps, of clinical behavioral intervention support based on need, compared to usual care. Intervention steps include peer parent coaching, telephone-based behavioral support, personalized psychological assessment and recommendations, and intensive assessment of T1D glycemic management with recommendations. Primary outcomes include children's glycemic control and parents' psychosocial functioning. Secondary outcomes include children's behavioral and psychosocial functioning. Exploratory analyses will evaluate demographic, disease-specific, and psychosocial factors related to progression in and response to each step of the intervention.

Introduction

Type 1 diabetes (T1D) is among the most prevalent chronic conditions diagnosed in childhood, occurring in one in every 400- 600 children in the United States [1]. Optimal T1D management requires completing a complex and time-consuming daily regimen, which includes frequent blood glucose (BG) monitoring, insulin dose calculation and administration, and diet/physical activity regulation [2,3]. Prevalence is increasing rapidly among young children, with 15-20% of new diagnoses occurring in children under age 5 [4–6]. Inherent developmental challenges of unpredictable diet, behavior, and physical activity patterns, increased insulin sensitivity, and social and behavioral development present significant challenges to optimal T1D management and outcomes in young children [7]. The current American Diabetes Association (ADA) glycemic target of <7.5% for children [8] is relatively conservative, and only approximately one-quarter of children meet this target [9,10]. Additionally, many parents of young children with T1D report elevated psychological distress [11–13], in part due to the relentless demands of T1D management on parents during this age range. Despite the growing prevalence and unique challenges of T1D management in early childhood, empirically supported patient education and counseling programs specifically designed for parents of young children with T1D are not widely available. The purpose of this paper is to describe the development and design of a randomized controlled trial to evaluate a stepped care behavioral intervention for parents of young children newly diagnosed with T1D.

Study Rationale

Barriers to Managing a New T1D Diagnosis in Early Childhood

At diagnosis of T1D, children are typically hospitalized for one or two days and parents spend six to eight hours during the hospitalization with a specially trained nurse educator learning ‘survival skills’ (i.e., administering insulin, BG monitoring, counting carbohydrates, and handling emergencies). After discharge, parents are asked to return within several weeks for more comprehensive outpatient diabetes education and medical follow-up. The period of time immediately after diagnosis is difficult, requiring parents to adjust to a rigorous daily regimen [12].

Approximately two-thirds of young children experience a temporary restoration of beta cell function (known as the “honeymoon period”) after T1D diagnosis as a result of insulin therapy, during which less insulin is required [14]. Yet, within the first year, 90% of young children no longer fall into this category, and parents must quickly adjust to both a new T1D regimen and changing physiological T1D needs [14]. Due to increased insulin sensitivity, young children are susceptible to hypoglycemia in the short-term and cognitive effects in the long-term, both of which contribute to daily BG management challenges and parent stress [15–18].

Young children are normally unable to reliably detect and/or report emerging symptoms of hypoglycemia. Parents have difficulty discriminating between behavioral cues signifying a low or high BG level and normal misbehavior [19], which can interfere with proper T1D management [11]. Disruptive mealtime behaviors are common in young children with T1D [20] and have been reported at higher rates than among children without T1D [21]. Unpredictable eating, in particular, is a source of stress for parents, as transient food preferences and food refusal are typical at this age [22]. Insulin dosing, which is largely based on anticipated food consumption, can be extremely difficult to predict and administer accurately in advance. Many parents choose to administer insulin after a meal to avoid worries of hypoglycemia if too little food is eaten after insulin is given, despite the fact that this can compromise achievement of BG targets [23]. Additional unique challenges of T1D management in early childhood include daytime naps and nighttime BG monitoring [24,25], unpredictable activity patterns, and finding appropriate/safe child care [7]. Many parents hesitate to leave their young child in someone else's care: relatives, daycare providers/teachers, and babysitters also express fear of caring for a young child with T1D [19,26–28]. Not surprisingly, children's T1D care burden falls heaviest on mothers, who report performing 79% of injections and 70% of BG checks [29].

Support for Parents of Young Children with T1D

Parents of young children are in need of targeted T1D support. T1D pervades nearly all of children's activities, including play, meals, sleep, sibling interactions, peer relationships, and school/daycare interactions [27,30–32]. Parents report being in a constant state of vigilance [28,33] and must rapidly develop expertise in their child's T1D management while simultaneously teaching others about proper care [34]. There are major psychosocial obstacles for parents to overcome in the first months post-diagnosis, including grieving the loss of their “healthy” young child, increased prevalence of post-traumatic stress-like symptoms [35,36] and increased risk for depression and elevated depressive symptoms [37]. Maternal distress at diagnosis predicts maternal distress even years later [12]. Therefore, addressing parents' initial psychosocial concerns at diagnosis may yield long-term returns on improving child health and parental well-being.

Despite the need for targeted support for parents of young children following a new T1D diagnosis, there remains a gap in intervention research for this group, as most T1D clinical behavioral interventions limit trial participation to school-aged children or adolescents diagnosed for at least 6-12 months [38]. A few small-scale clinical behavioral interventions for parents of young children with T1D exist that include parent social support via parent mentors (“parent coaches” [37]), and our own telephone-based psychoeducational/behavioral intervention focusing on teaching skills to manage child behavior and parental stress related to T1D [9]. These interventions have demonstrated increased perceived support, decreased family burden, and improved coping [46], as well as decreased parenting stress [39]. Participants in our previous studies with interventions delivered at least 6 months post- diagnosis indicated wanting this program close to the time of diagnosis [39], highlighting the need for targeted behavioral support earlier. Despite the promising behavioral and psychosocial outcomes of previous intervention research, concomitant improvements in children's glycemic outcomes remain elusive. Thus, new or intensified approaches may be necessary to impact both parental well- being and child glycemic outcomes during the very challenging period following diagnosis.

Study Aims

This paper reports on the development and study design of First STEPS: Study of Type 1 in Early childhood and Parenting Support. The aim is to determine the efficacy of a stepped care behavioral intervention for parents of young children over the first year following a new T1D diagnosis. The stepped care design provides participants with up to three intensity levels, or steps, of clinical behavioral intervention support based on need, compared to usual care. Parents are randomly allocated in a 3:1 ratio to either the First STEPS intervention or usual care condition; all participants receive standard diabetes education as part of usual care in both conditions. First STEPS intervention delivery is incrementally intensive with trained peer parent coaches (Step 1), master's level research telephone counselors (Step 2), and registered nurse or master's level Certified Diabetes Educators and a doctoral level clinical psychologist (Step 3) occurring in-person or by telemedicine, with movement through the steps based on need.

The First STEPS study has two primary aims and corresponding hypotheses:

1. Determine the longitudinal impact of the intervention on young children's glycemic control and parent psychosocial functioning. The primary hypothesis is that the children of parents in the intervention condition will demonstrate significantly better glycemic outcomes (i.e., lower HbA1c, less BG variability) and parents will demonstrate significantly better psychosocial functioning (e.g., fewer depressive symptoms, lower pediatric parenting stress) relative to participants in usual care.

2. Determine the longitudinal impact of the intervention on parents' T1D management behaviors and child psychosocial functioning. The second hypothesis is that those in the intervention condition will experience significantly fewer parent and child-driven negative mealtime behaviors (e.g., parent-coaxing child to eat, making multiple food options after refusal; child- food refusal, longer mealtime duration, externalizing symptoms) and significantly better child psychosocial functioning (e.g., fewer behavior problems, higher quality of life) relative to participants in usual care.

An exploratory aim is to identify factors related to progression in and response to the stepped care intervention, including sociodemographic, disease-specific, and psychosocial factors that predict which parents/children require and respond best to more intensive interventions offered beyond Step 1.

Study Design and Methods

Conceptual Framework

The First STEPS intervention is guided by Social Cognitive Theory, which addresses psychosocial factors influencing health behaviors and methods of promoting durable behavior change [80-82]. Core elements of this theory include knowledge of the risks and benefits of health behaviors, perceived self-efficacy that one can control one's health behavior, outcome expectations about the costs and benefits of health behaviors, goal-setting, and the social and environmental facilitators of adaptive change. Intervention components include cognitive and behavioral strategies related to parents' health behaviors, including enhancing parental self-efficacy, goal-setting, and problem-solving related to T1D and/or child behavior management.

Stepped care interventions are designed to provide the least intense treatment necessary for optimal outcomes by closely monitoring progress and enhancing, or ‘stepping up,’ the intensity of the intervention when indicated, based on an algorithm with predetermined outcomes or goals [40,41]. Establishing the minimum treatment necessary to effect change and identifying those who might benefit from increased levels of care holds promise to ensure that an effective yet least resource intensive treatment is delivered first. The UK's National Institute for Health and Care Excellence recommended the implementation of stepped care models for all mental health concerns [42], and non-pharmacological models have been used successfully for substance abuse and weight management interventions [43,44]. Stepped care interventions targeting stress and psychosocial adjustment have shown promising reductions in psychological distress among adults with chronic conditions [45–47]. Currently, ongoing studies are using stepped care to address parental distress among children with chronic illness, including with T1D [48].

As described in detail below (Section 4.4), participants in the intervention condition progress from the least intensive (Step 1) to most intensive (Step 3) interventions stages based on predetermined thresholds for the child's A1c and/or the parent's current level of self-reported depressive symptoms assessed approximately every three months (Figure 1). Each step progressively builds on the last, with increasing levels of intensity, personalization, and interventionist specialized training. The study received approval from the institutional review boards at each site.

Figure 1. Stepped Care Algorithm.

*Step allocation based on meeting either of the following criteria: 1) A1c > 8.0% AND/OR 2) CESD score ≥ 16

Recruitment

The target sample for this study is n=200 parents who self-identify as the primary caregivers for children ages 1-6 newly diagnosed with T1D (consent and baseline data complete within 4 weeks of diagnosis date). Although we anticipate that the majority (90%) of primary caregivers will be female (e.g., mothers, grandmothers), we include males who self-identify as the primary caregiver (e.g., fathers, grandfathers). Secondary caregivers (SC) are invited to participate, if available and interested, by completing study questionnaires. SCs are not required to participate and primary caregiver participation is not contingent on SC participation. Parents are the focus of this investigation and therefore cannot have serious mental illness (e.g., schizophrenia) or developmental disability that would limit participation. Similarly, children with T1D are eligible to be enrolled if they do not have a life-threatening disease (e.g., cancer, cystic fibrosis) or developmental disability (e.g., autism, mental retardation); these features are assessed via parent report at the time of enrollment screening. Parents must be able to adequately understand, speak, and read English to participate - monolingual non-English speakers are excluded from the trial given the unavailability of standardized measurement tools in other languages.

Study staff recruit participants for this study at the time of diagnosis at two trial sites at children's hospitals in major cities in the northeastern and southern United States. Participant identification is systematic and facilitated by review of electronic medical record documentation of new T1D diagnoses and by the administrative hospital staff notifying the study team when children within the target age range are diagnosed with T1D. Research staff approach potentially eligible parents during the hospitalization following T1D diagnosis to briefly introduce the purpose of the study (i.e., conduct a ‘meet and greet’) and schedule a time to follow-up by telephone for a more detailed explanation. Meet and greet visits are not always possible (e.g., when hospital admission and discharge occur over the weekend or if a newly diagnosed family is not admitted to the hospital). All potential participants receive an informational letter and flyer about the study mailed to their home, and study staff follow-up by telephone. SCs are also invited to participate in the study following consent of the primary caregiver. Parents immediately choosing to decline participation are asked to provide demographic data and reason for declining, which will permit participation bias analyses to be conducted.

Randomization

After agreeing to participate in the study during the in-person or telephone meeting, parents are sent a link to complete a baseline assessment via a secured web-based survey (via REDCap [49]) within four weeks of the date of T1D diagnosis (measures described in section 3.5). If participating, SCs complete an abbreviated baseline assessment battery via REDCap. Following completion of the baseline assessment battery, participants meet with a trained study staff for an in-person, individual orientation session. The orientation session serves two purposes: (1) a behavioral run-in to increase participant understanding of the requirements of study participation and prepare parents for the level of commitment required to successfully adhere to the trial's protocol, and (2) to facilitate rapport between the parent and the study team. The session takes place when and where the child is seen for their first routine outpatient T1D visit, approximately 1-4 months following the T1D diagnosis. During the orientation session, study staff review the study timeline and activities, answer participant questions, and obtain written informed consent (electronic consent is documented via REDCap prior to completion of baseline questionnaires). They also invite the participant to share their diagnosis story, show a video featuring other parents of children with T1D talking about their experiences, and give children a small backpack containing children's books about diabetes, crayons, and a stress ball imprinted with the study logo. At the end of the session, study staff complete the randomization process. Participants are randomly allocated to either the intervention or usual care condition on a 3:1 basis using pre-allocated random blocks maintained by the study biostatistician. Both the block size and choice permutation are chosen at random to ensure balance across arms by trial site, child ethnicity, and parent marital status.

Stepped Care Intervention

All participants in the intervention group begin in Step 1 immediately following randomization. Study staff meet with participants at the next two diabetes follow-up visits (typically scheduled approximately 2-4 months apart), at which time A1c values are obtained by medical staff and the participant completes a measure of depressive symptoms (Center for Epidemiological Studies – Depression Scale, described below). At each visit, participants may remain in their current step or may advance to a higher step if the child's A1c value is >8.0% and/or if the parent's depressive symptoms meet or exceed the measure's clinical cut-off of 16. Over the course of the 9-month intervention phase, participants in the intervention arm may stay in Step 1, may advance to Step 2 at the first or second follow-up, or may advance to Step 2 at the first follow-up and to Step 3 at the second follow-up. Participants never return to a lower step, but rather stay in their current step if they do not meet criteria to advance.

Step 1 – Parent Coaches

Current World Health Organization guidelines call for healthcare systems to rethink the delivery of mental health services through “task-shifting,” or training lay community members to deliver services [50]. This shift enables widespread, low-cost dissemination and has demonstrated efficacy across a number of populations, cultures, and presenting problems. Peer parent social support may aid parent adjustment to T1D by offering emotional, tangible, informational, and companionate support from a mentor with unique insights into the challenges of parenting a child with T1D [51]. Parent mentors have been successfully incorporated into clinical behavioral interventions for several pediatric chronic illness populations [52–55]. Parents and children demonstrate improved adjustment and increased social support when a parent mentor is available, [34,54–56] and there is some evidence of better health outcomes (e.g., reduced asthma symptoms [53]). For parents of young children with T1D, Sullivan-Bolyai and colleagues reported outcomes of a 6-month peer parent consultation model that included fewer parental concerns, more confidence, and more resource identification relative to usual care [55].

In First STEPS, medical team members nominate parents to be parent coaches, who are then screened for eligibility based on their child's current age and age at diagnosis, relative glycemic control (i.e., A1c < 9.0%), and parent characteristics (e.g., interpersonal warmth, positive coping strategies, flexible perspective toward diabetes management). Study staff train parent coaches and obtain informed consent prior to assignment of their first target parent. Training is comprised of a four-hour in-person group meeting that includes information about the study protocol, interactive training in active listening skills, didactic training regarding research ethics, instruction about the limitations and scope of the parent coach role (e.g., not to offer medical advice), provision of resources (e.g., medical team on-call contact information), and role-plays of emergency medical and psychosocial crises. Following the in-person training, parent coaches have at least monthly (more often initially) phone calls with study staff to review their activities as parent coaches, answer questions, and practice active listening skills. Parent coaches are instructed to make contact with their assigned parents weekly for the first three months after the participant is randomized to the intervention, tapering to monthly contact for the remaining six months of the intervention period (nine months total). Parent coaches are encouraged to make contact using text/SMS, phone, and email, and one in-person meeting is encouraged in the first month following assignment. Parent coaches complete monthly online surveys reporting on their contact with participants.

Step 2 - Psychosocial Telephone Counseling

As the first step-up for intensification of the intervention, Step 2 includes 5 telephone sessions that aim to provide behavioral support focused on common challenges experienced by parents of newly diagnosed young children, with examples and applications of the skills tailored to the participant's family and experiences. Telephone counseling is an effective and acceptable format for promoting positive psychological adjustment to pediatric disease and reducing depressive symptoms [39,52,57–62]. Telephone-based interventions with parents are effective at decreasing behavior problems and anxiety in children, in increasing positive parenting practices, and improving mood [63–66]. The Step 2 intervention for First STEPS was based on our previously tested telephone-based supportive intervention for parents of young children with T1D, which was shown to be feasible, have high satisfaction, and resulted in decreased pediatric parenting stress [39]. Step 2 includes 5 telephone calls delivered by a trained, master's level interventionist following a manual. Participants receive a printed booklet with written information and exercises that complement the content of each telephone session. The calls are designed to provide more detailed education about child development as it relates to T1D management, training in cognitive and behavioral stress management and problem-solving skills, and encouragement to seek social support (Table 1). Each telephone session lasts approximately 45-60 minutes, follows a manual, and is audio-recorded for review, supervision, and to evaluate fidelity.

Table 1. Step 2: Phone Intervention.

Session/ SCT Principle(s)	Topic/Theme	Content/Approach	Behavioral Tasks
1 Expectations Emotional coping/Mood	Challenges of diabetes and impact on parent mood Physical and cognitive tasks of child's development Caregiver job sharing	Unique challenges of parenting a young child with diabetes Stages/tasks for physical and impact of diabetes and impact of diabetes Relaxation technique Problem solving technique	Discuss child's diabetes diagnosis and unique challenges for this age group practice breathing and problem solving exercises
2 Self-Control Reinforcements	Glycemic goals and challenges Preparing to leave child with other caregivers and for school/daycare/camp Interrelationships among thoughts/ feelings/actions Hypoglycemia fear	Age-specific glycemic goals Leaving child with other caregivers Understanding contributions of thoughts and feelings to parenting and daily T1D care Rehearsal of problem-solving technique	Problem-solve diabetes management difficulties Make plan for other caregiver assistance Generate ideas to prepare for school/daycare Practice monitoring thoughts and feelings related to daily T1D care
3 Observational learning Self-efficacy	Parenting Routines (eating; timing of insulin administration, sleep) General child behavior Identification of family strengths	Developmental tasks for social and emotional function and eating and impact of diabetes Parenting strategies to promote good behavior and decrease misbehavior specific to mealtimes and diabetes management tasks; mealtime goals of administering insulin prior to or during meals	Practice differential attention and contingency management for eating behaviors and diabetes management tasks Discussion of setting consistent routines for health behaviors
4 Review of prior sessions	Impact of diabetes on family Getting tangible and emotional Impact of diabetes on family Getting tangible and emotional support for diabetes management Advocacy Self-care and future problem prevention Gratitude	Problem solve how to increase support and increase time for self Problem solve how to increase support and increase time for self and family while still promoting good glycemic control Parents' role as advocate for child and her/himself Finding time for self-care and behavioral activation	Problem solve how to increase support and advocate for Problem solve how to increase support and advocate for her/himself and daily management Generate ideas for self-care and gratitude exercises Engage in activity scheduling
Variable1 Observational learning Group phone call with other participants	Social support Sharing diabetes tasks with other caregivers	Network with other parents in similar circumstances Receive/provide emotional support Share tips on practical help for diabetes specific challenges related to daily tasks and mealtimes	Be active participant in group phone call Review skills of problem solving and parenting Review skills of problem solving and parenting

Note.

¹Participants become eligible for the group phone call session after completion of session 1; group calls are scheduled based on availability of other participants. Group phone calls must be completed within 12 weeks of the participant's entry into Step 2.

Step 3 – Individualized Support from Diabetes Clinical Team

The most intensive step of the intervention, Step 3 focuses on personalized assessment and recommendations related to both BG management and emotional/behavioral issues, delivered by highly trained professionals. To address BG, a Certified Diabetes Educator (CDE) places a blinded continuous glucose monitoring (CGM) sensor on the child and instructs the participant to manage T1D as usual and to use provided logs to keep track of BG values, food intake, activity, and insulin doses/administration. A blinded CGM is used because the purpose is to gather data for the CDE to give personalized recommendations, and parents having access to the near real-time CGM data using a non-blinded device would likely influence their T1D management behaviors rather than serve an assessment purpose. For children already using a personal CGM, data from their device are used and an additional device is not placed, but the family is instructed to complete the food, activity, and insulin logs. Any families who opt out of having a blinded CGM placed complete the logs and are instructed to do more frequent BG monitoring. One to two weeks after the blinded CGM is placed and logs are completed, the participant meets with the CDE a second time to review the glucose data and logs and the CDE gives individualized recommendations for T1D management regimen changes (e.g., insulin, nutrition, activity level). All recommendations are documented in the electronic health record to ensure the medical team is aware of any changes to the regimen.

Step 3 also involves an individualized consultation with a licensed clinical psychologist specializing in pediatric diabetes, either in person or by telemedicine based on the participant's location and schedule. In a 45-60 minute meeting, the psychologist assesses parent mood and adjustment related to the child's T1D diagnosis and management, the child's behavior and adjustment to living with T1D, and family functioning. Based on this conversation, the psychologist makes personalized recommendations and referrals for additional mental health or behavioral support as appropriate. Like the CDE, the psychologist documents a brief note about this encounter in the electronic health record.

Control Condition – Usual Care

Usual care is similar at both sites and includes monitoring by the clinical team nurses following discharge from the hospitalization for diagnosis, with families asked to call the clinic nurses daily for 1-2 weeks post-diagnosis. During these calls BG levels and management tasks are reviewed and changes are made as needed. Usual care also includes providing families with procedures for calling clinicians on-call, and outpatient visits with a nurse educator and physician are scheduled approximately every 3-4 months. There are psychologists on the diabetes teams at both sites, and it is estimated that <5% of children/parents are followed v appointments are recorded to examine potential confounding effects.

Measures

Assessment Schedule

Baseline assessments are completed within one month following diagnosis. Follow up assessments take place at 3-, 6-, 9-, and 15- months after diagnosis (Table 2). SCs complete questionnaires at baseline and follow-ups at 9- and 15-months after diagnosis. We attempt to collect follow-up data from any participants who may drop-out prior to or during the intervention to assess bias and conduct intent-to-treat analyses.

Table 2. Assessment schedule.

Measures	Baseline	First Check-In	Second Check-In	First Follow-up	Second Follow-Up
	Within 1 month of diagnosis	2-4 months after Randomization	6-8 months after Randomization	9-12 months after Randomization	15-18 months after Randomization
A1c	•	•	•	•	•
BG meter data		•	•	•	•
Complications		•	•	•	•
Demographics	•
BPFAS	•			•	•
CESD	•	•	•	•	•
DCQ	•			•	•
ECBI
Health literacy screen	•
HFS	•			•	•
PFS	•			•	•
PIP	•			•	•
PROMIS-A	•			•	•
PSQI	•			•	•
PSS	•			•	•
SED-P	•			•	•
SPSI-RS	•			•	•

Note. Abbreviations as follows: A1c – hemoglobin A1c, BG – blood glucose, BPFAS – Behavioral Pediatrics Feeding Assessment Scale, CES-D – Center for Epidemiologic Studies Depression Scale, DCQ – Diabetes-Sspecific Coparenting Questionnaire, ECBI – Eyberg Child Behavior Inventory, HFS – Hypoglycemia Fear Survey, PFS – Protective Factors Survey, PIP – Pediatric Inventory for Parents, PROMIS-A – Patient-Reported Outcomes Measurement Information System – Anxiety subscale, PSQI – Pittsburgh Sleep Quality Index, PSS – Perceived Stress Scale , SED-P – Self-Efficacy for Diabetes Scale for Parents, SPSI – Social Problem-Solving Inventory, Revised, Short-Form.

Diabetes Management

Glycemic outcomes

Glycosylated hemoglobin A1c is the most widely accepted measure of overall glycemic control. It is represents the average BG over the preceding three months; quarterly assessments are recommended by the ADA [8] and are obtained at every clinic visit. A1c values are extracted from the electronic medical record for this study. Study staff also download personal BG meters at each clinic visit to determine average BG level, frequency of BG checks, and percentage of time above, below, and within recommended ranges.

Diabetes Complications

Parents report on the frequency of episodes of severe hypoglycemia, ER visits, hospitalizations, and missed work/school. Medical chart review confirms reports of ER visits or hospitalizations.

Child Mealtime Behavior

Participants complete the Behavioral Pediatrics Feeding Assessment Scale (BPFAS [67]) a measure of parents' strategies for managing child behaviors related to food and mealtimes. The Parent Scale of the BPFAS contains ten items assessing the frequency with which the parent engages in actions to manage their child's behavior at mealtime (e.g., “I coax my child to get him/her to take a bite,” “If my child does not like what is served, I make something else.”), using a five-point Likert scale ranging from Never to Always. Parents also rate (yes/no) whether they consider the behavior to be a problem. This yields two scores regarding parents' behaviors: (1) frequency and (2) problem. The BPFAS has been successfully used with parents of young children with T1D with satisfactory internal consistency (α = .74-.88) [68,69].

Parental Quality of Life

Depressive Symptoms

The Center for Epidemiological Studies-Depression Scale (CES-D [70]) is a 20-item self-report measure of negative and positive affect, somatic symptoms, mood, and interpersonal relations. Items are rated on a four-point Likert-type scale. The CES-D has adequate test-retest reliability and correlates well with clinical ratings of depression severity. Internal consistency coefficients range from .85-.90 [70–72]. For participants in the intervention group, scores that meet or exceed the clinical cut-off of 16 at the check-in visits advance the participant to the next intervention step; this score has been associated with a marker of individuals at risk for clinical depression with good sensitivity and specificity and high internal consistency [73]. For all participants throughout the study, a licensed psychologist contacts parents with scores ≥20 (indicating elevated risk for depression) for additional assessment and provides a list of local mental health resources.

Anxiety Symptoms

Participants complete the Patient Reported Outcomes Measurement Information System (PROMIS) Emotional Distress-Anxiety-Short Form (PROMIS-A [74]). This seven-item measure assesses how often the participant has been bothered by symptoms such as feeling fearful, worried, and having difficulty concentrating during the past seven days. The PROMIS-A has demonstrated good internal consistency (α = .89; [75]) and initial evidence of validity.

Parenting Stress

The Pediatric Inventory for Parents (PIP [76]) and the Perceived Stress Scale (PSS [77]) measure two aspects of parenting stress. The PIP measures pediatric parenting stress, and participants rate the frequency of 42 experiences commonly reported by parents of children with medical conditions (e.g., feeling helpless about child's condition, having financial troubles, making decisions about medical care) and their difficulty, both using five-point Likert-type scales. Frequency and difficulty ratings are separately summed, with higher scores indicating greater pediatric parenting stress. The PIP has been used with parents of children with T1D (α = .80-.96; [78,79]) and other pediatric conditions [80–82]. The PSS is a widely-used instrument for measuring the degree to which participants perceive various situations to be stressful. Participants rate the frequency with which they experienced ten stress-related thoughts and feelings during the last month. The PSS has shown evidence of validity [77].

Psychological and Behavioral Mediators

Self-Efficacy

Participants complete the Self-Efficacy for Diabetes Scale for Parents (SED-P [83]), in which they rate their confidence in completing 22 daily T1D care tasks using a five-point Likert-type scale. The SED-P demonstrated good internal consistency (α = .87) and significant associations between self-efficacy and parenting stress in previous research with this population [79].

Fear of Hypoglycemia

The Hypoglycemia Fear Survey (HFS [84]) is a 13-item measure of parents' fear or worry regarding their child experiencing hypoglycemia. The HFS has been used with parents of children with T1D at various ages [85] with good internal consistency (α = .90) and associations with parental stress [79].

Parent Sleep

The Pittsburgh Sleep Quality Index (PSQI [86] measures parents' sleep quality, specifically sleep disturbance in the past month. Participants report on their sleep quality, latency, duration, efficiency, disturbances, medications, and daytime dysfunction. For this study, three additional questions were added to assess the impact of managing their child's diabetes on parents' sleep. Reliability estimates of the PSQI are good (α >.80).

Co-Parenting

The Diabetes-Specific Coparenting Questionnaire (DCQ [87]) assesses how parents work together specifically related to T1D management, in three coparenting dimensions: cooperation (five items), conflict (seven items) and triangulation (two items). Internal consistencies for the DCQ based on three different community samples were acceptable for all three scales (α range = .69 to .84).

Coping Resources

Participants complete the Social Problem Solving Inventory: Revised- Short Form (SPSI-R: Short [88]) and the Protective Factors Survey (PFS [89]) to assess parents' social, emotional, and behavioral resources. The SPSI-R: Short is a 25-item measure of an individual's problem-solving style and ability to generate solutions. It has demonstrated strong internal consistency and validity [88]). The PFS assesses parents' perceptions of their family and support on 4 scales: Family Functioning/Resiliency, Social Support, Concrete Support, Nuturing and Attachment, Child Development/Knowledge of Parenting. The PFS has demonstrated high internal consistency overall (α = .76 -.89) and across all subscales (α = .52 - .75) [89]

Child Behavior

The Eyberg Child Behavior Inventory (ECBI [90]) measures parents' perceptions of children's behaviors via ratings of the frequency of 36 common misbehaviors (e.g., argues with parents about rules, whining) and whether each item is a problem for the parent (yes/no); it yields 2 scores: (1) intensity rating and (2) problem rating. Reliability estimates are high [91].

Health Literacy

Parents complete a validated three-item measure of health literacy created by Chew and colleagues [92], which assesses the frequency with which the parent has trouble understanding written medical information, requires assistance reading medical materials, and their confidence in completing medical documents [92].

Clinical and Demographic Characteristics

Demographics

Parents complete a demographic questionnaire with information including race/ethnicity, number of parents and children in home, family medical/psychiatric history, socioeconomic and educational information, insurance coverage, and child's school/daycare status.

Clinical Characteristics

The child's illness characteristics (e.g., date of diagnosis, insulin regimen/dose, use of CGM) are assessed via medical record review.

Process Data

Following study completion, intervention participants provide feedback about their experiences to help our research team evaluate the acceptability, usability, and strengths and weaknesses of the intervention program and further refine it. Using a self- report treatment satisfaction measure adapted from prior intervention studies [93,94], participants in the intervention condition rate their perceptions about: (a) the usefulness of specific components of the intervention (e.g., peer parent coaches, education, print materials, session topics, meeting with CDE and Psychologist); (b) their frequency of use of strategies/skills learned in the intervention; and (c) their satisfaction with logistical arrangements of intervention (e.g., timing and length of sessions).

Parent Coach Outcomes

To explore whether serving as a parent coach in Step 1 of the intervention has any impact on the parent coaches themselves or on their own children with T1D, they complete a brief battery of questionnaires before being assigned their first participant and after completing their role with their final participant. Parent coaches complete the CES-D, PIP, and PFS (described above). They also complete the Parents Diabetes Quality of Life Questionnaire, a 42-item measure of parents' satisfaction with aspects of their life as parent of a child with T1D, degree to which their child's T1D causes problems in their daily life, and worries about their child with T1D [95]. Items are rated on a five-point Likert scale, and the measure includes three subscales for satisfaction, impact, and worry. The PDQOL has been reported to be reliable and valid (α = 0.64-0.89) [95].

Fidelity

For Step 1, parent coaches complete brief monthly surveys about their interactions with each participant they are assigned. The survey assesses the frequency and type (e.g., in person, phone, text message) of contacts, who initiated contacts, topics discussed, and their perceptions of the helpfulness and participant engagement in the contacts. At the end of the study, parent coaches provide feedback about their experiences in the study, including their most common and preferred methods of contacting participants, average time spent interacting with participants, perceived helpfulness of training and monthly calls with study staff, and perceived impact of being a parent coach on their confidence in managing their own child's T1D. Parent coaches are also asked to indicate areas they could have used more guidance and recommendations to improve the program.

Research staff members providing Step 2 sessions are trained using an intervention manual and must demonstrate initial competency through clinical ‘check-outs’ for intervention delivery. There are monthly group supervision conference calls for all interventionists with the goal of ensuring uniform delivery to minimize site differences. All Step 2 calls and Step 3 sessions are audio recorded and 10% of each will be randomly selected for two independent coders to monitor and record fidelity in an ongoing fashion.

Data Analyses

Power

The total sample size to achieve a power p=0.80 at α=0.05 level was estimated for the continuous and binary outcome measures. For pairwise comparison of intervention effect (e.g., intervention Step 1 vs. usual care) on a continuous outcome, about 34 cases per group in the model would achieve a statistical power of ≥0.80 at the 0.05 level to detect a small effect size of 0.35, assuming a modest ρ=0.20. For dichotomous outcomes, about 30 cases per group would reach a power of ≥0.80 at 0.05 level to detect a response proportion difference of 0.18 that is equivalent to a medium odds ratio (OR) of 2.13, assuming ρ=0.20. Our previous studies show a small attrition rate (up to 10%). Our sample of 200 will ensure large enough statistical power for intervention efficacy evaluation for our longitudinal analysis models.

Planned Statistical Analyses

Aim 1

We will first use a Generalized Estimating Equation (GEE) model [107,108] to test the hypothesis that children of parents in the intervention condition will demonstrate significantly lower glycemic outcomes and parents will demonstrate significant better psychosocial functioning relative to participants in usual care. Because the final intervention step status (i.e., those who end the intervention in Step 1, 2, or 3) will not be finalized until the third data collection point, it is treated as a time-varying categorical variable. Clinical and demographic characteristics (e.g., insulin regimen, diabetes duration) and study site will be controlled in the model. Various contrasts (i.e., overall intervention vs. control, Step 1 vs. Steps 2 and 3, and Steps 1 and 2 vs. Step 3) will be performed to compare each outcome by intervention status.

Aim 2

A sequential latent growth model (LGM) will be used to test the hypothesis that parents and children in the intervention condition will engage in significantly fewer negative mealtime behaviors and will report significantly better child psychosocial functioning than those in usual care. The growth trajectories of each outcome measure will each be modeled separately and using parallel-process LGM models [96,97].

Aim 3

A cross-sectional analysis of factors related to progression in and response to the stepped care intervention will be conducted when the intervention status is finalized. First,chi-square tests will be used to examine the binary relations between intervention status and individual sociodemographic, illness, and psychosocial factors. Then multinomial logit model with a four-level categorical outcome (control, Steps 1, 2, and 3 intervention) will be used to explore which factors predict the likelihood of being in each specific status of the stepped care intervention, controlling for other factors.

Discussion

Parents experience high levels of stress surrounding their child's T1D diagnosis, and providing support tailored to each individual family's needs may offer psychological and medical benefits. Strengths of the First STEPS study design include a randomized controlled trial of an intervention based in Social Cognitive Theory for an underserved target population of parents of young children newly diagnosed with T1D. By attending to the developmental needs of young children newly diagnosed with T1D and their parents, we hypothesize that this intervention can affect proximal (i.e., daily management behaviors, parental depressive symptoms), distal (i.e., glycemic control), and anticipated (i.e., prevention of later non-adherence and medical complications) outcomes. Teaching parents effective management skills and offering diabetes-related support is a preventive strategy designed to equip parents to be role models of optimal T1D care for their children and to lay the foundation for their later skill acquisition. Results of this trial will inform subsequent clinical research into parenting programs in this scientific area.

The innovative stepped care study design explores the use of a judicious and potentially cost-effective approach to match level of support to each parent's needs. Families receive the simplest form of care appropriate, with resources efficiently and effectively distributed. The First STEPS approach is aligned with advice from experts in the field of implementation science, who note that essential components of cost-effective, wide reach, and impactful interventions include: 1) novel methodological designs such as stepped care, 2) using paraprofessionals (such as parent consultants) to deliver services, and 3) technology to bring the intervention directly to the people in need [98,99]. This multimodal, personally-tailored parent intervention achieves these goals in its systematic approach to delivering the least intensive care necessary. Based on frequent assessment of children's glycemic data and parental depressive symptoms, participants progress through a series of clinical behavioral strategies to target parental well-being and children's glycemic outcomes on multiple levels. Each intervention step provides a more focused and individualized emphasis on diabetes education, counseling, and management, and the progression from intervention delivery via telephone support (Steps 1-2) to in-person (Step 3) systematically intensifies participant engagement and personalizes the content of the intervention.

In summary, more psychosocial support is needed as parents adjust to a new diagnosis of T1D in their young children, and intensified clinical strategies are needed to more effectively target children's glycemic outcomes. This multi-modal stepped care intervention study will evaluate the efficacy of providing the least intense treatment necessary for optimal outcomes by closely monitoring child and parent progress and advancing intervention delivery accordingly. First STEPS includes highly translatable methods including social support via parent mentor lay workers, brief phone-delivered cognitive and behavioral support via master's level interventionists, and culminates with individualized consultations with a certified diabetes educator and clinical psychologist for participants with the greatest need. We hypothesize families in the stepped care intervention will demonstrate better child glycemic control and behavior, as well as better parent psychosocial outcomes. Results have the significant potential to translate into clinical practice to enhance the diagnosis process for families with new-onset T1D through future dissemination studies.