Participants and Materials
The procedures and interventions for this project were described in a published protocol report [23]. This intervention trial was described according to the CONSORT 2010 guideline [24]. Prior to being enrolled in the study, potential participants were screened according to eligibility criteria, which are presented in Table 1.
Table 1
The inclusion and exclusion criteria.
Inclusion Criteria | Exclusion Criteria |
Ages between 40 to 75 years | Self-reported neurological diseases (e.g. Alzheimer’s disease, Parkinson’s disease, Traumatic Brain Injury, Stroke, Multiple Sclerosis) |
Self-reported diagnosis of type 2 diabetes | Self-reported untreated sleep disorders as well as: - Scored > 4 on Stop-Bang score - Failed to pass Restless Leg Syndrome Diagnostic Index |
Scored > 10 on Insomnia Severity Index and self-reported symptoms of insomnia at least 3 nights/week for the past 3 months | Scored ≥ 7 on Brief Pain Inventory |
Able to travel to the University of Kansas Medical Center to attend 6 sessions | Scored ≥ 21 on Beck Depression Scale |
Able to understand and follow verbal commands in English | Scored ≥ 15 on Generalized Anxiety Disorder-7 |
| Self-reported following medical issues: Chronic Fatigue Syndrome, Fibromyalgia, Bipolar, Seizure Disorders and Rheumatic Diseases, Dialysis, blindness, trans-femoral amputation, speech deficits, or significant auditory impairment |
| Performed night shift work |
| Heavy alcohol drinker (≥ 15 alcohol drinks per week for men and ≥ 8 alcohol drinks per week for women) |
| Reported being pregnant |
Study Design
This RCT had an allocation ratio of 1:1, and utilized a superiority framework to test the effectiveness of the CBT-I. Participants were randomly assigned to either the CBT-I group (n = 14) or the HE group (n = 14). We used age to stratify participants into either the older (63–75 years) or the younger (40–62 years) age group. This study was registered in the Clinical Trials Registry (NCT03713996) [25]. This study was approved by the Institutional Review Board and the Human Subjects Committee of the University of Kansas Medical Center. All participants signed a written informed consent before the assessment visit. Data collections and provided interventions were taken place at the University of Kansas Medical Center.
Outcomes
All participants completed outcome measures at baseline, and all participants completed the same outcome measures one week after completing the intervention. The primary outcome, insomnia severity, was included in the RCT Part I in which the power calculation was established and its preliminary data was published elsewhere [22].
Diabetes control measurement
A point-of-care instrument was used to assess HbA1c using a disposable finger stick HbA1c kit (A1CNow + test kit; Bayer Healthcare, Tarrytown, NY). This instrument measures the level of glycosylated hemoglobin via immunoassay, reflecting average glucose blood levels over the past 6 to 12 weeks [26]. During a previous diabetes management program, the A1CNow + provided accuracy and precision when performing a point-of-care, and a 0.05 reduction in HbA1c is considered clinically meaningful [27]. In addition, random blood glucose (RBG) levels were assessed by a glucose meter (FreeStyle Flash, Contour® Bayer Healthcare, Diagnostic Division, Tarrytown, NY). Participants were not asked to follow dietary restrictions prior to the RBG test. During the intervention, participants in the CBT-I group were asked to record their blood glucose level on their own right before bedtime and after awakening in the morning throughout the study period (i.e., 7 days/nights per week for 7 weeks).
Diabetes self-care behavior (DSCB)
Self-care was assessed using the Diabetes Care Profile (DCP), which is a validated survey that measures 13 psychosocial and educational factors [28, 29]. The 13 domains are associated with the management of diabetes, including understanding management of practice, support, control problems, social and personal factors, positive attitude, negative attitude, care ability, importance of care, self-care adherence, diet adherence, long-term care benefits, exercise barriers, and glucose monitoring barriers [29]. A standardized total DCP composite score was established to present all 13 domains that were scored according to the Fitzgerald et al. scoring criteria [29]. Next, each participant’s domain score was standardized using z-scores, and then averaged to create a standardized total DCP composite score. High scores on the DCP composite score indicate better DSCB.
Fatigue severity
Daily fatigue was measured using the Fatigue Severity Scale (FSS) that consists of 9 items developed to assess disabling fatigue on daily life. The FSS has been shown to be valid and reliable [30]. Each item was measured on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Mean item response for the completed FSS items was used for analysis.
Interventions
All participants in the CBT-I group and HE group attended 6 sessions that were scheduled consistently one session per week with the CBT-I provider. Neither the CBT-I provider nor the participants were blinded in this study. The protocol paper describes session by session of both interventions [23].
Cognitive Behavioral Therapy for Insomnia
Five main therapeutic techniques were provided during the 6-session including sleep restriction therapy, stimulus control therapy, sleep hygiene, relaxation techniques, and cognitive therapy. In order to monitor nightly sleep changes and issues, the CBT-I provider reviewed the sleep diary for each session. In addition, calculates of sleep changes were made to prescribe the sleep schedules for the following week.
Health Education
Five main health education materials were introduced during the 6-session including brief sleep hygiene, foot care, diabetes classifications, healthy diet, and physical activity. During the HE sessions, we provided informal face to face interview to engage the participants into the conversations. Participants’ comprehensive and experiences about the provided materials were facilitated through open questions.
Statistical analysis
All data analyses were performed using SPSS 23.0 for Mac (Chicago, IL) and R (https://www.R-project.org/) [31]. Descriptive statistics included means and standard deviations for the assessed variables. We used Shapiro–Wilk tests to assess the normality of residuals during model development. For the main analysis, we used Mann-Whitney U tests to examine the between-group differences of the CBT-I and HE groups in HbA1c, RBG, DSCB, and fatigue change scores. We also used Wilcoxon signed-rank tests to compare within-group changes for both groups. Effect sizes were calculated using Cohen’s d [32]. We calculated absolute percentage changes in all outcomes to graph the between-group differences. For secondary purpose, we used linear regression analyses to predict blood glucose levels (before bedtime and after awakening in the morning) based on 49 days across the course of the study including 6 weeks CBT-I and post-assessment. For all analyses, the alpha level was set at .05.