Research Design
This cross-sectional study was conducted as part of a larger project investigating the consistency of sleep schedules in people with T2DM. Portions of this project have been published elsewhere [5]. This study utilized data from people with T2DM only and people with both T2DM and insomnia symptoms to compare fatigue, daytime sleepiness, vitality, and physical function. Participants were stratified into two groups, those with insomnia (IN) and without insomnia (No-IN), using a cut-off score of >10 on the insomnia severity index (ISI) [18].
Data Collection Procedures
Participants were recruited through University Research Center’s Frontiers research subject registry [19], Cray Diabetes clinic, and campus advertisements, as well as through flyers distributed to the surrounding community. The study was approved by the University Research Center’s Institutional Review Board. Written informed consent was obtained from each participant during the first study visit.
Participants were enrolled in the study following telephone and in-person screening sessions. Individuals were included if they 1) had self-reported T2DM; 2) were 40-75 years old; 3) were able to understand English; and 4) were able to attend and finish the testing procedures. Individuals were excluded if they 1) were at risk of untreated OSA or RLS as determined by the Stop Bang and RLS Diagnostic Index; 2) reported being pregnant; 3) reported heavy alcohol use (i.e., ≥15 alcohol drinks per week for men and ≥8 for women; 4) had a self-reported history of neurological disease, bipolar disorder, seizure disorder, chronic fatigue syndrome, rheumatic disease, dialysis, blindness, or amputation; 5) currently performed night-shift work; 6) reported severe symptoms of pain, depression and anxiety, as evidenced by a score of ≥7 on the Brief Pain Inventory, ≥21 on the Beck Depression Scale, and ≥15 on the Generalized Anxiety Disorder–7 scale. These exclusion criteria were designed to minimize the potential effects of other comorbidities and sleep disorders on daytime functioning.
Participants were divided into either the IN group or the No-IN group based on their score on the ISI, with those scoring > 10 allocated to the IN group and those scoring ≤ 10 allocated to the No-IN group. In addition, people in the IN group reported the insomnia symptoms for at least 3 months. The ISI is a self-report measure designed to evaluate the nature, severity, and impact of insomnia [18]. Scores on the ISI range from 0 to 28, with higher scores indicating greater insomnia severity [18]. Previous research has shown this instrument to be an excellent screening tool to predict the diagnosis of insomnia [20], and a cut-off score >10 has been shown to result in high sensitivity (97.2%) and specificity (100%) for the detection of insomnia in a clinical sample [18].
Participants
A total of 60 participants with self-reported T2DM participated in the study and were included in the final analysis. The sample size was determined based on an ongoing project to study the night to night sleep variation and outcomes in people reported T2D with and without insomnia symptoms.
Measures
All measurements, including demographic and clinical variables, fatigue severity, daytime sleepiness symptoms, and QoL related vitality and physical function, were obtained during the course of a single visit.
Demographic and clinical variables: Information regarding age, sex, education and ethnicity was collected during the assessment visit. Body mass index was calculated via height and weight measurements. Positive Airway Pressure (PAP) machine utilization was determined through the use of a yes/no question (e.g., “Do you use a PAP machine?”). Compliance with PAP usage was assessed via a diary indicating the nights of sleep with using a PAP machine. Non-compliance was defined as non-usage on more than 2 out of 7 nights and/or < 4 hours per night in this project [21].
Daytime functioning: Each participant completed a comprehensive assessment of daytime functioning including fatigue severity, daytime sleepiness, vitality, physical function, and depression symptoms.
Fatigue severity was measured using the Fatigue Severity Scale (FSS), which is a 9-item questionnaire that has been validated in people with diabetes [22]. The FSS emphasizes the impact of daily functional fatigue accumulation during the past week on subscales of motivation, exercise, interference with work, family, or social life. These subscales are summed to yield with a score of <4 indicating no fatigue, scores between 4 and 4.9 indicating moderate fatigue, and a score ³5 indicating severe fatigue [22].
The Short Form-36 vitality (SF 36-vitality) subscale is widely used to measure energy in chronic disease groups [23]. Vitality represents the combination of fatigue and energy. The SF-36-vitality includes four questions, two related to fatigue and two related to energy over the past 4 weeks. These questions represent both positive (energetic) and negative (tired) states. Scoring criteria for SF-36-vitality was used for each item, then summed to range between zero (worse scores) to 100 (optimal scores) [24].
Physical function over the past 4 weeks was measured using the Short Form-36 physical subscale (SF 36-physical function) [24]. This subscale contains ten items rated from zero (very limited ability to perform daily physical activities) to 100 (able to perform all daily activities without limitations). SF 36 subscales were chosen based on the limitations in physical activities and symptoms related to fatiguability because of health problems
Daytime sleepiness symptoms were assessed through the Epworth Sleepiness Scale (ESS) which refers to usual lifestyle in recent times. The ESS consists of eight items rated on a 4-point Likert scale, with subjects rating how likely they would be to fall asleep in 8 different states of daily activity [25]. The ESS has demonstrated satisfactory test-retest reliability (r = .82) and internal consistency (α = .88) [25]. A cutoff score of ≥10 suggests pathological sleepiness [25].
Statistical Analysis
All data analyses were performed using SPSS 23.0 for Mac (Chicago, IL). Descriptive statistics included means and standard deviations, and frequencies were used for continuous variables and categorical variables, respectively. Skewness and kurtosis tests examined the normality of residuals during model development. Chi-square and independent sample t-test analyses were used to assess between-group differences in categorical and continuous variables, respectively. Pearson correlation was utilized to investigate the relationship between daytime functioning outcomes and ISI scores. A multivariate general linear model assessed the differences between groups in daytime functioning outcomes after controlling for covariates. Partial Pearson’s correlation tests were used to assess the relationship between daytime functioning outcomes and insomnia severity after controlling for demographic variables. Symptoms of depression, as assessed by the Beck Depression Inventory [26], was also included as a covariate. Cohen's guidelines were used to illustrate the direction and magnitude of correlations in which r = .1–.29 is small, r = .3–.49 is medium and r = .5–1.0 is large [27]. All tests were conducted at an alpha level of 0.05.