Assessing Sleep Concerns in Individuals With Acquired Brain Injury: The Feasibility of a Smartpad Sleep Tool

In this study, we utilized a cross-sectional study design comprising 68 participants. Participants were recruited at the Acquired Brain Injury Psychiatry Clinic at Johns Hopkins Bayview Medical Center. Inclusion criteria included proficiency in English, age ≥18, and history of ABI. Adults unable to consent were enrolled with the written consent of a legally authorized representative. Study visits were completed either before or at the conclusion of the clinic visit. All study staff completed basic sleep training in MySleep101 (an educational program), which covered the diagnostic criteria for sleep disorders, clinical sleep presentations, and treatment of common sleep disorders (6). Trained study staff then obtained written consent, completed the MySleepScript application with participants, and shared the results with their clinician. This study was approved by the Johns Hopkins Institutional Review Board. All included participants gave written informed consent for the collection of clinical data.

MySleepScript is an electronic, customizable application (on iPhone/iPad and licensed by Hopkins) developed by a team of interdisciplinary sleep experts at Johns Hopkins University. Details of MySleepScript and the algorithms used are described elsewhere (6). In brief, the application uses a tree algorithm to trigger the addition of relevant questionnaires based on patient responses and excludes redundant questions based on negative responses. Questionnaires administered to all participants included demographic surveys, the Pittsburgh Sleep Quality Index (PSQI; with a global score >5 indicating poor sleep quality) (7), and an internally developed Johns Hopkins Sleep Environment Inventory. Based on patient responses in MySleepScript, the following additional questionnaires were generated: the Insomnia Severity Index (ISI) to assess for insomnia (no clinical significance=0–7, subthreshold insomnia=8–14, clinically significant moderate insomnia=15–21, clinically significant severe insomnia=22–28) (8), the Patient Health Questionnaire–9 (PHQ-9) to assess for depression (mild=5–9, moderate=10–14, moderately severe=15–19, severe=>20) (9), the Berlin Questionnaire to assess for sleep apnea (positive score defined by at least two categories) (10), a single-question restless legs syndrome (RLS) screen (11), and the Primary Care (posttraumatic stress disorder) PTSD Screen for DSM-5 (PC-PTSD-5; cutoff score, ≥3) (12). The program then used a preset algorithm to make the following referrals: sleep clinic, sleep behavioral clinic, urology clinic, and headache clinic. MySleepScript also generated customized patient educational materials that study participants could elect to receive electronically. Educational topics included insomnia, circadian rhythm sleep disorder, RLS, sleep apnea, and PTSD, in addition to a generic sleep environment education hand-out. These educational materials gave patients general information about the symptoms that raised concerns and behavioral interventions that could be implemented at home to potentially improve sleep.

Descriptive statistics were performed for demographic and clinical characteristics. Continuous variables are presented as means and standard deviations, and categorical variables are presented as frequencies and percentages. Independent samples t test was used to compare mean questionnaire scores, and one-way analysis of variance was done to compare differences across several groups. Categorical variables were compared using the chi-square test and Fisher’s exact test. All tests were two-sided with a significance level of p-value <0.05. Analyses were performed using SPSS version 20 (IBM, Armonk, N.Y.) and STATA version 12 (StataCorp, College Station, Tex.); data did not violate the assumptions of the parametric tests used in analyses.

Demographic characteristics of study participants are displayed in Table 1. The mean age of the cohort was 46.3 years (SD=14.8, range: 19–73 years); the majority of participants were Caucasian (82.4%) and male (57.4%).

Characterization of brain injuries is summarized in Table 1. The mean time from the first incident of brain injury was 13.1 years (SD=12.9); 26% of the cohort had more than a single ABI. Traumatic brain injury (TBI) was the most prevalent (69.2%) type of ABI. A total of 57.5% of participants had a period of confirmed loss of consciousness.

A total of 48.3% of participants self-reported dissatisfaction with sleep, as indicated by a response of “dissatisfied” or “very dissatisfied” on the ISI (Table 2). On the basis of the PSQI score (score >5), 83.9% of individuals had poor sleep quality. The mean PSQI score was 9.2 (SD=4.7), and the mean ISI score was 11.8 (SD=7.8; absence of insomnia 0–7). The ISI scores indicated that 44.9% of study participants had moderate to severe clinical insomnia. PSQI and ISI scores were significantly correlated (r=0.82, p<0.001). Higher ISI scores were significantly associated with more minutes taken to fall asleep (r=0.47, p<0.001), greater frequency of difficulty falling asleep within 30 minutes (r=0.54, p<0.001), and greater frequency of waking up in the middle of the night or early in the morning (r=0.44, p<0.01). The Berlin Questionnaire for sleep apnea was positive for 36.5% of participants, and the RLS screen was positive for 39.3% of participants. On the basis of referral algorithms, 65% (N=42) of participants were referred to a sleep clinic and 34% (N=22) were referred to a behavioral sleep medicine clinic.

The majority of participants elected to complete the questionnaires with study staff assistance (69.1%). Assistance consisted of a staff member reading the questions on the application and inputting the patient’s response on the iPad. The mean time of application completion was 16.6 minutes (SD=5.4); mean time was not significantly correlated with whether the questionnaires were completed alone (17.8 minutes [SD=7.6]), with study staff assistance (15.9 minutes [SD=4.4]), or with caregiver assistance (19.8 minutes [SD=7.0], p=0.19). Longer time to completion was significantly associated with older age (r=0.43, p<0.01). Time to completion was also significantly different between groups on the basis of education (F=3.28, df=3, 51, p=0.028, η²=0.16). Specifically, Bonferroni-adjusted pairwise comparisons between the education groups indicated that patients with a high school education/GED or less (16.9 minutes [SD=3.6]) had significantly longer times to completion than patients who completed some college (12.7 minutes [SD=4.8], p=0.048). There were no significant group differences for time taken to complete the application for income levels (F=1.24, df=3, 50, p=0.305, η²=0.07) and type of ABI (F=1.85, df=4, 54, p=0.132, η²=0.12).

Less than half (44.6%) of participants elected to receive the generated information sheets on sleep education and hygiene. Individuals electing to receive information significantly differed on the basis of education levels (χ²=8.19, df=3, p=0.04); a higher proportion of patients with a high school education/GED or less (53.1%) were less likely to elect receiving sleep materials than patients with higher levels of education (some college: 21.9%; college graduate: 18.8%; professional degree: 6.2%). There were no age (t=0.59, df= 63, p=0.56), sex (χ²=0.58, df=1, p=0.45), or income (χ²=3.20, df=3, p=0.36) differences in electing to receive sleep materials.

Individuals with TBI had significantly higher PSQI scores than individuals with other types of ABI (mean=10.1 [SD=4.5] versus 7.4 [SD=4.7]; t=2.08, df=57, p=0.04, respectively). There were no significant associations between ISI scores (t=0.47, df=25, p=0.64) or Berlin scores (t=48=0.60, 48, p=0.55) and type of ABI. PSQI scores differed significantly across the different levels of consciousness following brain injury (p=0.05). However, this was not significant after Bonferroni correction (Table 2). There were significant differences in ISI scores and level of consciousness following brain injury. Insomnia severity scores were significantly different for participants based on the level of consciousness following injury, with higher scores in the group with loss of consciousness ranging from 0 to 30 minutes and in the group with unknown loss of consciousness compared with individuals in the medically induced coma group (15.7 [SD=6.6] versus 12.9 [SD=7.5] versus 2.0 [SD=1.6], respectively, p=0.003). Results from the Berlin Questionnaire were not associated with level of consciousness (p=0.260). The Berlin Questionnaire was positive for 77.6% of participants with a loss of consciousness ranging from 0 to 30 minutes and for 13.3% of participants who lost consciousness for >30 minutes and negative for participants placed in a medically induced coma.