Participants
We are recruiting participants through an aged persons’ community mental health services in regional catchments of Victoria, Australia. A highly inclusive approach will be used to maximise participation and increase generalisability. Participants will be eligible for the trial if they meet the following criteria: aged 65 years or above, been referred to an aged persons’ mental health service, current Comorbid Insomnia Disorder [
1], and past and/or current Major Depressive Disorder [
1].
Participants who meet criteria for Comorbid Insomnia Disorder will report (a) dissatisfaction with their sleep associated with either difficulty initiating sleep, difficulty maintaining sleep, or early morning awakening with the inability to return to sleep; (b) causing significant impairment of functioning; (c) at least 3 nights per week; and (d) is present for at least 3 months. Participants who meet criteria for past and/or current Major Depressive Disorder will report (a) depressed mood or lack of interest/pleasure consistently over a 2-week period; (b) five or more of the following symptoms, which include change of appetite/weight, sleep disturbance, psychomotor agitation/retardation, loss of energy, worthlessness/guilt, indecisiveness, or suicide ideation; and (c) the symptoms cause significant functional impairment.
Participants will be excluded if they are cognitively impaired (Mini Mental State Exam score below 24; [
21]); in a crisis stage of mental illness (for example, exhibiting psychotic features or demonstrating active suicidal intent/plan); on unstable doses of medication (this means eligible participants will need to be on the same doses of their prescribed medications for at least 1 month prior to commencing the intervention); currently participating in Electro-Convulsive Therapy (ECT); or currently participating in CBT with another psychotherapist.
Assessment
Potential participants will be provided with a trial invitation via the aged persons’ mental health service staff. Interested participants will be asked to sign an informed consent document prior to commencing the eligibility assessment. Once consent is obtained, participants will be invited to participate in a preliminary insomnia screen. This screen will correspond with the
Diagnostic and Statistical Manual for Mental Disorders 5
th
edition (DSM-V) criteria for Comorbid Insomnia Disorder [
1] and follow Morin and Benca’s [
40] insomnia assessment guidelines. Participants meeting comorbid insomnia criteria will then be invited to participate in a clinical interview eligibility assessment.
Demographic information will be collected, including age, sex, relationship status, education level, income source, accommodation type, religious beliefs, past occupation, medical history, and medications. Participants will be asked if they have been diagnosed with a sleep disorder, and if so, their treatment (for example, CPAP) and perceived treatment effectiveness. Participants will also be screened for sleep apnea and other comorbid sleep conditions (for example, restless legs) by administering the SLEEP-50 [
65]. It is important to note that participants with comorbid medical, sleep, and psychiatric conditions will be included in this trial.
The Mini International Neuropsychiatric Interview (MINI 6.0; [
58,
59]) will be administered to assess Major Depressive Disorder (past/current episode) and other high prevalence mental disorders. The MINI is a clinical psychiatric diagnostic tool with structured ‘yes’ or ‘no’ responses, (for example, ‘Were you ever depressed or down, most of the day, nearly every day, for two weeks?’). The MINI is more efficient to administer compared to other structured diagnostic instruments (for example, SCID), taking approximately 15 to 30 minutes to complete. The MINI corresponds with the
DSM-IV diagnostic criteria, which is commonly applied in research settings. The MINI has demonstrated inter-rater and test-retest reliability and validity across diverse populations (for example, [
58,
60]) including older adults [
51]. Since the MINI was standardised on the
DSM-IV, the principal investigator updated the items for each disorder to be consistent with the
DSM-V criteria.
The Mini-Mental State Examination (MMSE; [
21]) will be used to screen for possible cognitive impairment. The MMSE contains 30 items that assesses a range of cognitive functions, such as orientation, short-term memory, language, comprehension, attention and calculation [
21]. Total scores on the MMSE range from 0 to 30. Participants with scores of 23 or below are considered likely to have a cognitive impairment [
21]. The MMSE has shown high levels of reliability and validity; for instance, studies have reported internal consistencies of up to Cronbach’s α = .90, and test-retest reliabilities over a 24-hour period above
r = .85 [
36]. In addition, Mitchell [
36] found the MMSE demonstrated adequate sensitivity (71.1 % to 85.1 %) and specificity (81.3 % to 95.6 %) across a range of settings.
Measures
Once the eligible clinical assessment has been completed, eligible participants will be asked to complete a self-report questionnaire package at pre-treatment (week 0) to collect baseline outcome measures. Participants will also be asked to complete the outcome measures at post-intervention (week 8) and 3-month follow-up (week 20). Participants will be financially reimbursed ($20) each time they complete the questionnaire pack ($60 total) to acknowledge this additional time dedicated toward the study. Table
1 shows the schedule and frequency of the assessments during the trial.
Table 1
Assessments and administration frequency
Primary outcomes | Insomnia severity | ISI | x | x | x |
| Depression severity | GDS | x | x | x |
Secondary outcomes | Sleep quality | CSD | x | x | x |
| Diagnosis | MINI 6.0 | x | x | x |
| Anxiety | GAI-SF | x | x | x |
| Hopelessness | BHS | x | x | x |
| Beliefs about sleep | DBAS-10 | x | x | x |
| Sleep conditions | SLEEP-50 | x | x | x |
| Health | EQ-5D-3 L | x | x | x |
Eligibility | Insomnia diagnosis | DSM-V Insomnia Screen | x | x | x |
| Cognitive screen | MMSE | x | - | - |
Moderators | Demographics | Demographic Information | x | - | - |
Treatment | Expectations | TCI | x | - | - |
| Feedback | Reflective Focus Group | - | x | - |
The Insomnia Severity Index (ISI; [
38]) is a self-report instrument measuring the participant’s perception of his or her level of insomnia. The ISI assesses the subjective symptoms and consequences of insomnia, as well as the degree of concerns or distress caused by those difficulties during the previous two weeks. The ISI comprises seven items assessing the severity of sleep-onset and sleep maintenance difficulties (both nocturnal and early morning awakenings), satisfaction with current sleep pattern, interference with daily functioning, ability to notice the level of sleep impairment, and degree of distress or concern caused by the sleep problem. Each item is rated on a five-point scale (0 = none, to 4 = very). Total scores range from 0 to 28, with higher scores representing more severe levels of insomnia. Total scores of 0 to 7 indicate no clinically significant insomnia; scores of 8 to 14 indicate sub-threshold insomnia; and scores of 15 to 28 indicate moderate to severe levels of insomnia. Bastein et al. [
2] conducted a psychometric study of the ISI among older adults. It was found that the mean item-total correlations were
r = .56 at pre-treatment,
r = .69 at post-treatment, and
r = .72 at follow-up. The internal reliability coefficients remained stable from .76 at baseline to .78 at follow-up. Bastein et al. concluded that the ISI was a useful clinical tool for screening insomnia severity among older adults or as an outcome measure in insomnia treatment research.
The Consensus Sleep Diary (CSD; [
12]) allows individuals to record information about their nightly sleep pattern. A recent panel of international sleep experts [
12] developed the CSD as the current standardised sleep diary measure for insomnia research. The CSD contains nine items that were considered by the expert panel to represent the most critical sleep parameters (for example, ‘What time did you get into bed?’). The CSD was formatted so that 1 week of nightly sleep data could be recorded on a single diary page. The CSD instructions included general information, such as what to do if the respondent misses recording on a particular day, and an item-by-item instruction guide to enhance the likelihood of correct item interpretation. Additional instructions indicate that all items should be completed in the morning within 1 hour of getting out of bed. Previous research (for example, [
52]) has reported that sleep diaries were accurate when compared to polysomnographic data (kappa = .87; sensitivity = 92.3 %; specificity, 95.6 %). Researchers agree that having individuals prospectively self-monitor their sleep with a sleep diary is a useful psychometric tool for insomnia assessment and for examining treatment effects (for example, [
10]).
The SLEEP-50 [
65] is designed to measure the intensity of an individual’s subjective sleep complaints on a range of sleep conditions, including sleep apnea, insomnia, narcolepsy, restless legs, circadian rhythms, sleepwalking, nightmares, other factors influencing sleep, and functioning impairment. The SLEEP-50 is scored on a four-point Likert scale from 1 ‘not at all’ to 4 ‘very much’. An example sleep apnea item includes ‘I am told that I hold my breath when sleeping’. The SLEEP-50 is considered a reliable and valid screening tool for comorbid sleep conditions, Spoormaker et al. reporting a 3-week test-retest reliability of
r = .78 and Cronbach’s α of .85.
The Dysfunctional Beliefs and Attitudes About Sleep 10-Item Scale (DBAS-10; [
19]) measures the intensity of maladaptive beliefs about sleep (for example ‘When I have trouble getting to sleep, I should stay in bed and try harder’). Espie et al. redeveloped this measure into a shorter version of 10 items from the original 30-item DBAS scale of Morin [
38]. Participants complete each question using a 10-cm visual analogue scale, anchored with strongly disagree and strongly agree. Added together, the 10 item responses provide the final DBAS score. Total scores range from 0 to 100, with higher scores representing more rigid or stronger levels of dysfunctional beliefs and attitudes about sleep. The DBAS-10 was has been found be a reliable measure of dysfunctional beliefs about sleep among older adults (Cronbach’s
α = 0.88; [
55]).
The Geriatric Depression Scale (GDS; [
74]) is a 30-item clinician-rated questionnaire used to assess depression severity specifically for older adults. Participants are asked to respond ‘yes/no’ to each item, for example, ‘Over the past week have you felt that your life has been empty?’ Possible scores range from 0 to 30, with higher scores indicating the presence of more depressive symptomatology. A total score of 0 to 9 indicates normal levels of depression, 10 to 19 indicates mild to moderate levels of depression, and 20 to 30 indicates severe depression [
74]. Yesavage et al. reported that the GDS had a high degree of internal consistency (Cronbach’s α = .94) and weekly test-retest reliability (
r = .85), and displayed strong correlations with other well-validated depression measures (for example, Hamilton Rating Scale for Depression,
r = .83).
The Geriatric Anxiety Inventory Short Form (GAI-SF; [
11]) is a five-item version of the original 20-item Geriatric Anxiety Inventory [
46], which assesses anxiety severity among older adults. Respondents answer yes/no to the five items, for example, ‘I worry a lot of the time’. A score of three or more indicates probable anxiety disorder. At this cut off, sensitivity was 75 %, and specificity was 87 %. Internal consistency was also found to be high, Cronbach’s α at .81 [
11].
The Beck Hopelessness Scale (BHS; [
5]) is a self-report instrument, which entails 20 true/false statements designed to assess the degree to which an individual holds over negative beliefs about the future from the previous week. For example, a participant that answers ‘True’ to the follow question, ‘My future seems dark to me’, would score 1 point and represent a pessimistic response. Each of the 20 statements is scored 0 or 1, with the total being calculated by summing the pessimistic responses for the 20 items. The total BHS score ranges from 0 to 20, with higher scores reflecting higher levels of hopelessness. A total score ranging from 0 to 3 identifies minimal hopelessness, from 4 to 8 identifies mild hopelessness, from 9 to 14 identifies moderate hopelessness, and greater than 14 identifies severe hopelessness. The BHS has been used widely among older adult community samples (for example, [
68]). Test-retest reliability of the BHS over a 6-week period ranged from
r = .66 to
r = .69, and internal consistency was high, Cronbach’s α = .93 [
7].
The EQ-5D-3 L scale [
69] measures perceived health status. It comprises five items measuring mobility, self-care, usual activities, pain, and mood. Each item is rated on a three-point Likert scale (1 = no problems, 2 = some problems, 3 = extreme problems). The EQ-5D-3 L also contains an additional item, which asks respondents to rate their overall current health level using a visual analogue scale from ‘0 = worst imaginable health’ to ‘100 = best imaginable health’. The EQ-5D-3 L has been validated across several populations and countries (for example, [
24]) and is considered an appropriate outcome measure of health status [
9].
The Treatment Credibility Index (TCI; [
15]) measures treatment credibility and expectancy for use in clinical trials. It involves six items that cover two factors: thinking-based credibility (for example, ‘How logical does the therapy offered to you seem?’) and feeling-based expectancy (for example, ‘How much improvement in your symptoms do you really feel will occur?’). The TCI demonstrated high internal consistency (Cronbach’s α = .85) and good test-retest reliability (
r = .82 for expectancy and
r = .75 for credibility).
Sample size calculation
Statistical power analysis was based on an examination of treatment-time interactions in a repeated measures analysis of covariance (RMANCOVA) of the ISI measure [
38], with adjustment for confounders. We specified an initial value of 18 points in the adjusted mean of the ISI score at pre-intervention [
22]; a mean reduction of 10 points from pre- to post-intervention for the CBT-I treatment [
22]; and a target value of two points difference between the pre-post changes for each of the three treatment conditions (that is, -8, −10, and −12 points, respectively) [
43], with no further change at follow-up [
43], and assuming a ‘within-treatments’ SD of 3.4 points [
22]. This resulted in an effect size of 0.23. Under the assumptions of constant correlation over time (sphericity), with an estimated magnitude of
r = 0.6 based on reported test-retest reliabilities of 0.79 [
14] and 0.86 [
76], and allowance for a design effect [
16] of 1.07 due to clustering of participants within treatment groups (based on an assumed cluster size of five and an intra-class correlation of 0.017, [
22]), with a significance level of 5 % and 80 % power, the required sample size calculated using GPower software [
20] is
N = 35 (33 × 1.07). Maintaining the same specifications, but focusing solely on the CBT-I and CBT-I-D conditions, the specific target difference of two points between the pre-post changes for these two particular treatments corresponds to an effect size of 0.13. In order to have 80 % power to detect this particular difference in a post hoc pairwise comparison, the required sample size is increased to
N = 85 (78 × 1.07).
With regard to the GDS measure [
74], Secker and Brown [
57] reported pre- and post-intervention means of 7.0 and 4.4 in a treatment group, and 5.8 and 4.7 in a control group, a difference of 1.5 units [(7.0 - 4.4 = 2.6) - (5.8 - 4.7 = 1.1)] in the pre- and post-change, and an average within-treatment
SD of 3.5. Replacing the values of 2.0 and 3.4 in the ISI analysis with 1.5 and 3.5 results in effect sizes of 0.16 with respect to comparisons across the three conditions and 0.10 with respect to comparisons between the two conditions. This leads to required sample sizes of
N = 68 with respect to comparison of mean GDS across the three conditions, and
N = 150 with respect to comparison of mean GDS between two of the conditions. Because this is larger than
N = 85 from the ISI projection,
N = 150 becomes our target sample size, which equates to 50 participants or 10 groups per condition.