Abstract
Dementia-related sleep problems can be complex and challenging. Environmental interventions which resynchronise the sleep/wake cycle have been trialled with promising results for people with dementia in institutionalised settings. However, there is less research concerning community-dwelling people with dementia and their family carers. This study involved a five-week feasibility study including timed light therapy, exercise and sleep education. Sleep and physical and mental functioning were measured at the beginning and end of the trial using objective measures, standardised questionnaires and structured participant feedback. Of 15 community-dwelling pairs who participated, nine completed the trial. The case studies presented here reveal that it is feasible for this population to use non-pharmacological interventions, with positive outcomes. However, there are also issues that can mask benefits or prevent compliance. The options for treating dementia are limited. Environmental interventions may help manage dementia-related sleep problems and further trials would be worthwhile to improve compliance and evaluate effectiveness.
Keywords
Introduction
Life expectancy is rising, and the number of people with dementia (PWD) is also expected to increase (Alzheimers New Zealand Incorporated, 2012). Pharmacological options for managing dementia-related symptoms have limited effectiveness and some negative side effects (Ballard, Creese, Corbett, & Aarsland, 2011; Francis, Palmer, Snape, & Wilcock, 1999). Psychosocial approaches, which enable individuals to modify and manage their symptoms, have been highlighted as key for reducing the economic and social impact of dementia as well as contributing to the amount of time PWD can remain in their own home (Alzheimers New Zealand Incorporated, 2010; Jyrkkä, Enlund, Korhonen, Sulkava, & Hartikainen, 2009; Prince, Albanese, Guerchet, & Prina, 2014).
Sleep problems have been highlighted as one of the most challenging behavioural symptoms of dementia (Hope, Keene, Gedling, Fairburn, & Jacoby, 1998). Sleep disturbances and irregular sleep timing are associated with poorer cognitive performance (Moe, Vitiello, Larsen, & Prinz, 1995), poor physical and mental health (Foley et al., 1995) and mortality (Gehrman et al., 2004). Therefore, the appropriate management of dementia-related sleep problems could lead to subsequent improvements in waking symptoms and quality of life of PWD.
Carers of PWD are also more likely to report sleep problems. This is typically due to the sleep disturbance caused by the PWD, the effect of caregiver burden, and the co-morbidities of the caregiver (McCurry, Logsdon, Teri, & Vitiello, 2007). When the sleep of both the PWD and caregiver is affected, the overall care-giving situation appears to become more difficult (Gibson, Gander, & Jones, 2014). Therefore, it is important to investigate the sleep experience of PWD while they are still residing at home, along with the sleep of their family caregivers.
Previous trials in institutions, as well as the community, have revealed positive outcomes for PWD using a combination of bright light therapy with physical or social activities, as well as sleep hygiene education or behavioural modification. Such interventions are low-risk as they use environmental and behavioural factors rather than pharmaceuticals to promote more robust rhythms of sleep timing, a reduction in dementia-related sleep disturbances, as well as improvements in other psychosocial factors. For some, such improvements are greater than pharmacological alternatives (Alessi, Yoon, Schnelle, Al-Samanai, & Cruise, 1999; Martin, Marler, Harker, Josephson, & Alessi, 2007; McCurry, Gibbons, Logsdon, Vitiello, & Teri, 2005; McCurry et al., 2011).
Previous research in this area lacks the view point of the PWD because of its typically controlled and quantitative nature. However, there is a growing movement to actively include PWD in the research process (McKeown, Clarke, Ingleton, & Repper, 2010; Wilkinson, 2002). To achieve this, a ‘person-centred’ and flexible approach to research is required, so that PWD and their families are appropriately included in the process (Finnema, Dröes, Ribbe, & Van Tilburg, 2000; Innes, 2009). The aims of the present study were to pilot a community-based mixed-methods approach for gathering information on the sleep of PWD and carers in New Zealand; to design and evaluate practical non-pharmacological interventions to improve sleep and waking function; and to use a case-study design to present each pair's story thoroughly. The effectiveness of the interventions was evaluated by comparing objective and subjective sleep data of the pairs before the intervention and in the fifth week of the intervention. The waking symptoms of dementia and impact of caregiving were also compared between Times 1 and 2. Summaries of the 15 sets of results and comparisons of grouped data are reported elsewhere (Gibson, 2015). The present article describes case-study examples. This is an important approach because PWD and carers are a diverse population who benefit from and deserve assessment and treatment on a case-by-case basis.
Methods
Participants
A convenience sample of 15 pairs were recruited via Alzheimer's Wellington, a local organisation offering support and advocacy service for PWD living in the Wellington Region of New Zealand. Community advertisements, presentations and referrals from healthcare professionals were also used. Inclusion criteria were living together as a PWD/family carer pair in the Wellington community, the PWD being aged ≥ 65 years, and that they and/or their family carer had an interest in improving their sleep. Diagnosis of dementia was based on the participating pair's report of a medical diagnosis. Because of the use of bright light in the intervention, participants were screened for any eye conditions or medications which could cause photosensitivity (Wirz-Justice, Benedetti, & Terman, 2008). Those who were considered at increased risk used natural light rather than the artificial bright light therapy.
Participants with dementia provided written consent when able, otherwise proxy consent was obtained via the carer. Verbal assent was also gathered via audio recording during the study briefing. Participants' doctors were informed of their patient's participation and given the opportunity to notify the researcher of any further contraindications for bright light exposure. Approval was obtained from the Central Health and Disability Ethics Committee (CEN/11/01).
The interventions
The interventions used included bright light therapy, exercise and sleep hygiene education. A light box was provided (Day-Light; Uplift Technologies Inc., 2009) as an alternative to natural light. The Day-Light provides up to 10,000 lux of illumination at a comfortable sitting distance (30–33 cm; Wirz-Justice et al., 2008). The proposed exercise regime was 30–40 minutes of walking or yoga style stretches and movement. Participants were loaned a senior's exercise DVD (Sit and Be Fit; McLennan, 1999 ) and were given a pamphlet to promote safe warm down after exercise (Push Play Stretches; Sport and Recreation New Zealand, 2003).
A booklet was developed for the study that contained information on sleep with ageing and dementia, and sleep hygiene guidelines. Content of the ‘sleep support handbook’ was based on the current literature and other sleep hygiene guides for older adults (e.g., McCurry, Gibbons, Logsdon, Vitiello, & Teri, 2003; National Sleep Foundation, 2011).
Procedure
All study procedures took place within the participants’ homes. Delivery of the intervention was provided by the first author (‘the researcher’). PWD completed their questionnaire in the presence of the researcher and their family carer to support completion. In cases when the PWD was not cognitively able, questions were answered in an interview style, or the family carer acted as a proxy. Participants were telephoned by the researcher on a weekly basis to remind them of the study protocol and answer any questions regarding the study. They also had access to a free phone number to contact the researcher throughout for advice or assistance. Documents describing how to complete the interventions and with safety instructions were provided.
During the five-week trial, PWD were encouraged to get 30 minutes of bright light exposure every morning (or on as many mornings as possible) between 09:00 and 11:00, using either the light box or natural light. PWD were also asked to undertake the low intensity exercises on as many days as possible during the middle of the day (11:00–14:00). Carers were encouraged to join in with physical activities or allocate appropriate supervision as they saw necessary. Finally, participants were encouraged to try some of the ideas/solutions offered in the sleep support handbook, and note in the sleep diaries if and when they had tried new things.
Outcomes
The study protocol included a baseline week of sleep monitoring and questionnaire assessments (Time 1), followed by a five-week intervention with sleep monitoring during the fifth week and questionnaire assessments at the end of the intervention (Time 2). Sleep/wake activity of both PWD and carers was measured for seven days at Time 1 and again at Time 2 using the Actiwatch-2™ (Mini Mitter, Philips, Respironics) worn on the non-dominant wrist. The device has a button serving as an event marker to press at relevant times (e.g. at bedtime), and a sensor for measuring the amount and duration of ambient white light exposure (lux) from the wrist.
Activity counts were recorded in one-minute intervals and actigraphy records were analysed using custom software provided by the manufacturer, with the medium sensitivity threshold. Time in bed was defined using information from the event marker, light data, and 24-hour time lines in daily diaries (typically completed by carers). They were asked to note times of sleep start and end (to the nearest 15 minutes) for any sleep that was 10 minutes or longer. Diaries recording the sleep times and activities of the PWD were continued between Times 1 and 2 to help document compliance with the study protocol.
The PWDs’ questionnaire included demographic details (age, gender, type and date of dementia diagnosis), and validated questionnaires concerning sleep (the Pittsburgh Sleep Quality Index (PSQI); Buysse, Reynolds, Monk, Berman, & Kupfer, 1989, range =0–21, > 5 = ‘disturbed sleep’), and health (the Self-Administered Comorbidity Questionnaire (SCQ); Sangha, Stucki, Liang, Fossel, & Katz, 2003). They also completed the Mini Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975, range = 30–0, > 23 = ‘no impairment’, 19–23 = ‘mild cognitive impairment’, 10–18 = ‘moderate cognitive impairment’, and ≤ 9 = ‘severe cognitive impairment’) and the Quality of Life in Alzheimer's Disease: Patient and Caregiver Report (QOL-AD; Logsdon, Gibbons, McCurry, & Teri, 1999, range = 52 (excellent)–13 (poor)).
The carers’ questionnaire included demographic details (age, gender and caregiving timing and relief), questions about their own sleep and health (the PSQI, SCQ and MMSE), as well the Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983, range = 0–21, 0–7 = ‘normal’, 8–10 = ‘borderline’, 11–21 = ‘heightened risk’) and the Carers of Older People in Europe index (COPE; McKee et al., 2003, range for positive impact of caring = 15–0, < 12 = reduced coping. Range for negative impact of caring, >12 = reduced coping). Carers also completed standardised questionnaires concerning the PWD's sleep (the Sleep Disorders Inventory (SDI); Tractenberg, Singer, Cummings, & Thal, 2003, range = 0–12/‘severely disturbed sleep’) and QOL-AD. Dementia-related symptoms were measured using the Revised Memory and Behaviour Problems Checklist (RMBPC; Teri et al., 1992), which provides a global score (0–96) as well as sub-scores for memory (range = 0–28), depression (range = 0–36) and disruptive behaviours (range = 0–32), with higher scores indicating increased frequency of behaviours and negative carer affect. Medications were recorded for both PWD and carers and assessed for sleep-related side-effects.
Feedback about the methods and research processes was collected at the end of their involvement, regardless of whether they completed the entire protocol. This involved rating each aspect of the methods and intervention (easy, ok or difficult) as well as open-ended questions concerning the protocol and interventions.
Data analysis
Actigraphy records were visually examined and analysed with the diary data to identify rest periods. Sections of complete inactivity (due to the actiwatch not being worn or for unknown reasons) were excluded from analysis. Rest periods were categorised into night and day. Night was defined as the time the participant initially went to bed intending to sleep (bedtime) until the final morning rising (rise time), and therefore varied both between and within subjects. Note if participants had reported periods of wakefulness during the night, these periods were still scored as within the night sleep interval. The remaining time constituted day intervals. Key actigraphy variables presented here are total sleep duration at night (hours), sleep efficiency (percentage of time spent asleep whilst in bed at night) and sleep duration in day (minutes). Medians were derived for each variable for each participant at Times 1 and 2 and the difference between medians calculated. A purpose-built algorithm was used to calculate the percentage of monitored days on which a participant was asleep (defined by actigraphy) in each 10-minute interval across the 24-hour day. These ‘sleep propensity curves’, which summarise an individual's sleep/wake pattern, were calculated separately for Times 1 and 2.
Global scores were calculated for each questionnaire and any subscales. The PSQI was considered key, with scores of more than 5 indicating trouble sleeping (Buysse et al., 1989). Compliance with the intervention was evaluated from the diaries. The number of days with light or activity logged inside or outside the recommended time frames was calculated. Field notes were collected at the end of each home visit and after telephone conversations to record participants' experience with the protocol and any feedback. Effectiveness of the intervention and the feasibility of the protocol and intervention was evaluated based on the case studies (Blampied, 1999; Thomas, 2011). Pseudonyms are used throughout.
Results
Those who completed the trial
Key outcome variables and summary of improvements between Times 1 and 2 for PWD and carers (n = 9).
PSQI: Pittsburgh Sleep Quality Index; TST: Total Sleep Time; MMSE: Mini Mental State Examination; SDI: Sleep Disorders Inventory; QOL-AD: Quality of Life in Alzheimer's Disease; RMBPC: Revised Memory and Behaviour Problem Checklist; HADS-A: Hospital Anxiety and Depression Scale–Anxiety; HADS-D: Hospital Anxiety and Depression Scale–Depression; COPE: Carers of Older People in Europe index; ∨: decrease in scale; ^: increase in scale.
Case studies 1 and 2: Improvement across the protocol versus marked deterioration in the condition of the PWD
Cases 1 and 2 illustrate contrasting examples of improvement across the protocol versus marked deterioration in the condition of the PWD.
Pair 1: Adam and Claire
Adam and Claire were a married couple. Adam was 82 years old and had been diagnosed with Lewy Body dementia, diagnosed approximately five years prior to the study. He had an MMSE score of 5/30, indicating severe cognitive impairment. He also had Parkinson's disease and lung disease. Some of his medications had fatigue listed as a potential side effect, two listed insomnia and one also listed nightmares. Claire was aged 68 years at the time of the study. She had been providing dementia-related support for her husband for six years. She also worked a 40-hour week, during which time a relief carer assisted Adam. Claire had haemochromatosis and suspected arthritis. She took two medications which had drowsiness and fatigue listed as possible side effects, as well as (more rarely) dizziness, insomnia or nightmares.
Adam had a raised PSQI score (6/21). This was related to waking up in the middle of the night and early morning, instances of coughing or snoring, feeling too hot or cold, as well as bad dreams (three or more times per week). Claire also noted that he had some daytime sleepiness and, less often (once or twice per week), uncomfortable breathing which disrupted his sleep. Claire added that he had frequent symptoms of restless legs syndrome, would walk around at night, have episodes of confusion or disorientation, and would come and wake her up (three or more times per week). The SDI score also indicated that Adam's sleep was disturbed (score = 4.2/12), including wakefulness and wandering-type behaviours at night, and sleepiness in the daytime. Claire rated many of these behaviours as severely distressing. They slept in separate rooms.
According to Adam's diaries, he used the light box on 23 of the 35 days of the trial during the recommended time frame, and on a further four days, he used it later in the morning. Claire tended not to enforce using it on the mornings when Adam was more agitated. They did not use the exercise DVD because Adam was not physically able. He only walked inside and up and downstairs. Only one session of physical activity was documented across the five-week trial. Claire reported that the sleep support handbook was clear and informative. One aspect they reported trying was changing the time of day that Adam took one of his medications.
During the trial, Claire noticed that Adam was much more active during the days that he used the light. She commented that he would wake up more ‘sprightly’ and at a more convenient time in the morning. She also noticed that Adam was more lucid during the times of light exposure. For example, he would be more likely to participate in completing the crossword or feed himself breakfast. She commented that he seemed less disturbed at night and not as sleepy during the day. For them, this had meant being able to go out on social excursions and enjoy their time together more.
At Time 2, Claire still rated Adam's sleep as poor. His PSQI score increased to 7/21 because of an increase in hallucinations and talking during his sleep. However, the SDI score decreased to 2.9/12. This was related to Adam not being as sleepy in the daytime or waking so early. Furthermore, the severity ratings and carer-related distress associated with Adam's sleep disturbances reduced at Time 2. Claire reported that although Adam was still sleepy during the day, this had reduced and he seemed much more alert when he was awake, which enabled them to have more of a social relationship. Claire rated his quality of life more highly at Time 2 compared with Time 1 (QOL-AD = 34.0 vs. 27.0/52). The frequency of memory-related symptoms also improved between Times 1 and 2 (as indicated by the RMBPC memory score reducing from 24 to 15/28).
Adam's actigraphy data corroborated that he had very variable sleep timing and exceptionally disturbed sleep at Time 1 (Figure 1). His median night-time sleep interval was between 19:00 and 07:00 at Time 1, whereas at Time 2, it was between 18:08 and 09:20. There were clear improvements at Time 2, when he was getting almost 3.5 hours more sleep at night (median = 12.3 hours, range = 11.0–14.4 hours, vs. 8.5 hours, range = 6.6–11.7 hours). Also, despite being in bed for longer, he had fewer awakenings during the night at Time 2. This was reflected in the increase in his sleep efficiency from a median of 78% at Time 1 (range = 53–88%) to 81% at Time 2 (range = 69–97%). Adam's daytime sleep increased from a median of 78 minutes at Time 1 (range = 47–269 minutes) to 192 minutes at Time 2 (range = 58–198 minutes). However, as shown in Figure 1, his naps were more likely to occur later in the day and were more consolidated compared with Time 1.
Adam's sleep timing at Times 1 (black line) and 2 (grey line), six days. At Time 2, Adam was averaging more sleep during his night intervals and had more a more consistent pattern of daytime napping.
Actigraphically, Claire had reasonably good sleep at both Times 1 and 2. Her median night-time sleep interval was at 22:00–06:00 at Time 1 and 22:08–05:55 at Time 2. She was having a median of 8.0 hours sleep at Time 1 (range = 6.5–9.3 hours) and 6.7 hours at Time 2 (range = 6.1–8.5 hours). She spent a median of 92.4% of her time in bed at night asleep at Time 1 (range = 83–95%) and 91% at Time 2 (range = 82–97%). However, Claire's subjective ratings of sleep indicated symptoms of insomnia and awakenings at night at Time 1. Her PSQI score improved between Times 1 and 2 (9 vs. 6/21). This was due to reductions in the amount of time she felt it was taking her to get to sleep, being less frequently disturbed by Adam, and not needing to use a sleeping medication as often. Her HADS and COPE scores remained within normal range across the trial.
Pair 2: Jack and Ella
Jack and Ella were a married couple who were living in a motel while their home was undergoing renovations. Jack was 82 years old and had been diagnosed with vascular dementia 11 years prior to the study. He had an MMSE score of 13/30, indicating moderate cognitive impairment. His comorbidities included heart and lung disease as well as a colostomy and urinary incontinence. One of the medications he was taking had somnolence listed as a possible side effect. Ella was aged 74 years old at the time of the study. She had been providing dementia-related care for her husband for over 11 years. She had support from a relief carer for about eight hours per week. Ella had an arthritic condition and one of the medications she was taking had somnolence listed as a possible side effect.
Jack had a raised PSQI score (8/21). This was mostly related to getting up to use the bathroom in the night (three or more times per week) and daytime sleepiness (once or twice per week). In addition, Ella reported that Jack had occasional episodes of disorientation or confusion, loud snoring (occurring once or twice a week), as well as occasional apnoeas. His SDI score indicated some issues with getting to sleep as well as daytime sleepiness. However, these were not defined as severe or occurring every day and so the total score was low (0.2/12). Jack and Ella slept in the same bed.
In the diaries, Jack was documented as using the light box on 22 out of 35 days during the 9:00–11:00 time period (although some instances began prior to 09:00). Ella also recorded five episodes of natural light exposure for Jack during the allotted timeframe. Ella reported that they used the light successfully and that Jack enjoyed using it while reading. However, he sometimes fell asleep in front of it. This meant that Ella had to monitor him carefully and try and keep him awake during the light therapy time.
Because of their small living space, they chose not to attempt the exercise DVD. They already did some walking and were content to continue this form of exercise for the physical activity aspect of the intervention. However, the weather was cold during their intervention period and they admitted missing some of their scheduled exercise. Physical activity was recorded on 20 out of the 35 days between 11:00 and 14:00. Ella noted that the sleep support handbook was informative and that she had the impression that the intervention was helpful for them.
At the end of the trial, Ella commented that Jack's behaviour and memory had become much worse. This was reflected in a reduction in his MMSE score (from 8 to 5/30) and an increase frequency of dementia-related behaviours noted on the RMBPC (4.6–8.0). Ella commented that he had increased difficulty with language, would refuse to get ready for bed and was having more episodes of confused and agitated behaviour at night. He also had periods when he would resist care assistance. She also noted that he was still sleepy in the day and it was hard work encouraging him to do any activities. She felt she was so tired at night that it was likely that she was sleeping through some of Jack's awakenings and potentially under-reporting the degree of their sleep disturbance.
Jack still rated his sleep quality as ‘fairly good’ at Time 2, and his diary ratings reflected this. Jack's PSQI score improved to be within normal range (5.0/21). However, some of this reduction could be attributed to the pair ‘not knowing’ his sleep onset timing at Time 2. The SDI rating showed minor improvement to Jack's sleep (0.1/12). Jack's self-rated QOL-AD score remained mediocre between Times 1 and 2 (37 vs. 34/52, respectively), Ella also rated his quality of life as stable although lower than Jack did (23 vs. 25/52).
Jack's actigraphy data showed that he had very variable sleep timing and exceptionally disturbed sleep at both Times 1and 2 (Figure 2). His median night-time sleep interval was between 21:10 and 08:00 at Time 1, and at Time 2, it was between 22:20 and 07:40. At Time 2, he actually showed a reduction in his median total sleep time at night (median = 5.7 hours, range = 3.0–7.4 hours) compared with Time 1 (median = 6.7 hours, range = 4.3–7.6 hours). His sleep efficiency remained stable but low between Times 1 (median 63%, range = 61–78%) and 2 (median = 61%, range = 31–70%). His daytime sleep increased from a median of 27 minutes at Time 1 (range = 24–76 minutes) to 47 minutes at Time 2 (range = 23–78 minutes). However, as shown in Figure 2, he napped less during the mornings at Time 2 compared with Time 1.
Jack's sleep timing at Times 1 (black line) and 2 (grey line), six days. At Time 2, Jack showed minimal improvement to his sleep quantity or quality.
Actigraphically, Ella had disturbed sleep at both Times 1 and 2 with more variable sleep timing and efficiency at Time 2. Her median night-time sleep interval was at 23:10–07:00 at Time 1 and 23:10–07:00 at Time 2. She was having a median of 6.8 hours sleep at Time 1 (range = 6.1–8.7 hours) and 6.9 hours at Time 2 (range = 4.4–7.7 hours). She spent a median of 89% of her time in bed at night asleep at Time 1 (range = 84–90%) and 83% at Time 2 (range = 77–92%). However, Ella's subjective ratings of sleep indicated some improvements between Times 1 and 2. Her PSQI score was raised at Time 1 (8.0/21), which she related to her waking early, getting up to use the bathroom and coughing or snoring. However, this decreased to within normal ranges at Time 2 (3.0/21), despite her comments of Jack being more disruptive at night and her suffering from some anxiety associated with coping. Her HADS scores indicated increased likelihood for depression at both Times 1 and 2 (HADS-D = 9 and 8/21, respectively). Her anxiety score and COPE index were within the normal range across the trial.
Those who withdrew from the trial
Six of the pairs withdrew during the trial. These could be classified into two categories: first, those who had less severe symptoms of dementia with minor sleep problems, and who lost interest or were confused by the study protocol; and second, those who had more advanced symptoms of dementia with substantial sleep problems, but were admitted to hospital or a care facility during the trial period. Cases 3 and 4 illustrate these two types of pairs who withdrew from the trial.
Case studies 3 and 4: Loss of interest in study protocol versus rapid deterioration of illness
Pair 3: Georgina and Lisa
Georgina was 89 years old and was temporarily living with her daughter, Lisa and her family, while the family organised care assistance for her in her home town. She had an undefined type of dementia diagnosed approximately two years prior to the study. Her MMSE score was 23/30, which was within the normal range of cognitive functioning. However, her RMBPC indicated that she had some problems with memory, as well as some disruptive behaviours related to dementia (RMBPC global frequency score = 28/96). Lisa reported Georgina's quality of life as lower than she did 34 vs. 44/52, respectively). Georgina's comorbidities included heart disease and high blood pressure. She frequently took a sleeping tablet, which understandably had sedating effects as well as having other side effects on sleep. Georgina also took another medication, which had fatigue, insomnia and abnormal dreams listed as possible side effects.
Lisa was aged 55 years at the time of the study. She had been providing dementia-related care for her mother for six months. She was providing care for several hours a day, whilst also working a 40-hour week and caring for her teenage children. She had support from her husband. Lisa also took a sleeping medication regularly, which had sedating effects as well as having other side effects on sleep. Her HADS and COPE scores were within the normal range.
At the beginning of the study, Lisa had some concerns as Georgina was not fully aware of her dementia diagnosis, so she did not want the research process to be too distressing. However, they both had sleep problems and were interested in the trial. The researcher assured Lisa that the home visit would be approached very carefully, and that additional time and support would be made available. Georgina seemed anxious during the study briefing and data collection, and her behaviour was somewhat defensive towards the researcher, despite having declared that she was interested and happy to participate. Lisa reported that Georgina was a little upset after the initial home visit and appeared to consider the researcher as a clinician, despite her seeming to understand what the research was about. She often refused to wear the Actiwatch. Lisa commented that Georgina would forget why she had to wear it and remove it, and then would appear paranoid when reminded to put it back on again.
Georgina sand Lisa slept in separate rooms. Georgina rated her sleep as ‘very good’ and her PSQI score was within the normal range (4.0/21). She would get up to go to the bathroom (three or more times per week), and occasionally have trouble sleeping due to bad dreams. Lisa's score on the SDI indicated that Georgina had some additional issues with her sleep including snoring, getting up in the night, waking others and sleeping excessively in the day (however, most of these activities were occurring less than once per week and were not considered severe, hence a low score of 0.5/12). Actigraphically, Georgina was having reasonable sleep at night. Her median night interval ranged from 22:13 to 07:35, within which she had a median sleep duration of 8.0 hours (range = 7.1–8.3 hours). However, some nights she was spending over 2 hours awake, resulting in variable sleep efficiency (median = 85%, range = 75–90%). No daytime sleep was recorded.
In comparison with her mother, Lisa's sleep was poor. She had a PSQI score of 9/21, which was mainly associated with not being able to get to sleep within 30 minutes, waking in the night or early morning, getting up to use the bathroom and having bad dreams (once or twice per week). She regularly took sleeping medication (three or more times per week) and reported that she had trouble sleeping because she was stressed and restless. Her actigraphy records corroborated these sleep disturbances. Her median night interval ranged from 22:10 to 06:25; however, her median sleep duration at night was just 5.9 hours (range = 4.9–7.4 hours), and her median sleep efficiency was 73% (range = 64–90%).
Georgina and Lisa were issued with a light box and DVD; however, Georgina resisted using either. Lisa reasoned that this was due to a negative association her mother had made with the research process. They were reminded that they could withdraw at any time. They continued to trial the interventions over the subsequent weeks as Lisa believed that her mother would ‘come around’. Lisa reported that Georgina was having more walks outside and they felt there may have been some improvement to sleep. She also reported having more nice dreams. However, she still refused to use the light box. Lisa struggled to maintain the diaries after week 1, commenting that she lost the motivation without having the Actiwatch to remind her to complete them.
Towards the end of the trial, Lisa was concerned that going through the sleep monitoring process and questionnaires again might upset her mother. Georgina had refused to wear the Actiwatch when suggested at the end of the trial. Lisa felt that the study was tainted by her experience at the onset and did not want to add to their family's burden by pursuing the study any further.
In the feedback form, Lisa commented that despite being made fully aware of what the study entailed, she hadn't realised how much of a negative impact completing the questionnaire would have on her mother. There was a resistance to answering the questions, which Lisa interpreted as the thought and concentration required being too challenging as well as confronting. She also associated the negative impact of the research with paranoia, which Georgina typically exhibited with regards to clinical processes, rather than any problem with the researcher's approach.
Pair 4: Edward and Rose
Edward and Rose were a married couple. Edward was 73 years old and had vascular dementia diagnosed two years prior to the study. He had a MMSE score of 18/30, indicating moderate cognitive impairment. His comorbidities included high blood pressure, anaemia or blood disease, cancer, depression, back pain and a prostate problem. Three of his medications had tiredness or fatigue listed as possible side effects. Edward rated his quality of life more highly than Rose did (39 vs. 23/52, respectively). Rose was aged 55 years at the time of the study. She had been providing dementia-related care for her husband for one year. She was providing care all day and night and was in the process of organising some additional support. No comorbidities or medications were disclosed. Her HADS scores were within the normal range indicating low risk for anxiety or depression. However, her positive COPE score was reduced, indicating increased likelihood for carer burden. Edward and Rose had recently moved into the region because of downsizing their business and home to manage Edward's symptoms and be closer to family.
At the beginning of the study, Rose commented on how bad Edward's sleep was, giving the example that he would go to his room for a nap and return five minutes later believing he had slept for a long time. They would often be awake at night and Edward was also sleeping a lot during the day. She was very keen for them to try a non-pharmaceutical alternative for treating his sleep problems. Edward rated his sleep as ‘fairly good’. His PSQI score was within the normal range (4.0/21); however, he had trouble sleeping because of waking in the night or early morning, getting up to use the bathroom, coughing or snoring, pain or being woken by Rose (three or more times per week). Rose also noted that he had occasional apnoeas. The SDI score indicated that Edward had some additional, dementia-related sleep disturbances (2.0/12). These included episodes of disorientation or confusion, wandering and disruptive behaviours at night as well as sleeping excessively during the day. Rose typically rated these as moderate-marked and she found them distressing. Edward completed three days of actigraphy recording before withdrawing from the study. Results showed that his median night interval ranged from 20:30 to 07:00. However, he was sleeping just a median of 5.5 hours at night (range = 5.2–6.4 hours). He was spending between 3.5 and 6.7 hours of the night awake, so his sleep efficiency was very low (median = 61%, range = 44–62%).
Edward and Rose slept in the same bed. Rose rated her own sleep as ‘fairly bad’. Her PSQI score indicated some trouble sleeping. This was mostly related to not being able to get to sleep within 30 minutes, waking up in the middle of the night or early morning, getting up to use the bathroom, feeling too hot and being woken by Edward (three or more times per week). She also had some instances of restless legs (three or more times per week). Rose's actigraphic sleep timing was variable. Her median night interval ranged from 23:30 to 07:20. She was having a median of 7.4 hours of sleep at night (range = 6.6–7.9 hours); however, she was spending up to 1.9 hours awake, so her sleep efficiency sometimes affected (median = 90%, range = 80–92%).
Four days into actigraphy monitoring at Time 1, Edward had a fall (suspected seizure) and was admitted to hospital. They suspended their sleep monitoring at this time due to the hospital environment and changes to medications. On Edward's return from hospital, Rose commented that his sleep had become much worse. She was finding it so hard to manage that she had some sleeping tablets prescribed for Edward, but they didn't seem to be having any effect. She was still very enthusiastic about taking part in the trial and the researcher agreed that, if the baseline questionnaire data could be repeated, then they could still trial the interventions. However, the following week, Edward moved into a care facility and they withdrew from the study. Rose expressed her disappointment at not having had the opportunity to try the interventions. She was given the sleep support handbook. She mentioned that she would endeavour to help Edward exercise and get more sunlight whenever possible.
Discussion
The present study was unique as it was home-based and included both PWD and carers using a dyadic case-study approach. This allowed for factors contributing to the feasibility of the intervention and protocol to be assessed at a person-centred level. For example, it offered insights on the effects of progression of disease on coping with daily living and maintaining the study protocol, and participant's feelings towards their diagnosis of dementia and presence of the researcher, which in some cases affected their compliance or comfort with the protocol. The interventions were based on well-defined physiological and psychosocial processes affecting sleep, and have been successfully trialled with other populations including PWD, but with variable results regarding effectiveness of the interventions (Colenda, Cohen, McCall, & Rosenquist, 1997; Friedman et al., 2012; McCurry et al., 2005, 2011).
The set of case studies presented here illustrates the diversity of PWD and their carers' response to the intervention, as well as some of the challenges of conducting dementia-related sleep research in the community. They also show that, given the right timing and conditions, these non-pharmacological interventions for improving sleep are a welcomed option for PWD and their carers', and that they are feasible to use at home, with minimum risk and with the potential for positive outcomes.
For those with substantial but stable symptoms of dementia, as well as considerably disturbed sleep, the findings suggest that it can be feasible for PWD to use non-pharmacological interventions in their own home, with some positive effects for both PWD and carers. For example, Pair 1 found great benefit from using the light box. They enjoyed taking part in the study and were as compliant as possible considering Adam's severe dementia-related impairments. Sleep was described as a huge problem for them at the start of the study. They found the intervention beneficial for Adam's sleep quantity and quality, which was reflected in the actigraphic recordings as well as in Claire's reports. It appeared that Adam's sleep became more organised and predictable. This subsequently influenced how he functioned and behaved during the day, as well as improving how Claire felt she slept.
The other group who completed the full study were pairs who had substantial sleep disturbances, but there was rapid deterioration in the PWD's dementia-related symptoms or comorbidities between Times 1 and 2, over-riding any potential benefits from the intervention. For example, Pair 2 were reasonably compliant with the research protocol, despite Jack’s stage of dementia and their unusual living conditions. However, Jack's cognitive functioning and dementia-related behaviours deteriorated during the study period. In turn, Ella reported that his sleep had become more disturbed and that he had increased daytime sleepiness. She also reported deterioration of her own sleep due to providing more care at night and increased anxiety. The actigraphy recordings showed no improvement in objective sleep variables.
Six of the 15 who enrolled in the study withdrew before Time 2. This degree of attrition was not unexpected (Harris & Dyson, 2001; McCurry et al., 2011). However, it is important to understand the reasons behind withdrawal to try to improve study recruitment and compliance in future research. Some withdrew because they lost interest or were confused or upset by the study protocol. For example, Pair 3 had many barriers to completing the protocol. Georgina's anxiety around clinical processes and her not being fully aware of her dementia diagnosis made the consent process and data collection difficult. Her hostility towards the researcher transferred to any activities associated with the research. Her carer, Lisa, felt her own sleep was affected by her work commitments and stress, which she also admitted impacted their compliance with the research protocol.
Others withdrew due to unforeseen circumstances preventing them from completing, despite the participant's substantial sleep complaints and dedication to the study. For example, Pair 4 were potentially ideal candidates. They both had considerably disturbed sleep at night and Rose was very enthusiastic about trying something non-pharmaceutical for Edward's condition. However, Edward had a fall and was admitted to hospital, at which point his dementia deteriorated and medications changed. Soon after this episode, he moved into a care facility and withdrew from the study.
Such mixed findings have also been reported in previous studies, using various research designs (e.g., Brown et al., 2013; Forbes et al., 2009). Inconclusive results have been attributed to variability within the population of older people and PWD. Differences in demographic factors, comorbid health issues and medications, type and stage of dementia as well as the fluctuating nature of sleep problems and dementia-related behaviours, are all likely to be confounding variables (Foley, Ancoli-Israel, Britz, & Walsh, 2004; Phillips & Ancoli-Israel, 2001). Comparisons between studies are also difficult because of differences in the measures of sleep used and the specific characteristics of the interventions.
This was an exploratory study. The qualitative and case study approach offered valuable insights on the heterogeneous nature of the population. The person-centred approach helped the introduction of the intervention during a challenging time when participants were coping with the diagnosis, symptoms and prognosis of dementia. However, the small sample size and lack of a control group limited the ability to reliably measure the effectiveness of the interventions. To address this, ideally a randomised controlled trial with a more homogeneous sample would be undertaken. This is challenging with this population due to constraints around access, recruitment and retention of research participants (Bartlett & Martin, 2002); the variable nature of ageing and dementia (Eeles, 2006) and ethical constraints (Wilkinson, 2002).
Qualitative data offer a rich understanding of each pair's experience without over-generalising the findings. The psychosocial enquiry into the experience of sleep problems and how to manage them is also considered useful in the clinical setting (Finnema et al., 2000). Given the insufficient options for treating dementia-related diseases, non-pharmacological options such as timed light therapy, physical activity and sleep education are considered promising for healthcare professionals and support workers to consider when managing the sleep and wellbeing of PWD. Further studies would be worthwhile to improve compliance and evaluate effectiveness.
Footnotes
Acknowledgements
We would like to acknowledge the participants of this study.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Health Research Council of New Zealand feasibility (feasibility grant number: 11/562) and Maurice and Phyllis Paykel Trust project (project and equipment grant, 2010).