This study indicates that a combined intervention using bright light therapy and/or therapeutic exercise with sleep hygiene education is feasible for people with MCIoD and their family carers at home. Given the paucity of options for treating dementia, the intervention has potential for improving the sleep and wellbeing of older people at risk or diagnosed with dementia. The mixed-methods approach and use of case studies including both care recipients and care partners was novel, allowing investigation into the context of conducting such a study in the NZ community. The protocol was considered feasible and suggestions for future trials were obtained. The small heterogeneous sample limited the ability to measure the clinical effectiveness of the interventions, or to determine the characteristics of those most likely to benefit. However, some pairs did report benefits to their sleep and waking wellbeing. In other cases, a rapid deterioration of health masked any possible improvements or prevented participants from being able to complete the trial. These mixed findings are concordant with previous research with people living in institutions or the community, using both cross-sectional studies and randomised controlled trials 25. They are representative of the clinical reality of families affected by dementia.
The interventions offer a low-risk, drug-free option for managing dementia-related sleep disturbances. They were designed to address well-defined physiological and psychosocial processes affecting sleep and waking function (most notably stabilising the function of the circadian master clock). 6,7 The protocol in the present study differed from previous community-based trials e.g.13, 26 in that bright light was recommended at mid-morning, using a higher intensity and for a shorter duration. The exercise component also had a recommended time frame. Delivering the interventions this way was considered more suitable for the participants, since dementia-related symptoms and behaviours often exacerbate towards evening, which is also a time of high care demands. 6 The timing of the light and exercise exposures was also anticipated to minimise the likelihood of the interventions having an arousing effect on the care recipient before bedtime. 11,12
Compared to previous community-based studies,13,26 the present sample of care recipients had more varied types of dementia-related impairment. They also had less problematic sleep at baseline, as indicated by both actigraphic and subjective reports as well as higher baseline light exposure (more than 2,000 lux). Similarly, the sample of carers had relatively good actigraphic sleep compared to those in previous samples. However, their subjective ratings indicated a high prevalence of sleep problems which was more comparable.8 Discrepancies between subjective and objective sleep data of carers is not unusual and likely associated with the psychosocial impact of caring on the symptoms of insomnia. 27 Carers in the present sample appeared to have reduced likelihood of anxiety or depression compared to previous samples of NZ carers 23, which may have contributed to different sleep outcomes.
Variable improvement in sleep problems associated with MCIoD may be attributable to a range of factors. These include variability in comorbid diseases and demographic factors, the fluctuating nature of sleep and dementia-related behaviours, as well as response bias, possible placebo effects, and the impact of researcher support. Comparisons between studies are also confounded by differences in sleep measures and intervention characteristics. The course of cognitive deterioration and factors necessary for stable sleep timing and efficiency (including damage to the eye, pineal gland, and circadian system as well as presence of another primary sleep disorder) are also expected to affect intervention effectiveness. 25
Based on the present case studies, it is recommended that future trials are tailored towards care recipients who have limited access to bright light and physical activity, as well those with poor sleep status at baseline. A larger randomised controlled trial would enable greater monitoring of the impact of the intervention while controlling for biases. The Inclusion of qualitative elements in such trials is recommended to strengthen our understanding of the psychosocial elements of sleep health with ageing, MCI and dementia as well as providing a person-centred perspective on the effectiveness of the interventions in this highly variable population.
Considering the variation between families affected by dementia as well as issues with recruitment, compliance, and attrition in the present study, future trials may benefit from a more personally tailored approach to engagement and delivery as well as a reduced research load on the participants. This could be achieved through automated rather than diary-assisted scoring of actigraphy data (removing the need for keeping dairies), 28 customising the interventions, and delivering them via primary healthcare or pre-existing home care services with a support person to facilitate compliance with the intervention(s). Such an approach has been used with other health promotion interventions for older people e.g. 29 and would likely facilitate recruitment and retention of participants, as well enabling better evaluation of the impact of the intervention and its broader use in aged-care services.