Getting older is generally associated with more sleep disturbances, with 20%-70% of 50-80 year olds reporting a sleep problem.1-4Older women (aged 550 years) are more likely to report disturbed sleep with insomnia-type symptoms, whereas older men typically report more sleep-disordered breathing and daytime sleepiness.2,3,5 Studies that have measured sleep objectively corroborate these reports, showing that sleep quality and duration generally decrease with age.6 However, limited research has focused on the prevalence of sleep problems among people of advanced age (>80 years). A recent US-based study found that successful ageing (ie, living with good quality physical and mental health status despite physiological ageing) might be associated with a reduction in reporting sleep problems.7Self-reported sleep problems and disorders are common in New Zealand.8 In a 2001 representative national survey of 4,000 adults aged 20-59 years (72.5% response rate), the estimated population prevalence of chronic sleep problems (lasting at least 6 months) was 28.6% for Mori and 24.6% for non-Mori (p=0.033).9 There are also consistent relationships between poorer sleep health and greater socioeconomic deprivation as measured by the New Zealand Deprivation Index 2006 (NZDep 06, an area-based measure of socioeconomic deprivation).8,10Sleep problems are negatively associated with self-rated physical and mental wellbeing, mood, and quality of life.1,4,9 Poor sleep degrades daytime alertness and performance, thereby increasing the risk of incidents such as road traffic accidents and falls.11,12 The likelihood of sleep disturbances related to health problems increases with age.13 These include depression, pain, respiratory and cardio-vascular diseases, cognitive impairment and dementia.2,5 When older people are also required to care for someone with an illness, sleep problems are likely to be exacerbated.14,15 The proportion of older people in the population is increasing, and it is becoming more common for people to provide care to their family members at home,16 making the sleep of older carers of interest. People of advanced age are likely to have had exceptionally good health,17 therefore the predictors and consequences of sleep problems may not be the same as for younger populations.The current study used data from the inception interviews (Wave 1) of Te Puwaitanga o Ng Tapuwae Kia Ora Tonu. Life and Living in Advanced Age: A Cohort Study in New Zealand (LiLACS NZ). This is the first study of its kind, using face-to-face surveys to collect information on factors that contribute to successful ageing by Mori and non-Mori of advanced age.18 Sleep was not the primary focus of LiLACS NZ, but some general questions on sleep were included. These provide the first insights into sleep health during advanced age, which could contribute to better health services to recognise and manage geriatric sleep problems.The aims of these analyses were to estimate the prevalence of self-reported sleep problems among Mori and non-Mori of advanced age; to investigate the independent associations between demographic and health-related factors, and self-reported sleep problems; and to determine whether self-reported current sleep problems increased the likelihood of reporting other adverse health-related outcomes.MethodsThis project was approved by the Northern X Regional Ethics Committee (NXT 09/09/088). The LiLACS NZ Wave 1 survey was completed with 421 Mori (aged 79-90 years) and 516 non-Mori (aged 84-86 years). The Mori cohort included participants who identified themselves as Mori, either alone or as one of multiple ethnicities. They had a broader age range than the non-Mori cohort to account for known differences in longevity between the two populations.19,20 The participants were living in the Bay of Plenty and Lakes District Health Board areas in 2010. All Mori born between 1 January 1 1920 and 31 December 31 1930, and all non-Mori born between 1 January and 31 December 1925, were identified from the electoral roll, primary care databases, word-of-mouth, Mori tribal networks, and through local publicity. The extensive use of local organisations meant that more difficult to reach populations (eg, those who had disabilities) were able to be contacted and invited to participate by someone known to them.18 Those who gave informed consent completed the face-to-face survey in their own home with a trained interviewer using standardised techniques. An overall response rate of 57% was achieved.19 The populations recruited approximated the age and sex distribution of the underlying population, except that women were over-represented.18The sample used for the present analyses included all participants who completed the question, Do you have trouble with your sleeping (on at least 3 nights per week) such that it interferes with your activities the following day (eg, unrefreshed in the morning, fatigue, poor concentration, or irritability)? This question gave an indication of current sleeping problems. Participants who answered yes to this question were asked to indicate the types of problems they were experiencing from a list of eight symptoms (waking up in the early hours of the morning; taking a long time to get to sleep; lying awake for most of the night; getting up at night to go to the toilet; worry keeping you awake at night; snoring; sleep walking/sleep talking; or other sleeping problems). Participants were also asked, How much trouble did you have with sleeping when you were young? Answers were dichotomised into yes (a little, some, or a lot) vs no (none at all) to give an indication of past sleeping problems.Study procedures allowed for participants to complete a shortened core survey that did not include the sleep questions if, for example, they were considered by members of a whnau (extended family) to not be able to manage answering the interview questions for themselves or they were very disabled in residential care.21 Of the 937 participants in Wave 1, 261 completed the core survey only, and 5 did not complete either version of the survey. There were also 22 instances of participants who completed the full survey, but did not answer the sleep questions, giving a total of 288 participants who did not answer the sleep questions. The analyses presented here focus on the remaining 251 Mori and 398 non-Mori (the sleep sample ).Demographic variables included the following variables provided in the LiLACS NZ database: ethnicity (Mori vs non-Mori); sex (male vs female), age (year increments); and socioeconomic deprivation, as measured using the NZDep 06 (decile 1 = least deprived to decile10 = most deprived). Caregiving was defined by the question, How often do you currently provide care or assistance for other people? For this study, carers included those responding occasionally, less than once a week, once a week, two to five times weekly, or daily (six to seven times weekly). Non-carers were defined by answering never. Residential status (defined as living alone, living with a spouse only, living with family including or not including spouse, or living in residential care) was also described.The physical and mental health variables considered for multivariate analyses included scores from the following standardised scales which were provided in the LiLACS NZ database: the Short-Form Health Survey (SF-12),22 as a measure of physical and mental quality of life, as well as pain that interferes with daytime functioning; the Mini Mental Status Exam (MMSE),23 as a measure of cognitive functioning; the Geriatric Depression Scale (GDS-15),24 as a measure of self-reported depression symptoms; the Pearlin Mastery Scale (PMS),25 as a measure of perceived control; and the Physical Activity Scale for the Elderly (PASE).26 Additional measures included: doctors records confirming a diagnosis of depression; a global score from a five-point Likert scale rating coping in different situations (times of loss, financial hardship, on-going health problems, times of trouble for family and friends, and overall); a five-point Likert scale rating the experience of ageing (On the whole has growing older been a positive or negative experience for you?); and single-item questions to define those who had fallen and how often, those who were a current or past smoker, and those who drank alcohol (four or more times a week versus monthly or less).AnalysisThe sleep sample and the participants who did not answer the sleep questions were compared by ethnicity, sex, age, and NZDep 06 using sequential logistic regression modelling with 98.3% of the total observations (n=922).Differences in the proportion of Mori and non-Mori who reported each type of sleep problem were investigated using chi-square tests. A multivariable logistic regression model was used to investigate the independent associations between self-reported current sleep problems and a range of demographic, physical and mental health factors, and reporting a past sleep problem. The demographic variables of ethnicity, sex, age and NZDep 06 were included in the models, based on a priori evidence. This model included 89.7% of all observations. Due to limited available power, only those health-related factors that were significantly associated with the outcome at the univariate level (p<0.1) were included. A version of the model was run with the interaction term sex X ethnicity, but the interaction was non-significant.A series of multivariate logistic regression models were used to determine whether reporting a current sleeping problem was associated with reporting poorer physical or mental health, after controlling for ethnicity, age, sex, and NZDep 06. In these models, the outcome variables were dichotomised using validated cut-off scores (eg, <25 on the MMSE, and >8 on the GDS) or, in instances when validated cut-offs were unavailable, by scoring within the 25th percentile of the particular scale. Models included between 93.7% and 98.8% of all observations, due to missing values for some variables. Adjusted odds ratios (ORs) and 95% confidence intervals (95%CIs) were estimated for groups of interest. Analyses were undertaken in SAS\u00ae (2011, Version 9.3, Cary NC).ResultsDemographicsOf the 649 participants who responded to the question concerning a current sleep problem, 38.7% were Mori and 61.3% were non-Mori. Their demographic characteristics are summarised in Table 1. Sequential logistic regression analyses revealed that, after controlling for sex and NZDep 06, Mori were more likely to be within the group who did not answer the sleep question compared to non-Mori (OR=3.00, 95% CI=2.16-4.16, p=<0.0001). The likelihood of not answering the sleep question also increased with age (OR=1.18, 95% CI=1.10-1.27, p=<0.0001).\r\nTable 1: Demographic characteristics of Mori and non-Mori participants in the sleep sample. \r\n \r\n \r\n Variable\r\n \r\n Mori (n=251) n (%)\r\n \r\n Non-Mori (n=398) n (%)\r\n \r\n \r\n \r\n Females\r\n \r\n 153\r\n \r\n (61.0%)\r\n \r\n 211\r\n \r\n (53.0%)\r\n \r\n \r\n \r\n Carers\u2020\r\n \r\n 84\r\n \r\n (33.5%)\r\n \r\n 107\r\n \r\n (26.9%)\r\n \r\n \r\n \r\n NZDep 06\r\n \r\n \r\n \r\n Decile 1 (Least deprived)\r\n \r\n 4\r\n \r\n (1.6%)\r\n \r\n 12\r\n \r\n (3.0%)\r\n \r\n \r\n \r\n Decile 2\r\n \r\n 2\r\n \r\n (0.8%)\r\n \r\n 13\r\n \r\n (3.3%)\r\n \r\n \r\n \r\n Decile 3\r\n \r\n 6\r\n \r\n (2.4%)\r\n \r\n 28\r\n \r\n (7.0%)\r\n \r\n \r\n \r\n Decile 4\r\n \r\n 24\r\n \r\n (9.6%)\r\n \r\n 47\r\n \r\n (11.8%)\r\n \r\n \r\n \r\n Decile 5\r\n \r\n 11\r\n \r\n (4.4%)\r\n \r\n 16\r\n \r\n (4.0%)\r\n \r\n \r\n \r\n Decile 6\r\n \r\n 31\r\n \r\n (12.4%)\r\n \r\n 85\r\n \r\n (21.4%)\r\n \r\n \r\n \r\n Decile 7\r\n \r\n 20\r\n \r\n (8.0%)\r\n \r\n 67\r\n \r\n (16.9%)\r\n \r\n \r\n \r\n Decile 8\r\n \r\n 34\r\n \r\n (13.6%)\r\n \r\n 59\r\n \r\n (14.8%)\r\n \r\n \r\n \r\n Decile 9\r\n \r\n 34\r\n \r\n (13.6%)\r\n \r\n 37\r\n \r\n (9.3%)\r\n \r\n \r\n \r\n Decile 10 (Most deprived)\r\n \r\n 85\r\n \r\n (33.9%)\r\n \r\n 34\r\n \r\n (8.5%)\r\n \r\n \r\n \r\n Residential status:\r\n \r\n \r\n \r\n Lives alone\r\n \r\n 104\r\n \r\n (41.4%)\r\n \r\n 193\r\n \r\n (48.5%)\r\n \r\n \r\n \r\n Lives with spouse only\r\n \r\n 66\r\n \r\n (26.3%)\r\n \r\n 153\r\n \r\n (38.4%)\r\n \r\n \r\n \r\n Lives with family (+/- spouse)\r\n \r\n 81\r\n \r\n (32.3%)\r\n \r\n 52\r\n \r\n (13.1%)\r\n \r\n \r\n \r\n Lives in residential care\r\n \r\n 2\r\n \r\n (0.8%)\r\n \r\n 19\r\n \r\n (4.8%)\r\n \r\n \r\n \r\n\u2020 n=249 for Mori carers, and 386 for non-Mori.\r\nReporting sleep problemsIn the sleep sample, 26.3% of Mori and 31.7% of non-Mori reported having a current sleep problem (\u03c72=2.13, p=0.145). Women were more likely to report a current sleep problem than men (33.0% vs 25.3%, \u03c72=4.55, p=0.033). Both Mori and non-Mori problem sleepers endorsed a median of three sleep symptoms. Of those who were Mori, 50.0% endorsed more than three symptoms compared to 34.9% of non-Mori (\u03c72=4.10, p=0.043). All of the participants reporting a current sleep problem endorsed at least one symptom of insomnia (waking up too early, taking a long time to get to sleep, and/or lying awake for most of the night). Figure 1 shows the proportions of Mori and non-Mori reporting a sleep problem who endorsed each sleep symptom. Other sleeping problems included physical aches and pains, symptoms of restless legs, sleep disordered breathing, hallucinations, taking medicines, and providing care for others. Of the full sleep sample, 5.2% of Mori and 6.8% of non-Mori reported having a past sleep problem.\r\nFigure 1: The proportion of sleep symptoms endorsed by Mori (n=66) and non-Mori (n=126) participants who reported having a current sleep problem.* = p (chi-square) <0.05, ** = p (chi-square) <0.01.\r\nTable 2: Independent predictors of self-reported current sleep problems within LILACS NZ (n=582). \r\n \r\n \r\n Independent variables\r\n \r\n Categories/range\r\n \r\n Adjusted OR\r\n \r\n 95%CI\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n Mori\r\n \r\n 0.52*\r\n \r\n 0.30-0.90\r\n \r\n \r\n \r\n \r\n \r\n Non Mori (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n Sex\r\n \r\n Females\r\n \r\n 1.48\r\n \r\n 0.98-2.24\r\n \r\n \r\n \r\n \r\n \r\n Males (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n Age\r\n \r\n 79-90\r\n \r\n 0.94\r\n \r\n 0.82-1.07\r\n \r\n \r\n \r\n Deprivation (NZDep 06)\r\n \r\n 1-10\r\n \r\n 1.05\r\n \r\n 0.96-1.14\r\n \r\n \r\n \r\n Caregiving\r\n \r\n Carer\r\n \r\n 1.52\r\n \r\n 0.99-2.32\r\n \r\n \r\n \r\n \r\n \r\n Non Carer (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n Past sleep problem\r\n \r\n Yes\r\n \r\n 2.67*\r\n \r\n 1.25-5.72\r\n \r\n \r\n \r\n \r\n \r\n No (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n Mental health (SF-12)\r\n \r\n 14.9-79.3\r\n \r\n 0.95***\r\n \r\n 0.92-0.98\r\n \r\n \r\n \r\n Cognition (MMSE)\r\n \r\n 2-30\r\n \r\n 0.96\r\n \r\n 0.89-1.03\r\n \r\n \r\n \r\n Depression rating (GDS)\r\n \r\n 0-14\r\n \r\n 1.09\r\n \r\n 0.96-1.24\r\n \r\n \r\n \r\n Depression diagnosis\r\n \r\n Yes\r\n \r\n 1.37\r\n \r\n 0.73-2.54\r\n \r\n \r\n \r\n \r\n \r\n No (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n Perceived coping (PMS)\r\n \r\n 5-25\r\n \r\n 1.04\r\n \r\n 0.96-1.14\r\n \r\n \r\n \r\n Perceived control\r\n \r\n 9-35\r\n \r\n 0.98\r\n \r\n 0.92-1.05\r\n \r\n \r\n \r\n Physical health (SF-12)\r\n \r\n 6.0-65.8\r\n \r\n 0.97*\r\n \r\n 0.94-1.00\r\n \r\n \r\n \r\n Pain interference (SF-12)\r\n \r\n 0-100\r\n \r\n 1.00\r\n \r\n 0.99-1.01\r\n \r\n \r\n \r\n Number of Falls\r\n \r\n 0-3\r\n \r\n 1.14\r\n \r\n 0.92-1.42\r\n \r\n \r\n \r\n Alcohol\r\n \r\n Drinker\r\n \r\n 0.81\r\n \r\n 0.53-1.24\r\n \r\n \r\n \r\n \r\n \r\n Non drinker (Ref)\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n* p<0.05, ** p<0.01, *** p <0.001Covariates: ethnicity, age, sex, and NZDep 06.\r\nAbbreviations: SF-12 (Short Form 12 Item Survey), MMSE (Mini Mental State Exam), GDS (Geriatric Depression Scale), PMS (Pearlin Mastery Scale), OR (Odds Ratio),CI (confidence Interval), Ref (reference group).\r\nTable 2 shows the results of the multiple logistic regression analyses investigating factors associated with reporting a current sleep problem. This shows reduced likelihood for participants who were Mori vs non-Mori (OR=0.52, 95% CI=0.30-0.90). Those who also reported a past sleep problem were more likely to also report a current sleep problem than those who did not (OR=2.67 95% CI=1.25-5.72). With each point of increase on the SF-12 mental health scale, there was a 6% reduction in the likelihood of reporting a sleep problem. With each point increase on the SF-12 physical health scale, there was a 3% reduction in the likelihood of reporting a sleep problem.\r\nTable 3: Independent associations between reporting current sleep problems and poorer health. \r\n \r\n \r\n Model\r\n \r\n Health outcome variable\r\n \r\n n, (%) with condition\r\n \r\n Observations\r\n \r\n Adjusted OR\r\n \r\n 95%CI\r\n \r\n \r\n \r\n 1\r\n \r\n Depression, (GDS)\r\n \r\n 49 (7.6%)\r\n \r\n 641\r\n \r\n 4.42***\r\n \r\n 2.40-8.14\r\n \r\n \r\n \r\n 3\r\n \r\n Depression diagnosis\r\n \r\n 62 (9.8%)\r\n \r\n 631\r\n \r\n 1.96*\r\n \r\n 1.15-3.37\r\n \r\n \r\n \r\n 5\r\n \r\n Cognitive impairment\r\n \r\n 76 (12.1%)\r\n \r\n 629\r\n \r\n 1.49\r\n \r\n 0.89-2.61\r\n \r\n \r\n \r\n 7\r\n \r\n Poor mental health\r\n \r\n 155 (25.5%)\r\n \r\n 608\r\n \r\n 2.45***\r\n \r\n 1.65-3.63\r\n \r\n \r\n \r\n 8\r\n \r\n Poor coping\r\n \r\n 144 (23.5%)\r\n \r\n 612\r\n \r\n 1.80**\r\n \r\n 1.20-2.71\r\n \r\n \r\n \r\n 9\r\n \r\n Poor control\r\n \r\n 167 (27.4%)\r\n \r\n 610\r\n \r\n 1.82**\r\n \r\n 1.23-2.68\r\n \r\n \r\n \r\n 10\r\n \r\n Poor physical health\r\n \r\n 148 (24.3%)\r\n \r\n 608\r\n \r\n 1.64*\r\n \r\n 1.10-2.46\r\n \r\n \r\n \r\n 11\r\n \r\n Fallen in last year\r\n \r\n 215 (34.5%)\r\n \r\n 624\r\n \r\n 1.64**\r\n \r\n 1.14-2.36\r\n \r\n \r\n \r\n 12\r\n \r\n Reduced physical activity\r\n \r\n 147 (23.7%)\r\n \r\n 624\r\n \r\n 1.36\r\n \r\n 0.89-2.06\r\n \r\n \r\n \r\n 13\r\n \r\n Increased pain interference\r\n \r\n 181 (29.2%)\r\n \r\n 619\r\n \r\n 1.89**\r\n \r\n 1.30-2.76\r\n \r\n \r\n \r\nTable 3 summarises the findings from the logistic regression models investigating whether current sleep problems are independent risk factors for poorer health outcomes. These show significant associations between sleep problems and health outcomes. For example, those reporting a current sleep problem were more likely to also score highly for symptoms of depression on the GDS than those who did not report a current sleep problem (OR=4.42, 95% CI=2.40-8.14).DiscussionThis is the first study to estimate the prevalence of reporting sleep problems among Mori and non-Mori of advanced age. Among 251 Mori and 398 non-Mori aged 79-90 years, we found that 25.5% of Mori and of 31.7% non-Mori reported a sleep problem. Reporting a current sleep problem was more likely among non-Mori and among those who reported a past sleep problem, or poorer self-rated physical or mental health, after adjusting for demographic variables. The associations between sleep problems and health are likely to be bi-directional,27 highlighting the importance of sleep for the health-related quality of life of our rapidly ageing population.In contrast to previous research involving 20-59 year olds,8,9 the present study found that Mori were less likely to report a current sleep problem than non-Mori. The reasons for this difference are unknown. However, it should be noted that, of those reporting a sleep problem, non-Maori were more likely to volunteer other sleep symptoms not specifically asked for in the study compared to Mori. This suggests that how sleep problems are understood and reported might differ between the two samples. Response biases may have also been a factor, since Mori were three times more likely than non-Mori to be amongst those who did not respond to the sleep questions, and people with poorer health were less likely to complete the full questionnaire with the sleep questions. The lower prevalence of reporting sleep problems among Mori could also reflect the fact that the LiLACS NZ sample includes exceptionally long-lived Mori, the majority of whom strongly identify with their culture (eg, many being fluent in speaking te reo Mori [Mori language], and frequently visiting a marae [sacred meeting place of a tribe]) compared to younger generations.21 Previous LiLACS NZ research has identified cultural engagement (but not socioeconomic status) as an independent predictor of better physical health-related quality of life.21 Further research is required to clarify whether or not lifestyle and cultural engagement are associated with reporting sleep problems.In further contrast to previous research,8 in the current study NZDep 06 was not associated with reporting sleep problems in the fully-adjusted models. The Mori participants in the present study were over-represented in the most deprived NZDep 06 deciles, whereas the non-Mori sample had a more normally distributed NZDep 06 profile. These distributions are similar in the general population of Mori and non-Mori aged over 50 years.28 However, they differ from previous studies of 20-59 year olds, where non-Mori were over represented in the least deprived deciles.8 It is also possible that NZDep 06 alone is not the most reliable way to measure socioeconomic position in people of advanced age, who are more likely to live with family members or in rest homes or institutions that may not reflect their true socioeconomic position.29 Alternatively, socioeconomic position may be less closely related to health-related quality of life in advanced age.21 Additional research using different methods of defining socioeconomic position is needed to clarify this.The prevalence of reported sleep problems is lower in the LiLACS NZ sample than in previous studies of adults aged 60-80 years.2,4 Grandner et al7 propose that this pattern could be related to the exceptional health that leads to living longer, or a reduction in the effects of life or societal stressors on sleep in advanced age. Expectations and perceptions of good or acceptable health are also thought to change with age,30 and older people as well as their family members may downplay, or be less likely to complain of sleep disruptions.15 It is therefore important for health-care professionals to actively ask their older patients about sleep.4More symptoms were endorsed by those reporting sleep problems (a median of three for Mori and non-Mori) compared to previous studies.4 However, the list included trips to the toilet and waking early, which alone may not be considered problematic. All of the participants reporting a current sleep problem reported at least one insomnia symptom. Insomnia is common with ageing and reflects physiological and psychological changes that increase the likelihood of sleeplessness.4Carers are considered likely to have disturbed sleep.14,15 However in the present study, caregiving (including carers providing any type or amount of care) was not associated with reporting a current sleep problem. Future research should consider gathering more detailed information on the type of care being provided and the time spent caregiving.Those who reported having a past sleep problem were more likely to report a current sleep problem, although the definition of younger used in the survey is unclear. This relationship has been reported throughout the lifespan and highlights the importance of early diagnosis and treatment of sleep problems as a way of preventing any negative impact in later life.The first set of r