WHAT IS IT ABOUT SLEEP? POPULATION RESEARCH ON SLEEP QUALITY AND SOCIAL DETERMINANTS - PowerPoint Presentation

1. WHAT IS IT ABOUT SLEEP? POPULATION RESEARCH ON SLEEP QUALITY AND SOCIAL DETERMINANTSDiane S. Lauderdale, PhDLouis Block Professor of Public Health Sciences and the CollegeChair, Public Health Sciences The National Social Life, Health, and Aging Project (NSHAP) at NORC at the University of Chicago is funded by the National Institute on Aging (NIA): Wave 1 (R01 AG021487), Wave 2 (R37 AG030481), Wave 2 Partner Data Collection (R01 AG033903), Wave 3 NSHAP (R01 AG043538), and New Cohort (R01 AG048511).NSHAP sleep research is supported by AG042164 from NIA and the Basic Behavioral and Social Sciences Research Opportunity Network (OppNet), NIH.

2. OutlineEpidemiology and Sleep ResearchMeasuring SleepConcordance of survey and objective sleep measuresSocial correlates of population variation in sleep SociodemographicsLoneliness and Social Disconnectedness2

3. A Little History of EpidemiologyEpidemics and HygienePost WW2 epidemiologic transitionNoncommunicable diseases and risk factor epidemiologyClinical and biological factors (eg, lipids, BP, weight)Behavioral & Social Factors (eg, smoking, SES, alcohol)Renewed interest in communicable diseases3Presentation Title Here |

4. Sleep Enters EpidemiologyAlameda County Human Population Laboratory Survey, 19651Alameda 7: excessive alcohol, smoking, obesity, sleeping < or > than 7-8 hours, physical inactivity, snacks, and not eating breakfastAmerican Cancer Society Cohort, 19602,3Those reporting 7 h of sleep demonstrated the lowest mortality rate.41. Belloc NB: Relationship of health practices and mortality. Prev Med 1973;2:67-812. Hammond EC. Some Preliminary findings on physical complaints from a prospective study of 1,064,004 men and women. Am J Public Health Nations Health Jan 1964;54:11–23.3. Kripke DF, Simons RN, Garfinkel L, Hammond EC. Short and long sleep and sleeping pills. Is increased mortality associated? Arch Gen Psychiatry Jan 1979;36(1):103–16. .

5. Kripke et al. Arch Gen Psych 1979 -- ACS I 5

6. Impact of sleep debt on metabolic and endocrine function. Spiegel , Leproult , Van Cauter; Lancet 199911 men aged 18-278 h per night for 3 nights; 4 h per night for 6 nights; 12 h in bed per night for 6 nightsthey STAYED in bed but kept awake on the last two days of each condition during the non-sleep hours.Last day of sleep restriction: glucose metabolism, diurnal cortisol pattern, immune response, and appetite-regulating hormones all worse. IMPLICATION: “Voluntary sleep curtailment has become common.” “Irrespective of the underlying mechanisms, when compared with the sleep-recovery condition, sleep debt, which is experienced by a substantial proportion of people in more-developed countries, is clearly associated with metabolic and endocrine alterations that may have physiopathological consequences in the long term.”6

7. 2 QuestionsHow accurate is self-reported sleep duration?Noisy?Biased?Could there be social confounders of the sleep & death association?Social correlates of sleep variation7

8. How can sleep be measured?Survey questionsDurationQuality; Insomnia symptomsSleep LogsMeasuring sleep through brain activity: Polysomnography (PSG) Estimating sleep by arm motion: Actigraphy8Sleep Measurement |

9. How can sleep be measured?Survey questionsDurationQuality; Insomnia symptomsSleep LogsMeasuring sleep through brain activity: Polysomnography (PSG) Estimating sleep by arm motion: Actigraphy9Sleep Measurement |

10. Why are survey questions about duration challenging?Sounds simpleComplicated, multi-stepEstimate average bedtime and wake timeSubtract, often around midnight(instead just ask bedtime & take time?)For those who balk at this challenging process, where do they get an answer?A reasonable, socially desirable, or normative answer10Sleep & Die |

11. 8 Hours is the right answer“A day and night together lasts 24 hours. It is sufficient that a person sleep one third of that time, which is eight hours.” Maimonides (1135-1204)Who has time to sleep?“What is it to "redeem the time" from sleep? It is, in general, to take that measure of sleep every night which nature requires, and no more; that measure which is the most conducive to the health and vigour both of the body and mind.” John Wesley (1703-1791)11What is the socially desirable answer?

12. Actigraphy: objective option for population researchResearch-grade activity trackersAccess raw data, manufacturers’ or own algorithmsNon-invasive method of monitoring rest and activityRecords # and intensity of movements per epochDistinguish sleep from wakePattern across epochs What you get:DurationConsolidation (WASO, fragmentation, wake bouts, % sleep) => “Quality”Circadian timingTimingNo first-night effectDuration Correlation with PSG .8 to .9 (but…)12Sleep & Die |

13. 13

14. 14

15. How similar are survey and actigraph durations?Among middle-aged adults we added actigraphy to CARDIA & compared 3 nights actigraphy with self-report durationCorrelation 0.47, with self-reported sleep averaging 48 minutes longerPersons with mortality risk factors (fair/poor SRH, obesity or more depressive symptoms) systematically reported shorter sleep at the same level of measured sleepImplies reporting bias could generate associations between shorter reported sleep and poor health/mortality due to systematic bias15Self-reported and measured sleep duration: how similar are they?Lauderdale, Knutson, Yan, Liu, Rathouz, Epidemiology. 2008

16. 2 QuestionsHow accurate is self-reported sleep duration?Noisy?Biased?Could there be social confounders of the sleep & death association?Social correlates of sleep variation16

17. CARDIA: Actigraph DurationBy Education Group17%Objectively measured sleep characteristics among early-middle-aged adults: the CARDIA studyLauderdale, et al. American Journal of Epidemiology 2006

18. CARDIA: Actigraph DurationBy Race-Sex Group18%Objectively measured sleep characteristics among early-middle-aged adults: the CARDIA studyLauderdale, et al. American Journal of Epidemiology 2006

19. National probability sample of adults born 1920-1947, enrolled in 2005/6 later augmented with birthyears 1948-1964In-home interviews + biomeasuresUniversity of Chicago sociology, statistics, medicine, psychology, epidemiologyCollected by NORC at the University of ChicagoParticularly rich in measures of social interactions

20. 20

21. NSHAP Wave 2 Participants(n = 3,377)Randomly Selected for Sleep Study(n = 1,117) Consented to Participate(n = 897)Refused Sleep Study(n = 220)Baseline Sleep Data (n = 759)Actigraph and sleep log (n = 723)Only actigraph (n = 16)Only sleep log (n = 20)Not Age Eligible(n = 42) Usable Sleep Data(n = 801)Could Not Be Recontacted or Did Not Return Any Usable Data (n = 96)Random 1/3 of Wave 2 participants asked to participate in a substudy Actigraphy: 3 nights (72 consecutive hours) Booklet additional sleep questions 3-night sleep logSleep Substudy began in 2010-2011

22. Sleep Survey QuestionsNSHAP Core questions:Usual bedtimes & wake times, asked for weekdays and non-workdays How often do you feel rested upon waking in the morning?Sleep Booklet questions:How many hours do you usually sleep at night?How often do you have trouble falling asleep?How often do you have trouble with waking up during the night and not being able to fall back asleep?How often do you have trouble with waking up too early?3 nights of sleep log22

23. Sleep Survey QuestionsNSHAP Core questions:Usual bedtimes & wake times, asked for weekdays and non-workdays How often do you feel rested upon waking in the morningSleep Booklet questions:How many hours do you usually sleep at night?How often do you have trouble falling asleep?How often do you have trouble with waking up during the night and not being able to fall back asleep?How often do you have trouble with waking up too early?3 nights of sleep log23

24. Sleep Duration MeasuresSURVEY SLEEP HOURS: “How many hours do you usually sleep at night?” SURVEY CALCULATED SLEEP TIME: “What time do you usually go to bed and start trying to fall asleep?” and “What time do you usually wake up?” weekdays & weekendsSLEEP LOG TIME (3 night average)ACTIGRAPH TOTAL SLEEP TIME (3 night average) 24Correlations12341. Survey 1 question12 Calculated bed & wake0.3913 Sleep Log0.460.5614 Actigraph0.290.380.521

25. Loneliness and Social IsolationCoauthorsJade BensonEloesa McSorleyLouise Hawkley Loneliness refers to the distressing feeling that a person’s desired social needs are not being met Social isolation summarizes observable social connections Loneliness and social isolation are distinct and have a low correlation25Presentation Title Here |

26. Previous Work on Social Isolation, Loneliness & SleepLoneliness and social isolation predict increased risk for morbidity and mortalityBoth loneliness and social isolation have been previously linked to worse sleep qualityAlmost all used self-reported sleep measuresLoneliness was associated with greater sleep fragmentation using actigraphy among a group with uniformly high level of social connections: Hutterites.26Loneliness Is Associated with Sleep Fragmentation in a Communal Society. Kurina, Knutson, Hawkley, Caccioppo, Lauderdale, Ober. Sleep 2011

27. Measure of Social Isolation 4 measures of social network Network size, network range, proportion of network living in household, frequency of interaction with network 4 measures of social participation Frequency of: group meetings, socializing with friends/relatives, volunteering, religious services Number of friends Each variable standardized, averaged, and then standardized and inversed 27

28. Measure of Loneliness 3-item version of UCLA loneliness scale Frequency of feeling: lack companionshipleft outisolated Range 0 – 6 Spearman’s correlation between loneliness & social isolation in NSHAP = 0.20 (p < 0.001) 28

29. Research Questions Are loneliness and social isolation both& similarly associated with sleep quality? Do these associations differ depending on how sleep was measured? 29

30. 30Mean (SD) or %Age72.6 (7.4)Female53.2%Education < High School 17.9% High School 23.8% Some College 35.3% College Degree or Higher 22.9%Ethnicity (n = 756) White73.4% Black11.6% Hispanic11.5% Other3.4%Marriage Married/Living with Partner69.8% Divorced/Separated7.9% Widowed20.2% Never Married 2.1%Study Population (N=759).

31. Sleep Measures Actigraph sleep quality (standardized)WASO Percent sleep 4 duration measures (hours)Actigraph (TST)Single-question Sleep log calculated duration Calculated time in bed from core questions Insomnia symptom score (0 – 8) Having trouble falling asleep Waking up during the night and not being able to fall back asleep Waking up too early Feeling rested upon waking in the morning31

32. Sleep Measures32

33. Statistical Analysis Each sleep measure was modeled as an outcome in separate models – ordinal logistic regression for the insomnia symptom score; OLS for all others 2 models run for each of seven sleep outcomesSocial isolation as predictor, controlled for demographics Loneliness as predictor, controlled demographicsComplex survey33

34. SOCIAL ISOLATIONBeta coefficients, 95% CIs, and p-values for social isolation index from 7 regression models in which each sleep outcome is separately regressed on social isolation, adjusted for demographics: age, gender, race/ethnicity, and education 34 ß 95% CIp-valueActigraph Sleep Measures     WASO 0.130.05, 0.210.002 Percent Sleep -0.11-0.18, 0.030.007 Actigraph TST 0.03-0.08, 0.140.59Self-Reported Sleep Measures    Single-Question Duration -0.05-0.18, 0.080.43 Sleep Log Calculated Duration0.170.04, 0.290.008 Calculated Sleep Interval 0.140.02, 0.270.03 Insomnia Symptom ScoreOR = 1.000.87, 1.160.96

35. LONELINESSBeta coefficients, 95% CI, and p-values for loneliness from seven regression models in which each sleep outcome is separately regressed on lonelinessadjusted for demographics: age, gender, race/ethnicity, and education 35 ß 95% CIp-valueActigraph Sleep Measures     WASO 0.080.01, 0.150.02 Percent Sleep -0.07-0.13, -0.010.03 Actigraph TST -0.05-0.13, 0.030.25Self-Reported Sleep Measures    Single-Question Duration -0.11-0.20, -0.010.02 Sleep Log Calculated Duration-0.05-0.14, 0.040.28 Calculated Sleep Interval 0.02-0.07, 0.120.67 Insomnia Symptom ScoreOR = 1.311.17, 1.47<0.001

36. Further confounding?Cognitive impairmentFrailtyPainObesitySleep, social isolation & loneliness and these health conditions are all measured at the same time.Possibility of bidirectional associations makes it impossible to sort out confounding from mediation. 36

37. Summary for Social Isolation & LonelinessLoneliness & Social Isolation are both related to worse actigraph sleep quality and neither are related to actigraph duration.Loneliness is related to self-reported sleep duration and insomnia symptoms.Social isolation is related to longer duration from bedtimes & waketimes and the sleep log => later waketimes.37

38. ImplicationsDissatisfaction with sleep (short self report duration and insomnia) is associated with loneliness, but not lack of social connections.Both loneliness and social isolation are related to more disrupted sleep, which we repeatedly find to be the most salient dimension of sleep for health outcomes: cognitive decline, poor olfaction, worse cortisol profile.38