Fam Community Health Vol. 37, No. 4, pp. 317–326 C 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins Copyright
The Impact of Mid- and Late-Life Loss on Insomnia Findings From the Health and Retirement Study, 2010 Cohort Cherie Simpson, PhD, APRN, CNS-BC; Joseph C. Allegra, MPH; Amara E. Ezeamama, PhD; Jennifer Elkins, PhD; Toni Miles, MD, PhD Bereavement and insomnia are both well-documented risk factors for illness. We use cohort data to estimate risk of insomnia after death of a family member among adults aged 50 to 70 years. Each day, 6700 persons die in the United States. During the next 20 years, this number will increase. In this cohort, any loss increases the likelihood of insomnia. The highest rates of insomnia occur among women aged 50 to 59 years; men aged 65 to 70 years, and persons reporting death of a spouse/partner or child. Physical activity reduces this risk by one-third. Bereavement is a public health issue requiring a targeted response. Key words: insomnia, late-life bereavement, loss, sleep
E
STIMATES suggest that 50 to 70 million Americans suffer from sleep disorders or sleep deprivation, an unmet public health problem with high costs both to individuals and to society.1 Sleep disorders such as insomnia hinder the performance of complex functions such as driving, for example, and increase the risk for workplace accidents.2,3 Insomnia alone has been estimated to cost $6.6 billion in direct costs of health care utilization, transportation, and medications, as well as indirect costs of absenteeism and lost work productivity.4 Insomnia can also have adverse effects on health,5 and a growing body of Author Affiliations: School of Nursing, The University of Texas at Austin, Austin (Dr Simpson); and Department of Epidemiology and Biostatistics (Mr Allegra and Drs Ezeamama and Miles) and School of Social Work (Dr Elkins), University of Georgia, Athens. The authors declare no conflict of interest. Correspondence: Cherie Simpson, PhD, APRN, CNSBC, School of Nursing, The University of Texas at Austin, 1700 Red River, D0100, Austin, TX 78701 (csimpson@ mail.nur.utexas.edu). DOI: 10.1097/FCH.0000000000000039
literature has linked insomnia to individuals’ increased risk for cardiovascular disease, diabetes, anxiety, and depression.6-8 Increasing the public’s knowledge about sleep health through education about sleep and sleep disorders such as insomnia is, therefore, an important goal.9 Insomnia has been described by both behavioral and cognitive models; its etiology and pathophysiology are complex.10 Lifestyle behaviors such as alcohol consumption or physical activity can hurt or help sleep,11 but a known trigger for insomnia is the loss of a loved one. During the bereavement and grief that follow upon such a loss, individuals are especially vulnerable for insomnia’s onset. The cognitive model of insomnia would posit this loss as a precipitating factor for the disorder; the loss is then followed by rumination and worries that contribute to the disorder’s perpetuation. A literature review of late-life spousal bereavement reminds us that the severe emotional strain of the loss of a loved one is accompanied by profound changes in lifestyle and status, often including reductions in financial security, perceived personal 317
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
318
FAMILY & COMMUNITY HEALTH/OCTOBER–DECEMBER 2014
safety, and freedom of action.12 All of these facets of the situation are likely to lead to sleep disruption. In an aging society, loss and its potential impact on the occurrence of insomnia become increasingly important. Loss of a partner is not the only life stage loss experienced by middle-aged and older adults. Other losses include the deaths of parents, siblings, and adult children. Although US mortality rates are falling overall, the number of deaths has increased each year for the past 75 years.13 Middle-aged and older adults have large social networks that extend well beyond their families to include friends and colleagues. Such connections also propagate grief after loss.14 Experience of the loss of a close relative or spouse is common, reaching 86% by the age of 70 years. Population data have shown persistent, elevated risk for mortality of adults across the age span for a 10-year period after these losses.15 What is not clear is whether the elevated mortality risk is purely driven by age or whether the experience of loss is an independent contributor to the elevated risk of mortality. One of the goals of this study is to explore whether loss has an ageindependent association with insomnia—a morbidity linked with a range of adverse outcomes including mortality. So far, however, the literature on sleep and bereavement is limited. Spousal bereavement research has shown that complaints about sleep are common.16 The severity of grief contributes to poor sleep quality independently of depressive symptoms, but concurrent depression increases poor sleep further.17 Even bereaved persons who fail to meet a formal diagnosis of depression have measurable sleep impairment.18 The literature clearly indicates that sleep disruption and insomnia occur after the loss of a loved one. But the studies cited were done with small convenience samples, usually of widowed women. The impact of loss is not well understood at the population level. From a public health perspective, the question is, what is the impact of loss on insomnia in the general population? Who is at most risk for postloss insomnia? This study will explore these questions by examining the
risk of insomnia after exposure to loss in midand late-life adults using the 2010 Health and Retirement Survey cohort. The research questions are as follows: 1. What is the association between the experience of a loss and the prevalence of insomnia symptoms in a nationally representative sample of US adults 50 to 70 years of age? 2. If insomnia prevalence is significantly elevated among person with and without loss, are there disparate effects on insomnia severity depending on the type of loss? 3. Does the association between loss and insomnia symptoms differ for specific age cohorts? 4. Does the association between insomnia and loss differ between men and women? METHODS Data source and study population This study is based on data from the Health and Retirement Study (HRS), a survey designed to follow a representative sample of Americans older than 50 years for every 2 years. Since its inception in 1992, the HRS has enrolled 5 additional cohorts and has captured data on changes in the labor force as well as on the health transitions that individuals undergo toward the end of their working lives and into the years that follow.19 The overall response rate for each of the follow-up waves is higher than 80%. Sampling weights are provided on all HRS data sets to compensate for the unequal probabilities of selection between core and oversample domains (blacks, Hispanics, and residents of Florida).20 The HRS weighted sample is representative of all noninstitutionalized individuals in the US population in the age-eligible range. The HRS design and implementation details have been discussed elsewhere.21 For purposes of these analyses, the sample (n = 12 759) comprised respondents 50 to 70 years of age to capture those in mid- to late life who participated in
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
The Impact of Mid- and Late-Life Loss on Insomnia the 2010 wave (February 2010 to November 2011) and who provided information regarding the primary predictor of exposure to loss as well as the outcome of insomnia. Measures Insomnia symptoms In 2010, respondents were asked 4 questions to determine their sleep status. They were asked how often they had: (1) trouble falling asleep, (2) trouble with waking up in the middle of the night, and (3) trouble waking up too early, as well as (4) how often they felt rested in the morning. The possible responses were “most of the time,” “some of the time,” or “rarely or never.” Previous research based on HRS data defined individuals as experiencing insomnia if they answered “most of the time” or “sometimes” to the first 3 questions and “rarely or never” or “sometimes” to the fourth question (feeling rested in the morning).22 The number of insomnia symptoms was summed, and an index was created from 0 to 4. In addition to the index, an insomnia scale was created by dividing the number of reported insomnia symptoms by the number of possible insomnia symptoms and multiplying by 100. This ensured reliable interpretation of the β coefficients as a percent increase in insomnia score per unit increase in total loss. To better describe the sample, the participants were then grouped as: (1) those with no insomnia symptoms, (2) a subclinical insomnia group with 1 to 2 symptoms, and (3) a clinical insomnia group with 3 to 4 symptoms. Loss In the 2010 HRS survey, the mortality of respondents’ mothers and fathers was assessed with the following questions: “Is your mother alive?” and “Is your father alive?” In addition, the mortality of the respondents’ mothers and fathers was captured as a wave preloaded response (“previous wave mother living” and “previous wave father living”). The responses for questions regarding respondents’ mothers’ mortality were combined, and a new
319
variable was created to encapsulate the mortality status of the respondent’s mother. The same procedure was used for the mortality status of the father. For the mortality status of a spouse/partner, a single variable was created by combining responses to the following questions: “previous wave spouse/partner die during study?” and “divorce/widow since previous wave?” The death of a sibling was enumerated by examining the question “year sibling died?” If a year was given, the death of the respondent’s sibling was grouped as “yes” (“no” otherwise). Finally, 2 questions were used to determine the loss of a child: “number of children ever?” and “number of children living?” If the difference between these 2 questions was not zero, the loss of a child was said to have occurred. For purposes of these analyses, we accumulated all losses independently of the dates of their occurrences. Respondents’ losses were dichotomized as any loss versus no loss. Respondents were further categorized as having no losses, 1 loss, 2 losses, or 3 or more losses. Demographic characteristics To capture the experience of adults in the periretirement period, the sample was restricted to persons between the ages of 50 and 70 years during the 2010 interview cycle. For analysis, age was treated as a categorical variable: 50 to 54, 55 to 59, 60 to 64, and 65 to 70 years. Other categorical variables included gender, race, and education (30 kg/m2 ), smoking status (yes vs no), and alcohol use (any vs none) as known lifestyle factors that can influence sleep. In the
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
320
FAMILY & COMMUNITY HEALTH/OCTOBER–DECEMBER 2014
HRS, physical activity was ascertained by the following questions: “How often vigorous activity?” “How often moderate activity?” “How often mild activity?” Possible responses were “more than once a week,” “once or twice a week,” “1 to 3 times a month,” or “hardly or never.” If the response was “hardly or never” on all 3 questions of physical activity, it was assigned to the “none” category. Any other response besides “hardly or never” to any physical activity question was categorized as physically active. Smoking was assessed by the question “Do you smoke cigarettes now?” If the answer was “yes,” the participant was assigned to the smoker category (“no” for nonsmokers). Likewise, for alcohol use, if the respondent answered “yes” to the question, “Do you ever drink alcohol such as beer, wine, or liquor?” the respondent was assigned to the “drinker” category (“no” for nondrinkers). Body mass index was calculated on the basis of self-reported height and weight. Depression Depression was assessed using a short form of the Center for Epidemiologic StudiesDepression Scale.23 Respondents were asked 8 questions about the presence or absence of depressive symptoms within the past week. Consistent with prior research,24 the study categorized individuals with 4 or more symptoms as having elevated depressive symptoms. Statistical analyses Exploratory analysis was performed to determine the distribution of insomnia in the sample population across demographic groups and covariates. Estimated risk differences (RDs) and 95% confidence intervals (CIs) for the association between loss and the number of insomnia symptoms (0 [reference], 1, 2, 3, 4) were calculated using multiple linear models. The first model looked at the crude association between the exposure to loss and insomnia symptoms. Model II controlled for age and gender. Model III additionally controlled for physical activity, BMI,
smoking status, and alcohol use. Model IV added depression to all covariates in model III. To explore the potential heterogeneity of the association between loss and insomnia by age and sex, stratified linear regression models were built within strata of these covariates (model not shown.) All analyses were conducted using Stata, version 12 (StataCorp, College Station, Texas). The a priori significance level for all statistical tests was set at α < .05. RESULTS Table 1 shows characteristics of 2010 HRS participants between the ages of 50 and 70 years. In this sample of 12 759 participants, 52.3% were female, 76% were non-Hispanic white, and 58% had a college or graduate education. A total of 2453 (19.2%) reported no insomnia symptoms, 5201 (40.8%) reported 1 or 2 symptoms of insomnia, and 5105 (40%) reported 3 or 4. There was a high rate of some level of physical activity among participants (93%). The majority were overweight (56.5%, with BMI >30 kg/m2 ), and 19% reported that they currently smoked. Thirtynine percent reported that they never drank alcohol. Based on Center for Epidemiologic Studies-Depression Scale scores of 4 or more symptoms, 15.7% of the participants could be classified as clinically depressed. Table 2 shows the intensity of this cohort’s experience with the death of family members by category. In this presentation, the term loss captures the perspective of the survivor. Loss of a spouse, for example, emphasizes the perspective of the widow or widower. As expected, the majority of these adults had experienced a loss (87.2%). In many cases (52%), there were reports of 2 losses or more. Death of a parent was common: 80% had experienced the loss of a father and 62% had experienced the loss of a mother. Approximately 6.2% had experienced the loss of their spouse or partner and 2.26 experienced the loss of a child. (These data are not shown.) Table 2 suggests that there is a positive relationship between the number of insomnia
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
The Impact of Mid- and Late-Life Loss on Insomnia
321
Table 1. Weighted Participant Characteristics for Age, Gender, Lifestyle Factors, Depression, and Loss Experience Compared With Number of Insomnia Symptoms, Health and Retirement Study 2010 Cohorta
Number of Insomnia Symptoms
Total Sample (n = 12 759) Age, y 50-54 (n = 3290) 55-59 (n = 3725) 60-64 (n = 3094) 65-70 (n = 2650) Gender Female (n = 6675) Male (n = 6084) Physical activity None (n = 901) Active (n = 11 858) BMI ≤30 kg/m2 (n = 5553) >30 kg/m2 (n = 7206) Smoking status Nonsmoker (n = 4826) Smoker (n = 2438) Alcohol use Nondrinker (n = 5094) Drinker (n = 8396) Depression Not depressed (n = 10 665) Depressed (n = 2004)
0 Symptoms %
Subclinical Insomnia 1-2 Symptoms %
Clinical Insomnia 3-4 Symptoms %
17.7 19.0 19.9 20.6
38.4 39.8 41.3 44.4
43.9 41.2 38.8 35.0
16.3 22.5
38.3 43.4
45.4 34.1
14.6 19.6
37.8 41.0
47.5 39.4
20.9 17.9
41.0 40.6
38.1 41.5
19.6 18.0
41.9 37.7
38.5 44.3
16.8 20.5
38.7 41.8
44.5 37.7
22.1 4.8
44.1 23.6
33.8 71.6
χ2 P
The Impact of Mid- and Late-Life Loss on Insomnia Findings From the Health and Retirement Study, 2010 Cohort Cherie Simpson, PhD, APRN, CNS-BC; Joseph C. Allegra, MPH; Amara E. Ezeamama, PhD; Jennifer Elkins, PhD; Toni Miles, MD, PhD Bereavement and insomnia are both well-documented risk factors for illness. We use cohort data to estimate risk of insomnia after death of a family member among adults aged 50 to 70 years. Each day, 6700 persons die in the United States. During the next 20 years, this number will increase. In this cohort, any loss increases the likelihood of insomnia. The highest rates of insomnia occur among women aged 50 to 59 years; men aged 65 to 70 years, and persons reporting death of a spouse/partner or child. Physical activity reduces this risk by one-third. Bereavement is a public health issue requiring a targeted response. Key words: insomnia, late-life bereavement, loss, sleep
E
STIMATES suggest that 50 to 70 million Americans suffer from sleep disorders or sleep deprivation, an unmet public health problem with high costs both to individuals and to society.1 Sleep disorders such as insomnia hinder the performance of complex functions such as driving, for example, and increase the risk for workplace accidents.2,3 Insomnia alone has been estimated to cost $6.6 billion in direct costs of health care utilization, transportation, and medications, as well as indirect costs of absenteeism and lost work productivity.4 Insomnia can also have adverse effects on health,5 and a growing body of Author Affiliations: School of Nursing, The University of Texas at Austin, Austin (Dr Simpson); and Department of Epidemiology and Biostatistics (Mr Allegra and Drs Ezeamama and Miles) and School of Social Work (Dr Elkins), University of Georgia, Athens. The authors declare no conflict of interest. Correspondence: Cherie Simpson, PhD, APRN, CNSBC, School of Nursing, The University of Texas at Austin, 1700 Red River, D0100, Austin, TX 78701 (csimpson@ mail.nur.utexas.edu). DOI: 10.1097/FCH.0000000000000039
literature has linked insomnia to individuals’ increased risk for cardiovascular disease, diabetes, anxiety, and depression.6-8 Increasing the public’s knowledge about sleep health through education about sleep and sleep disorders such as insomnia is, therefore, an important goal.9 Insomnia has been described by both behavioral and cognitive models; its etiology and pathophysiology are complex.10 Lifestyle behaviors such as alcohol consumption or physical activity can hurt or help sleep,11 but a known trigger for insomnia is the loss of a loved one. During the bereavement and grief that follow upon such a loss, individuals are especially vulnerable for insomnia’s onset. The cognitive model of insomnia would posit this loss as a precipitating factor for the disorder; the loss is then followed by rumination and worries that contribute to the disorder’s perpetuation. A literature review of late-life spousal bereavement reminds us that the severe emotional strain of the loss of a loved one is accompanied by profound changes in lifestyle and status, often including reductions in financial security, perceived personal 317
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
318
FAMILY & COMMUNITY HEALTH/OCTOBER–DECEMBER 2014
safety, and freedom of action.12 All of these facets of the situation are likely to lead to sleep disruption. In an aging society, loss and its potential impact on the occurrence of insomnia become increasingly important. Loss of a partner is not the only life stage loss experienced by middle-aged and older adults. Other losses include the deaths of parents, siblings, and adult children. Although US mortality rates are falling overall, the number of deaths has increased each year for the past 75 years.13 Middle-aged and older adults have large social networks that extend well beyond their families to include friends and colleagues. Such connections also propagate grief after loss.14 Experience of the loss of a close relative or spouse is common, reaching 86% by the age of 70 years. Population data have shown persistent, elevated risk for mortality of adults across the age span for a 10-year period after these losses.15 What is not clear is whether the elevated mortality risk is purely driven by age or whether the experience of loss is an independent contributor to the elevated risk of mortality. One of the goals of this study is to explore whether loss has an ageindependent association with insomnia—a morbidity linked with a range of adverse outcomes including mortality. So far, however, the literature on sleep and bereavement is limited. Spousal bereavement research has shown that complaints about sleep are common.16 The severity of grief contributes to poor sleep quality independently of depressive symptoms, but concurrent depression increases poor sleep further.17 Even bereaved persons who fail to meet a formal diagnosis of depression have measurable sleep impairment.18 The literature clearly indicates that sleep disruption and insomnia occur after the loss of a loved one. But the studies cited were done with small convenience samples, usually of widowed women. The impact of loss is not well understood at the population level. From a public health perspective, the question is, what is the impact of loss on insomnia in the general population? Who is at most risk for postloss insomnia? This study will explore these questions by examining the
risk of insomnia after exposure to loss in midand late-life adults using the 2010 Health and Retirement Survey cohort. The research questions are as follows: 1. What is the association between the experience of a loss and the prevalence of insomnia symptoms in a nationally representative sample of US adults 50 to 70 years of age? 2. If insomnia prevalence is significantly elevated among person with and without loss, are there disparate effects on insomnia severity depending on the type of loss? 3. Does the association between loss and insomnia symptoms differ for specific age cohorts? 4. Does the association between insomnia and loss differ between men and women? METHODS Data source and study population This study is based on data from the Health and Retirement Study (HRS), a survey designed to follow a representative sample of Americans older than 50 years for every 2 years. Since its inception in 1992, the HRS has enrolled 5 additional cohorts and has captured data on changes in the labor force as well as on the health transitions that individuals undergo toward the end of their working lives and into the years that follow.19 The overall response rate for each of the follow-up waves is higher than 80%. Sampling weights are provided on all HRS data sets to compensate for the unequal probabilities of selection between core and oversample domains (blacks, Hispanics, and residents of Florida).20 The HRS weighted sample is representative of all noninstitutionalized individuals in the US population in the age-eligible range. The HRS design and implementation details have been discussed elsewhere.21 For purposes of these analyses, the sample (n = 12 759) comprised respondents 50 to 70 years of age to capture those in mid- to late life who participated in
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
The Impact of Mid- and Late-Life Loss on Insomnia the 2010 wave (February 2010 to November 2011) and who provided information regarding the primary predictor of exposure to loss as well as the outcome of insomnia. Measures Insomnia symptoms In 2010, respondents were asked 4 questions to determine their sleep status. They were asked how often they had: (1) trouble falling asleep, (2) trouble with waking up in the middle of the night, and (3) trouble waking up too early, as well as (4) how often they felt rested in the morning. The possible responses were “most of the time,” “some of the time,” or “rarely or never.” Previous research based on HRS data defined individuals as experiencing insomnia if they answered “most of the time” or “sometimes” to the first 3 questions and “rarely or never” or “sometimes” to the fourth question (feeling rested in the morning).22 The number of insomnia symptoms was summed, and an index was created from 0 to 4. In addition to the index, an insomnia scale was created by dividing the number of reported insomnia symptoms by the number of possible insomnia symptoms and multiplying by 100. This ensured reliable interpretation of the β coefficients as a percent increase in insomnia score per unit increase in total loss. To better describe the sample, the participants were then grouped as: (1) those with no insomnia symptoms, (2) a subclinical insomnia group with 1 to 2 symptoms, and (3) a clinical insomnia group with 3 to 4 symptoms. Loss In the 2010 HRS survey, the mortality of respondents’ mothers and fathers was assessed with the following questions: “Is your mother alive?” and “Is your father alive?” In addition, the mortality of the respondents’ mothers and fathers was captured as a wave preloaded response (“previous wave mother living” and “previous wave father living”). The responses for questions regarding respondents’ mothers’ mortality were combined, and a new
319
variable was created to encapsulate the mortality status of the respondent’s mother. The same procedure was used for the mortality status of the father. For the mortality status of a spouse/partner, a single variable was created by combining responses to the following questions: “previous wave spouse/partner die during study?” and “divorce/widow since previous wave?” The death of a sibling was enumerated by examining the question “year sibling died?” If a year was given, the death of the respondent’s sibling was grouped as “yes” (“no” otherwise). Finally, 2 questions were used to determine the loss of a child: “number of children ever?” and “number of children living?” If the difference between these 2 questions was not zero, the loss of a child was said to have occurred. For purposes of these analyses, we accumulated all losses independently of the dates of their occurrences. Respondents’ losses were dichotomized as any loss versus no loss. Respondents were further categorized as having no losses, 1 loss, 2 losses, or 3 or more losses. Demographic characteristics To capture the experience of adults in the periretirement period, the sample was restricted to persons between the ages of 50 and 70 years during the 2010 interview cycle. For analysis, age was treated as a categorical variable: 50 to 54, 55 to 59, 60 to 64, and 65 to 70 years. Other categorical variables included gender, race, and education (30 kg/m2 ), smoking status (yes vs no), and alcohol use (any vs none) as known lifestyle factors that can influence sleep. In the
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
320
FAMILY & COMMUNITY HEALTH/OCTOBER–DECEMBER 2014
HRS, physical activity was ascertained by the following questions: “How often vigorous activity?” “How often moderate activity?” “How often mild activity?” Possible responses were “more than once a week,” “once or twice a week,” “1 to 3 times a month,” or “hardly or never.” If the response was “hardly or never” on all 3 questions of physical activity, it was assigned to the “none” category. Any other response besides “hardly or never” to any physical activity question was categorized as physically active. Smoking was assessed by the question “Do you smoke cigarettes now?” If the answer was “yes,” the participant was assigned to the smoker category (“no” for nonsmokers). Likewise, for alcohol use, if the respondent answered “yes” to the question, “Do you ever drink alcohol such as beer, wine, or liquor?” the respondent was assigned to the “drinker” category (“no” for nondrinkers). Body mass index was calculated on the basis of self-reported height and weight. Depression Depression was assessed using a short form of the Center for Epidemiologic StudiesDepression Scale.23 Respondents were asked 8 questions about the presence or absence of depressive symptoms within the past week. Consistent with prior research,24 the study categorized individuals with 4 or more symptoms as having elevated depressive symptoms. Statistical analyses Exploratory analysis was performed to determine the distribution of insomnia in the sample population across demographic groups and covariates. Estimated risk differences (RDs) and 95% confidence intervals (CIs) for the association between loss and the number of insomnia symptoms (0 [reference], 1, 2, 3, 4) were calculated using multiple linear models. The first model looked at the crude association between the exposure to loss and insomnia symptoms. Model II controlled for age and gender. Model III additionally controlled for physical activity, BMI,
smoking status, and alcohol use. Model IV added depression to all covariates in model III. To explore the potential heterogeneity of the association between loss and insomnia by age and sex, stratified linear regression models were built within strata of these covariates (model not shown.) All analyses were conducted using Stata, version 12 (StataCorp, College Station, Texas). The a priori significance level for all statistical tests was set at α < .05. RESULTS Table 1 shows characteristics of 2010 HRS participants between the ages of 50 and 70 years. In this sample of 12 759 participants, 52.3% were female, 76% were non-Hispanic white, and 58% had a college or graduate education. A total of 2453 (19.2%) reported no insomnia symptoms, 5201 (40.8%) reported 1 or 2 symptoms of insomnia, and 5105 (40%) reported 3 or 4. There was a high rate of some level of physical activity among participants (93%). The majority were overweight (56.5%, with BMI >30 kg/m2 ), and 19% reported that they currently smoked. Thirtynine percent reported that they never drank alcohol. Based on Center for Epidemiologic Studies-Depression Scale scores of 4 or more symptoms, 15.7% of the participants could be classified as clinically depressed. Table 2 shows the intensity of this cohort’s experience with the death of family members by category. In this presentation, the term loss captures the perspective of the survivor. Loss of a spouse, for example, emphasizes the perspective of the widow or widower. As expected, the majority of these adults had experienced a loss (87.2%). In many cases (52%), there were reports of 2 losses or more. Death of a parent was common: 80% had experienced the loss of a father and 62% had experienced the loss of a mother. Approximately 6.2% had experienced the loss of their spouse or partner and 2.26 experienced the loss of a child. (These data are not shown.) Table 2 suggests that there is a positive relationship between the number of insomnia
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
The Impact of Mid- and Late-Life Loss on Insomnia
321
Table 1. Weighted Participant Characteristics for Age, Gender, Lifestyle Factors, Depression, and Loss Experience Compared With Number of Insomnia Symptoms, Health and Retirement Study 2010 Cohorta
Number of Insomnia Symptoms
Total Sample (n = 12 759) Age, y 50-54 (n = 3290) 55-59 (n = 3725) 60-64 (n = 3094) 65-70 (n = 2650) Gender Female (n = 6675) Male (n = 6084) Physical activity None (n = 901) Active (n = 11 858) BMI ≤30 kg/m2 (n = 5553) >30 kg/m2 (n = 7206) Smoking status Nonsmoker (n = 4826) Smoker (n = 2438) Alcohol use Nondrinker (n = 5094) Drinker (n = 8396) Depression Not depressed (n = 10 665) Depressed (n = 2004)
0 Symptoms %
Subclinical Insomnia 1-2 Symptoms %
Clinical Insomnia 3-4 Symptoms %
17.7 19.0 19.9 20.6
38.4 39.8 41.3 44.4
43.9 41.2 38.8 35.0
16.3 22.5
38.3 43.4
45.4 34.1
14.6 19.6
37.8 41.0
47.5 39.4
20.9 17.9
41.0 40.6
38.1 41.5
19.6 18.0
41.9 37.7
38.5 44.3
16.8 20.5
38.7 41.8
44.5 37.7
22.1 4.8
44.1 23.6
33.8 71.6
χ2 P
