NHATS has been designed to capture a detailed picture of how functioning in daily life changes with age (Freedman 2009). Below is a description of the sample and several key measures used in this report. Those taken from Freedman & Spillman (2013) are designated by an asterisk (*).
Sample. The first round of NHATS took place in 2011 with a national sample of older adults drawn from the Medicare enrollment file (Montaquila et al. 2012a). African Americans and respondents at older ages were oversampled. In all, 8245 interviews were completed. Respondents living in the community and residential care settings other than nursing homes received a two-hour in-person interview that included self-reports and performance-based measures of disability. For respondents in residential care settings including nursing homes a facility respondent provided information about services available and the type of place.
This report draws on information from respondents who completed the sample person interview (N=7609) and excludes those living in nursing homes (N=468). A large proportion of nursing home residents are known to have dementia (ranging from one-half to three-quarters; Magaziner et al. 2000; Spillman 2011). Excluding the over 1 million nursing home residents in 2011 biases our estimate of the proportion of persons with probable dementia downward by about 1.5-2 percentage points.
NSOC was conducted with caregivers to NHATS participants. Interviews with caregivers gathered information about the caregiving experience, caregiving support, and demographic, socioeconomic, and family characteristics. NSOC eligibility was a two-stage process. First NHATS respondents were eligible if they reported help with mobility, self-care or household activities for health/functioning reasons, or lived in residential care settings. Second, caregivers were eligible if they were a relative or an unpaid caregiver helping with a broad set of activities that included mobility, self-care, or household activities or other activities such as transportation or medical visits. For more details on the NSOC design see Kasper et al. (2013b) and Spillman et al. (2014).
Key Measures from NHATS for this Report. We constructed several key measures for this report including dementia status, activity limitations, assistance, the potential and actual care network, hours of care, and residence in supportive care places.
Dementia status. Three types of information collected by NHATS were used to create a classification indicating probable dementia, possible dementia, and no dementia: (1) a report that a doctor told the sample person he/she had dementia or Alzheimer's disease; (2) a score indicating probable dementia on a screening instrument (the AD8; Galvin 2005, 2006) administered to all proxy respondents; and (3) results from cognitive tests that evaluate memory, orientation and executive function.
Details on construction of this classification and results of sensitivity and specificity of the NHATS measure in a sample of persons clinically assessed for dementia are available elsewhere (Kasper et al. 2013a). Using this classification approach, NHATS estimates 14.8% of persons 71 and older have probable dementia (11.2% of persons ages 65 and older). This estimate aligns well with others. The Aging, Memory and Demographic Study, found 13.9% of persons 71 or older had dementia (Plassman et al. 2007). A recent study using the Health and Retirement Survey, gave an estimate of 14.7% for persons 70 and older (Hurd et al. 2013).
This report focuses on the non-nursing home population ages 65 and older, 9.7% of whom have probable dementia.
Activity limitations.* We constructed a three-category measure reflecting limitations in self-care, mobility, or household-related activities. Self-care activities include bathing, dressing, eating, and toileting. Mobility-related activities include getting out of bed, getting around inside one's home or building, and leaving one's home or building. Taken together, self-care and mobility activities correspond to activities of daily living (ADLs). Household activities are laundry, hot meals, shopping for personal items, paying bills/banking, and handling medications, corresponding to instrumental activities of daily living (IADLs). Three hierarchical categories were created summarizing how activities are carried out: (1) without difficulty and without assistance from another person; (2) with difficulty when carried out alone and with whatever accommodations the individual has made; and (3) with assistance from another person, which for household activities must be for health-related or functioning-related reasons. See Freedman & Spillman (2013) Appendix I for further details.
Assistance.* Individuals receiving assistance with 3+ self-care or mobility activities are considered separately from those receiving assistance with 1-2 self-care or mobility activities. The third category consists of individuals who receive assistance with household activities for health or functioning reasons but do not receive help with self-care or mobility. Combined, these three groups form the non-nursing home population receiving help with any self-care, mobility or household activities (analogous to any ADL or IADL assistance).
Actual care network.* Actual networks include persons who helped in the last month with self-care or mobility tasks, household tasks, or selected other tasks (driving, seeing the doctor, less common money matters, and health insurance matters). If the respondent lived in a residential care setting, staff members were not included in counts of network members, but were considered a source of paid help in prevalence estimates. Other non-staff paid and unpaid persons assisting respondents in residential care were counted individually. In this report we focus on actual networks for sample persons who received help in the last month with any self-care or mobility task or any household task for health or functioning reasons.
Supportive care environments.* NHATS distinguishes among care settings that are nursing homes, supportive care settings other than nursing homes, and all other community settings. Nursing homes and other supportive care places were confirmed through an interview with a facility staff member. Such an interview was triggered by questions in the sample person interview about whether the place where they lived offered group meals, bathing and dressing care, or had different levels of care the sample person could move to if he/she needed care.1 In cases where an interviewer conducted the facility interview first, the type of setting was first confirmed through the facility interview, and a sample person interview attempted if the respondent was found to be in a residential care setting other than a nursing home. Supportive care included the following place types (as reported by the facility respondent): adult family care homes, group homes, board and care homes, personal care homes, assisted part of a multi-level place, and enriched housing (housing with services).
Hours of unpaid and paid care.* For older adults living outside the nursing home, NHATS provides estimates of hours of unpaid and paid care in the last month provided by each caregiver, excepting hours provided by staff at residential care settings.
Hours were missing for one or more caregivers for 12% of sample persons receiving help. We imputed missing paid and/or unpaid hours for each caregiver based on the NHATS respondent's age, sex, level of assistance, and, for informal caregivers, the caregiver's relationship to the sample person. We then summed hours across all caregivers for each sample person to create total hours. See Freedman & Spillman (2013) Appendix II for details.
We include all hours provided by an individual identified as a (non-staff) caregiver for the activities identified earlier, including self-care, mobility, household tasks, driving, seeing the doctor, and help with insurance matters and other less common financial tasks. NHATS respondents report only total hours provided by each caregiver identified, so that hours cannot be parsed into those devoted to particular activities.
Key Measures from NSOC Constructed for this Report. Measures from NSOC that were taken from Spillman, Wolff, Freedman and Kasper (2014) are designated by two asterisks (**).
Types and amount of help.** Each NSOC respondent reported the total hours of care they provided in the month prior to interview. NSOC respondents also reported whether they helped with activities in five domains of care:
- self-care and mobility activities;
- household activities;
- health or medical care activities; and
- interactions relating to the health care system and providers.
Work and family.** Each caregiver's age, relationship to the care-recipient, gender, and whether they worked for pay was collected in NSOC. For employed caregivers, hours worked in the last week was obtained. For those who were employed but did not work in the last week, hours were assessed for the most recent week when they did work.
Positive and negative aspects of caregiving.** The primary measures used to characterize positive and negative aspects of caregiving draw on four items assessing perceived gains from caregiving and four items assessing perceived negative aspects of caregiving (Pearlin et al. 1990, Lawton et al. 1989). Each item has three response categories: very much, somewhat, or not so much. Both sets of questions were preceded by a neutral introduction in which the interviewer asked the respondent to listen to the statements and "answer whether this describes your situation ". Values were rescaled so that each item was valued from 0 (not so much) to 2 (very much). Factor analysis indicated that the two sets of items represented two factors, one positive with loadings of 0.52 or higher (alpha=0.70) and the other negative with loadings of 0.58 or higher (alpha=0.75). Each set of items was then combined to create positive and negative scales valued 0-8. We also examined three items in which respondents were asked where they experienced financial, emotional, and physical "difficulty," and if so, to rate the level of difficulty of each type on a scale of 1-5. Only 94 (<5%) of the 1996 respondents in the analysis sample was missing any of the 11 items. Missing values were imputed separately for positive, negative, and difficulty items using a vector approach that matched donors and recipients by residential setting, the reported values for the vector of positive (negative, or difficulty) items, relationship to care-recipient, gender, and age. Donors were respondents with the respective vectors fully reported.
Support available to and sought by caregivers. NSOC interviewed persons named as helpers (unpaid or relatives) to NHATS participants receiving assistance with mobility, self-care or household activities for health or functioning reasons. Details on NSOC eligibility, content, and data are available in the NSOC User Guide (2013) at http://www.nhats.org. The items used in this report asked caregivers about: having friends or family to help care for the NHATS participant; whether in the last year they went to a support group for people who give care; whether in the last year they used a service that took care of the NHATS participant so the caregiver could take some time away; whether in the last year they received any training to help them take care of the NHATS participant; whether in the last year they found financial help for the sample person including helping apply for Medicaid; and whether in the last year they looked for any of the support services they did not report using.
Weighted Percentages and Population Estimates. For all estimates we use analytic weights that take into account differential probabilities of selection and non-response (Montaquila et al. 2012). For population estimates, we further adjust totals to the age distribution of the sample frame.