The dependent variable for our analysis was a three level hierarchical measure which differentiated those who were dependent in activities of daily living (ADL), those who were dependent in mobility or instrumental activities of daily living (IADL), and those who were not dependent in either. Individuals were classified into their highest level of dependency defined as follows:
ADL DEPENDENT: Elderly individuals residing in the community, who, because of a health or physical problem, reported that at the time of the survey they had difficulty with and received human assistance with eating, transferring, toileting, dressing or bathing.
MOBILITY/IADL DEPENDENT: Elderly individuals residing in the community, who at the time of the survey were not ADL dependent, but because of a health or physical problem reported difficulty with and received human assistance with inside mobility, outside mobility, meal preparation, grocery shopping, money management, housework (light and heavy) or telephone usage.
INDEPENDENT: Elderly individuals residing in the community who at the time of the survey were neither ADL nor IADL dependent.
Given the construction of the 1984 NHIS-SOA, it had to be assumed that an individual who received help or supervision with any ADL or Mobility/IADL item was actually in need of such assistance. In addition, incontinence, though not mentioned in the above definition, was captured by other ADL measures. That is, we elected to exclude from our definition of ADL dependency individuals who were suffering from stress incontinence only. These are individuals who, though incontinent, do not require human assistance, nor report the need for assistance, in any one of the other five ADLs. Such individuals have no bearing on manpower estimates. Those who were incontinent and did need help were included in the ADL definition by virtue of needing help in one or more of the remaining ADL functions, e.g. dressing.