National Invitational Conference on Long-Term Care Data Bases Conference Package (continued)

IV. The 1985 National Nursing Home Survey

TABLE OF CONTENTS

1985 NATIONAL NURSING HOME SURVEY: FACT SHEET

SPEAKER COMMENTS

USE OF NURSING HOMES BY THE ELDERLY: PRELIMINARY DATA FROM THE 1985 NATIONAL NURSING HOME SURVEY

DISCHARGES FROM NURSING HOMES: PRELIMINARY DATA FROM THE 1985 NATIONAL NURSING HOME SURVEY

NOTES

1985 NATIONAL NURSING HOME SURVEY: FACT SHEET

National Center for Health Statistics (March 1987)

I. BACKGROUND

As part of its continuing program to provide information on the health of the Nation and the utilization of its health resources, the National Center for Health Statistics (NCHS) periodically conducts a nationwide survey of nursing facilities. The 1985 National Nursing Home Survey (NNHS), the third in a series, is authorized under Section 306 (42 USC 242k) of the Public Health Services Act. Facilities covered in the survey are those providing some level of nursing or personal care without regard to licensure status or to certification status under Medicare or Medicaid. Participation is voluntary.

II. PURPOSE

The purpose of the NNHS is to collect baseline and trend statistics about nursing facilities, their services, residents, discharges, and staff. The resulting published statistics will describe the Nation's nursing facilities and the health status of their residents. These data are used for studying the utilization of nursing facilities, for supporting research directed at finding effective means for treatment of long-term health problems, and for setting national policies and priorities.

III. CONFIDENTIALITY

Confidentiality is provided to all respondents in the NNHS as assured by Section 308(d) of the Public Health Services Act (42 USC 242m) which states that: "Information...which would permit identification of any individual or establishment...will be held in strict confidence, will be used only for the purposes stated for this study, and will not be disclosed or released to others without the consent of the individual or establishment."

IV. PROCEDURES FOR DATA COLLECTION

Data were collected from a nationally representative sample of 1,220 nursing and related-care homes using a combination of personal interview and self-enumeration techniques. Information about the facility (e.g., number of beds, certification status, number and kinds of staff) was collected through a personal interview with the administrator or designee. With the administrator's permission, a questionnaire was sent to the facility's accountant to obtain basic expense and revenue information and to a maximum of four registered nurses (RNs) to obtain information related to job retention. Through interviews with appropriate nursing staff, information was collected on maximum samples of five current residents and six recent dis- charges. In addition to basic demographic information, data were collected about the sample patients' medical conditions, impairments, functional limitations, services received and sources of payment. A family member of the patient was contacted by telephone to obtain data on socioeconomic status and prior episodes of health care -- information which generally is not available at the facility.

V. PLANS FOR DATA RELEASE

The results of the 1985 NNHS will be released in publications and public use computer tapes. As noted in the above section on confidentiality, no information will be released which identifies individuals or establishments. Publication plans include pamphlets presenting preliminary data, a summary volume presenting detailed tabulations, and individual analytical reports on special topics such as utilization measures and resident characteristics. Release of data will begin in 1987.

SPEAKER COMMENTS

Genevieve Strahan, National Center for Health Statistics

Esther Hing, National Center for Health Statistics

Edward S. Sekscenski, National Center for Health Statistics

I. INTRODUCTION

Facilities participating in the 1985 NNHS were selected from a universe of over 20,000 nursing and related-care homes. Of the 1,220 facilities selected, six were identified as having been included in the pretest phase of the survey. It was decided by NCHS not to recontact these same facilities but instead to transcribe data from the pretest instruments to the national survey instruments. During the fielding effort of the remaining 1,214 facilities, 56 were identified as out-of-scope. Of the remaining in-scope facilities 1,079 participated in the survey for a response rate of 93 percent.

First contact to the facility was made in May 1985 prior to the beginning of the survey. A telephone prescreening procedure was performed to verify contact information for facilities selected in the sample. This prescreening was designed to update facility data concerning facility name, address, telephone number, and the administrator's name.

The next contact made to the sample facility was in the form of an introductory information packet to the administrator. The packet contained a letter from the Director of NCHS explaining the importance of the survey and informing the administrator that an interviewer would be calling for an appointment. The packet also included letters of endorsement from professional health organizations. About a week after the packet should have been received, the interviewer contacted the administrator to set up the appointment to conduct the survey. Depending on the size of the facility, one interviewer or a team of two or three interviewers visited the facility.

A part of the facility visit included the administration of three questionnaires: (1) Facility Questionnaire, (2) Expense Questionnaire, and (3) Nursing Staff Questionnaire.

II. FACILITY FILE

The Facility Questionnaire (FQ), printed in canary yellow, was completed by the interviewer in a face-to-face interview with the administrator or his/her designee. Collected on the FQ was basic information about the facility: ownership, certification status, bed size, number of admissions and inpatient days of care, services provided to residents and nonresidents, and number of nonresidents served. Staffing in several occupational categories was collected for full-time and part-time employees. Full-time equivalent employees for each category were tabulated utilizing the number of hours worked which was collected for all part-time employees. Thirty-five hours of part-time work are taken to equal that of one full-time employee. The survey collected for the first time in 1985, per diem rates for routine care set by nursing homes. These rates were collected separately for Medicare, Medicaid, and private pay patients. Per diem rates will be one of the key units of analysis from the facility file. By matching the unique facility ID number from all documents completed in a sample home, information about residents collected in two other components of the survey can be described by characteristics of the facility. For example, estimates of current residents can be tabulated by ownership of the facility.

The administrator did not always have all the data required for the FQ at hand and needed to consult records or staff in other offices. Questions that required specific numerical data were printed on a separate sheet, referred to as the FQ work sheet. The interviewer gave this work sheet to the administrator at the end of the interview to be completed later. The interviewer picked up the work sheet at the end of the day or at a later date.

In 1985, the typical nursing home was independently and privately owned. It had about 85 beds -- most of which had some form of certification. This typical home had 71 employees per 100 beds. The estimated 19,100 nursing homes set average rates of $61 for skilled private pay daily care and $62 for Medicare skilled care.

These data and more are included in Advance Data report, Number 131, "Nursing Home Characteristics -- Preliminary Data from the 1985 National Nursing Home Survey." Data from the facility file along with data from four other components of the NNHS will be included in a special report to be published by the end of this year.

III. EXPENSE DATA FILE

Upon completion of the FQ, the Expense Questionnaire (EQ), and its accompanying Definition Booklet, printed in green, were presented to the administrator for completion. In many facilities, the administrator completed the EQ; in others, he referred the interviewer to an accountant, a bookkeeper, or a central office. This instrument was completed by a respondent at his or her convenience. A postage-paid return envelope was provided for the return of the EQ. The EQ collected data on two major topics: expenses and revenues. Expense data included payroll, health care services, insurance, taxes, food, utilities, maintenance, and drug expenses. Revenue data included sources of income from patient and nonpatient sources such as contributions. In lieu of a completed EQ, each facility was offered the option of providing the interviewer with a recent financial statement.

IV. NURSING STAFF FILE

After obtaining the financial statement or the name and address of the anticipated respondent for any necessary follow-up, the interviewer introduced the nursing staff component of the NNHS. These two documents (the Nursing Staff Sampling List and the Nursing Staff Questionnaire) were used to collect data on RNs working in nursing homes.

The Nursing Staff Sampling List, printed in blue, was completed by the interviewer in collaboration with a staff member designated to help. For the preparation of this list, it was necessary to divide employment status of all facility RNs into one of three categories: (1) those who are employed on the staff of the facility, (2) those scheduled to work who were retained through a special contractual relationship, and (3) those scheduled to work who were retained through a temporary service. Three columns were provided in which to list separately persons in each category. The sampling list provided the universe of RNs separated into the three groupings. With the introduction of the Nursing Staff Sampling List came the first need to use sampling tables. Each interviewer received a pack of sampling tables. The pack consisted of ten independent sets of three different kinds of tables, which were numbered and color-coded according to the component to which they applied. Table 1 was blue and was used to select the nursing staff sample.

In order to ensure random in-facility samples, Table 1 had ten versions, numbered 0 to 9. The fourth digit of a facility ID number determined which version of Table 1 was to be used to select the RN sample at that facility. For example, if the facility ID number were 1234-00-7, the interviewer would consult version 4 or Tables 1-4. This method of assignment assured a fairly even distribution of facilities among all the versions of the sampling tables.

After finding the total number of RNs recorded on the Nursing Staff Sampling List, the interviewer referred to the version of Table 1 mandated by the fourth digit of the facility ID number. After locating the correct total in the "Total # Listed" column, the interviewer read across to find out which sample line numbers on the sampling list determined the individuals chosen for questionnaire completion. The selection of up to four RNs from each sampled nursing home yielded a sample of 3,349 nurses.

The Nursing Staff Questionnaire (NSQ) also printed in blue, was, when at all possible, personally distributed by the interviewer to those RNs selected on the sampling list. The NSQ was self-administered. When personal delivery was not possible, the questionnaire was either mailed to a home address or left at the facility. A postage-paid business reply envelope was provided for return of the completed questionnaire. If a questionnaire was not received within 28 days of the facility visit, a reminder letter and duplicate instrument were sent.

The NSQ gathered information on the work experience, hours, activities, education, training, salary, and opinions about recruitment and retention issues of RNs working in nursing homes. Basic demographics about each RN were also collected.

Data were collected from 2,763 of the sampled RNs for a 80 percent response rate.

The typical RN working in a nursing home was prepared to work as an RN in a diploma program and has been employed as an RN for more than ten years. She (98 percent are female) worked full-time on a nonrotating day shift. She is white, married with either no children living at home, or the children are of school age (6 to 18). She is scheduled to work an average of 32.5 hours per week and earns about $334 per week.

An Advance Data report should be released this year, reporting characteristics of RNs in nursing homes. Future reports will provide detailed information about RNs working in nursing homes and will be published in both Series 13 and Series 14 reports. Data on RNs will also appear in the special report that combines data from several other components.

Now Esther Hing will talk about the current resident component of the 1985 National Nursing Home Survey.

V. DESCRIPTION OF CURRENT RESIDENT QUESTIONNAIRE

Data from the Current Resident Component of the NNHS are cross sectional and are representative of nursing home residents in the United States as of the night before the survey. To draw the sample of residents, lists of residents in the facility were constructed at the time of the survey. In nearly half of the sample homes, the nursing home provided photocopied or computer generated lists of current residents. In the remaining homes, the lists had to be constructed by copying the names of residents from ledgers, or other lists of patients. A sample of five or fewer residents were selected per sample home resulting in an overall sample of 5,395 current residents.

The Current Resident Questionnaire (CRQ) was used to collect data on the sample of residents. This questionnaire was administered by personal interview with a knowledgeable staff member who referred to the residents' medical record when necessary. The most frequent respondent to the CRQ was a nurse (55 percent), followed by the administrator or owner of the nursing home (17 percent). In about 3 percent of the cases, no staff was available and the interviewer abstracted the data from the medical records. Participation for this questionnaire was very high, the response rate was 97 percent. Item response rate in this questionnaire were also high. This result is by design, since items with low response rates in our pretest of the NNHS were not included on the final questionnaire. Item response rates from the national study were similar to those found in the pretest.

Follow-up information on the sample of current residents was also obtained in a telephone interview with the residents' next-of-kin. The residents' next-of-kin, friends or guardian may have been contacted. Only sample residents with no next-of-kin or other known contacts were ineligible to have this telephone follow-up. The instrument used to collect the follow-up data was called the Next-of-Kin Questionnaire.

The CRQ collects information about the demographic, medical and other utilization characteristics of the nursing home population. These items are summarized in Table 1. Demographic variables include age, sex, race, hispanic origin, and martial status. Medical data include the diagnoses at admission and currently. Up to eight diagnoses were listed for each time period. The data were coded according to the Clinical Modification of the Ninth Revision of the International Classification of Diseases. Other medical data collected include vision and hearing status and prevalence of mental disorders. Utilization data collected include the length of stay since admission and the total monthly charge last month.

Table 1 shows items collected for the first time in the NNHS. These items have asterisks to the left. These items include marital status at admission, presence of living children, diagnoses-related group data for hospital transfers, hospital stays while a resident, history of other nursing home stays, instrumental activities of daily living (this involves the need for help in such activities as caring for personal possessions, handling money, securing personal items and using the telephone), disorientation or memory impairment and sources of payment at admission.

This table also shows that the tape for the CRQ will include the facility number and the resident weight. The facility number uniquely identifies each facility in the survey. By matching the facility number on the CRQ with the facility number on the FQ, information from the FQ such as bed size or ownership type can be moved to the CRQ for further analysis. The resident weight is used to inflate the sample data to national estimates. The weights associated with each file and how they were computed are discussed below.

One of the principal strengths of the current resident data is that it provides national estimates of the population in nursing homes. This is useful to health planners and policy makers who need descriptive data on the utilization of nursing homes.

In particular, several of the new items collected in the CRQ were added to shed light on long-term care policy issues. The items on sources of payment at admission and last month, for example, provides estimates of nursing home residents who had to "spend-down" before becoming eligible for medical assistance from Medicaid. A question was also added to the CRQ on the diagnoses-related group code for all persons transferred to the nursing home from short-stay hospitals. This data, along with other variables from the survey, may be used to assess the impact of the Medicare prospective payment system (PPS) on nursing home care since its implementation in 1983.

Data from the CRQ, however, have certain limitations. Because of the resident sample is selected from patients currently residing in the facility, the length of stay for respondents is incomplete and underestimates the true length of stay that would be achieved at some point in the future. Residents with long length of stay, however, are over-represented in the current resident sample because of the short-time frame -- overnight -- of the sample. As a result, a person admitted to the nursing home for a short stay, for example, one day, has fewer chances of being included in the sample than a person with a stay of one year. Because of these limitations, the current resident data is inappropriate for examining the flow of patients in and out of nursing homes. The best data for investigating this issue would be a longitudinal study of a cohort of persons admitted to nursing homes. Longitudinal surveys, however, are expensive to conduct.

To date one report on the use of nursing homes by the elderly has been published using the current resident data. This report discussed the utilization rate or number of residents per 1,000 population, 65 years and over by age, sex, and race. Selected health and socioeconomic characteristics were also examined. The report found that about 5 percent of the elderly resided in nursing homes on any given day during the survey period of the 1985 NNHS. Use of nursing homes increased with age for both sexes but was greater for females than males. Use of nursing homes was lower for elderly persons who were black or of other races than for white persons. For the most part, these trends have not changed since 1973-74, when the first NNHS was conducted; however, there were some exceptions. There was an increase in the use of nursing homes by elderly black persons and a decrease in use by those 85 years and over. If any one is interested in receiving a copy of this report, they can write to us.

As Genevieve has mentioned, the next NNHS report to be published will be a summary report presenting data from most components of the 1985 NNHS. This report will include current resident tabulations covering all topics covered on the questionnaire. This report will probably be released at the end of the year.

After the summary report, the next scheduled report using current resident data will be a study of the impact of the Medicare PPS on nursing home care.

And now Ted will discuss the data on discharged residents.

VI. DATA PROCESSING

Once data were collected, a series of checks were performed to assure that all responses were accurate, consistent, logical and complete. Manual edits were performed to check the completeness, format and consistency of the data. For example, sampling lists for current residents were checked to determine that the sample was correctly selected. Following the manual edit, diagnostic data were coded according to the Clinical Modification of the Ninth Revision of the International Classification of Diseases. Range checks and checks of identifiers were also performed at the time of keying. At all steps of data preparation and data entry, quality control procedures were taken to minimize processing errors. Once the data were entered, separate files for each questionnaire were created, and extensive computer edits were performed. Computer edits performed were basically of two types: (1) data cleaning based on consistency tests and (2) data flagging for imputation. Data flagged as "missing" during the editing process were then replaced with "good data" from a randomly picked similar responding case. Once the data base was edited and missing data imputed, weights were assigned, and constructed variables such as, length of stay and age of resident were computed. At this point, national estimates may be produced from the data tapes.

Data processing of the next-of-kin file followed a different track, since the data were basically keyed during the telephone interview. Data cleaning of the next-of-kin file was not as extensive as data obtained from the nursing home, because the computer assisted telephone interview automatically followed correct skip patterns. Quality control procedures for the next-of-kin interviews included silent monitoring of calls, review of complete and incomplete cases, and nonresponse conversion efforts. At the conclusion of interviewing for the next-of-kin, the relationship of the respondent to sampled current or discharged resident was coded. Then weights were assigned and constructed variables or recodes were computed.

And now I will talk about how data from the NNHS are weighted to produce national estimates.

VII. WEIGHTING

The design of the 1985 NNHS is a complex multistage probability sample survey. For the sample data to reflect national estimates, the data needs to be inflated by a weighting factor. The weights for the 1985 NNHS estimators include three basic components:

1. Inflation by the reciprocal of the probability of selection,
2. Adjustment for nonresponse, and
3. A first-stage ratio adjustment to total beds in the sampling frame.

For facility level estimates such as the number of nursing homes, number of beds, or total cost of providing care, the probability of selection is the product of the facilities' probability of being included in the sampling frame times the probability of its being selected from the frame. Only homes from the Complement Survey had a probability of being included in the sampling frame of less than 1. For second-stage estimates of current and discharged residents, and RNs, the probability of selection is the product of the probability of facility selection times the secondstage probability of selection for these sampling units.

The nonresponse adjustment factor brings estimates based on the responding cases up to the level that would have been achieved if all eligible cases had responded. The effect of the first-stage bed ratio adjustment is to bring the sample in closer agreement with the known universe of beds.

All three components were used to estimate facility characteristics correlated with bed size, and estimates of current residents, discharged residents, and RNs. The first-stage bed size ratio adjustment, however, was not included in estimates of nursing homes and facility characteristics uncorrelated with bed size.

Weighting factors used to estimate the number of residents and discharges with next-of-kin are similar to the weights for current and discharged residents with the exception of an additional nonresponse adjustment factor for nonresponse to the question requesting the names of next-of-kin and an adjustment factor for the existence of next-of-kin or other contacts for sample residents and discharges.

As a result, estimates of residents and discharges from the next-of-kin file will be less than the overall estimates of residents and discharges.

It should be noted that caution should be used when producing estimates by metropolitan status since the sample was not specifically designed to produce detailed estimates by this characteristic.

VIII. DISCHARGE RESIDENTS

My presentation deals exclusively with the discharged resident component of the 1985 NNHS. Comparisons made with other data files, including previous NNHS, are illustrative and not exhaustive; continuities do exist with many of the data items in the 1973-74 and 1977 discharged resident segments of the NNHS although some items have not been repeated and a number of new items have been added to the 1985 NNHS. This section will outline some but not all of the similarities and differences between the 1973-74, 1977 and 1985 surveys, and hope to cover all items on the 1985 survey. It is also possible to cross a number of the data items available from other components of the 1985 NNHS to yield further information on the discharged population.

Data in the discharged resident file of the 1985 NNHS were obtained from personal interviews conducted in the sample nursing homes with employees deemed most knowledgeable of the discharge residents' health status and conditions during their stay at the sample home. In most cases the interviewee was either a nurse or medical records person who consulted with the available medical records of the discharged resident during the course of the interview. As was true in both previous NNHSs and in the current resident segment of the 1985 survey, no residents were consulted personally in the discharge component of the 1985 survey.

Unlike the 1973-74 and 1977 surveys, the 12-month reference period from which the discharged resident's sample was drawn for the 1985 survey, ended on the date immediately preceding the survey date. Previous survey reference periods for discharges were the calendar year 1972 and 1976. The survey's reference period was changed for the 1985 survey in an attempt to obtain both more current and readily available data and to provide for information on the utilization of nursing homes by both residents and discharges over a more closely related period of time. However, data from the 1985 NNHS for the discharged resident population and current resident population continue to differ in several major areas.

Briefly, while the discharged resident estimates represent all discharges over a 12-month period, the current resident population is estimated for a single night, that immediately prior to the survey date. The discharge sample, therefore, may underestimate those nursing home residents who tend to stay for very lengthy durations, while the current resident population may underestimate those persons with very short durations of stay. While the current resident file provides for what may be considered a "snapshot" of nursing home residents on any given day, the discharged resident file provides for some indication of the over-the-year changes in the nursing home population at least, this is, in terms of whom is being discharged from the nation's approximately 20,000 nursing and related-care homes.

A sample of six or fewer discharged residents were selected per sample home resulting in an overall sample of 6,023 discharged residents. The Discharged Resident Questionnaire (DRQ) collected data on the discharged residents' demographic characteristics (including age, sex, race, Hispanic origin, and marital status), their discharge diagnoses, and the discharge destinations of live discharges, whether or not the resident had difficulty in controlling his/her bowel and whether he/she was bedfast or chairfast during the seven days prior to being discharged from the nursing home. Also obtained was information on the primary sources of payment for the month of discharge (although unlike in the CRQ, no charge data were obtained on the discharged residents). All of these above data items provide continuity with similar data obtained in the 1977 NNHS.

New to the 1985 discharged resident component of the NNHS are data items on the primary diagnoses of discharged residents at admission, categorical information on prior living arrangements immediately preceding admission, and primary source of payment data for the month of admission. Also new were questions on the discharged resident's history of other stays in the sample and other nursing homes, including dates of admission and discharge, and the total number of homes in which the discharged resident had been a resident patient. These questions' data will begin to provide some evidence of patterns of nursing home utilization over a lengthier period of time than a single stay in a single nursing home.

I have a limited number of copies of the 1985 Discharge Resident Questionnaires available for anyone who would like to peruse them after the session today. I will be open for other questions on the DRQ of the 1985 NNHS also at that time.

Publications from the discharged resident component of the 1985 NNHS will include an Advancedata report, scheduled to be released later this summer, and a Series 13 report to be released in 1988.

USE OF NURSING HOMES BY THE ELDERLY: PRELIMINARY DATA FROM THE 1985 NATIONAL NURSING HOME SURVEY 1

Esther Hing, Division of Health Care Statistics, National Center for Health Statistics

I. INTRODUCTION

Most elderly people are not in nursing homes. Of an estimated 28.5 million Americans aged 65 years and over in the United States, only 5 percent were residents of nursing homes on any given day from August 1985 through January 1986. This finding from the 1985 NNHS is consistent with findings from previous NNHSs conducted in 1973-74 and 1977.2 In these surveys also it was found that about 5 percent of the elderly were residents of nursing homes.

Differences, however, exist in the use of nursing homes by age, sex, and race subgroups. In this report, these differences in use rates are examined. Differences in the health and socio-economic characteristics examined in this report are functional dependencies in the basic activities of daily living (ADLs)--bathing, dressing, using the toilet room, transferring from a bed or chair, continence, and eating; cognitive functioning (disorientation or memory impairment and senile dementia or chronic organic brain syndrome); marital status at admission; whether residents had living children; living arrangements prior to admission to the nursing home; and primary source of payment at admission. The focus of this report will be a comparison of the characteristics of the elderly who reside in nursing homes with characteristics of those who reside in the community.

The data presented in this report are from the 1985 NNHS, a nationwide sample survey of nursing homes, their residents, discharges, and staff conducted by NCHS. The survey, which was conducted from August 1985 through January 1986, was the third of a continuing series of nursing home surveys. The first survey was conducted from August 1973 through April 1974, and the second was conducted from May through December 1977.

Facilities included in the 1985 NNHS were nursing and related care homes in the conterminous United States that had three or more beds set up and staffed for use by residents and that routinely provided nursing and personal care services. A facility could be free standing or could be a nursing care unit of a hospital, retirement center, or similar institution as long as the unit maintained financial and employee records separate from the parent institution. Placed providing only room and board were excluded, as were places serving only persons with specific health problems (for example, mental retardation or alcoholism).

The sampling frame for the 1985 NNHS consisted of the following components:

• The 1982 NMFI, a census of nursing and related care homes conducted by NCHS.
• Homes identified in the 1982 Complement Survey of the NMFI as "missing" from the 1982 NMFI.3
• Nursing homes opened for business from 1982 through June 1984.
• Hospital-based nursing homes identified in records of HCFA.

The resulting frame contained 20,749 nursing homes. In this report, the terms "nursing homes" and "nursing and related care homes" are used interchangeably.

Estimates in this report are based on a sample of 4,646 elderly residents of the 1,079 nursing homes participating in the survey. A fixed sample of five or fewer residents per sample facility was selected. Residents included in the sample were those on the nursing home's roster the night before data collection began. Data were collected by interviewing knowledgeable nursing home staff members, who referred to the residents' medical records when necessary. Additional followup information on the sample residents was collected by telephone interview with the residents' next-of-kin. (A resident's guardian or friends were contacted if there was no next-ofkin.) Data collected from the next-of-kin focused on the circumstances and reasons for the resident's nursing home admission. In this report, only data obtained from the nursing home staff are presented. In later reports estimates from the next-of-kin component will be included.

Data presented in this report are preliminary and may differ slightly from estimates presented in later reports because of further data editing. Another report presenting preliminary estimates of nursing homes and utilization characteristics of homes has already been published.4

Although data on residents reported by the nursing home staff were collected in a similar manner in earlier NNHSs as in the 1985 survey, note should be taken of some differences. First, personal care and domiciliary care homes were excluded from the scope of the 1973-74 NNHS but included in the two later surveys. The effect of this difference, however, is small because only about 2 percent of all nursing homes in 1973 were personal care or domiciliary care homes and they housed only about 1 percent of the beds and residents.5 Second, certain variables presented in this and later reports were not available from the previous surveys. Data on some variables discussed in this report--marital status at admission, the presence of living children, ability to transfer in or out of a bed or chair, and primary source of payment at admission--were not collected as a single item in the 1973-74 and 1977 surveys but as separate items in 1985. This difference should be considered when comparing data by race from the 1985 NNHS and previous surveys.

Because data in this report are national estimates based on a sample, they are subject to sampling errors. Information on sampling variability is presented in the Technical notes.

II. UTILIZATION RATES

In 1985 an estimated 1,491,400 residents lived in 19,100 nursing homes nationwide. Of these residents, 1,315,800, or 88 percent, were 65 years of age and over. The number of elderly residents in nursing homes increased 17 percent from 1977 to 1985. Residents aged 85 years and over comprised the largest age group (45 percent), followed by those aged 74-85 years (39 percent) and 65-74 years (16 percent). Because of the preponderance of the very old in nursing homes, those aged 85 years and over accounted for 76 percent of the increase in elderly residents from 1977 to 1985. The proportion of elderly residents who were aged 85 years and over increased from 40 percent in 1977 to 45 percent in 1985.

Not only were nursing home residents typically very old but they also tended to be female and white. Seventy-five percent of elderly residents were female. Similarly, 93 percent of elderly residents were white. Only 6 percent were black, and less than 1 percent were other races (a category that includes Asian and Pacific Islanders, American Indians, and Alaska Natives). On the average, elderly females were older than their male counterparts (84 versus 81 years). Elderly white residents, who had an average age of 83 years, also tended to be slightly older than elderly black residents (81 years) and other residents (80 years).

As measured by the percent of elderly residing in nursing homes, the patterns of nursing home utilization mirrored the distributions of residents by age, sex, and race. On any given day during the survey period, 5 percent of the population aged 65 years and over resided in nursing homes (table 1). The rate of nursing home use increased sharply from 1 percent of those aged 65-74 years to 22 percent of those 85 years and over. Elderly females were twice as likely as elderly males to be residents of nursing homes. Six percent of elderly females were in nursing homes, compared with 3 percent of elderly males. Although use of nursing homes increased with advancing age for both sexes, women used nursing homes at significantly higher rates than men did regardless of the age group. One in four women 85 years of age and over resided in nursing homes, compared with one in seven men of the same age (figure 1). This greater utilization by elderly women than men is a reflection of women's longer life expectancy.6 It is also a reflection of a greater tendency among persons without spouses and with poor health to enter nursing homes.

Elderly white persons are more likely to reside in nursing homes than black persons and those of other races are. In 1985, 5 percent of the elderly white population, compared with 4 and 2 percent of the population of black and other races, respectively, resided in nursing homes. The greater likelihood of elderly white people to reside in nursing homes was particularly true in the oldest age group. Of the population 85 years and over, 23 percent of white people, compared with 14 percent of black people, resided in nursing homes.

This lower use of elderly black people and those of other races may result from substitution of informal care at home for formal nursing home care. According to data from the 1982 National Long-Term Care Survey (NLTCS), a higher proportion of elderly black people and people of other races than elderly white people were functionally impaired and remained in the community. Overall, 29 percent of the noninstitutionalized elderly who were of black or other races were functionally impaired in ADLs or home management activities for at least three months, compared with only 19 percent of white people.7 Thus, elderly persons who were of black or other races were over represented among the noninstitutionalized most at risk of needing nursing home care. This finding suggests "the use of a more extended support system among black persons than among white persons."7 Other studies have shown that elderly black persons are more likely than elderly white persons to receive care at home.8

The proportion of the elderly residing in nursing homes has not changed since the period 1973-74, when the first NNHS was conducted (figure 2). An exception to this trend is the increase in the proportion of elderly black persons using nursing homes. During the period 1973-74, 2 percent of the elderly black population resided in nursing homes; in 1985, the proportion was nearly 4 percent. In contrast, the proportion of the elderly in nursing homes did not change from 1973-74 to 1985 for persons who are white or of other races. About 5 percent of elderly white persons and 2 percent of elderly persons of other races were residents of nursing homes throughout this period. The percent of elderly males and females as well as the percent of the elderly aged 65-74 and 75-84 years who resided in nursing homes also remained the same. The percent of persons 85 years and over, however, decreased: 25 percent of persons aged 85 years and over resided in nursing homes in the period 1973-74, compared with 22 percent in 1985.

III. FUNCTIONAL DEPENDENCIES

Because of the preponderance of very old residents in nursing homes, it is not surprising that many residents required assistance in performing or did not perform the basic ADLs, which are needed for independent living. In 1985, 91 percent of elderly residents required assistance in bathing; 78 percent received assistance in dressing; 63 percent required assistance in using the toilet room; 63 percent required assistance in transferring from a bed or chair; 55 percent were incontinent (bowels, bladder, or both); and 40 percent required assistance in eating (table 2). These findings are consistent with earlier studies by Katz and Akpom, in which it was shown that loss of independence is most likely to occur in bathing and least likely to occur in eating.9

In general, elderly residents in nursing homes were more dependent in performing the ADLs in 1985 than in 1977. A larger proportion of elderly residents required assistance or had difficulty with bathing, using the toilet room, continence, and eating in 1985 than 1977 (table 2 and table 3). The exception to this trend was for dressing. The proportion of elderly residents requiring assistance in this ADL remained the same in both years. (Information about transferring from a bed or chair is not available from the 1977 NNHS.)

A partial explanation of the increased level of functional dependency is the shift in the age dis- tribution of nursing home residents to the very old age group (85 years and over), noted earlier. However, as table 2 and table 3 show, the proportion of residents functionally dependent in each ADL was generally higher in 1985 than in 1977 even when age was held constant. For example, a larger proportion of residents aged 85 years and over were dependent in bathing, dressing, using the toilet room, continence, and eating in 1985 than in 1977. Another explanation is the impact of Medicare policy on nursing home care. Under the Medicare prospective payment system (PPS), instituted in 1983, hospitals are encouraged to reduce patient length of stay. Patients released earlier under this new system may require a higher level of care in the nursing home than they would have needed if they had stayed longer in the hospital.10

In general, dependency in ADLs increases with age. In 1985, the percent of residents requiring assistance in bathing increased from 85 percent for residents 65-74 years to 94 percent for residents 85 years and over. Similarly, difficulty with bowel or bladder control increased from 43 percent for residents 65-74 years to 58 percent for residents 85 years and over. Because female residents were older, on the average, than male residents, they tended to require assistance in ADLs more often than males did. A greater proportion of female than male elderly residents needed assistance in bathing, dressing, using the toilet room, transferring from a bed or chair, and eating. There was no statistically significant difference in the percent incontinent by sex. Elderly black residents also needed assistance in ADLs more often than elderly white residents did. This was the case in five of the six ADLs. There was no statistically significant difference in the percent incontinent by race.

The six ADLs may be summarized into a single measure of ADL dependency by summing the number of activities in which a resident required assistance.9 In 1985, 30 percent of elderly residents required assistance in all six ADLs, and only 8 percent were independent in all six activities. The mean number of dependencies was 3.9. The mean number of ADL dependencies increased with age from an average of 3.4 dependencies among residents 65-74 years to 4.2 dependencies among those 85 years and over. Females tended to be more functionally dependent than males. Overall, elderly females had an average of 4.0 ADL dependencies, and elderly males had an average of 3.6. Elderly black residents also tended to be more functionally dependent than elderly white residents. The average number of ADL dependencies was 4.2 among elderly black residents, compared with 3.9 among elderly white residents. Thus, the data show a greater need for care in nursing homes among female and black residents. In the case of females, this is correlated with higher use of nursing homes. This is not the case, however, for elderly black persons.

Although it is possible that nursing home policy may preclude the resident from performing ADLs without assistance, the overwhelming need for assistance in ADLs among nursing home residents suggests that this dependency may have been a reason for entering the nursing home. (The importance of functional status as a reason for nursing home admission was also found in a study of Medicare recipients.11) In contrast, the need for such assistance is minimal among the noninstitutionalized elderly. According to data from the Supplement on Aging to the 1984 National Health Interview Survey, 6 percent of the noninstitutionalized elderly received assistance in bathing; 4 percent, in dressing; 2 percent, in using the toilet room; 3 percent, in transferring from a bed or chair; and 1 percent, in eating (table 4). Data from the 1982 NLTCS, which covered noninstitutionalized Medicare enrollees most at risk of needing long-term care (LTC) (people functionally impaired in ADLs or the instrument activities of daily living for at least three months), indicate a lower need for assistance in ADLs than was found among nursing home residents. In 1982, 42 percent of the elderly impaired living in the community required assistance in bathing, 20 percent required assistance in dressing, 21 percent required assistance in using the toilet room, 26 percent required assistance in transferring from a bed or chair, and 6 percent required assistance in eating.7 Additional insights should be provided on the reasons for admission when data from the next-of-kin component are available.

IV. COGNITIVE IMPAIRMENT

Another reason for nursing home placement that is cited in the literature is deteriorating cognitive functioning.12 In 1985, 63 percent of elderly residents were disoriented or memory impaired to such a degree that performance of the basic ADLs, mobility, and other tasks were impaired nearly every day. Disorientation or memory impairment was defined as being unable to remember dates or time, unable to identify familiar locations or people, unable to recall important aspects of recent events, or unable to make straightforward judgments. Major causes of disorientation or memory impairment in the elderly are senile dementia and chronic organic brain syndrome. In 1985, 47 percent of elderly residents were reported to have at least one of these conditions (table 5). Sixty-six percent of elderly residents who were disoriented or memory impaired were also reported to have senile dementia or chronic organic brain syndrome.

In general, disorientation or memory impairment increased with age: 56 percent of residents 65- 74 years of age had memory impairment or disorientation, compared with 67 percent of those 85 years and over. Elderly female residents were memory impaired or disoriented more often than elderly male residents were--64 percent and 59 percent, respectively. This finding may be related to females' greater longevity. Although it appears that elderly black residents were memory impaired more often than elderly white residents were (70 percent, compared with 62 percent of elderly white residents), the difference was not statistically significant. Similar patterns were also found for residents with senile dementia or chronic organic brain syndrome when examined by age, sex, and race.

V. MARTIAL STATUS AT ADMISSION

The marital status of residents may have influenced the decision to enter the nursing home because persons without spouses may not have anyone living with them to provide personal care services that would allow them to stay in the community longer. In 1985 the majority of elderly residents were without spouses at the time of admission to the nursing home: 65 percent were widowed, 6 percent were divorced or separated, and 14 percent had never married (table 6). In contrast, only 16 percent of elderly residents were married at the time of admission. The likelihood of being widowed increased with age, and the proportion who were married decreased with age. In the group 65-74 years, 36 percent of residents were widowed; 77 percent of residents 85 years and over were widowed. Elderly female residents were more likely to be widowed (74 percent) than elderly male residents (37 percent). Elderly males were more likely to be married (33 percent) than elderly female residents (11 percent).

The tendency of persons without spouses to enter nursing homes is highlighted by comparing the marital status of the functionally impaired elderly living in the community with that of elderly nursing home residents. The proportion married was larger among the functionally impaired elderly living in the community (44 percent) than among elderly nursing home residents (16 percent). Thus, 84 percent of the elderly in nursing homes were without spouses, compared with 56 percent of the functionally impaired living in the community.7

VI. PRESENCE OF LIVING CHILDREN

Data on whether nursing home residents had living children were collected for the first time in the 1985 NNHS. Among elderly nursing home residents, the majority (63 percent) had living children. The proportion of residents with children increased with age and was greater for female residents (66 percent) than male residents (56 percent). The trends among residents with children mirror the increasing utilization rates by age and the greater nursing home use by elderly women. Additionally, these trends appear to contradict the notion that the lack of children, which is a proxy measure for the lack of a social support network, is a risk factor for nursing home institutionalization. The finding that most elderly residents had children does not explain by itself why people enter nursing homes because this variable is confounded by several factors. First, it is not known whether the residents' children lived close enough to provide care and, if they did, whether they were physically able to provide care. Although 69 percent of residents 85 years and over had children, their children were probably in their sixties and may not have been physically able to care for their aging parents. Furthermore, for many residents, admission to the current nursing home was not from the community but from another health institution. As will be discussed in the next section, more than one-half of elderly residents were transferred to the nursing home from another health facility. For these residents, obtaining appropriate continuing care was a deciding factor in entering the current nursing home. Their children may not have been able to provide adequate informal care in the home. Further insights on this issue should be gained when the next-of-kin data on the sample residents are available.

There was one exception to this trend. Only 42 percent of elderly black residents had children, compared with 65 percent of elderly white residents. In the 1982 NLTCS it was found that noninstitutionalized elderly black persons who were functionally impaired were more likely to live with children than functionally impaired elderly white persons were.7

VII. LIVING ARRANGEMENTS PRIOR TO NURSING HOME ADMISSION

The living arrangements of residents prior to admission reflect both the amount of support given in the environment in which they previously lived and their health. A majority of the residents (57 percent) were transferred from another health facility (table 7). The most common type of health facility transferred from was a short-stay hospital (39 percent). Only 12 percent of residents were transferred from another nursing home, and 3 percent were transferred from some type of mental facility (mental hospital, facility for the mentally retarded, psychiatric unit of a short-stay hospital, or mental health center). The proportion of elderly residents admitted from a short-stay hospital in 1985 (39 percent) was a significant increase from the proportion in 1977 (34 percent). This finding may also be related to the introduction of the Medicare PPS, under which hospitals have a strong incentive for early discharge of patients needing LTC services.10 Further analysis of this issue will be presented in a later report.

The increasing proportion of residents transferred from short-stay hospitals to nursing homes also reflects the increasing role hospitals play in the provision of care to the elderly. For example, in the 1985 NNHS it was found that 22 percent of elderly residents were hospitalized for acute episodes of illness while still a resident of the nursing home. Thirteen percent of elderly residents had only one hospitalization, and 9 percent had two or more hospital stays (table 8). Hospitalizations of elderly residents were less likely among those who were admitted to the nursing home from a short-stay hospital, only 14 percent of whom had a subsequent hospitalization while a resident. In contrast, 27 percent of elderly residents not admitted from a short-stay hospital were hospitalized while a resident of the home.

When examined by age, the proportion of elderly residents transferred from a short-stay hospital or nursing home did not vary. Elderly female residents, however, were more likely to be admitted from a short-stay hospital (40 percent) than elderly male residents were (35 percent). In addition, a higher proportion of elderly black residents (50 percent) than elderly white residents (38 percent) were transferred from a short-stay hospital. These findings appear to be correlated with the generally more dependent functional status of elderly women and black residents.

Forty percent of elderly residents were admitted from a private or semiprivate residence; 15 percent had lived alone prior to the nursing home admission, 19 percent lived with family members, and 3 percent lived with persons who were not family members. Residents 75 years and over were more likely than those 65-74 years to have lived alone prior to being admitted to the nursing home. Elderly female residents were more likely to have lived alone (16 percent) than elderly male residents (12 percent). Elderly black residents were less likely (7 percent) than elderly white residents (15 percent) to have lived alone prior to admission. This may result from the tendency of functionally impaired elderly black people to "draw on a more extended range of relationships in their living arrangement than white persons."7

VIII. PRIMARY SOURCE OF PAYMENT AT ADMISSION

Data on sources of funds used to pay for nursing home care provide a rough measure of residents' income sources because public funds for nursing home care under certain government programs are available only to those who cannot afford to pay for such care. The Medicaid program, for example, is a joint Federal-State program providing medical benefits to persons who qualify for welfare and to some of the "medically needy" (those who would be on welfare if their incomes were a little lower). The State-set criteria for Medicaid eligibility vary from State to State but cover most poor people in the United States.13

Information on the payment sources used during the month of admission was collected for the first time in the 1985 NNHS. Table 9 shows the primary payment source used by elderly residents in the home one month or more. One-half of elderly residents relied primarily on their own income or family support to pay for the first month in the nursing home, and 40 percent relied primarily on the Medicaid program to pay for care. Medicaid finances both skilled nursing and intermediate care services in nursing homes. At the time of admission, 26 percent of elderly residents received intermediate care and 14 percent received skilled nursing care through the Medicaid program. Only 5 percent of elderly residents relied on Medicare. Extended care benefits under Medicare are limited to 100 days following a hospital stay of at least three days. Three percent of elderly residents relied on other government assistance or welfare, and another 3 percent relied on other payment sources. Overall, 48 percent of elderly residents relied on some form of public funds to pay for their stay at the time of admission.

There were differences in primary payment source by age. Residents 75 years of age and over were more likely to use their own income or family support for primary payment than were residents aged 65-74 years. Of residents 75-84 years and 85 years and over, 51 and 52 percent, respectively, relied on their own income or family support to pay for care, compared with 39 percent of residents 65-74 years. Medicaid was the primary payment source for a larger proportion of residents 65-74 years (45 percent) than residents aged 75-84 years (39 percent) or 85 years and over (39 percent). The primary payment source also varied by sex. A larger proportion of elderly females (42 percent) than elderly males (36 percent) relied on Medicaid for payment.

There were major differences in the patterns of payment at admission by race. Elderly black residents were almost twice as likely to use Medicaid as the primary source of payment (70 percent) as elderly white residents were (38 percent). Conversely, elderly white residents were more likely to use their own income or family support as their primary payment source (52 percent) than elderly black residents were (17 percent). The differences in payment source by sex and race reflect the generally low income of elderly women and elderly black people in the noninstitutionalized population,14 and in particular among the functionally impaired elderly living in the community. In 1982, 46 percent of elderly females who were functionally impaired and living in the community had family incomes of less than $7,000, compared with 31 percent of their male counterparts. Similarly, 61 percent of functionally impaired elderly black persons had family incomes of less than $7,000, compared with 37 percent of functionally impaired elderly white persons. (Family income included income of the functionally impaired individual and all members living with him or her.)7

IX. CONCLUSIONS

On any given day during the survey period for the 1985 NNHS, about 5 percent of the elderly were residents of nursing homes. Use of nursing homes increased with age for both sexes but was greater for females than for males, especially in the older age groups. Use of nursing homes was lower for elderly persons who were black or of other races than for white persons. These trends have remained constant since the period 1973-74, when the first NNHS was conducted, with the exception of an increase in the use of nursing homes by elderly black persons and a decrease in use by those aged 85 years and over.

Examination of some health and social characteristics revealed that dependency in ADLs was widely prevalent among elderly nursing home residents but much rarer among the noninstitutionalized elderly. The lack of available caregivers may have been a confounding factor for the preponderance of persons without spouses in nursing homes. The role of the residents' children or their living arrangements prior to admission in the decision to enter a nursing home is not clear from the data examined. The need for continuing care in a nursing home and the availability and willingness of the residents' children to provide informal home care are issues that need further examination before conclusions can be drawn. There issues will be examined in future reports in which data from the next-of-kin component are presented. The lower use of nursing homes by elderly black persons appears to be related to a greater substitution of informal care at home for formal nursing home care.

In this report, data on the primary source of payment for care used by residents during the month of admission were also presented. The data show that in 1985 one-half of elderly residents relied primarily on their own income or family support to pay for the first month in the nursing home, and 40 percent relied primarily on the Medicaid program to pay for care. Residents 75 years of age and over were more likely to use their own income or family support to pay for care at admission than residents aged 65-74 years were. Residents aged 65-74 years were more likely to use Medicaid. Elderly black residents were almost twice as likely as elderly white residents to use Medicaid as the primary payment source at admission. Overall, 48 percent of elderly residents relied on some form of public funds (Medicaid, Medicare, other government assistance, or welfare) to pay for their stay at the time of admission.

TECHNICAL NOTES

Because the statistics presented in this report are based on a sample, they will differ somewhat from figures that would have been obtained if a complete census had been taken using the same schedules, instructions, and procedures. The standard error is primarily a measure of the variability that occurs by chance because only a sample, rather than the entire universe, is surveyed. The standard error also reflects part of the measurement error, but it does not measure any systematic biases in the data. The chances are 95 out of 100 that an estimate from the sample differs from the value that would be obtained from a complete census by less than twice the standard error.

The standard errors used in this report were approximated using the balanced repeated-replication procedure. This method yields overall variability through observation of variability among random subsamples of the total sample. A description of the development and evaluation of the replication technique for error estimation has been published.15, 16

Although exact standard error estiamtes were used in tests of significance, it is impractical to present exact standard error estimates for all statistics used in this report. Thus, a generalized variance function was produced for aggregated resident estimates by fitting the data presented in this report into a curve using the empirically determined relationship between the size of an estimate X and its relative variance (rel var X). This relationship is expressed as:

rel var X = S2x/X2 = a + b/X

where a and b are regression estimates determined by an iterative procedure. Preliminary estimates of standard errors for the percents of the estimated number of residents are presented in table 10.

The Z-test with a 0.05 level of significance was used to test all comparisons mentioned in this report. Not all observed differences were tested, so lack of comment in the text does not mean that the difference was not statistically significant.

DISCHARGES FROM NURSING HOMES: PRELIMINARY DATA FROM THE 1985 NATIONAL NURSING HOME SURVEY 17

Edward S. Sekscenski, M.P.H., Division of Health Care Statistics

This report presents information on discharged residents of nursing and related-care homes based on preliminary estimates from the 1985 National Nursing Home Survey (NNHS). The 1985 NNHS is the third in an ongoing series of sample surveys designed to provide a variety of data on nursing homes in the conterminous United States and is conducted periodically by the National Center for Health Statistics (NCHS). Previous surveys were conducted in 1973-74 (NCHS, 1977) and 1977 (NCHS, 1979).

The data presented in this report were collected between August 1985 and January 1986 and deal specifically with demographic, health, and other characteristics of persons formally discharged from nursing homes during the 12-month period immediately prior to the survey date. Other reports already published present information on nursing home residents (NCHS, 1987a) and facilities (NCHS, 1987b) based on national estimates from the same survey. Two other reports resulting from the 1985 NNHS will provide information on registered nurses employed at nursing homes and on current and discharged nursing home residents. The latter report will be based on a followup survey of the next of kin of the sample population. A summary report presenting data from all five components of the survey also will be prepared by NCHS. Because data in this report are preliminary, they may differ slightly from those published later after further edits are conducted.

Facilities included in the 1985 NNHS were nursing and related-care homes in the conterminous United States that had three beds or more set up and staffed for use by residents and that routinely provided nursing and personal care services. A facility could be freestanding or could be a nursing care unit of a hospital, retirement center, or similar institution as long as the unit maintained financial and employee records separate from the parent institution. Facilities providing only room and board were excluded, as were those serving only persons with specific health problems (for example, mental retardation or alcoholism).

The sampling frame for the 1985 NNHS consisted of the following components:

• The 1982 National Master Facility Inventory (NMFI) (NCHS, 1986), a census of nursing and related-care homes conducted by NCHS.
• Homes identified in the 1982 Complement Survey of the NMFI as "missing" from the 1982 NMFI.
• Nursing homes opened for business from 1982 through June 1984 and identified by the NCHS Agency Reporting System (NCHS, 1968).
• Hospital-based nursing homes identified in records of the Health Care Financing Administration.

The resulting frame contained about 20,500 nursing homes, and a sample of 1,220 homes was selected. In this report, the terms "nursing homes" and "nursing and related-care homes" are used interchangeably.

Estimates in this report are based on a sample of 6,023 discharges from the 1,079 nursing homes participating in the survey. A more detailed description of the survey design, data collection methodology, and estimation procedures for the NNHS has been published elsewhere (Shimizu, 1987). A brief discussion of the standard errors associated with these data is presented in the Technical notes to this report. For convenience, this report uses the terms "discharges" and "discharged residents" interchangeably.

I. BACKGROUND AND TYPE OF DATA

Data in this report were obtained from personal interviews conducted in the sample nursing homes with the employees deemed most knowledgeable of the medical records of the discharged residents. In most cases the interviewee was either a nurse or medical records person who consulted with the available medical records of the discharged resident during the interview. As was true in the NNHS of previous years, no discharges were consulted personally in this component of the survey. The full sample consisted of six or fewer discharges from each nursing home whose discharge dates fell within the 12 months prior to the survey date.

The 12-month reference period from which the discharge residents' sample was drawn for the 1985 survey ended on the date immediately preceding the survey date. Previous survey reference periods for discharges were the calendar years 1972 and 1976. The reference period of the 1985 survey was changed in an attempt to obtain more current and readily available data and to provide information on the utilization of nursing homes by both residents and discharges over a more closely related period of time. However, data from the 1985 NNHS for the discharged resident population and current resident population differ in several major areas. These differences are discussed in more detail in other NCHS publications (NCHS, 1978). Briefly, while the discharged resident estimates represent all discharges over a 12-month period, the current resident population is estimated for a single night, that immediately prior to the survey date. The discharge sample, therefore, may underestimate those nursing home residents who tend to stay for very lengthy periods, while the current resident population may underestimate those persons with very short durations of stay. While the current resident file provides for what may be considered a "snapshot" of nursing home residents on any given day, the discharged resident file provides for some indication of the over-the-year changes in the nursing home population.

Because the methodology for counting discharged residents from the 1973-74 NNHS differed from that of the 1977 (NCHS, 1981) and 1985 surveys, no comparisons will be made in this report between estimates from the 1973-74 survey and those derived from the 1985 NNHS. The 1973-74 NNHS estimated the total number of discharges from each nursing home in the sample from one question in the facility component of the survey. The 1985 NNHS obtained a complete listing of all discharges from the sample nursing home. Comparisons will be presented of estimates from teh 1977 and 1985 discharged resident components of the NNHS where appropriate.

II. DEMOGRAPHIC CHARACTERISTICS, DEPENDENCY, AND DURATION OF STAY

The 1985 NNHS found that an estimated 1,223,500 persons were discharged from an estiamted 19,100 nursing and related-care homes during the 12 months prior to the survey date. Because the survey was conducted between August 1985 and January 1986, the 12-month reference period could have fallen anywhere beginning August 1984 and ending January 1986. The preliminary 1985 estimate represents about a 9.5 percent increase over the 1,117,500 discharges estimated by the 1977 NNHS. Of the recent total, about 37 percent were men while 63 percent were women, roughly the same as was found in the 1977 survey (see table 1). In contrast to the discharge population of 8 years earlier, however, the distribution of discharges in the 1985 survey was more heavily weighted with persons aged 85 years and over and by persons more dependent on the nursing home staff in terms of performance of selected activities of daily living.

Although nearly 9 of every 10 discharges in both surveys were aged 65 years and over, the proportion aged 85 years or over rose from 30 to 38 percent between 1976 and 1984-85. Partly as a result of the aging of the discharge population, the proportion of all discharges who were not dependent in either mobility or continence decreased during the 8-year period from 40 to 31 percent while the proportion who were dependent in both of these functions increased drom 35 to 45 percent. The proportion of all discharges who were totally bedfast also rose between survey from about 21 to 35 percent and the proportion who were chairfast remained about 25 percent. Although in both the 1977 and 1985 surveys older discharges tended to be more dependent than were younger discharges (NCHS, 1981), increased dependencies were evident in all major age groups between surveys (see table 2 and table 3).

In the 1977 and 1985 surveys, persons who were discharged at older ages were more likely to have had lengthier durations of stay in the nursing home than persons discharged at younger ages. This was as true for men as it was for women. The median duration of stay for all discharges was 82 days according to the 1985 survey; for persons aged 85 years and over, however, it was 145 days (see table 4). Women discharges, who tend to be older than discharged men (overall median ages, 83 and 79 years, respectively), also had a longer median duration of stay, 93 as compared with 66 days, according to the 1985 survey. Older women, however, also tended to stay longer in nursing homes than older men. At least half of all women over 84 years of age had been confined to the sample nursing home for more than 4 months according to the 1985 survey, while comparable older men had a median duration of stay of little over 3 months.

Although the estimated overall median durations of stay for all discharges, as well as those for all men and all women in the 1985 survey show observable increases over comparable estimates from the 1977 survey, none of these increases is statistically significant (according to a Z test with 0.05 level of significance). Similarly, none of the differences between surveys in the proportional distribution of discharges by similar duration-of-stay categories was significant. Nearly two-thirds of all discharges in either survey had stays of less than 6 months. About 31 percent in the 1985 survey had been discharged after stays of 1 month to less than 6 months. The remaining 37 percent of discharges in the 1985 survey had been confined to the nursing home for 6 months or more (see table 4).

Because these data represent durations of stay in a nursing home identified with a single discharge, they tend to underestimate the overall duration of stay for persons who may have had a series of admissions and discharges to the same or multiple nursing homes over one episode or more of illness. Definitions of nursing home stays used in this report coincide with those used in the 1977 NNHS. The 1985 NNHS also attempted collection of information on multiple stays in nursing homes of the discharged residents with histories of other nursing home stays. These data will be presented in forthcoming publications on the 1985 NNHS.

The 1985 NNHS was the first in the series to obtain race and Hispanic origin information on discharged residents. According to the 1985 survey, about 92.8 percent of all discharged residents were white persons, while only 6.7 percent were black persons. Another half percent were of other racial groups including Asian and Pacific Islander, American Indian, and Alaskan native. About 3 percent of the total were known to have been of Hispanic origin, an ethnicity designation distinct from race (see table 1). These distributions are similar to the distributions by race and Hispanic origin of current nursing home residents in the 1985 survey (NCHS, 1987a). Although differences in overall durations of stay are suggested in the median estimates of white and black discharged residents, these differences are not statistically significant at the 0.05 level of significance. Similarly, no statistically significant difference exists between the median duration of stay of Hispanic persons and that for all discharges in the 1985 survey. Discharged residents of Hispanic origin, however, had a male-to-female ratio nearly the reverse of that of the overall discharged population, 66 to 34 percent.

The distribution of discharged residents by marital status did not change appreciably between the 1977 and 1985 surveys. It appears, however, that factors associated with a person being married at the time of discharge impact favorably on shorter durations of stay in a nursing home. Other studies have also found that the availability of a spouse as home caregiver is one factor in decreasing the likelihood of admission to a nursing home (for example, Butler and Newacheck, 1981), and previous NNHS's have found similar favorable impacts on short durations of stay for nursing home discharges.

Widowed persons constituted the majority of all discharges, 55 percent in the 1985 survey. their median duration of stay was 107 days (see table 4). By contrast, the median duration of stay of married discharges, who constituted the next largest marital group, 22 percent, was only 41 days. Discharges who were never married, however, as well as divorced or separated discharges also had relatively lengthy median stays (see table 4).

Not surprisingly, widowed discharged residents, noted above as having relatively long stays, were also the oldest of the marital group, with an overall median age of 85 years. However, married discharged residents, who as a group has relatively short durations of stay, had an older median age, 78 years, than discharges who were divorced or separated, 70 years.

The effects of age do appear to explain many of the differences in the abilities of discharged residents to perform selected activities of daily living during their final week in the nursing home. While about 40 percent of persons who were aged 65-74 years at discharge had been dependent in both mobility and continence, about half of all discharges older than 84 years were dependent in both categories. In terms of specific dependencies, about one-third of discharges between ages 65 and 84 years were bedfast in their last week in the nursing home, while about 4 in 10 aged 85 years or over were bedfast (see table 2).

Bladder and bowel incontinence was also related to age at discharge. About half of all discharges aged 75-84 years were incontinent of bladder in their last 7 days in the nursing home. Among persons aged 85 years and over, this proportion rises to about 59 percent. Similarly, while about 39 percent of discharges aged 65-74 years were incontinent of bowel in their last week in the nursing home, the comparable proportion rises to 52 percent for persons aged 85 years or over. As might be expected, median duration of stay was longer for discharges who were dependent in both continence and mobility, 108 days, than for those not dependent in either of these daily activities, 64 days.

Differences in functional statuses in selected activities of daily living for discharges in the 1977 and 1985 NNHS are summarized in table 3. As is noted above, discharges in the 1985 survey were generally less mobile and more likely to have been incontinent of bowel, bladder, or both in their last 7 days in the nursing home than were discharges in the 1977 survey. These general increases in dependencies are partially a function of the increased proportion of discharges aged 85 years and over, who as a group are more dependent in these activities than are younger discharges. However, there were also increases in the proportions of discharges who were dependent in both mobility and continence among those under 65 years, 65-74 years, and 75-84 years, as well as those aged 85 years and over (see table 3).

III. LIVING ARRANGEMENTS BEFORE ADMISSION AND AFTER DISCHARGE

The 1985 NNHS collected information on the living arrangements of all discharged residents for the periods immediately prior to admission and, for live discharges, immediately after discharge. The 1977 survey obtained comparable data only for the living arrangements after discharge. Information on both prestay and poststay living arrangements of discharged nursing home residents provides for a more comprehensive understanding from a wider perspective of the population that utilizes nursing homes.

A minority, about 28 percent of all discharged residents, had been admitted to the nursing home from a private or semi-private residence (see table 5). Slightly over half of these discharged residents had been living with family members at the time of their admission.

About 69 percent of all nursing home discharges had been admitted directly from another health facility, with 8 of every 10 of them representing transfers from general or short-stay hospitals. A slightly higher proportion of female discharges had been admitted from general or short-stay hospitals than had men, 57 versus 51 percent. However, another 7 percent of the male discharges had been admitted directly from a veterans hospital. About 1 in every 10 discharges who had been admitted from another health facility came from another nursing home. The porportions were about the same for both men and women.

The median duration of stay in the sample nursing home was far longer for those discharges who had been admitted from a private or semiprivate residence, 118 days, than for those admitted from a hospital, 57 days. This was partially dur to the differences in ages of those in either group. Among those discharges admitted from a residence, about 42 percent were over age 84 years at their discharges. About 37 percent of those admitted from a hospital were aged 85 years or over.

Discharges who had originally been admitted from another nursing home also tended to have long durations of stay. According to the 1985 survey, their median duration of stay was 263 days. The proportion of those discharged over 84 was comparable to that of persons admitted from private or semiprivate residences, 43 percent.

The proportion of live discharges going to private or semiprivate residences immediately following their nursing home stay decreased between the 1977 and 1985 surveys from 37 to 30 percent (see table 6). As a corollary, the proportion of live discharges who were discharged to another health facility increased from 59 to 68 percent. The latter was almost entirely the result of an increase in the proportion of live discharges going to general or short-stay hospitals, from 41 to 49 percent. (Unknown living arrangements following discharge remained about 2-4 percent of the total.)

The increase in live discharges to hospitals, although partially a result of the increased proportion of older persons among all discharged residents, is not fully explained by this shift in demographics. While the proportion of discharges aged 85 years or over going directly to hospitals is slightly larger than is the comparable proportion for discharges aged 65-84 years in both the 1977 and 1985 surveys, the increase in either proportion between surveys is greater among the younger age group. Among live discharges aged 85 years or over, the proportion discharged to hospitals did not rise significantly between the 1977 and 1985 surveys. In 1977 it was 52 percent and in 1985 it was 54 percent. Among live discharges aged 65-84 years old, however, the proportion discharged directly to hospitals increased from 39 to 50 percent over the same period.

The median duration of stay was longer for those persons discharged to another health facility, 113 days, than for those discharged to a private or semiprivate residence, 36 days. Among the former, those who were discharged to a general or short-stay hospital had a median duration of stay of 130 days. In contrast, among those discharged to a private or semiprivate residence, those who went to live with family members had a median duration of stay of 34 days.

IV. PRIMARY SOURCE OF PAYMENT AT ADMISSION AND DISCHARGE

For the first time, the 1985 NNHS collected information on the primary sources of payment for all discharges for the month in which they were admitted to the sample nursing home as well as for the month in which they were discharged. The 1977 NNHS obtained primary source of payment data only for the month of discharge from the nursing home. As might be expected, primary payment sources often differed depending on whether the payment was for the admissino or the discharge month. These differences generally are greater the longer the duration of stay. When observation is made of the total discharge population as a whole, much less shifting among various payment sources is evident, partially due to the large proportion of persons with relatively short durations of stay. However, patterns are evident in shifts of primary payment sources, especially among discharges who shift to medicaid at some time during their stay.

For the month of admission, own income or family support was the primary source of payment for the largest proportion of discharges regardless of their eventual duration of stay. About 4 of every 10 discharges relied primarily on this source to pay for nursing home care in the month of admission, a ratio that was the same whether the completed stay was of short, medium, or lengthy duration (see table 7). The median duration of stay for persons whose primary source of payment for their admission month was own income or family support, 77 days, was similar to that of the overall discharge population. Their distribution by duration of stay was also similar to that for all discharges.

According to the 1985 surveys, the proportion of all discharges who relied on medicaid as the primary payment source in their month of admission totaled about 35 percent. Medicaid coverage for nursing home care is dividedinto two categories, skilled and intermediate, depending on the certification status of the nursing home. While about 15.5 percent of all discharges relied on medicaid skilled funds in their admission month, another 19.6 percent relied on medicaid intermediate care funds. Unlike the proportion of discharges relying on own income to pay for care in their admission month, the proportion of discharges relying primarily on medicaid differed by the eventual durations of stay. Discharges whose completed stays were relatively lengthy were more likely to have relied on a type of medicaid in their admission month than were those whose stays were relatively short (see table 7).

For example, while 12 percent of all discharges whose stays were less than 1 month in duration relied primarily on medicaid skilled care funds to pay for their nursing home care, 19 percent of those whose stays were 6 months or longer relied primarily on this source in their admission months. Comparable proportions for discharges who relied on medicaid intermediate care funds were 11 percent among those whose stays were less than 1 month and 27 percent for those whose completed stays were 6 months or more.

The median durations of stay of discharges who relied on either medicaid skilled or medicaid intermediate funds to pay for nursing home care in their admission months were 145 and 187 days, respectively, each of which is significantly above the median for the discharge population as a whole.

Medicare accounted for a smaller proportion of all discharges' primary sources of payment in their admission months than either their own income or family support, or the combined total of medicaid. Medicare, however, varied quite widely as a primary admission month payment source according to eventual completed duration of stay. Unlike similar differences outlined above for those relying on medicaid, the proportion of all discharges relying on medicare as their primary source of payment in their admission month was greater among discharges with relatively short durations of stay and smaller for those with longer completed stays. About 18 percent of all discharges relied primarily on medicare for payment for nursing home care in their admission months. But, while the proportion was 30 percent among discharges whose stays were less than 1 month, for discharges whose completed stays were 6 months or more, only 6 percent had relied primarily on medicare in their admission month. The median duration of stay was 29 days for all discharges whose primary source of payment in the month of admission was medicare, significantly below the median for all discharges.

All other sources of payment, including other government assistance or welfare, religious organizations, volunteer agencies, Veterans Administration contracts, initial payment-for-life care funds, and others accounted for about 5 percent of all discharges' primary sources of payment for month of admission. This proportion did not vary significantly by completed duration of stay. Discharges relying on these other sources, however, tended to be younger than those whose primary payment sources were medicare, medicaid, or own income. Only about 22 percent were over age 84 years at their discharges, which is significantly below the comparable proportion for all discharges.

For the month of discharge, own income or family support was also the primary source of payment for about 4 of every 10 discharges. Although some variability exists in this ratio by duration of stay, as many as 38 percent of all discharges whose stays were 6 months or more relied primarily on this source for payment of nursing home care in their discharge month as opposed to 45 percent among discharges whose stays were 1 month to less than 6 months in length.

Medicaid, skilled and intermediate care funds combined, accounted for another 40 percent of all discharges' primary payment sources in their discharge months. The overall proportion who relied primarily on medicaid, however, was larger for those with longer stays than for those with relatively short stays. For example, while a total of 22 percent of discharges with stays of less than 1 month relied on some form of medicaid as their primary payment source, among discharges whose stays were 6 months or longer, a total of 56 percent relied on medicaid in their discharge months. About 25 percent of those who stayed 6 months or longer relied on medicaid skilled care funds, and another 31 percent relied on medicaid intermediate care funds as the primary payment sources in their discharge months.

The proportion of discharges who relied on medicare as the primary payment source in their discharge month is a reflection of the limitations of coverage for nursing home care imposed by this Federal health care program. Medicare is limited to the first 100 days of nursing home care for residents who had been admitted directly from a general, short-stay hospital. The resident must also require specific medical assistance according to criteria established by the Federal Health Care Financing Administration (Health Care Financing Administration, 1986). Reliance on medicare as the primary source of payment for the discharge month, therefore, is restricted to discharges with relatively short durations of stay.

Among all dischrages, about 12 percent used medicare as their primary source of payment in their discharge months. Among those whose stays were less than 1 month, however, about 29 percent relied primarily on medicare, as opposed to about 9 percent with stays of from 1 month to less than 6 months in length.

Changes in primary sources of payment between admission and discharge months are summarized in table 8 for all discharges with a duration of stay of 1 month or more. The percent distributions show that except for those entering with medicare as their primary payment source, more than 8 of every 10 discharges relied on the same primary source of payment in their discharge month as they had utilized in their admission month. For example, among persons using primarily their own income or family support in their admission month, 85 percent relied primarily on this source also in their discharge month. The comparable proportion for medicaid (skilled and intermediate combined) is about 90 percent, while about 87 percent who primarily used other sources in their admission month also relied on those other sources in their discharge month.

Among all persons with durations of stay of 1 month or more who utilized medicare as their primary payment source in their admission month, however, only about 37 percent relied primarily on medicare in their discharge montn as well. This was largely the result of the 100-day limitation for medicare coverage of nursing home care. About 32 percent of discharges who used primarily medicare in their admission month shifted to their own income or family support as primary payment source in their discharge month, while another 28 percent shifted to some form of medicaid.

As noted above, while the overall proportion of discharges relying primarily on medicare decreased between admission and discharge months, the proportion using some form of medicare rose. Shifts to medicaid as the primary source of payment varied by both duration of stay and primary payment source in admission month (see table 9). About 11 percent of persons who entered with other than medicaid as their primary payment source shifted to medicaid by the month of their discharge. The proportions of discharges shifting in this manner varied from 10 percent for those with stays of 1 month to less than 6 months to about 22 percent for those with stays of 6 months or more in duration.

Persons entering with medicare as the primary payment source in their admission month were more likely to shift to medicaid than persons entering with own income or family support. This was especially true for discharges whose durations of stay were beyond the 100-day limit imposed by the medicare program. About 10 percent of discharges who had used their own income in their admission month shifted to medicaid by their discharge month, while 15 percent of those relying primarily on medicare in the admission month converted to medicaid. About 8 percent of persons who entered using primarily their own income or family support and had stays of from 1 month to less than 6 months shifted to medicaid, as opposed to 19 percent of those with equal durations of stay who relied primarily on medicare in their admission month. Among discharges entering with medicare whose durations of stay were 6 months or longer, 52 percent shifted to some form of medicaid by their discharge months.

It is not possible from the discharged resident data to pinpoint, however, when during a discharged resident's stay a shift from one payment source to another may have occurred. Differences in primary sources of payment in admission and discharge months are indicative only of a change between two point in time. While a pattern is suggested in the differential proportions of discharges shifting from one primary payment source to another, especially for discharges shifting to medicaid, it is not discernible from the data when these shifts occurred. Although the disaggregation of discharges who shift to medicaid by various duration of stay categories provides some evidence of a "spend down" to medicaid, more detailed data are required to determine when during a resident's stay this shift actually occurs and, for those with multiple stays, in which stay it occurred. Data on the latter issue are available from the next-of-kin component of the survey. Data from the next-of-kin component of the 1985 NNHS will be published in a forthcoming report from NCHS.

V. SUMMARY AND HIGHLIGHTS OF DATA

The 1.22 million nursing home discharges in the 1985 NNHS represent about a 9.5-percent increase from the 1977 survey. Dependencies in both mobility and continence were more prevalent among all age groups in the most recent survey while there was also an increase of from about 30 to 38 percent in the proportion of discharges aged 85 years and over. While the overall median duration of stay, as well as those of men and women, showed observed increases between the 1977 and 1985 surveys, none of these increases is statistically significant. The rise from 41 to 49 percent in the proportion of live discharges going to a hospital, however, is statistically significant. The increase is largely the result of increased hopsitalization of live nursing home discharges aged 65 to 84 years, although the proportion of discharges to a hospital remains larger among those aged 85 years and over.

About 4 of every 10 discharges used own income or family support as primary payment source in both admission and discharge months. The proportion using medicaid, however, generally rose with duration of stay, while only discharges with relatively short stays relied primarily on medicare, due to the limitations on coverage for nursing home care by the medicare program.

More detailed information from the 1985 NNHS, especially on sources of payment, diagnoses at admission and discharge, and duration of stay by admission and discharge characteristics, will be forthcoming in subsequent publication from NCHS.

VIII. REFERENCES

TECHNICAL NOTES

Because the statistics presented in this report are based on a sample, they will differ somewhat from figures that would have been obtained if a complete census had been taken using the same schedules, instructions, and procedures. The standard error is primarily a measure of the variability that occurs by chance because only a sample, rather than the entire universe, is surveyed. The standard error also reflects part of the measurement error, but it does not measure any systematic biases in the data. The chances are 95 out of 100 that an estimate from the sample differs from the value that would be obtained from a complete census by less than twice the standard error.

The standard errors used in this report were approximated using the balanced repeated-replication procedure. This method yields overall variability through observation of variability among random subsamples of the total sample. A description of the development and evaluation of the replication technique for error estimation has been published.15, 16

Although exact standard error estiamtes were used in tests of significance, it is impractical to present exact standard error estimates for all statistics used in this report. Thus, a generalized variance function was produced for aggregated resident estimates by fitting the data presented in this report into a curve using the empirically determined relationship between the size of an estimate X and its relative variance (rel var X). This relationship is expressed as:

rel var X = S2x/X2 = a + b/X

where a and b are regression estimates determined by an iterative procedure. Preliminary estimates of standard errors for the percents of the estimated number of residents are presented in table 10.

The Z-test with a 0.05 level of significance was used to test all comparisons mentioned in this report. Not all observed differences were tested, so lack of comment in the text does not mean that the difference was not statistically significant.

NOTES

1. Reproduced for the conference package from the NCHS Advancedata, Number 135, May 14, 1987, DHHS Pub.No.(PHS)87-1250. Copies can be obtained from U.S. Public Health Service, National Center for Health Statistics, 3700 East-West Highway, Hyattsville, MD 20782, (301)436-8500.

2. National Center for Health Statistics, E. Hing: Characteristics of nursing home residents, health status, and care received, National Nursing Home Survey, United States, May-December 1977. Vital and Health Statistics, Series 13, No.51, DHHS Pub.No.(PHS)81-1712. Public Health Services, Washington, U.S. Government Printing Office, Apr. 1981.

3. National Center for Health Statistics, D. Roper: Nursing and related care homes as reported from the 1982 National Master Facility Inventory Survey. Vital and Health Statistics, Series 14, No.32, DHHS Pub.No.(PHS)86-1827. Public Health Service, Washington, U.S. Government Printing Office, Sept. 1986.

4. National Center for Health Statistics, G. Strahan: Nursing home characteristics, preliminary data from the 1985 National Nursing Home Survey. Advance Data From Vital and Health Statistics, No.131, DHHS Pub.No.(PHS)87-1250. Hyattsville, Md., Mar. 27, 1987.

5. National Center for Health Statistics, A. Sirrocco: Inpatient health facilities as reported from the 1973 MFI Survey. Vital and Health Statistics, Series 14, No.16, DHEW Pub.No.(HRA)76-1811. Health Resources Administration, Washington, U.S. Government Printing Office, May 1976.

6. National Center for Health Statistics: Advance report, final mortality statistics, 1984. Monthly Vital Statistics Report, Vol.35, No.6, Supp.2, DHHS Pub.No.(PHS)82-1120. Public Health Service, Hyattsville, Md., Sept. 26, 1986.

7. C. Macken: A profile of functionally impaired elderly persons living in the community. Health Care Financing Review, Vol.7, No.4, HCFA Pub.No.03223. Office of Research and Demonstrations, Health Care Financing Administration, Washington, U.S. Government Printing Office, Summer 1986.

8. Institute of Medicine: Racial Differences in Use of Nursing Homes, in Health Care in a Context of Civil Rights. Washington, D.C., National Academy Press, 1981.

9. S. Katz and C.A. Akpom: Measure of primary sociobiological functions. Int. J. Health Serv. 6(3):493-508, 1976.

10. M. Meiners and R. Coffey: Hospital DRGs and the need for long-term care services, an empirical analysis. Health Serv. Res. 20(3):359-384, Aug. 1985.

11. M. Cohen, E. Tell, and S. Wallack: Client-related risk factors of nursing home entry among elderly adults. J. Gerontol. 20(6):785-792, Nov. 1986.

12. R. Kane, R. Matthias, and S. Sampson: The risk of placement in a nursing home after acute hospitalization. Med. Care 21(11): 1055-1061, Nov. 1983.

13. K. Davis and C. Schoen: Health and the War on Poverty, A Ten-Year Appraisal. Washington, D.C. The Brookings Institution, 1978.

14. U.S. Bureau of the Census: Characteristics of the population below the poverty level, 1982. Current Population Reports, Series P-60, No.144. Washington, U.S. Government Printing Office, 1984.

15. National Center for Health Statistics, P.J. McCarthy: Replication, an approach to the analysis of data from complex surveys. Vital and Health Statistics, Series 2, No.14, PHS Pub.No.1000. Public Health Service, Washington, U.S. Government Printing Office, Apr. 1966.

16. National Center for Health Statistics, P.J. McCarthy: Pseudoreplication, further evaluation and application of the balance half-sample technique. Vital and Health Statistics, Series 2, No.31, DHEW Pub.No.(HSM)73-1270. Health Services and Mental Health Administration, Washington, U.S. Government Printing Office, Jan. 1969.

17. Reproduced for the conference package from the NCHS Advancedata, Number 142, September 30, 1987, DHHS Pub.No.(PHS)87-1250. Copies can be obtained from U.S. Public Health Service, National Center for Health Statistics, 3700 East-West Highway, Hyattsville, MD 20782, (301)436-8500.