Future Directions for Residential Long-term Care Health Services Research

You Are Here: AHRQ Home > Research Findings > Long-term Care > Future Directions for Residential Long-term Care Health Services Research

Table 4. Comparison of Four National Data Sources¹

Design, Measure, Characteristics	1997 Medicare Current Beneficiary Survey (MCBS)*	1999 National Nursing Home Survey (NNHS)	Administrative Data from HCFA	1996 Medical Expenditure Survey—Nursing Home Component (MEPS NHC)
Sample Design
*Design*	Person-based national probability design with a rotating longitudinal panel	Periodic cross-sectional multi-stage design with facilities sampled in first stage and persons in the last stage	-Person-specific health status (Minimum Data Set; MDS) data -Medicare enrollment and claims data -Medicaid claims data -Facility-specific state survey inspection (OSCAR) data and SNF cost reports	Multi-stage probability facilities sampled in the first stage and persons in the last stage
Target population and sample size	Current Medicare Beneficiaries, from the HCFA enrollment files: 11,884 persons (for 1997): -Persons with community-only data: 10,398, no SNF NH days; 239, with SNF NH days -1,015 persons in LTC only -232 persons in community & LTC	Nursing homes, NH residents and NH discharges: 8,056 persons (for 1999): -Residents -Discharges 1,496 nursing homes	-MDS: All NH residents in certified NHs -All Medicare Beneficiaries and their claims -Medicaid enrollees and their claims for persons in the 35 states with SMRF data -OSCAR: All certified NHs -SNF cost reports: SNF certified NHs	Nursing homes and persons who used NH during year: 5,899 persons (for 1996): -3,209 Jan. 1 residents -2,690 persons admitted to NH 815 nursing homes
*Residential facility type(s) in population*	All LTC residential facilities	Licensed or certified NHs²	Certified NHs	Certified NHs or licensed NH w/ 24 hr/7 day/week nursing staff ³
*Reference period*	Three calendar years' worth of use and expenditures data for a person	A single cross-section for a person; some items refer to "last month" or at admission.	-MDS: at NH admission, quarterly, annually and when there is a significant change; 1999 first year with Complete repository data. -Claims: Ongoing -OSCAR: Annually and quarterly cross-sections -SNF Cost Report: Yearly	A calendar year's worth of use and expenditures data
*Frequency of (survey) administration*	Ongoing since 1991. Since 1997 LTC facility questionnaire has mirrored much of the MEPS NHC instrument	Six times since 1973/74; scheduled again for 2001	Ongoing	Conducted for 1987 and 1996
*Rounds and mode of data collection*	12 Rounds of computer-assisted personal interviewing, over 4 years	A round of in-person interviewing with a paper instrument	Self-administered paper or electronic filing; transmitted to HCFA in a standardized data format	3 rounds of computer-assisted personal interviewing, over 1 1/2 years
*Respondents and data sources*	-In LTC facility: Administrators and staff, medical records, and NH billing department sources -Community: Sampled person or their proxy -For Medicare HMO's enrollees: medical provider sources - Medicare claims data	In the NH: Administrators and staff, medical records and NH billing department data⁴	-MDS: Care providers in the NH -Claims: Medical providers and providers' staff, medical records and NH billing departments -OSCAR: State inspection surveyors -SNF Cost Report: NH administrators and billing department staff	-In the NH: Administrators and staff, medical records and NH billing sources -Community residing next-of-kin -Can be linked to Medicare claims data
Person Measures
*Changes in health status*	Mirrors MDS items, but collected annually for 4 years; for persons admitted to LTC during year, health status is also measured upon admission and 90 days later.	Cross-sectional measures not directly comparable to MDS items, no measure of change/outcomes	MDS: Measured at multiple points in time (at least quarterly)	Mirrors MDS items; collected for two points in time. The occurrence of infections, pressure ulcers, fractures are measured over a calendar year
*Expenditures and sources of payments (SOPs)*	-Amounts billed, amounts paid and sources of payment, by 10 SOP types, for services in the LTC facility and in the community -Data linked to Medicare claims to improve estimates	Total charges billed for care last month, and primary and secondary SOPs; not collected are amounts paid, source, and data for community-based care.	Only Medicare/ Medicaid reimbursements	-For care provided in the NH, amounts billed, amounts paid and sources of payment, by 10 SOP types -Community-based expenditures not collected -Linkage of NH residents to Medicare claims data possible
*Income and assets*	A question on gross total income	None	None	Detailed income and asset questions
*Follows persons across care settings*	Follows persons across all settings, but sample of LTC persons who transition is very limited: -232 who moved between community and LTC, and with data from both sources -239 with short SNF stay, but no facility reported data	For residents, admission source For discharges, discharge source	Limited	Information on NH transitions between: -NH units (general to Alzheimer's) -Different NHs -NHs and the hospital -NHs and the community -NH admission and discharge sources
*Use of services provided while a resident in the facility*	3 years of medical provider use while resident in LTC settings; includes use provided by trained staff and frequency of use (for 13 service types), hospital-based use (inpatient, outpatient, and ER) and prescribed medicine use	Services provided to person last month, by type of service (for 17 service types); no measure of frequency or use of hospital-based services (inpatient, outpatient or ER)	Services reimbursed by Medicare or Medicaid; eventually prescribed medicine use	1 year of medical provider use while resident in NH; includes use of services provided by trained staff and frequency (for 15 service types), hospital-based use (inpatient, outpatient and ER) and prescribed medicine use
*Process measures*	Limited, could measure: - Appropriateness of prescribed medicine (PMED) use - Linkage of LTC services use, PMED use, -Occurrence of preventive services (i.e., mammogram, chest x-ray, and pap smear)	Limited: Admini-stration of flu shots, pneumococcal vaccine, and Tetanus-Diphtheria (Td) Toxoid booster	In future, could measure, for NH residents, appropriateness of prescribed medicine use in conjunction with a person's health status characteristics	Limited, could measure: -Appropriateness of prescribed medicine (PMED) use -Linkage of NH services use, PMED and person characteristics (e.g., conditions)
Facility Characteristics
Basic characteristics, e.g., number of beds, ownership, certification	Facility characteristics can be linked to persons; facility-level estimates not possible	Facility-level estimates possible; cannot link facility data to persons on recent public use files	Yes (for certified facilities)	Facility-level estimates possible; can link facility data to persons in the MEPS Data Center environment
*Characteristics of facility units (e.g., special care units)*	-Unit type (e.g., personal care) and number of beds; for special care units (e.g., sub-acute) also captures whether unit has dedicated staff & proportion of residents with Medicare/Medicaid as an SOP -Persons can be linked to their specific unit(s) - Facility-level analysis not possible	-Facility-level analysis possible - Type of unit (for 12 unit types) and number of unit beds; no specific mention of personal care beds - A person cannot be linked to a specific unit.	-For certified NHs type of special care unit (for 12 types) and number of beds in the unit -Facility-level analysis possible -Facility data can be linked to person-level MDS data.	-Unit type (e.g., personal care) and number of beds; for special care units (e.g., sub-acute) also captures whether unit has dedicated staff & portion of residents with Medicare/Medicaid as an SOP - Persons can be linked to their specific unit(s) -Facility-level analysis possible
*Facility staffing levels*	Not currently collected	Number of FTEs, by type, that work in the NH (for 13 types)	Number of FTEs that work in the facility, by type (for 13+ types) and employer (facility or contractor)	For two points in time: nursing staff FTEs, by type (RN, LPN, aide) and employer (NH or contract), their wage rates, staff turnover rates
*Facility cost reports*	Not collected, but linkage to HCFA SNF cost report data theoretically possible	Attempted in 1995; due to low response rates not repeated in 1997 or 1999	SNF cost reports available; linkage to persons theoretically possible	Attempted, high levels of non-response; linkage to persons
*Linkage of facility to market characteristics (e.g., county-specific data)*	Not with public use files	Not with public use files	Not with public use files	Could be done in the MEPS Data Center environment

Notes: ER=emergency room; FTE=full time equivalent; HCFA=Health Care Financing Administration; LTC=long term care; MDS=Minimum Data Set; MEPS=Medical Expenditure Panel Survey; NH=nursing home; OSCAR=On-Line Survey Certification and Reporting System; PMED=prescribed medicine; SMRF=State Medicaid Research File System; SNF=skilled nursing facility.
¹ This table has been modified from an earlier version in order to reflect updated information presented at the meetings.
² The 1991 frame included the universe of board and care homes.
³ The 1987 institutional expenditure survey also included persons in personal care homes and facilities for persons with mental retardation.
⁴ The 1985 NNHS included community residing next-of-kin as respondents.

* While the MCBS collects data on persons living in the community and in long-term settings, the items described here are for persons while residing in a long-term care facility, and not to data items collected while living in the community, unless otherwise specified.

Return to Document

Table 4. Comparison of Four National Data Sources1

Table 4. Comparison of Four National Data Sources¹