Use of medicare services before and after introduction of the prospective payment system

Health Services Research, August, 1993 by Kenneth G. Manton, Max A. Woodbury, James C. Vertrees, Eric Stallard

In 1984, 25,401 individuals were selected by subsampling 12,100 (45.45 percent) of 26,623 persons not eligible for an interview in 1982, because a large portion of the 6,393 1,992 = 8,385 persons disabled or institutionalized in 1982 were expected to be alive in 1984 and interviewed. A total of 4,916 persons passing age 65 between surveys were sampled, so the entire 1984 U.S. population 65 is represented. Of 4,864 surviving to the 1984 screen date 4,759 (97.8 percent) responded. Of 20,485 persons selected from the 1982 sample for screening in 1984, 20,219 were alive; and of 19,889 (98.4 percent) screened in 1982, 17,484 survived to 1984 and 17,117 (97.9 percent) were screened (80 percent by phone and 20 percent in person). The 1982-1984 combined sample includes all 6,393 chronically disabled community residents and 1,992 institutional residents in 1982: these were persons eligible to be interviewed in 1984, regardless of current disability status.

Of 22,674 persons alive on April 1, 1984 (2,646 died between April 1, 1982 and April 1, 1984), 22,348 were alive at the 1984 screen date, and 21,876 (97.9 percent) were screened. For the interview, 6,264 were determined eligible from screening the "age-in" sample (N = 4,864), the 1982 nondisabled/nonrespondent surviving subsample (N = 11,151), and survivors of the 1982 disabled (5,107) and institutional populations (1,226). Of this 6,264 total, 96 died between the screen and the interview attempt. Of the 6,168 alive, 5,934 (96.2 percent) responded. Four hundred eighty-seven persons disabled in 1982 were not disabled in 1984; of the 1,836 institutional residents identified at the screen, 1,778 were alive at the time of the institutional interview, and 1,728 (97.2 percent) of them responded. Of 7,946 persons eligible for a community (6,168) or institutional (1,778) interview, 7,662 (96.4 percent) responded. The two-stage response was 94.4 percent, excluding deaths.

Since NLTCS samples are drawn from Medicare files, the vital and residence statuses of all 25,401 persons were traced and Medicare Part A use was recorded. Consequently, nationally representative Medicare and mortality experience was available for periods before, and after, the introduction of PPS. Because the NLTCS provides health data on all chronically disabled respondents, mortality and health changes for persons not using hospitals (nor any Medicare service) can be determined.

Medicare Part A

Medicare Part A data are bills and bill summaries from hospitals, HHAs, and SNFs for care covered by Medicare for all persons in the NLTCS (whether interviewed or not) from October 1, 1982 through September 30, 1986. Since nursing homes are also reimbursed privately or by state Medicaid, along with some home health care, Medicare is a partial, but important, component of post-acute care use.

In 1982, anticipating PPS, the Health Care Financing Administration (HCFA) required coding diagnoses, procedures, age, sex, charges, LOS, and admission and discharge dates for 100 percent of hospital bills. Since data quality improved as PPS phased in, the average patient's health can appear to have changed (i.e., changes in DRG coding make hospital case mix appear to increase in severity) while reduced hospital admission rates actually increased the average severity. Thus, independent health data, comparably measured before and after PPS, are needed. Since health in the 1982 and 1984 NLTCS is consistently measured (Manton, Stallard, and Woodbury 1991), and vital status is ascertained for all persons, independent measures of health are available for pre- and post-PPS comparisons.