Scope and Methodology

Our work was done primarily at HCFA headquarters. We also visited three
regional home health intermediaries (in Chicago, Illinois; Milwaukee,
Wisconsin; and Clearwater, Florida) and two HCFA regional offices
(Chicago and Atlanta) to obtain workload and performance data,
information concerning RHHI claims review operations, and an update on
HCFA's implemented and planned program changes. We also interviewed
officials at HHS' Office of the Inspector General in Baltimore and Atlanta.

We reviewed pertinent laws, regulations, court decisions, and HCFA policies to identify changes in eligibility determination and medical review practices. And we reviewed studies related to home health benefit utilization and control issues.

To identify home health growth patterns and variations in utilization, we analyzed data from Medicare's Provider of Service and Home Health National Claims History files. These data include information on all paid claims for the period 1989 through 1993.54 We used data from the Provider of Service file to determine agency growth through time and across geographic regions and to identify provider ownership type. And we used the Medicare claims data to calculate mean and median home health visits, by total and by each type of service, broken out by geographic area55 and HHA ownership types.

While the average visits per year provides a general indication of
variations in utilization of home health services, it does not indicate the
length of each individual's episode of care nor does it provide a picture of
the intensity of services provided during this time. To obtain a more
in-depth look at variations in practice patterns, both across regions and
among various types of HHAS, we conducted an episode-of-care analysis for
four diagnoses: diabetes, heart failure, hypertension, and hip fracture. 56
The first three diagnoses were selected because they are among the most
common primary diagnoses associated with home health care.57 Hip
fracture was selected because it is generally regarded as having a more
predictable pattern of treatment with a more finite end point. We selected
beneficiaries with one of the above primary diagnoses who began

54Most recent data available.

55 Geographic breakdowns were based on the state of beneficiary residence.

56 Because home health episodes are not clearly defined by admission and discharge dates, we defined episodes of care as a series of home health visits preceded and followed by a 60-day period with no visits.

57Approximately 22 percent of Medicare home health patients have one of these four primary
diagnoses.

data systems, which are subject to periodic HCFA revie examinations. HCFA relies on the data obtained from th evidence of Medicare-covered services and expenditur management and budgetary decisions. For this reason independently examine the internal and automatic dat controls for automated systems from which we obtain analyses. With this exception, we conducted our work generally accepted government auditing standards bet December 1995.

58 Because some episodes of care continued indefinitely, that is, without a visits, we selected a cutoff point that allowed us to analyze utilization of point of 210 days allowed us to look at full episodes for 99.22 percent of h 94.60 percent of heart failure beneficiaries, 94.59 percent of hypertension 91.54 percent of diabetes beneficiaries.