Skip to Content
Interactive Textbook on Clinical Symptom Research Logo


Home Button

Administrative Data and Hospice Care
Author Bios
Introduction
Health Insurance Data
Basis for Payment Data
Hospice Claims Data
The Medicare Model
Claims Data Uses
Hospice & Palliative Care
Statistical Challenges
Correct Denominators
Starting the Clock
Currently selected section: Costs of EOL Care
Conclusions


Chapter 18: Using Adminstrative Data to Study Hospice Care: Studying Costs of Care at the End of Life
         

Administrative data include information on charges and reimbursements (See page 3 to review these terms). They do not include information about non-reimbursed expenses to the patient or his/her family with the exception of co-payments and deductibles. Aggregating reimbursements can be problematic for a variety of reasons. As with survival, deciding when to "start the clock" for aggregating expenses can be tricky.

Pure administrative data can be supplemented with other data to increase the range of inference that can be drawn. These include:

  1. Cause of death information as coded on death certificates. This can be obtained from the National Death Index (NDI) through the National Center for Health Statistics (NCHS), or from State health departments, and is included in the Surveillance Epidemiology and End Results (SEER)-Medicare linked tumor registry/claims data.
  2. Location of death. Location of death can be estimated in terms of in-hospital/not in hospital from Medicare claims by looking at vital status at discharge. Most states' death certificates contain a field noting location of death. From claims, researchers can discern persons receiving hospice care at death, but knowing whether someone died while in hospice does not provide information about their physical location because hospice is a service not a location. Thus, hospice users may die at home, in a nursing home, in a hospital, or in an inpatient hospice center. Medicare claims do not allow researchers to distinguish among these.

  3. Date of diagnosis is useful for studies of survival with a disease. The first date a diagnosis code appeared or from SEER-Medicare linked data is a reasonable estimate for many purposes. However, that some diagnoses may occur prior to Medicare eligibility (leading to an under-estimate of survival time), SEER codes the diagnosis date as the date the condition was recognized or could have been recognized.
  4. For cancer patients, stage at diagnosis is sometimes important, in combination with time from diagnosis to death. SEER-Medicare data are the best source of this information. There is no good way to determine stage at diagnosis for cancer patients (Cooper, 1999) or severity of illness for non-cancer patients using claims data alone.
Page 12 of 13
Previous Section