Volume 52 | Number 1 | February 2017

Abstract List

Amy S. Kelley M.D., M.S.H.S., Kenneth E. Covinsky M.D., M.P.H., Rebecca J. Gorges M.A., Karen McKendrick M.A., Evan Bollens‐Lund M.A., R. Sean Morrison M.D., Christine S. Ritchie M.D., M.S.P.H.


To create and test three prospective, increasingly restrictive definitions of serious illness.

Data Sources

Health and Retirement Study, 2000–2012.

Study Design

We evaluated subjects’ 1‐year outcomes from the interview date when they first met each definition: (A) one or more severe medical conditions (Condition) and/or receiving assistance with activities of daily living (Functional Limitation); (B) Condition and/or Functional Limitation hospital admission in the last 12 months and/or residing in a nursing home (Utilization); and (C) Condition Functional Limitation Utilization. Definitions are increasingly restrictive, but not mutually exclusive.

Data Collection

Of 11,577 eligible subjects, 5,297 met definition A; 3,151 definition B; and 1,447 definition C.

Principal Findings

One‐year outcomes were as follows: hospitalization 33 percent (A), 44 percent (B), 47 percent (C); total average Medicare costs $20,566 (A), $26,349 (B), and $30,828 (C); and mortality 13 percent (A), 19 percent (B), 28 percent (C). In comparison, among those meeting no definition, 12 percent had hospitalizations, total Medicare costs averaged $7,789, and 2 percent died.


Prospective identification of older adults with serious illness is feasible using clinically accessible criteria and may be a critical step toward improving health care value. These definitions may aid clinicians and health systems in targeting patients who could benefit from additional services.