NIOSHTIC-2 Publications Search
A comparison of two methods for calculating expected mortality.
Steenland-K; Beaumont-J; Schulte-P; Hornung-R; Rinsky-R; Clayton-JF; Price-WH; Anderson-JR; Anderson-KM; Hartz-A; Hoffmann-RG
Stat Med 1985 Jan; 4(1):105-109
A collection of three letters to the editor, and a reply to them, were presented. The appropriateness of the conclusion that the traditional person/years (PY) method used in occupational mortality studies was not justified was challenged. It was argued that there was no bias in the PY method versus the prospective method (PM), called the actuarial method, since each method estimated a different parameter, one based on person/time, the other on cumulative risks. The authors conclude that the PY method remained an appropriate method for long term occupational mortality estimations. A second letter challenged the rejection of the PY method of expected death calculations. They justified the PY method basing their logic on the PM model itself, suggested the term exposure to death be preferred over expected death, and stated that the PY method had the added advantage of simplicity. The third letter stated that the PM method was at best only nearly as efficient as the PY method, and even became quite poor when follow up became complex. The authors reply asserted that the PM method can be used to determine expected mortality while the PY method can only estimate it. Added uses are in deriving baseline mortality and in study design. The PY method continues to be a useful technique, but the PM method may provide useful information in some studies.
NIOSH-Author; Long-term-study; Mathematical-models; Mortality-data; Occupational-safety-programs; Risk-analysis; Statistical-analysis; Epidemiology
Issue of Publication
Statistics in Medicine
Page last reviewed: March 11, 2019
Content source: National Institute for Occupational Safety and Health Education and Information Division