Death waits for no one: deferred demises take a couple of hits.

Two new reports challenge the idea, which has been promoted in a series of high-profile studies, that elderly people suffering from serious physical illnesses can prolong their lives just long enough to experience a personally meaningful event, such as a birthday or a religious holiday.

An analysis of California death records from 1985 through 2000, conducted by economist Gary Smith of Pomona College in Claremont, Calif., indicates that elderly Asian immigrants don't put off dying until the week after the Harvest Moon Festival, a major annual event for them. That result counters a 1990 study, based on California data from 1960 through 1984. The earlier investigation found that mortality rates of Chinese women at least 75 years of age dipped in the week before the Harvest Moon Festival and rose in the week after.

Smith's data analysis reveals no sign of death postponement before the Harvest Moon Festival for Chinese-, Korean-, and Vietnamese-Americans. This result held, regardless of whether he defined elderly as being a minimum of 65 years old or 75 years old. It also made no difference whether deaths on the day of the festival were classified as occurring before or after the event.

Moreover, Smith found that the original data from 1960 to 1984 exhibit a death-postponement pattern only if deaths on the festival day are classified as having occurred after the festival. That statistical partition makes no sense, he argues, because the festival's central ritual--a family meal--takes place at midnight at the end of the holiday.

Other prior investigations of this alleged delayed-death effect are also suspect, contend Judith A. Skala and Kenneth E. Freedland, both of Washington University School of Medicine in St. Louis. They reviewed 18 such studies published between 1973 and 2001.

For example, a 1987 report found a 20 percent rise in deaths shortly after Christmas in Ohio but no corresponding decline in deaths before Christmas. Reanalysis of the data indicated that the surge in deaths actually began 5 days before the holiday and peaked on Christmas Day, the researchers say.

"Research ... has failed to provide convincing evidence that psychological phenomena such as 'giving up' or 'holding on' can influence the timing of death" Skala and Freedland conclude.

Smith's findings and those of Skala and Freeland appear in the May/June Psychosomatic Medicine.

In a commentary published with the new reports, Ellen L. Idler and Stanislav Kasl, both of Rutgers University in New Brunswick, N.J., argue that there is still reason to suspect that deaths occur less frequently before major religious holidays than after them. In a 1992 study of elderly residents of New Haven, Conn., Idler and Kasl reported a death-postponement pattern for observant Jews around Yom Kippur and Passover and for observant Christians around Christmas and Easter.