It's widely believed that when one spouse dies the other tends to die sooner than he/she would otherwise. While this is plausible, I have a hunch it's one of those beliefs--like the disproportionate number of babies supposedly born nine months after a blackout, or toilet flushing during Super Bowl halftime emptying a city's water supply--that's nifty but untrue. So tell me, if (say) a 70-year-old man dies, is his wife likely to die sooner than she would if the old guy lived? --Steven Goldberg, via e-mail Hold on, ace. While Super Bowl bathroom breaks have never been shown to play havoc with a town's H2O, the "super flush effect" is real: as we've previously established, New York City occasionally experiences water-supply surges due to pressure drops during or after popular telecasts, most notably the two-and-a-half-hour M*A*S*H finale in 1983. Likewise, dying of grief isn't merely romantic legend but a well-documented phenomenon, verified in a range of studies over the last 20 or 30 years in several countries. For this you can thank actuaries, who'll stop at nothing to "substitute facts for appearances and demonstrations for impressions," a pithy line from John Ruskin that the Society of Actuaries has adopted as its motto, and a fair description of our theory of operation at the Straight Dope.
Analyzing spouse deaths is a straightforward exercise in data mining, the data in this case usually being the mortality info collected by insurance companies. Most studies compare the death rate for a group of widows/widowers to that of a control group. Generally, the ratio of the two death rates is around 1.4--that is, all else being equal, 140 widowed persons die compared to 100 nonwidowed.
A different approach used by Heekyung Youn and Arkady Shemyakin analyzed data over a five-year period from joint and last-survivor annuities (where payments are made so long at least one partner is still alive). If the deaths of spouses are independent, the probability of both dying should be basically the male death rate for that demographic times the female death rate. In fact, it's not: the incidence of both spouses dying is roughly four times what you'd expect.
In short, while the magnitude of the risk varies with the methodology, there's ample evidence that widows/widowers exhibit what actuaries call "excess mortality." Death rates are highest within six months after bereavement and continue high for about three years. Bereavement produces higher mortality among widowers than widows, and among younger ages than older (although the age-based differences aren't always significant). Socioeconomic status doesn't seem to matter.
Most experts agree there are three main reasons for this phenomenon, although the statistics don't tell you which is most important:
aDeath-related events affecting both spouses at the same time, such as both being in the same accident. Duh, but you can't overlook the obvious.
aCommon risk environment--for example, both spouses live in the same neighborhood subject to the same environmental factors such as pollution. Similarly, both spouses may engage in common health-related behaviors such as smoking or drinking.
aDirect dependence, where the death of one spouse changes the other spouse's life, also called "broken-heart" syndrome. This is the kind we're mainly interested in, but it can be tough to distinguish from the first two even if you've got the resources for a case-by-case analysis.
Suicide rates are significantly higher for the recently widowed--this was demonstrated by pioneering sociologist Emile Durkheim in a classic 1897 monograph. Suicide is most prevalent during the first week after the death of a spouse. There are also higher rates of death due to heart disease and conditions such as cirrhosis.
The mortality risk is higher for widows/widowers under age 55 when the bereavement was sudden, as from accident. For widows/widowers in the 50-to-65-year range, the highest risk is when the spouse has died of chronic illness. Several studies indicate the risk is lower when the widow/widower has family or other social support and higher when the bereaved is socially isolated.
Why are the bereaved at higher risk? The two broad explanations are grief and stress. Grief can lead to depression and its consequences. Stress can impair the immune system, increasing susceptibility to disease, and trigger dangerous behavioral changes such as higher cigarette and alcohol use. Most data is based on married couples because it's easier to obtain, but there are indications of a similar mortality increase following the death of children, siblings, and parents. Sure, grief is natural and proper, but it's healthier for the survivors to move on.
Art accompanying story in printed newspaper (not available in this archive): illustration/Slug Signorino.