On the occasion of Barbara Bush's death (from COPD) & President Bush's hospitalization (for sepsis): some detailed thoughts re the WIDOWHOOD EFFECT ("dying of a broken heart"), based on analyses using national samples of ~400,000 elderly couples and detailed health records. 1/
We did detailed analyses of cause-specific effects. Having your wife die of COPD raises your own risk of death by about 30%, possibly more than any other cause of death. (Other disease and gender pairings in the paper.) @AMJPublicHealth ajph.aphapublications.org/doi/abs/10.210… 2/ 
And, for the surviving husband, the risk of death from sepsis is especially high, compared to other things that might kill you -- about 20% higher. (Other disease and gender pairings in the paper.) 3/
In other work, we showed that the risk of dying of a broken heart is modified by the prevalence of other widow/ers in your neighborhood. Possibly, if you can share your grief with others who have faced the same fate, your own risk may be lower. ncbi.nlm.nih.gov/pubmed/18178300 4/
Also, you are much more likely to die if your *wife* dies than if your *ex-wife* dies. link.springer.com/article/10.135… This supports a causal effect of widowhood. (This is the part of my 'splainer that will amuse @popehat the most; it feels good to repay all his legal 'splainers.) 5/
We also find that white couples have a much higher widowhood effect than black couples, for complex reasons. Effects in inter-racial marriages are fascinating (if Barbara Bush had been black, George would not have have higher risk). journals.sagepub.com/doi/abs/10.117… via @asr_journal 6/
Almost instantly after a man's wife dies, his risk of hospitalization spikes (and other health quality indicators also show problems). This is illustrated by the case with President Bush at the moment. link.springer.com/article/10.135… 7/
Widowed people get lesser quality of care in another study of ours. But this is not likely to be an issue for President Bush: sciencedirect.com/science/articl… 8/
What can be done? Well, hospice care can attenuate the widowhood effect. Taking better care of the dying may reduce the risk of death (and dying of a broken heart) of the surviving spouse. This is a classic network effect. ncbi.nlm.nih.gov/pubmed/12791489 9/
Finally, it's not just widowhood that takes its toll, it's caregiving to your spouse even before they die, as we showed in this 2006 paper in @NEJM nejm.org/doi/full/10.10… (which also prompted my interest in network effects!) 10/
My thanks to all my co-authors in this sustained decade of work on widowhood, only some of whom are on twitter: @iwashyna, @jeremyfreese, Paul Allison (@StatHorizons), Felix Elwert, Lei Jin, Subu Subramanian 11/
And how could I forget: We found that widowhood was associated with an increased risk of sexually transmitted diseases for men (but not women). Largest effects found 6-12 months after a wife's death. And this was especially so after release of Viagra. ajph.aphapublications.org/doi/abs/10.210… 12/
