Day 100: Length of Stay and Years of Life Lost

It hasn't been a hundred days of solitude yet; the PlagueBlog count is rather of 100 days of infection in Massachusetts, since our first infected student arrived ex-Wuhan. In today's milestones, the US hit 80,000 deaths, and Massachusetts had only a 1% uptick in cases.

I was inspired by a comment to look into some real numbers for length of stay in nursing homes. Back in 2010, a paper in the Journal of the American Geriatrics Society pegged the median length of stay at 5 months, with 53% dying within six months. There's a long tail that makes the average length of stay more than a year, and a significant sex difference that makes the median only 3 months for men but 8 months for women. Surprisingly, chronic conditions were not a significant factor.

The numbers were even older than the paper, so I looked for some newer data, perhaps reflecting improvements in geriatric care. The closest thing I found was a more recent paper in the Journal of Post-Acute and Long-Term Care, but the numbers were not directly comparable. They considered only medicare patients entering nursing homes for the first time. Even with those restrictions, the median length of institutionalization was still about 5 months. While some patients were ultimately discharged (rather than dying in care or moving between hospitals, short-term care, and back to long-term care), they often died soon after discharge. The mortality rate over the course of the year after first admission was 35%. (Note that the admissions year they used was a fairly average flu season.)

If you compare the reality of nursing home care to some journalism out there about years of life lost (YLL), the numbers don't quite add up. Just for example, a study out of Scotland calculates YLL from the usual sorts of actuarial tables, which indeed indicate another 10 years of expected life for the elderly age groups most susceptible to coronavirus—on average. However, the life expectancy of persons already in nursing homes is somewhere around 14 months from admission, and usually less than that by the time they catch coronavirus. So where the actual YLL lies between 0.6 and 10 depends on how many victims are coming from nursing homes.

Numbers coming out of Europe estimate nursing ("care") home deaths at about 50% of coronavirus deaths overall. In the US, the picture is more complicated, but in leading states like New Jersey, Massachusetts, Connecticut, and Pennsylvania, well over 50% of deaths have occurred among nursing homes patients (plus some staff).