What are the odds you’ll need assistance to help with personal activities such as bathing or dressing before you die? For those about to turn 65, it’s about 50/50, according to an important new study. On average, you can expect to need this high level of care for about two years. But one-in-five Americans will need such assistance for less than one year while about one in seven will need extensive help with daily living for five years or more.

On average, this high level of long-term supports and services will cost about $140,000 (in 2015 dollars) from age 65 until you die. To think of it another way, if a 65-year-old wanted to have enough money to pay for this care for the rest of her life, she’d have to invest about $70,000.

But among those who use paid care (excluding those who use none), the average lifetime cost is more than $265,000. To afford that, you’d need to put aside about $134,000 at 65.

Those averages can be misleading. It turns out there are big differences by gender: Women are more likely to need extensive care, and for a longer period of time, than men. Similarly, low-income people will typically need high levels of care for nearly three years while those with high incomes will need such assistance for about a year-and-a-half.

Overall, about 6.3 million Americans over 65 need this extensive level of care. But by mid-century their numbers will more than double—to 15.3 million. Why? Mostly because the older you are, the more likely it is you will need very high levels of care. And the population of those 85+ is growing rapidly.

This important new analysis comes from a team of my Urban Institute colleagues led by Melissa Favreault. It’s described in a new paper by Melissa and Judith Dey of the Office of the Assistant Secretary for Planning and Evaluation (ASPE) of the federal department of Health and Human Services and funded by ASPE.

The study found that on average, people who need extensive long-term care pay for about half of their costs out-of-pocket. Medicaid pays for more than a third while private insurance funds less than 3 percent.

While this study is one of the most detailed ever done on long-term care needs and costs, it focuses only on those age 65 and older (because data are very limited on younger people with disabilities). In addition, it looks primarily at paid care and excludes assistance by family and friends which, for many older adults, is the most common form of assistance.

Previous reports estimated that Medicaid pays about 42 percent of all long-term care costs, compared to a little more than one-third in this paper: The reason: Earlier studies excluded non-care costs such as rent for those living in residential care facilities.

Other studies also have reached different conclusions about the likelihood of needing long-term care. But, for the most part, that’s because they are measuring different levels of need and focusing on all assistance, not just paid care.

For instance, a widely-cited 2005 paper by Peter Kemper, Harriet Komisar, and Lisa Alecxih projected that a then-65-year-old had nearly a seven-in-ten chance of needing long-term care. But they looked at those who required a broader-range of long-term supports and services, not just high levels of assistance.

Similarly, a study published last year by Vicki Freedman of the University of Michigan and my Urban Institute colleague Brenda Spillman found that about 18 million seniors need some help with personal tasks—about three times the number in the new research. But they included family care as well as help with activities such as housekeeping, transportation, or balancing a checkbook.

By contrast, Melissa focused on those with more serious functional limitations or severe cognitive impairment (from Alzheimer’s disease or other dementias). This is often described as needing assistance with at least two activities of daily living such as bathing, eating, dressing, or going to the bathroom; and it is usually the level of need at which someone could qualify for long-term care insurance benefits or Medicaid.

This is an enormously important study. It gives consumers a valuable picture of how much help they are likely to need in old age and, thus, how they should protect themselves against long-term care risk. And it gives policymakers a key baseline against which to measure new ideas for financing long-term supports and services.