If you want to know the expected future survival rates of old ladies you should look at a life table, which shows that directly.
ONS» has three-year average tables for the whole population, for download from
this page.
The latest table is for 2018-2020, and gives numbers per 100,00 live births. This is the bottom right-hand bit of the table (women now aged 95 -99; there are no annual data for age 100 and over):
age | mx | qx | lx | dx | ex |
95 | 0.260177 | 0.230227 | 12458.8 | 2868.4 | 3.06 |
96 | 0.289863 | 0.253171 | 9590.5 | 2428.0 | 2.83 |
97 | 0.322798 | 0.277939 | 7162.4 | 1990.7 | 2.62 |
98 | 0.352456 | 0.299649 | 5171.7 | 1549.7 | 2.43 |
99 | 0.380825 | 0.319910 | 3622.0 | 1158.7 | 2.26 |
The columns show (using standard life table notation) probabilities as number, and population numbers per 100,000:
age exact years at the start of each year of age
mx central mortality rate - number of deaths from 'age' to 'age'+1 divided by the average population during that year
qx mortality rate - number of deaths from 'age' to 'age'+1 divided by the population at 'age'
lx survivors - how many of 100,000 live births would be still alive at exact 'age' based on qx in this table
dx is the difference in lx: the number of the original 100,000 dying at 'age' to 'age'+1
ex is the period life expectancy - how many years of life there are on average at 'age'
These are worked out for each year's population data and then averaged over the three years 2018-2020.
These are population figures - based on all the people alive at the same time born over a long period, not cohort figures - based on those born at the same time, reaching any given age in different years.
So for 100 97-year old women, of average health for the country as a whole, 28 will not reach age 98. They will on average live 2.62 years, and have a less than evens chance of reaching 100.