Time to Conceive looked at 750 women aged between 30 and 44 who were living with a male partner not known to be infertile, had no diagnosis of infertility of their own, and had recently started trying to become pregnant.
The team took blood and urine samples from these volunteers and measured levels of three hormones often examined by fertility tests.
They then followed each volunteer for a year.
The upshot, published in 2017, was that hormone levels were uncorrelated with pregnancy within the 12-month window through which the researchers were looking.
But perhaps, Dr. Steiner subsequently speculated, that window was too narrow.
In 2020 she therefore got back in touch with the original participants for a follow-up.
She asked them if they would fill out a questionnaire about how many children they had had, how long it had taken to get pregnant and whether they had been diagnosed with infertility.
Some 336 of them agreed to participate.
Among these there had been 239 pregnancies, resulting in 225 live births.
More sadly, 73 participants were infertile.
But the hormone levels in the tests carried out in the original study did not predict these outcomes.
There was no difference, the researchers found, between women with poor results and those with normal ones.
In the case of one substance, for example - anti-Mullerian hormone, which is thought predictive because it is produced by cells in the egg-bearing follicles of the ovaries, and is thus believed to reflect the number of those follicles - 79% of those with low levels of it went on to give birth.
That was statistically indistinguishable from the 71% of mothers with normal levels.
The decline in fertility, says Dr. Steiner, is thus clearly not related to the decline in the quantity of eggs, but presumably to their quality.
And how to measure that remains unknown.