Four years of Dobbs data: one prediction in four survived contact with the numbers

Score the forecasts and one lands cleanly. The physician writing in STAT reports a relative increase of 5.6% in infant mortality in heavily restricted states, citing one study ^[3]. The other three predictions he held in 2022 do not appear in the aggregate data ^[1].

The asymmetry is not mysterious. Infant deaths in this setting sit at the end of a short causal chain: states banned abortion for congenital or genetic anomalies including those incompatible with survival, so more infants are born with those conditions and more die, and patients with high-risk conditions are being carried to term ^[4]. Maternal deaths are rarer events measured against a noisier background. The bans arrived alongside the tail of Covid-19 mortality, which complicates any attempt to separate overlapping effects ^[6]. Ban states already had worse maternal mortality before Dobbs, and those gaps have on average held steady since ^[7]. A null result assembled from those ingredients is not evidence that nothing happened, which is roughly the author's point when he notes that a stable national figure can conceal deterioration inside restrictive states or among minoritized groups ^[8], and that patients have died as direct results of bans regardless of what the national line does ^[9].

The abortion count is the more interesting failure, because the reason it failed is itself the story. Law is one input among economics, contraception and social change ^[12], and the legal change was met by activists, advocates and clinicians working to preserve access, often through telehealth and sometimes at personal legal risk ^[13]. The national number, in other words, is partly a measurement of the countermeasure rather than of the restriction. That also tells you how durable it is: much of the preserved access depends on mifepristone, which faces legal peril ^[14]. A count held up by one drug has one place to break. And national averages say nothing about the people in restrictive states who still are not reaching care ^[15].

On physicians, the estimates the author cites put net movement of OB-GYNs from restrictive to unrestrictive places at zero, or 1% to 2% ^[16]. Net is a difference between two flows, and he is careful to separate the individual doctors who did relocate from the nationwide exodus that was expected and did not arrive ^[17].

What the piece does with all this is refuse the obvious trade. The author's conclusion is that the failure of the worst predictions is not a reason for anyone to revise their view of abortion restrictions, if the guiding principles are autonomy and freedom ^[18]. He adds that in the country with the worst maternal and pregnancy-associated mortality among high income nations ^[10], an unchanged rate is a tragedy and not a case for celebrating stability ^[19]. That is a bill coming due for outcome-based advocacy on both sides. Arguments staked on forecasts inherit whatever the data eventually say, and four years is long enough for some of it to arrive ^[2].