What a birth certificate can and cannot tell you
If you plan a home birth, transfer in labor and deliver in hospital, the national data counts you as a hospital birth. That single fact bends the American numbers.
Almost every American number in the previous topic comes from birth certificates. You should know what a birth certificate can and cannot record, because a reader who finds this out somewhere else will stop trusting everything she was told.
Three different things get confused constantly, including by people arguing on both sides.
| The variable | What it means | What the US certificate does |
|---|---|---|
| Planned place of birth | Where you intended to be when labor started | Asked only of women who delivered outside a hospital |
| Actual place of birth | Where the baby was actually born | Recorded directly |
| Attendant | Who was there: physician, nurse-midwife, other midwife | Recorded, with known inconsistency in the other midwife category |
Here is the consequence, and it is the central problem in this whole field. A woman who plans a home birth, transfers to hospital in labor, and gives birth there is counted in the national data as a hospital birth. Transferred women are exactly the women in whom problems concentrate. Their outcomes are therefore credited to the hospital group. This bias runs in favor of planned home birth, and it makes the planned home birth group look cleaner than it is.
ACOG lists the same problem among others, in its own words: Observational studies of planned home birth often are limited by methodological problems, including small sample sizes; lack of an appropriate control group; reliance on birth certificate data with inherent ascertainment problems; reliance on voluntary submission of data or self-reporting; limited ability to distinguish accurately between planned and unplanned home births; variation in the skill, training, and certification of the birth attendant; and an inability to account for and accurately attribute adverse outcomes associated with antepartum or intrapartum transfers.
One American state fixed it and showed how much it mattered. Oregon revised its birth certificate in 2012 to ask every woman delivering in hospital: Did you go into labor planning to deliver at home or at a freestanding birthing center?
That made it possible to put transferred women back where they belonged. Among 79,727 Oregon births in 2012 and 2013:
| Outcome | Before counting transfers properly | After counting transfers properly |
|---|---|---|
| Stillbirth, out of hospital vs in hospital | 0.6 vs 1.3 per 1,000 | 2.4 vs 1.2 per 1,000 |
| Perinatal death | no significant difference | 3.9 vs 1.8 per 1,000 |
| Death in the first month | no significant difference | 1.6 vs 0.6 per 1,000 |
The adjusted difference for perinatal death was 1.52 more per 1,000 (95% confidence interval 0.51 to 2.54); for death in the first month it was 0.63 more per 1,000 (0.03 to 1.24). Admission to newborn intensive care was lower after a planned out-of-hospital birth in the same study; the published result for that outcome is a ratio without the underlying rates, so no figure for it is printed here.
The authors' own one-sentence summary is the fairest single statement in this literature, and both halves of it belong to you: Perinatal mortality was higher with planned out-of-hospital birth than with planned in-hospital birth, but the absolute risk of death was low in both settings.
The women differ too. Between 2004 and 2014, women giving birth outside a hospital were less likely to be obese before pregnancy (12.5% against 25.0%), less likely to smoke (2.8% against 8.5%), and more likely to have graduated college (39.3% against 30.0%). Every one of those differences pushes the out-of-hospital numbers in the direction that makes home birth look better than a like-for-like comparison would.
And there is no randomised trial to settle any of it. The Cochrane review on this question found one trial with eleven participants, and reported that five of its seven main outcomes were not reported at all. It is not evidence that home birth is safe, and it is not evidence that it is unsafe. It is evidence that the study has never been done. ACOG says the same: To date, there have been no adequate randomized clinical trials of planned home birth.
"Is that number about where I planned to give birth, or where I ended up?"
"Was that a planned home birth group, or does it include unplanned ones?"