Place of birth and perinatal mortality

Published in MT E-News 1:29, July 16, 1999

Tew M. Place of birth and perinatal mortality. J R coll Gen Pract 1985; 35(277): 390-394.

Using the raw perinatal mortality rates (PMR) from a 1970 British national survey, the hospital PMR was 27.8 per 1000 births versus 5.4 per 1000 for homebirths/general practitioner units (GPU). This was not because hospitals handled more high-risk births. When PMRs were standardized based on age, parity, hypertension/toxemia, prenatal risk prediction score, method of delivery and birth weight, adjusted hospital PMRs for each category ranged from 22.7 per 1000 to 27.8 per 1000 while homebirth/GPU rates ranged from 5.4 per 1000 to 10.5 per 1000.

The 1970 survey assigned a prenatal risk score to predict the likelihood of problems during labor. When PMRs for hospital versus home/GPU for the same level of risk (very low, low, moderate, high, very high) are compared, the hospital PMR was lower only at the very highest risk level. All differences, except in the "very high risk" category, were significant. The PMR for high-risk births in home/GPUs (15.5/1000) was slightly lower than that for low-risk births in the hospital (17.9/1000). Moreover, the PMRs in home/GPUs for very low, low, and moderate risk births were all similar, but hospital PMRs increased twofold between categories, which suggests that hospital labor management actually intensified risks.

The percentage of infants born with breathing difficulties (9.3% versus 3.3%), the death rate associated with breathing difficulties (0.94% versus 0.19%), and the transfer rate to neonatal intensive care units for infants with breathing problems who survived six hours (62.0% versus 26.2%) were all higher in the hospital (all p < 0.001), further evidence that hospital interventions do not avert poor outcomes.

Although no national study has been undertaken since, smaller studies confirm that increasing use of hospital confinement is not the reason for the overall drop in PMR since 1970. In fact, those years when the proportional increase in hospital births was greatest were the years when the PMR declined least and vice versa.

Henci Goer, Obstetric Myths Versus Research Realities, A Guide to the Medical Literature, Bergin & Garvey, 1995

Tew reported 1985 British data that confirmed the older data. At a lower level of risk, PMR was seven times higher in hospital. At a higher level of risk the perinatal death rate was four and a half times higher in hospital. Out of the hospital, "birth was very much safer to mothers in both the lower and higher risk groups." She examined 1986 birth data from the Netherlands, where one-third of births are at home. The country as a whole has excellent maternity care and outcomes. However, PMR for hospital births was six times higher than for homebirths. The Dutch data also allowed birth attendant comparisons. Looking at pregnancies of normal length, PMR was ten times higher for obstetricians than for midwives. For obstetricians "the average risk status [of women] at delivery was not much higher than that of midwives' deliveries. It could not possibly account for a PMR ten times as high for obstetricians as for midwives," decided Tew.

"The British and American experience, now powerfully supported by the Dutch results, tells us convincingly that homebirth and midwives are indeed 'safer than we thought.' Together they offer the safest option. The danger of home as a place of birth does not lie in its threat to the healthy survival of mothers and babies, but in its threat to the healthy survival of obstetricians and obstetric practice."

Diana Korte & Roberta Scaer, A Good Birth, a Safe Birth, Harvard Common Press, 1995

Sites Internet et associations francophones pour une approche « citoyenne » de la naissance