To investigate the factors determining the choice of maternity unit in
an outer London district where half of all births occur outside the d
istrict, a self-administered questionnaire was distributed by communit
y midwives at postnatal home visits. The main outcome Measures were: (
1) proportion of mothers able to deliver in unit of first choice; (2)
proportion of mothers making a personal choice rather than taking advi
ce from their GP; (3) reasons for not choosing district unit, by parit
y. It was found that 28/166 (17 per cent) of mothers were unable to de
liver in their first choice of unit, and 3/166 (2 per cent) were unabl
e to obtain their second choice. Of the women who did obtain their fir
st choice, 84/136 (62 per cent) had made a personal decision rather th
an being advised or told by their GP; 58 of these 84 personal deciders
went outside the district, but for 36/58 (62 per cent) this was eithe
r because the other unit was nearer their home or because they had del
ivered there before. It is concluded that, although GPs are theoretica
lly free to refer to any maternity unit, a sizeable minority of women
are unable to realize their own choices. GPs seem to regulate flows by
advising women without strong personal views to attend the local unit
. Differences in the quality of care as perceived either by women or t
heir GPs do not seem particularly important in determining choices. Th
e introduction of an internal market in maternity care seems unlikely
to result in improved quality of care.