Analysis
Statistical analysis
Responses to the surveys were analysed using SPSS versions
16.0,18 and 19. Where women took part in both antenatal and
postnatal interviews, their responses were linked and the analyses
reported here are based on these linked datasets. The linked data
records were compared with the full set of antenatal interviews to
check for response bias.
Cross-tabulations were used to explore relationships between
women's expectations, experiences, the care they received and
their degree of satisfaction with aspects of their care. Statistical
tests were carried out using CIA Confidence Interval Analysis and
Open Epi Version 3.01. χ2 tests were used for contingency tables
and 95 per cent confidence intervals were calculated for differences between proportions. If adjoining cells contained small
numbers, they were combined for statistical testing.
The key results are based on women interviewed both antenatally and postnatally. Although the aim was to analyse the data on
an ‘intention to treat’ basis, this was modified to some extent. As
recruitment was based on criteria applied in late pregnancy
analyses of antenatal care, and some analyses of care at the on set
of labour and method of childbirth and general views of care were
based on the original choice of place of birth given in the antenatal
interview. Because of selective referral of women's bookings from
the birth centre to the hospital for elective caesarean section or
induction, information about labour care was also tabulated by
planned place of birth at the beginning of labour care and
restricted to women whose labour started spontaneously. Women
who had emergency caesareans were excluded from tabulations
about care at birth and management of the third stage. Women
who transferred from the birth centre to the hospital in labour
were asked about the care they received in both places. Each
individual table specifies which women were included.