Both intervention and control groups received regular antenatal care, usually consisting of about 9 visits with the midwife throughout the pregnancy. Regular antenatal care is a midwife support during pregnancy and ends with the post partum visit 2–3 months after childbirth. Apart from laboratory tests and measuring weight, the woman's total health is taken into account. The midwife can give life style advice and when needed refer the woman to psychologist, social worker, physiotherapist or dietitian. The intervention in the project consisted of 60 minutes of extra time with the midwife, including two extra 30-minute appointments early in pregnancy. During the rest of the pregnancy, about 5 minutes of each 30-minute appointment were dedicated to lifestyle follow-up. The women's weight was checked at every appointment. The control women's weight was checked accord- ing to the regular antenatal programme, i.e. at enrolment, at weeks 25 and 37 and at the postnatal check-up which was the natural endpoint as the study was carried out at PC level.
The women in the intervention group were offered individua- lised dietary advice, a prescription for physical activity, walking poles and pedometers, participation in food discussion groups with a dietitian and information about community health centres offering lifestyle education and lighter exercise.
The dietary advice was based on the Swedish Nutritional Recom- mendations and adjusted to the needs of obese pregnant women. The food groups were offered a combination of lectures and interactive discussions on nutritional and dietary challenges and weight pro- blems, in three 90-minute sessions. The women who preferred not to participate in group discussions had the possibility to see the dietitian
individually. The midwives and the women agreed on the type, frequency and duration of the planned physical activity. Examples of activities were walking, swimming, aqua-aerobics, gym sessions and activities that could be performed at home, i.e. dancing, playing with their children and repeatedly walking up- and downstairs.
After including a woman in the intervention group, the midwife received a kit with material, and more information and leaflets could be found on the AHC homepage. A log book was introduced and the woman and the midwife mapped current diet and activity status. The woman was prescribed physical activity and given dietary advice. The log was used throughout the pregnancy to register weight, changes in activity and food, use of walking poles and pedometers, as well as participation in food discussion groups, exercise classes or other matters of interest. At the postnatal check-up, the post partum weight was inserted and the log was completed and returned to the project leader.
Before the start of the project, the midwives who would be engaged with the intervention women were given education about obesity and nutrition and physical activity during pregnancy. They were also trained in Motivational Interviewing (MI) (Emmons and Rollnick, 2001) and learned how to use the log book. A network with the surrounding community was formed and HC providers and doulas (coaching the woman during pregnancy and labour) were contacted to find areas for interaction and support. Cooperation was initiated with community health centres that organised walks and exercise classes especially created for pregnant women, and some municipal sports and recreation facilities offered price reductions at fitness centres and public swimming pools.
Statistical analysis
The descriptive results in the tables are presented as numbers, means (7 standard deviation (SD)) and medians (quartiles). Para- metric statistics were applied (Student's t-test) for comparisons between the intervention and control groups, in which numeric data were normally distributed. Where numeric data were not normally distributed, non-parametric statistics were applied (Mann–Whitney U-test, Wilcoxon). For categorical data, Pearson's χ2 test and Fisher's exact test were applied.
Multiple regression analysis was used to account for confounders in analyses on the effect of the intervention on the primary outcome of weight gain. The significance level was set at 0.05. IBM SPSS Statistics 19 and Microsoft Excel 2010 were used for analysing the data.
Findings
The approach rate, i.e. proportion invited to participate, for the project was 65% and the consent rate was 62%. Descriptive data concerning the participating women's baseline characteristics