Pregnancy and Chronic Disease: The Effect of a Midwife-coordinated Maternity Care Intervention

Pregnant women with chronic disease such as epilepsy, sclerosis, hypertension, and thyroid disorders have a significant increased risk of pregnancy complications such as spontaneous abortion, preterm delivery, and caesarian section. Children born by mothers affected by chronic disease have a higher risk of low birthweight, prematurity and birth defects. Moreover women affected by chronic disease have a higher incidence of post-natal depression than women with no history of chronic disease.

Previous studies have found that pregnant some women with chronic disease experienced a lack of continuity in the maternity care: they felt that their different care providers were giving contradictory information on issues such as mode of delivery, breastfeeding and medications. This lack of continuity led to dissatisfaction with care and insecurity with decision-making.

Previous studies that evaluate maternity care interventions, have shown that the elements of the interventions that have a positive effect on pregnancy outcomes such as lowering pregnancy and birth complications, health costs and patient satisfaction are: increased care-coordination, continuity of care providers, specially trained care providers (e.g. nurses and midwives), specialized teams, focus on transitions from e.g. one care setting to another and individualized women-centered care.

However, none of these studies have been targeted pregnant women with chronic disease and how to optimize their course of maternity care. Moreover, systematic reviews point out that studies that examine the effect of the interventions on patient satisfaction, health costs and psychological well-being.

On this background, the purpose of this study is to examine the effect of a midwife-coordinated individualized and specialized maternity care intervention for pregnant women with chronic disease on the length of hospitalization (during pregnancy and after delivery). Secondarily, the purpose is to examine the effect of the intervention on psychological well-being and patient satisfaction.

Patient enrollment The patients will be recruited from the Department of Obstetrics, Rigshospitalet, Denmark from October 2018 to February 2020. Participants will be selected on the basis of 1) the journal information provided from the women's general practitioner and 2) an electronic questionnaire that all pregnant women at Rigshospitalet are sent by e-mail and more than 90% return before 10 weeks of gestation. The information from the questionnaire is routinely transferred to the women's medical journals. Potential participants receive patient information about the study at around week 9-17 of gestation depending on how early the hospital, receive the information about the pregnancy.

The first contact to the patient will be by phone between 12-18 weeks of gestation as a part of the standard care when a midwife from the visitation will call the woman with information about shedueled appointments and programme for high-risk pregnancies at the department. The midwife will ask if information about the study can be sent til the woman and if a research midwife is allowed to follow up on this information with a phone call within a week. If the woman agree, a research midwife will phone the woman within a week and ask if she wants more information about the study and if she wants to participate. If so the woman is given the information and offered the choice between a personal meeting or information over the phone about the study and participation. If the woman agerrs to participate informed concent is obtained, and randomization is performed by a specially designed computer programme.

Data collection. All participants complete an electronic baseline questionnaire after informed consent is obtained and before randomization. All data from the electronically distributed questionnaires are stored in a secured database.

Questionnaire number two (33-37 weeks of gestation) and questionnaire three (two months after delivery) will be sent to the participants by e-mail and the returned questionnaires will be stored in a secured database.

Other data will be obtained from patient records.

Sample size calculation

A total of 258 women will be enrolled in the study. The sample size was determined using unpublished data on 426 pregnant women with chronic diseases hospitalized at Rigshospitalet during pregnancy and in the post-partum period in 2017. The average length of hospital stay (LOS) during pregnancy was 3.9 days; however, due to a right-skewed distribution, log-transformed LOS was used as the outcome for the sample size calculation. Therefore, the effect of the intervention is expressed as a ratio. We expect a reduction of 25% in LOS in the intervention group compared to the control group. This corresponds approximately to a mean reduction of 1 day of hospitalization and to an improvement of log (1-0.25) = -0.29 on the log scale, which we consider a realistic and clinically significant effect size. The SD of number of days in hospital on log scale was 0.80. With a difference of 0.29, SD = 0.80, a power of 0.8 and a significance level of 0.05, an analysis by the two-sample t-test requires 123 women in each group. We expect a maximum of 5% to be lost to follow-up, leading to a total of 129 women needing to be enrolled in each group.

Analysis plan

Analyses are to be performed based on the intention-to-treat principle. The log-transformed number of days in hospital in the two groups will be compared in the primary unadjusted analysis and in a secondary analysis, adjustments will be made for parity and age. Analysis will be performed using the general linear model unless we find a generalized linear model (e.g. Poisson or negative binomial) to fit the distribution of the number of days in hospital. Furthermore, the two groups will be compared with respect to the proportion of women with three or fewer days in hospital. Comparison of secondary and exploratory outcomes will similarly be performed using the general or generalized linear models as appropriate. In case of an uneven allocation of pregnancies with congenital malformations or chromosomal abnormalities to the two groups, analyses will be performed in the sub group of women without pregnancies with congenital malformations or chromosomal abnormalities.