Variation in intrapartum referral rates in primary midwifery care in the Netherlands: a discrete choice experiment.

ABSTRACT

OBJECTIVE: in midwife-led care models of maternity care, midwives are responsible for intrapartum referrals to the obstetrician or obstetric unit, in order to give their clients access to secondary obstetric care. This study explores the influence of risk perception, policy on routine labour management, and other midwife related factors on intrapartum referral decisions of Dutch midwives. DESIGN: a questionnaire was used, in which a referral decision was asked in 14 early labour scenarios (Discrete Choice Experiment or DCE). The scenarios varied in woman characteristics (BMI, gestational age, the preferred birth location, adequate support by a partner, language problems and coping) and in clinical labour characteristics (cervical dilatation, estimated head-to-cervix pressure, and descent of the head). SETTING: primary care midwives in the Netherlands. PARTICIPANTS: a systematic random selection of 243 practicing primary care midwives. The response rate was 48 per cent (117/243). MEASUREMENTS the Impact Factor of the characteristics in the DCE was calculated using a conjoint analysis. The number of intrapartum referrals to secondary obstetric care in the 14 scenarios of the DCE was calculated as the individual referral score. Risk perception was assessed by respondents׳ estimates of the probability of eight birth outcomes. The associations between midwives׳ policy on management of physiological labour, personal characteristics, workload in the practice, number of midwives in the practice, and referral score were explored. FINDINGS: the estimated head-to-cervix pressure and descent of the head had the largest impact on referral decisions in the DCE. The median referral score was five (range 0-14). Estimates of probability on birth outcomes were predominantly overestimating actual risks. Factors significantly associated with a high referral score were a low estimated probability of a spontaneous vaginal birth (p=0.007), adhering to the active management policy Proactive Support of Labour (PSOL) (p=0.047), and a practice situated in a rural area or small city (p=0.016). KEY CONCLUSIONS: there is considerable variation in referral decisions among midwives that cannot be explained by woman characteristics or clinical factors in early labour. A realistic perception of the possibility of a spontaneous vaginal birth and adhering to expectant management can contribute to the prevention of unwarranted medicalisation of physiological childbirth. IMPLICATIONS FOR PRACTICE awareness of variation in referrals and the associated midwife-related factors can stimulate midwives to reflect on their referral behavior. To diminish unwarranted variation, high quality research on the optimal management of a physiological first stage of labour should be performed.