The development and testing of an algorithm to support midwives’ diagnosis of active labour in primiparous women

Abstract

The research in this thesis aimed to develop an algorithm to support midwives’ diagnosis of active labour in primiparous women and to compare the effectiveness of the algorithm with standard care in terms of maternal and neonatal outcomes. Four linked studies are presented following the template suggested by the Medical Research Council (MRC 2000) Framework for development and evaluation of randomised controlled trials (RCT) for complex interventions to improve health.

Study one Aim: To develop an algorithm for diagnosis of active labour in primiparous women. Methods: An informal telephone survey was conducted with senior midwives to assess the need for a decision support tool for the diagnosis of active labour. A literature review identified the key cues for inclusion in the algorithm which was then drafted. Focus group interviews were conducted with midwives to ascertain the cues used by midwives in diagnosing active labour. Findings: Thirteen midwives took part in focus groups. They described using informational cues which could be separated into two categories: those arising from the woman (Physical signs, Distress and coping, Woman's expectations and Social factors) and those from the institution (Midwifery care, Organisational factors and Justifying actions). Study Two Aim: Preliminary testing of the algorithm Methods: Vignettes and questionnaires were used to test the consistency of midwives’ judgements (inter-rater reliability), the content of the algorithm and its acceptability to midwives (face and content validity). The study was conducted in two stages: the first stage (23 midwives) involved vignettes and questionnaires and the second stage (20 midwives) involved vignettes only. Findings: In the first stage a Kappa score of 0.45 indicated only moderate agreement between midwives using the algorithm. After modifying the algorithm, the Kappa score in stage two was 0.86, indicating a high level of agreement. While the majority of the midwives reported that the algorithm was easy to complete, most were able to identify snags or make suggestions for its improvement. Based on the findings of this study the algorithm was modified and the final version was developed.

Study three Aim: To assess the feasibility of carrying out a cluster randomised trial (CRT) of the algorithm, in Scotland. Specifically, to identify maternity units potentially willing to participate in a CRT, to test the implementation strategy for the trial and to collect baseline data to inform the sample size calculation. Methods: A questionnaire and interviews were used. The CRT methods were piloted in two maternity units and the algorithm was used for a three-month period in order to test its acceptability and provide estimates of compliance and consent rates. Results: All maternity units surveyed expressed an interest in the proposed study. Midwives’ compliance with study protocol differed between units, although the consent rate of women was high (89% and 84%). Ultimately, one unit achieved 100% of the required sample and the other 60%. The midwives reported that the algorithm was acceptable and was a useful tool, particularly for teaching inexperienced midwives.

Study four Aim: To compare the effectiveness of the algorithm for diagnosis of active labour in primiparous women with standard care in terms of maternal and neonatal outcomes. Method: A cluster randomised trial Participants: Fourteen maternity units in Scotland. Midwives in experimental sites used the algorithm to assist their diagnosis of active labour. Seven experimental units collected data from 1029 women at baseline and 896 post intervention. The seven control units had 1291 women at baseline and 1287 after study implementation. Outcomes: The primary outcome was the percentage use of oxytocin for augmentation of labour. Secondary outcomes were medical interventions in labour, labour admission management, unplanned out of hospital births and clinical outcomes for mothers and babies. Results: There was no significant difference between groups in percentage use of oxytocin for augmentation of labour or for the use of medical interventions in labour. Women in the algorithm group were more likely to be discharged from the labour suite following their first labour assessment and subsequently have more pre-labour admissions.

Conclusion The studies presented in this thesis represent the full process of developing and testing a complex healthcare intervention (the algorithm). The final study, a national cluster randomised trial, demonstrated that the use of the algorithm did not result in a reduction in the number of women who received oxytocin for augmentation or the use of medical interventions in labour. The results suggest that misdiagnosis of labour is not the main reason for higher rates of intervention experienced by women admitted to labour wards while not yet in active labour. These studies contribute significantly to the debate on care of women in early labour, the organisation of maternity care and to maternity care research.