An investigation of labour ward care to inform the design of a computerised decision support system for the management of childbirth

Abstract

Patient monitoring is a complex task, particularly during childbirth, where assessment of the baby's condition is inferred from the continuous electronic recording of the baby's heart rate pattern and maternal uterine contractions (CTG). Computerised decision support has long been advocated, as difficulties in the interpretation of the CTG have led to failure to intervene and unnecessary intervention. The problem is large, for obstetric litigation now accounts for 80% of the UK National Health Service litigation bill. The Plymouth Perinatal Research Group has developed a computerised decision support system for patient monitoring during childbirth and the UK Medical Research Council has agreed to fund a multicentre randomised trial. The work of this thesis was an investigation of the labour ward care system to inform the human-centred design of the decision support system for patient monitoring in childbirth, prior to the clinical trial. It was recognised that many decision support systems have failed to gain clinical acceptance, as conventional design models were inadequate. Lack of attention to the organisational context of the care system and the process of the direct patient care led to the design of inflexible 'expert' systems, which constrained working practices. A pilot ethnographic study of an existing decision support system, used for the analysis of umbilical cord blood samples, was undertaken to clarify the research approach required for the main study. It was found that barriers to effective use within the wider work system included inadequate implementation and lack of organisational support. A case study approach produced a more comprehensive account of the context and process of the use of the computer system. The main study combined qualitative with quantitative techniques to investigate the system of care in childbirth, both outside and within the delivery room, to provide a unique, holistic perspective. The organisational context of the labour ward was investigated by direct observation of clinicians over the course of their work for 220 hours. Observations were documented and transcribed to computer text files. Patterns of actions and events were coded using ATLAS(ti) data analysis software. The codes were counted and tabulated to model the main features of this labour ward care system, which was expressed in the form of a rich picture diagram. These findings were confirmed by a limited study of five other UK labour wards. The core qualitative categories, derived from the observation data, found a complex and problematic relationship between communication, decision making and accountability. Decisions were often made outside the delivery room and were subject to misinterpretation and bias. The organisational hierarchy made it difficult for junior staff to question clinical management decisions. A system of tacit practice, external demands upon clinicians and transient allocation of junior midwives to labour ward militated against teamwork. This increased the vulnerability of the care of mothers to error. The process of direct patient care, within the individual delivery room, of 20 mothers in labour was captured in a novel audio-video observation study. The 111 hours of first stage labour and 12 hours of second stage labour were recorded and digitised to computer files. Recurrent actions and patterns of behaviour were coded both quantitatively and qualitatively using ATLAS(ti) data analysis software. Midwives left the room on average every 15 minutes to be absent for 27% of the first stage of labour. Record keeping occurred on average every 10 minutes and accounted for 19% of midwives' time. Midwives had little time to talk with mothers and only sat down at the bedside for 15% of the time. Psychosocial support was not given priority. Parents were generally excluded from communication between clinicians yet 108 clinicians took part in the care of the 20 women. Pressures from medicolegal directives and task-orientated imperatives overshadowed meaningful interaction with parents and caused spurious care priorities. This work has revealed the need for a critical reassessment of the type of support that is required for monitoring situations in all areas of medicine. A range of functions, such as shared information displays and models, have been suggested to augment roles and relationships between clinicians and parents to support patient-centred care. The present work has revealed that a combination of computer-based technology and changes to working practice can support the parents, their individual carers and their various roles. In this way the system of care can be more aligned to the objective of a safe and emotionally satisfying birth experience for parents and staff. A further programme of research is required to follow-up the existing studies, develop these new forms of interaction between technology and clinicians, and evaluate their effectiveness. The research methods employed in the present work will provide a more comprehensive evaluation of the decision support system in the forthcoming multicentre trial. The methods of investigation have also been shown to be of relevance to patient safety research, service delivery and training.