a b s t r a c tRandomised controlled trials are the 'gold standard' for evaluating the effectiveness of interventions in health-care settings. However, in midwifery care, many interventions are 'complex', comprising a number of different elements which may have an effect on the impact of the intervention in health-care settings. In this paper we reflect on our experience of designing and evaluating a complex intervention (a decision tool to assist with the diagnosis of labour in midwifery care), examining some of the issues that our study raises for future research in complex interventions.& 2009 Elsevier Ltd. All rights reserved.
IntroductionThe randomised controlled trial (RCT) as a research design remains the gold standard for testing 'what works' in health-care settings. RCTs are as applicable for testing interventions that may be delivered by nursing, midwifery and allied health professionals (NMAHPs) (such as targeted health promotion advice or a structured rehabilitation programme) as they are for testing drug therapies or new surgical techniques. NMAHP interventions are often complex, in that they are 'built up from a number of components, which may act both independently and inter-dependently' (MRC, 2000), and are often introduced into equally complex health-care environments. This complexity makes the evaluation of NMAHP interventions, using a design such as the RCT, challenging; in this paper, we reflect on our experience of designing and evaluating a complex intervention (a decision tool to assist with the diagnosis of labour in midwifery care), examining some of the issues that our study raises for future research in complex interventions.
What is a complex intervention?Complex interventions are essentially complicated (Shiell et al., 2008). Rather than comprising a single active ingredient which may be tested (such as a single drug therapy), they are composed of a number of different interconnected parts which form the intervention to be tested. If we use our study as an example, we developed a decision tool that could be used by midwives to assist them to identify whether or not a woman attending a labour ward was in active labour (Cheyne et al., 2008a, b). In this example, the decision tool (a paper-based algorithm) involved the midwife collecting different items of information and using this information in a structured way to reach a judgement. The success (or not) of the intervention depended on how the midwife interacted with the woman and the decision tool to inform clinical judgements and management decisions. In addition, during the main trial, which was a cluster randomised trial (CRT), the decision tool was implemented within different maternity units, each of which had its own unique organisational characteristics that influenced how the tool was used in practice. It is this combination of multifaceted interventions, plus the influence of different social and organisational contexts, which makes complex interventions methodologically challenging to evaluate (Oakley et al., 2006).
Evalua...