Assessing the benefits of health research: lessons from research into the use of antenatal corticosteroids for the prevention of neonatal respiratory distress syndrome
Introduction
Do the benefits from health research justify the resources devoted to it? This question is increasingly asked, especially when the funding could otherwise be spent directly on providing health care. Various approaches to, and reasons for, assessing the benefits from health research have been advanced (Drummond, Davies, & Ferris, 1992; National Institutes of Health ((NIH), 1993; Buxton & Hanney, 1996; Grant, Cottrell, Cluzeau, & Fawcett, 2000; Smith, 2001; Croxson, Hanney, & Buxton, 2001). Some reasons revolve around the increasing demands for accountability for existing research funds, the desire to provide justification for current levels of expenditure and advocacy for extended funding. Others relate to the potential use of assessments in helping both to increase understanding of the processes involved and assist identification of how research systems can best be organised to enhance utilisation and benefits, especially for patients. Finally, it is claimed that assessment of utilisation could provide incentives for greater attention to be given to activities aimed at enhancing utilisation.
Previous modelling of research utilisation and payback assessment recognises that only sometimes do flows of health research knowledge make a direct impact on policy and practice: more often they simply add to the pool of knowledge (Kogan & Henkel, 1983; Hanney, Packwood, & Buxton, 2000). This increases the difficulties of assessing the benefits from research. Not only do the types of benefits that can flow from research have to be clarified, and methods selected for assessing whether these paybacks are accruing, but it is also important to consider how to identify which research is responsible for any payback achieved.
Few topics have figured so prominently in the research utilisation debate as the use of antenatal corticosteroids to prevent neonatal respiratory distress syndrome (RDS) when preterm delivery is expected. Indeed, this issue became a cause celebre among those concerned with encouraging greater research utilisation (Department of Health, 1993; Haines & Jones, 1994).
Section snippets
Models of research utilisation and payback assessment
We draw on models of research utilisation and frameworks for payback analysis in order to present and organise a wide range of available evidence. Theories of diffusion of innovations generally examine patterns of adoption of new findings. For example, Rogers (1995) developed the concept of the S-shaped adoption curve which helps inform analysis of how far uptake of research findings will occur ‘spontaneously’ and how far specific steps are necessary to encourage implementation. Rogers’
Methods for studying payback
For the analysis described here we were able to draw on the considerable existing literature about the use of antenatal corticosteroids. Some of this had been undertaken, in previous studies, by the authors of this article: detailed economic and payback analysis (Mugford, Piercy, & Chalmers, 1991; Mugford, 1993); key informant interviews (Hanney, 1994); surveys of potential users (Hanney, Soper, & Buxton, 2003b); and bibliometric analysis (Grant, Green, & Mason, 2003). The latter analysis
The development of the primary science
The identification of a specific starting point for an exercise such as this is often problematic but here there are strong arguments for starting with the work of Liggins (Liggins, 1969; Liggins & Howie, 1972). At the end of the 1960s he examined how, when glucocorticoids triggered the onset of labour in pregnant sheep, the lambs born prematurely had well-aerated lungs, while many of the control animals died of RDS (Liggins,1969). Further experiments using a range of animals were continued by
Secondary research
As early as 1981 Crowley undertook a meta-analysis of these trials but this review was first published in a structured form in the Oxford Database of Perinatal Trials and Effective Care in Pregnancy and Childbirth in 1989 and published in a journal a year later (Crowley et al., 1990). Just 12 trials, out of 23 examined, met the predefined criteria of research quality necessary for inclusion in the systematic review. The results were clear
Data from 12 controlled trials, involving over 3,000
Recommendations for use
Recommendations for the use of corticosteroids, for that small proportion of women for whom preterm delivery is expected, were made in a variety of documents with differing degrees of authority. These include reviews, editorials and textbooks. As early as 1979, a review article in The Lancet referenced Liggins and Howie, (1972) and stated that ‘Corticosteroids should be given to the mother intramuscularly over 48 h: they reduce the incidence and severity of idiopathic respiratory distress
Assessing and encouraging adoption/application rates
The Buxton and Hanney model suggests there can be a flow of research impacts. This goes from the primary outputs such as publications containing research findings, to the secondary outputs such as clinical policies and guidelines and then into adoption or application by practitioners. The model recognises, however, that the reality is rarely a simple linear sequence. Assessments of use in relation to corticosteroids show a somewhat curious pattern: in at least some countries there was quite a
Assessing benefits from Liggins’ research
Applying some elements of the multidimensional categorisation and an historical perspective, we start by looking at the benefits from the original work from Liggins, initially funded by a grant of less than £20000 from the Wellcome Trust (Grant et al., 2003). In terms of knowledge production, the original work on sheep resulted in Liggins’ (1969) paper that has been cited over 500 times. By itself such basic work is most unlikely to result in applications; instead, it is more likely to inform
Discussion and conclusion
This analysis shows that the body of research examined led to important knowledge production, or primary outputs, and resulted in a string of significant secondary outputs in the form of guidelines. But above all, and key to providing ‘good stories’ to assist with justifying research funds, this example provides an account of how health research can lead to health gains in terms of reduced mortality and morbidity.
The analysis presented here has limitations. It is based on a single case study of
Acknowledgements
The authors gratefully acknowledge the encouragement and helpful comments on earlier drafts of this paper received from Iain Chalmers, Ross Howie, Shyama Kuruvilla, Graham Liggins, Bryony Soper, two referees, and participants at the Scientific Basis of Health Services Conference, 2003, where some of these ideas were presented. We are also grateful for the contribution to the Wellcome Trust's Comroe and Dripps study made by Barbara Mason and Liz Green and for funding from the Department of
References (59)
The argument for prenatal administration of dexamethasone to prevent respiratory distress syndrome
Journal of Pediatrics
(1984)Randomization of the first patient
Medical Clinics of North America
(1975)- et al.
Assessing the costs and benefits of medical researchThe diabetic retinopathy study
Social Science & Medicine
(1992) - et al.
An exploration of opinion and practice patterns affecting low use of antenatal corticosteroids
American Journal of Obstetrics and Gynecology
(1995) - et al.
Prematurity
The Lancet
(1979) - et al.
A comparison of results of meta-analyses of randomized control trials and recommendations of clinical experts. Treatments for myocardial infarction
Journal of the American Medical Association
(1992) Pharmacological approaches to the acceleration of fetal lung maturation
British Medical Bulletin
(1975)- Benson, R. (1977). Premature labor and the small-for-dates infant. In D.N. Danforth (Ed.), Obstetrics and gynecology,...
- et al.
Understanding the relation between research and clinical policyA study of clinicians’ views
Quality in Health Care
(1997) - et al.
Are we prescribing multiple courses of antenatal corticosteroids? A survey of practice in the UK
British Journal of Obstetrics and Gynaecology
(1999)