The use of the Delphi technique in nursing and health research has increased dramatically over recent years and shows no signs of abating. The technique has been frequently used to identify the research priorities in many different areas of nursing and health research. Bond and Bond (1982) used the technique to establish clinical nursing research priorities as did many others (Lindeman 1975, Alderson et al. 1992, Forte et al. 1997, Lynn et al. 1998, Daniels & Ascough 1999, Soanes et al. 2000, Cohen et al. 2004, Annells et al. 2005).
It has also been used to identify the research priorities in other areas of health care including HIV/AIDS (Sowell 2000); occupational health (van der Beek et al. 1997, Sadhra et al. 2001); occupational medicine (Harrington 1994); health sector library and information services (Dwyer 1999); oncology (Browne et al. 2002); paediatric haematology, oncology, immunology and infectious diseases (Soanes et al. 2003) emergency care (Bayley et al. 1994, 2004; Rodger et al. 2004); midwifery (Fenwick et al. 2006); orthopaedic nursing (Salmond 1994); paediatric cancer nursing (Monterosso 2001); health informatics (Brender et al. 1999); dentistry (Palmer & Batchelor 2006); urologic nursing (Demi et al. 1996) and public health (Misener et al. 1994).
Back-Pettersson et al.'s (2008) study reports on research priorities identified by Swedish nurse using the Delphi technique. An expert panel of 118 clinicians was recruited and in total 338 nursing research areas were identified by the panel. This is a large number of research areas and illustrates the range of responses that can be generated by the Delphi technique and the wide range of research that is perceived by nurses to be required. Perhaps this is why most research priority studies that use the Delphi now tend to focus on a specific area of nursing, such as cancer nursing (McIlfatrick & Keeney 2003) or school nursing (Edwards 2002), rather than nursing in general.
Other studies that took this broad perspective to the identification of research priorities include Ross et al. (2004) in the UK, Eriksson (1995) in Norway, Hamrin (1995) in Sweden and MacMillan et al. (1989) in Scotland. As with Hamrin’s work, Back-Pettersson et al. (2008) reported on nursing research priorities in Sweden.
A large number of nurses were invited to participant in the study, 2665 in total. It would have been interesting to know the inclusion criteria by which the expert panel were selected, but this is not detailed. One can only assume that the criteria must have been wide-ranging to include such a large sample of nurses. The authors state that the expert panel was a ‘convenient group of informed individuals and specialists with the requisite expert knowledge’ (p. 3). Considering that the panel should be ‘experts’, it could be argued that every nurse meets this criteria and that this is why the invited numbers were so large. Delphi studies benefit greatly from a well-defined inclusion criteria so as to select panel members with the right expertise; this would have been essential if the research priorities were for a more specific nursing specialty.
The study maintained a good response rate throughout the three rounds. This is laudable as the Delphi is notorious for having high dropout rates. Considering that the 118 panellists were self-selecting from a very large population of 2665, this keenness to be involved is probably the reason for panel sustainability.
While the study successful ranked 41 research areas in order of priority, not much consideration is given in the paper to the level of consensus on these priorities. The authors acknowledge that the re-wording of 338 research areas in 41 statements may have accounted for some panellists’ willingness to complete all three rounds. However, there is a difference between the extent to which each expert panel member ranks a research area as a high priority and the extent to which he or she agrees with other members. In Delphi studies where the focus is priority setting consideration should be given to using a level of consensus among expert panel members and ranking only those statements which have reached that level.
This study set its findings in the context of some international work such as that of Hinshaw (1995). However, a discussion of how nursing research priorities in Sweden compare with those in other countries would have been an interesting and enlightening addition to this paper. The study also raised the issues which Delphi researchers must be mindful of such as the importance of consensus, inclusion criteria and response rates. With the emergence of technological real-time Delphi’s, e-Delphi’s and online approaches the method is constantly evolving. Nonetheless, the main focus on agreement and prioritisation must remain foremost in the minds of Delphi researchers no matter which type of Delphi is being used.
When the many research priority Delphi studies are examined, there is often little mention of the importance of bringing the findings to the attention of grant awarding bodies. In fact it was not possible to find a publication on how the Delphi findings led specifically to increased research funding. Therefore, it is crucial that Back-Pettersson et al. ensure that research funders in Sweden are aware of their findings. This could also inform the strategic thinking of Swedish nursing organisations. Only then will the study have the desired ‘ripple effect’ on nursing research policy.