The state of the science—the impact of declining response rates by nurses in nursing research projects

Abstract

Nursing research, nursing science and indeed the profession of nursing are at an important juncture within Europe. In many countries, the professional development of nursing, supported by the integration of the discipline within universities, remains relatively novel (Bressan et al., 2016; Colson et al., 2021). This professionalisation has heralded far-reaching changes, including the requirement for university diplomas and degrees as prerequisite entry to the profession, and an increased understanding, use and conduct of research (Evans et al., 2007). As such nursing increasingly supports and contributes to the evidence-based practice movement in health care, by further assurance of excellence in patient care, through the application of research and other evidence to inform care. However, as part of our engagement with this work, we have identified a fundamental issue that is impacting on research that seeks to collect data from practitioners—the often low response rates to survey type research projects by nurses. Nursing research relies very heavily on cross-sectional survey type research. While it is prudent, as the profession advances, to augment this type of research with interventions or longitudinal studies to examine topics in greater depth, there is still room for description, particularly of attitudes and practices to enhance understanding and inform practice improvements. However, robust, accurate measurement within this methodological approach requires not only attention to the validity and reliability of the instruments used to collect data, but also an adequate response rate. However, achieving adequate response rates to surveys used in nursing research is a perennial challenge in Europe, something that has not been clearly articulated or exposed in the published literature. While on the one hand this can be attributed to inadequate project planning, with simplistic solutions such as reminders emerging, it is a problem of great significance to the profession, as much of what we consider evidence may be reliant on less than adequate response rates. An average response rate is around 60% (at best) (Badger & Werrett, 2005; Sammut et al., 2021). While this is considered adequate, an uncertainty remains about the views of the those who did not respond (40% in this case). Moreover, survey response rates among the nursing profession, and indeed many health professionals, are notoriously low (Cook et al., 2009), and it is not surprising to find response rates of 10%–15% (Badger & Werrett, 2005; Sammut et al., 2021). The challenge with this underrepresentation relates to a potential bias in research, one which is often not acknowledged (Badger & Werrett, 2005). Moreover, we find that in popular topics, in our experience, this majority view is accepted as ‘fact’. However, when one explores more contentious issues the 40% who did not respond become important. Reviewers or respondents are more likely to question the authenticity of the findings based on the fact that more than one third did not respond. There is of course legitimacy in this view. However, while response rate might appear a straightforward project management issue (Sammut et al., 2021), our experience is that it can be topic related and also linked to contextual issues. That which the ‘audience’ find interesting or compelling will illicit or excite responses, whereas that which is viewed as less important, or indeed where the respondent simply does not understand the questions (such as surveys that test knowledge of research for example), receive much less attention. In our experience, conducting cross-sectional survey research involving nurses (and other health professionals) led to challenges in reaching an adequate response rate. At an early study phase, in a large cross-sectional study, we experienced very a poor uptake across Italian hospitals. This was an important study in our view, the aim of which was to describe staffing levels, work environment, aspects of well-being, and then correlate these findings with specific patient outcomes. We used tactics to improve this, such as those outlined by Sammut et al. (2021), and there was a small improvement, but not substantial. This experience found, from discussions with nurses, that they simply could not dedicate more time during their shift to fill out these documents, or to participate in surveys even if this only required 10–20 min. They often said they had a lot of work to do during their shift and did not have time to fill out and participate in surveys. Moreover, as this research is viewed as ‘work’ and as such completing these outside of hours would not be desirable. Another matter that came to our attention was how much nurses enjoyed being interviewed about important topics. While the response rate to surveys can be low, for the same topic, and in the same study, nurses were welcoming of the opportunity to talk on a one-to-one basis. This experience was in the context of a cross-sectional study exploring workplace violence. It is perhaps the sensitivity, and emotional effect of the subject that lends itself more to interviews, and perhaps this is something that researchers will need to consider for the future. Other aspects of this challenge that are important for researchers are the incentivisation and the distal relationship of the researcher. Contemporary ethical requirements shy away from any direct researcher contact with participants, and while ethical in terms of ensuring self-determination, can deter from that personal encouragement that often increased responses to survey research in the past. Providing personal information sessions can help, and is within ethical boundaries, and this was certainly our experience, but yet the distance of the researcher, and the lack of personal contact is problematic. Another matter that raises ethical issues is incentivisation. Certainly, Sammut et al. (2021) clearly demonstrated the benefit of this in terms of substantially raising response rates, in some countries this would not be permitted (from an ethical perspective). Moreover, most projects, especially in countries, such as Italy where the profession is still developing, would not have the resources necessary to support this approach. It is also true that European nurses are bombarded with (in particular online) surveys, and this too negates against a satisfactory response. One further issue is the length of the questionnaire. Although Sammut et al. (2021) did not draw particular attention to this in their review of response rate improvement, many contemporary surveys are long and perhaps off putting. Certainly, one of our own studies investigating workplace contexts comprised several questionaries related to nurse staffing, organisational support for nursing practice, job satisfaction, burnout and patient safety. Furthermore, the answers are not necessarily straightforward and easy to answer. For example, when exploring nurse staffing, nurses were asked to provide details on their last shift (day, evening or night) and the number of patients they were assigned. These answers require reflection and might take time to complete. Advancing nursing science is challenging, especially in countries with limited representation of nursing at professoriate level and where professional development is still in its infancy. Standard survey approaches to research, even those exploring topics of global interest such as staff, are increasingly challenged by low response rates. This raises additional challenges for researchers in such countries who are struggling to get the research off the ground. While standard approaches to address this, such as frequent reminders can yield better responses (Brtnikova et al., 2018), overall, the volume of survey research requests received in health care (Brtnikova et al., 2018), together with the business of the healthcare environment serves to mitigate against cooperation. It is time perhaps to consider a strategic approach to addressing these challenges. Perhaps firstly by limiting the volume of survey type research permitted within organisations and across patient groups. Coalescing the research within key themes, and ensuring that research groups work together, perhaps with a battery of instruments from various projects, being used at one key time point, might serve as a more effective way of managing participants' time. For nurses and other healthcare workers, it might be worth considering dedicated time within the working day that is set aside to attend to research participation tasks. This would certainly be useful for research related to matters of global or national importance, such as the aforementioned staffing project. None. No funding was received for this study. The authors declare that they have no conflict of interest. They also declare that they agree with the content of this manuscript, which has not been published or being considered for publication elsewhere. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Topics & Concepts