Reviewing the evidence base for nurse staffing and quality of care in nursing homes
- Correspondence to Nicholas G Castle
Graduate School of Public Health, University of Pittsburgh, A610 Crabtree Hall, Pittsburgh 15261, Pennsylvania, USA;
Implications for practice and research
■ There is some evidence that the registered nurse to nursing auxiliary ratio has an impact on the quality of care but further research is needed.
■ There is a need to consider other factors such as staff turnover and the use of agency staff when considering the quality of care in nursing homes (NHs).
■ Further research is needed to identify quality indicators that take into account what is important to residents and their families.
■ There are a number of methodological challenges to undertaking research in this area.
This research comes at a time when NHs are being increasingly asked to do more with less. Revenues for NHs do not appear to be keeping up with inflation. Given that staffing accounts for the largest cost in operating an NH, cuts to staffing levels are a constant threat. As such, evidence of whether nurse staffing levels directly influence quality and especially, at what level is badly needed.
This study examines the literature from 1980 to 2008. The staffing measures and quality indicators used in prior studies are categorised, examined and discussed.
This review did not identify a strong association between staffing levels and quality. This is despite the extremely large number of studies in this area. The authors, for example, included 83 studies in this review. This conclusion parallels those of previous literature reviews.1
It would seem somewhat incongruous to discern so little from so many studies. However, as the authors describe, there are considerable challenges to identifying what, on the face of things, seems to be such a self-evident concept (that staffing influences quality). These challenges include issues surrounding three concepts: what are NHs?, what do staffing levels measure? and what quality indicators are most appropriate?
In reality, NHs can be very diverse settings. Individual facilities can specialise in specific care (such as pressure ulcers) or given populations (such as the deaf). Within facilities, individual units can provide very specific care (such as Alzheimer's care). Thus, when large numbers of NHs are examined, even with case-mix adjustment, there may not be adequate controls for the diversity of the populations. What is certainly not happening is linking staff time with specific residents. In all studies, overall staffing levels are used. The link that is really needed is the staff time provided to each resident.
With respect to staffing levels themselves, these may be weak proxies for the actual care provided. Moreover, staffing levels are notoriously inaccurate. In addition, ‘staffing’ itself is not a uniform concept. A mix of registered nurses, licensed practical nurses and nurse aides exists. This mix of staff may be extremely important for some areas of quality of care. We should also not forget the potential influence of other staff (such as administrators, directors of nursing, medical directors, social workers). Analytically, this leads to many opportunities for interactions between staff in relationships with quality that may be non-linear.
Even with accurate staffing levels, considerable confounders exist with linking these levels to quality. Some staff may simply work harder than others, whereas other staff may work more efficiently than others. The training and experience of staff may dictate quality of care in addition to staffing levels. Even training and experience can be further expanded. For example, with experience, this could be an experience with facility routines and/or experience with the resident. The authors mentioned that agency staff may be important as these staff may not have either type of experience – yet they do increase staffing levels. A further example would be the influence of consistent assignment. This would mean that staff will have more experience with the same residents, and theoretically should provide higher quality of care.
The authors noted that the indicators used in the analyses are highly varied, although they do tend to generally neglect measures of satisfaction. A review of quality indicators notes the difficulties involved in using appropriate measures.2 For staffing, one has to more carefully consider the indicators used and whether a strong or weak influence is likely and the time period over which we expect this influence to manifest itself.
With these challenges in mind, clearly no single study represents a ‘gold standard’ in this area. An examination of more recent studies and the most methodologically advanced studies could provide evidence of a stronger association between NH staffing levels and quality.3 Our inability, in this area of research, to identify strong links between staffing levels and quality of care has important consequences. Resident care may be suboptimal. Providing practitioners and policy makers with this evidence is one of the most important challenges that exists in NH research.