Article Text

other Versions

Discourse analysis
Changes in patient outcomes coincide with increased nursing hours
  1. Adrian MacKenzie,
  2. Gail Tomblin Murphy
  1. WHO Collaborating Centre on Health Workforce Planning and Research, Dalhousie University, Halifax, Canada
  1. Correspondence to Gail Tomblin Murphy
    WHO Collaborating Centre on Health Workforce Planning and Research, Dalhousie University, 905-1660 Hollis Street, Halifax, Nova Scotia, B3J 1V7, Canada; gail.tomblin.murphy{at}

Statistics from

Commentary on: OpenUrlCrossRefPubMed

Implications for practice and research

  • The paper's findings align with previous evidence establishing a link between levels of nurse staffing and patient outcomes, and are important for policy makers and other stakeholders to consider.

  • Future evaluations should include all patients who may have been impacted by such policies, and use more rigorous methods to investigate a possible causal link between the policy and those outcomes.


Research has indicated that nurse staffing levels are often related to patient outcomes and this has led some to propose that staffing levels should be set at a statutory minimum; it is not clear, however, whether such policies are effective. Ongoing monitoring and evaluation of policies and practices are essential to the efficiency of healthcare planning and service delivery. An investigation of the impact of the nursing hours per patient day (NHPPD) staffing policy following its implementation in Western Australia in 2002, particularly as it pertains to outcomes for patients, is therefore warranted.


The authors seek to determine the impact of a new nurse staffing model by comparing 14 nurse-sensitive patient outcomes before and after its introduction and measuring concurrent levels of nurse staffing. This retrospective analysis was conducted using data from three adult tertiary hospitals in the state capital.


The paper's results show that the amount of NHPPD was higher in each hospital following the implementation of the policy. Also, a number of patient outcomes were different following the policy's implementation. The changes in outcomes were not consistent across the four hospitals or the 14 outcomes, and hence there may be an interaction between the type of environment and the impact of increased nurse staffing hours.


The methods reported in the paper make creative use of administrative data and are clearly explained. In addition, the reported changes in patient outcomes are of clear relevance to many stakeholders. However, some of the conclusions the authors draw do not appear justified by the evidence presented, and additional work may be required for the study to achieve its stated objective.

While the authors state that, “This study demonstrates that the increases in nurse hours after implementation of the NHPPD staffing method… improved a number of patient outcomes.” The study, however, establishes an association rather than a causal link between changes in nursing hours and patient outcomes.

The statement that ‘NHPPD benefits patient safety’ is not justified by the paper's results. Although certainly believable, the paper does not show that the NHPPD policy itself is responsible for the increase in nursing hours measured; one can only guess how much of the measured increases are a result of the policy. Without such a link, there is nothing in the paper to connect the policy itself with the outcomes measured. Even taking such a link for granted, such a statement seems overly conclusive for several reasons.

While the authors checked the correlation of successive outcome measurements over time and controlled for within-ward and within-hospital effects, they report no other steps taken to rule out the many factors other than nursing hours which influence patient outcomes and thus could have influenced at least some of the reported changes. Including additional variables in the study models, for example, would allow for the explicit consideration and measurement of the effects of other factors influencing patient outcomes. Analysing the association between nursing hours and patient outcomes at a series of narrower (eg, monthly) time intervals would allow, through the use of quasi-experimental design, for a more thorough investigation of the coincidence of changes in nursing hours and patient outcomes, particularly if the policy was not fully implemented simultaneously on all units at all facilities; changes in outcomes on units that implemented the policy later or did so gradually, could be compared with those on units that implemented the policy fully and immediately. Qualitative data from nurses and managers on the policy's implementation and any associated patient outcomes would provide complementary context to the quantitative analysis.

Importantly, the paper reports that some of the patient outcomes measured worsened following the policy's implementation; although not as many as the number that improved. Some consideration of these opposing findings is warranted. If the authors have weighed the outcomes that improved against those that worsened in some objective fashion, they have not reported as such.

An issue of generalisability is around the exclusion of patients outside tertiary care, and of maternal, newborn, paediatric, mental health and substance abuse patients within that sector; presumably changes to staffing were also made in these areas and thus may impact patient outcomes there as well. Some analysis of its impacts on these patients, who no doubt make up a substantial portion of those using the hospitals—not to mention those using nursing services outside of tertiary care—would thus be justified before broadly claiming that the policy improves patient care.

In summary, the paper reports an association between increased nurse staffing levels after implementation of a State policy, and changes in several patient outcomes. However it does not establish a causal link between the NHPPD policy and these outcomes, nor does it measure those outcomes for all of the patients who may be impacted by the policy.

View Abstract


  • Competing interests None.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.