Statistics from Altmetric.com
We were very interested to read the article by Callaghan and colleagues.1 They report a decline in mortality with less nurses caring for high risk infants over the first three days of life. This is a surprising finding, which is counter-intuitive to established neonatal wisdom. Indeed the authors quote a smaller study by Hamilton and colleagues,2 which found an increase in mortality with a reduced ratio of nurses to infants. Callaghan and colleagues’ findings may be true, with the most likely explanation of the deaths being excessive handling. Clearly, if this finding is replicated, then establishing the optimum number of nurses could lead to improved outcomes for high risk infants. It certainly warrants further study within the NHS and the United Kingdom.
It is widely assumed that increased numbers of nurses in the UK will improve the outcome of neonatal intensive care. Currently, in the National Health Service there is difficulty in maintaining adequate numbers of neonatal nurses, with many units having nurse staffing levels substantially below those recommended by the British Association of Perinatal Medicine.3 Unfortunately, this recommendation for more staff is not based on a great deal of evidence, and the authors are to be praised for studying this topic.
Their results should, however be interpreted with caution. The health systems of the UK and Australia are different, most particularly in the proportion of centralised care and the ratio of nurses to infants. Callaghan et al make the point that the UK has a ratio of two very low birth weight (VLBW) infants to one nurse, whereas in Australia the ratio is approximately one to one. The UK Neonatal Staffing Study has recently looked at 13 500 infants from 54 randomly selected units throughout the UK.4 This study did not show a clear relation between staff establishment and outcome, although it did show a linear relation between mortality and occupancy rates and a trend to increased risk of mortality with a lower nurse:infant ratio.
Callaghan et al discuss some of the weaknesses of their own study. There are also two factors that we wish to highlight. The first is that the authors have not looked in detail at the quality and abilities of the nursing staff. There is a wide variation in the abilities of staff, particularly when nursing agencies are used to provide nurses. As these staff may not work full-time or have much experience of the individual unit, they may be less efficient or able when compared to those full-time staff based on the unit.
The second factor is the method of determining nurse workload. Measuring the ratio of babies to staff is not an accurate assessment of nurse activity; a large number of well babies often need less care than a small number of sicker babies. It is not clear from the paper how the authors dealt with the term infants, and whether these are included in calculating the ratio. Did the authors use the number of nurses per VLBW infant or per all babies in the unit? In addition, large babies can also generate a substantial workload if they are very unwell (for example, babies with persistent pulmonary hypertension of the newborn or congenital diaphragmatic hernia). Further studies measuring the true overall workload may give a better indication of the relation with outcomes.