Nurses' retention and hospital characteristics in New South Wales, CHERE Discussion Paper No 52

NURSES’ RETENTION AND HOSPITAL CHARACTERISTICS IN NEW SOUTH WALES

The hospital characteristics come from the NSW Public Hospitals Comparison Data Book 1996/1997.
The information in this publication is provided by Area Health Services and is compiled from various
data sources. Numerous hospital statistics are provided and we construct a set of variables
measuring cross hospital variations in size, staffing levels, expenditures, complexity and intensity
of work.

Probit regressions are estimated with the observed retention status in 1997 as dependent variable
and personal, job, and hospital characteristics as explanatory variables. Various specifications are
estimated and tested. Conditional on personal and hospital characteristics, hospital type fixed effects
do not help explain retention either through the conditional mean of the probability or the variance of
the error. Hospital characteristics however do play a significant role.

Hospital size has ambiguous effects depending on the type of cases (acute versus non-acute) and
the type of patients. The number of babies tends to increase retention while non-acute cases tend to
reduce retention. The impact of the number of acute cases varies depending on the specific measure
used. Overall size effects tend to be imprecisely measured. Complexity of procedures performed also
has differing effects: emergency admissions increase retention while high cost procedures and the
average ANDRG (see Appendix 3 for definition) weight reduce retention. These effects are large and
significant. These results are consistent with the information coming from surveys of job satisfaction
among nurses. Complexity can represent a positive aspect of the job in the form of challenge and
learning possibilities. It can also mean higher stress levels and a removal from the decision-making
level of the health care.

Higher expenditures (at constant staffing levels) increase retention except for expenditures on visiting
medical officers (VMO) which reduce retention probabilities. Higher expenditures generally can
indicate greater non-wage benefits (e.g. staff development) as well as the presence of senior (i.e. high
salaried) staff specialists who can provide support and training. Anecdotal evidence also suggests
that increased usage of VMOs tends to increase the stress level of nurses and reduces the training
component of the nurses’ job. Again these effects are large and significant.

It is often argued that a major factor in high nursing quit rates is high workloads. The results on
nursing staff levels show that the type of workload matters as much as size. In particular, simply
increasing the nursing staff (hence reducing workload) while keeping workload mix constant does
raise the probability of retention but only by a small and insignificant amount.

Our results also suggest interesting and significant variations in retention based on personal and job
characteristics. Being male or being UK born substantially decreases the probability of retention
while being foreign born (excluding UK and NZ) increases retention. In general older nurses are
much more likely to stay in nursing and so do those with post basic qualifications and those who
have been promoted. Longer hours of work tend to increase retention at least over the relevant range
of observed hours per week.

In explaining the differences between the predicted retention probabilities for the bottom
and top quartiles in the sample it is found that the most important contributors are age and
hospital characteristics.

The paper is organized as follows. Section 2 describes the data sets while section 3 discusses
the econometric modelling and the results. This is followed by a further analysis of the estimation
results in the form of profiles of nurses based on their predicted retention probabilities. Concluding
remarks are offered in Section 5.

<<   4   5   6   7   8   9   10   >>

More intriguing information