Introduction: In hospitals, nurses frequently have to work in shifts that cover the 24-hour day.
There is a wealth of evidence that links shift work with negative outcomes for nurses,
and as such, health organisations have a duty to promote safe working hours and monitor
the wellbeing of their nursing workforce. One useful outcome for monitoring includes
staff sickness absence, as documented by hospital administrative records. While previous
research analyzing records of shifts and sickness absence have shown increased rates
when nurses are working certain shift types, there is a gap in understanding the effects
of more complex shift patterns, particularly those that occur over multiple days (e.g.,
working long and/or night shifts consecutively, having fewer than 48 hours rest between
ending a night shift and starting a day shift, and frequent shift rotations).
Methods: We analyzed registered nurses’ historical shift and sickness absence data, as recorded
by electronic staff rostering systems from acute inpatient wards in two large NHS
hospital Trusts in England. From this data, we created a series of variables that
correspond with variation in shift types/patterns, including long working hours, night
work, consecutive working days, and recovery time between shifts. Each variable was
defined by an exposure period, i.e., the shift configurations worked in the 7 and
28 days prior to each worked shift and sickness absence episode. We then used logistic
mixed regression models to estimate the relationships of these variables with sickness
absence, in terms of the change in odds of a shift being cancelled due to sickness.
Results: The final dataset contained 1,367,497 worked shifts and 19,876 sickness absence episodes
from 7,515 registered nurses across 95 wards. The majority of shifts were from nurses
working full-time (60%), and sickness episodes lasted a median of 4 days long (IQR
2–8 days). In the 7-day exposure multivariable model, intense consecutive spells,
quick returns, and shift rotations significantly increased the odds of sickness, with
quick returns and shift rotations also showing longer term effects in the 28-day multivariable
model. Nonlinear analyses of the proportion of long and night shifts worked revealed
that higher proportions (≥80%) were significantly associated with the greatest odds
of sickness absence in both 7-day and 28-day lookback windows.
Conclusion: This longitudinal analysis of routinely collected roster records provides new objective
insight into how shift patterns and working hours are linked with nurse wellbeing.
Analysis of 1.4 million records revealed that long hours, night work, consecutive
working spells, and inadequate rest periods significantly increased the odds of sickness
absence in weekly and monthly exposure windows. These findings help to inform future
research on how nurses’ shift patterns can be improved, particularly in terms of optimizing
ward rosters in ways that prioritize staff wellbeing. Support: The study dataset was
derived from a project funded by the UK National Institute for Health and Care Research
(NIHR) Health Services and Delivery Research Program (award No. NIHR128056) and the
NIHR Applied Research Collaboration (Wessex). The primary author was supported by
the UKRI Economic and Social Research Council South Coast Doctoral Training Partnership
(Grant Number ES/P000673/1).
