The impact of extending nurse working hours on staff sickness absence: Evidence from a large mental health hospital in England

Abstract

Background: A pressing international concern is the issue of mental health workforce capacity, which is also of concern in England where staff attrition rates are significantly higher than in physical health services. Increasing demand for mental health services has led to severe financial pressures resulting in staff shortages, increased workloads, and work-related stress, with health care providers testing new models of care to reduce cost. Previous evidence suggests shift work can negatively affect health and wellbeing (increased accidents, fatigue, absenteeism) but can be perceived as beneficial by both employers and employees (fewer handovers, less overtime, cost savings).

Objective: This study reports an evaluation of the impact of extending the shifts of nurses and health care assistants from 8 to 12 hours. Using data before and after the policy change, the effect of extended working hours on short term sickness (< 7 days) on staff is examined.

Setting: The setting is six inpatient wards within a large mental health hospital in England where the shift extension took place between June and October 2017. The Data come from wards administrative records and the analysis is performed using weekly data (N=463).

Methods: Causal inference methods (Interrupted Time Series and Difference-in-Difference) are used to compare staff sickness rates before and after the implementation, where the outcome variable is defined as the ratio of total sickness hours over the total scheduled working hours (full time equivalents) in a given week. Patient casemix, staff demographics, ward and time variables are included as controls.

Results: Estimation results establish that the extended shifts are associated with an increased percentage of sickness hours per week of between 0.73% and 0.98%, the equivalent of a complete shift per week per ward.

Conclusion: This is the first study to use causal inference to measure the impact of longer shifts on sickness absences for mental health workforce. The analysis is relevant to other providers which may increasingly look towards these shift patterns as a means of cost saving.

Keywords: England; health workforce; mental health providers; nurses; shift patterns; sickness absence.

Fig. 1

Change in policy timing per ward. The solid horizontal line is the local polynomial smooth of the dotted scatter values of percentage of hours of sickness absence up to 7 days using a triangle kernel function; the shaded area represents the 95% confidence interval around it and the vertical line indicates the timing of the introduction of the policy.

Fig. 2

Sickness absence in percentage up to 7 days, before and after policy implementation. Policy timing cut-off standardised at time zero. The solid horizontal line is the local polynomial smooth of the dotted scatter values using a triangle kernel function; the shaded area represents the 95% confidence interval around it and the vertical line indicates the timing of the introduction of the policy.

Fig. 3

Difference-in-Differences identification strategy: The policy timing varies, three wards (A, B and C) introduced the 12-hour shifts in June, two in September (wards D and E) and one in October 2017 (ward F). Between June 2017 to September 2017 there are three wards affected by the policy and three wards unaffected by the policy. The solid horizontal line is the local polynomial smooth of the dotted scatter values using a triangle kernel function; the shaded area represents the 95% confidence interval around it and the vertical lines indicate the timing of the introduction of the policy.

Fig. A.1

Placebo test, introduction of the policy one year before. The solid horizontal line is the local polynomial smooth of the dotted scatter values using a triangle kernel function; the shaded area represents the 95% confidence interval around it and the vertical lines indicate the timing of the introduction of the policy.