Professional Documents
Culture Documents
Patient safety climate and worker safety behaviours in acute hospitals in Scotland
Cakil Agnew, Rhona Flin , Kathryn Mearns
Industrial Psychology Research Centre, School of Psychology, University of Aberdeen, Aberdeen AB24 3UB, Scotland, UK
a r t i c l e
i n f o
Article history:
Received 17 May 2012
Received in revised form 13 December 2012
Accepted 24 January 2013
Available online 11 February 2013
Keywords:
Safety climate
Safety compliance
Safety participation
Worker and patient injuries
a b s t r a c t
Objectives: To obtain a measure of hospital safety climate from a sample of National Health Service (NHS)
acute hospitals in Scotland and to test whether these scores were associated with worker safety behaviors,
and patient and worker injuries. Methods: Data were from 1,866 NHS clinical staff in six Scottish acute
hospitals. A Scottish Hospital Safety Questionnaire measured hospital safety climate (Hospital Survey on
Patient Safety Culture), worker safety behaviors, and worker and patient injuries. The associations between
the hospital safety climate scores and the outcome measures (safety behaviors, worker and patient injury
rates) were examined. Results: Hospital safety climate scores were signicantly correlated with clinical
workers safety behavior and patient and worker injury measures, although the effect sizes were smaller
for the latter. Regression analyses revealed that perceptions of stafng levels and managerial commitment
were signicant predictors for all the safety outcome measures. Both patient-specic and more generic safety
climate items were found to have signicant impacts on safety outcome measures. Conclusion: This study
demonstrated the inuences of different aspects of hospital safety climate on both patient and worker safety
outcomes. Moreover, it has been shown that in a hospital setting, a safety climate supporting safer patient
care would also help to ensure worker safety. Impact on industry: The Scottish Hospital Safety Questionnaire
has proved to be a usable method of measuring both hospital safety climate as well as patient and worker
safety outcomes.
2013 National Safety Council and Elsevier Ltd. All rights reserved.
1. Background
1.1. Patient safety in Scotland
Recent research in healthcare has tended to focus on iatrogenic
injury to hospital patients but signicant numbers of healthcare
staff can also experience workplace injuries. In 20102011, 1,649
major injuries and 9,741over-3-day injuries to UK healthcare
employees were reported (HSE, n.d.). A National Health Service
(NHS) staff survey revealed that 19% of staff reported seeing at least
one error or incident that could have hurt staff, and 25% of staff had
witnessed at least one error or near miss that could have hurt patients
(Healthcare Commission, 2007).
Scotland has an NHS, similar to that in England, with rates of
adverse events for patients in acute hospitals of approximately 8%
(Williams et al., 2008), comparable to other countries. In 2007, the
Health Department launched a Scottish Patient Safety Alliance, a
national initiative to improve patient safety in acute hospitals, with
the aim of reducing adverse events by 30% and deaths by 15% in a
four year period, along with clinical targets (SPSP, n.d.) This was one
of the rst initiatives targeting patient safety on a national scale and a
key objective was, to drive a change in the safety culture in NHS orga-
0022-4375/$ see front matter 2013 National Safety Council and Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jsr.2013.01.008
96
Vlahov, Felknor, 1995; Naveh, Katz-Navon, & Stern, 2005; Neal et al.,
2000; Smith et al., 2010) or for the safety of patients (Sorra & Nieva,
2006). Huang et al. (2010) found an association, across 30 ICU units,
where lower safety climate scores related to increased length of stay
for patients, and less favorable perceptions of management by staff
were related to higher patient mortality rates. However in a survey of
staff in 30 hospitals in the USA, no evidence of relationships between
senior management's engagement or unit managers support for patient safety and patient safety indicators (e.g. hospital discharge data)
were observed (Rosen, Singer, Zhao, 2010).
Zohar, Livne, Tenne-Gazit, Admi, and Donchin (2007) assessed 955
Israeli hospital nurses safety behaviors through observational techniques, and showed both unit and hospital level safety climates (using
both generic and patient focused items) were predictors of workers safety behaviors. Similarly, a study of 789 hospital workers in the USA, found
that when senior management support for worker safety, safety feedback
and training were perceived favorably, workers experienced fewer blood
and body uid exposure incidents (Gershon et al., 2000). In Japan, a more
positive safety climate was associated with safety of workers such as
reduced needlestick and sharp injuries (Smith et al., 2010).
Thus the inuence of hospital safety climate on patient and worker safety outcomes is not entirely clear, even though they appear to
have common causal factors (Flin, 2007). Staff perceptions of generic
safety climate (without a specic focus on patient care) were related
to treatment errors in one Israeli acute hospital (Naveh et al., 2005).
Few studies have measured both patient iatrogenic injuries and
staff occupational injuries, but common associations are beginning
to emerge. Hofmann and Mark (2006) in a study of 1127 nurses
from 42 hospitals in the USA found that safety climate predicted
both patient outcomes (medication errors, urinary tract infections)
and nurse outcomes (back injuries, needlestick). The complexity of
patient conditions exerted a moderating effect. More recently,
Taylor et al. (2012) studied 723 nurses from 29 units in one hospital
and found that two safety climate factors were associated with
nurse injuries and patient adverse events (decubitus ulcer). A stafng
factor (turnover) was found to be a particular risk factor. Therefore, it
appears that when the safety climate is associated with safer patient
care, it may also be associated with better safety for workers.
1.3. Aim
The rst aim of this study was to test which dimensions of hospital
safety climate were associated with patient and worker safety outcome
measures for a Scottish sample. Three outcome measures were used:
(a) clinical workers self-reported safety behaviors, specically, safety
compliance and participation; (b) self-reports of worker errors affecting
patients (i.e. patient injury); and (c) self reports of worker injuries.
Positive associations between safety climate scores and self-reports of
desirable workers safety behaviors were expected, as well as negative
associations between safety climate scales and both worker and patient
injury rates. The second aim was to examine the inuence of hospital
climate perceptions relating to patient care versus more general safety
aspects, separately for both patient and worker-related safety outcomes. We expected that a more favorable safety climate focusing on
patient care would be associated with, not only reduced patient injuries,
but also reduced worker injuries. Similarly, both patient-specic safety
perceptions and generic safety climate would be positively related to
workers safety-related behaviors.
2. Method
2.1. Procedure
All 14 NHS Health Boards in Scotland were contacted and asked
to provide an acute hospital for the study, and eight agreed to
participate. One had a low response, another had a high rate of
incomplete data, and they were both excluded. Therefore, the questionnaire survey was conducted with six hospitals from different
regions of Scotland (during 2009). Paper questionnaires (plus covering letter and envelope for return) were sent to each participating
hospital. Participants were instructed to return the questionnaires
to the research team or to the collection point within the hospital
unit. No names were requested to enhance anonymity.
Advice obtained from the UK National Research Ethics Service (NRES)
was that this study was a Service Evaluation and therefore would not
require an NRES ethics application. Ethical approval was obtained from
the ethical committee in the authors academic department.
2.2. Sample
The sample consisted of 1866 clinical staff from six NHS acute
hospitals in Scotland, with an estimated 23% response rate (Table 1).
Although the numbers of questionnaires sent to each hospital were
known and used as the denominator, it was not clear how many questionnaires were actually distributed. The calculated response rate per
hospital ranged from 12% to 31% (probably an underestimate in some
cases, as it later transpired that not all the delivered questionnaires
had been distributed to staff in some units).
2.3. Measure
Scottish Hospital Safety Questionnaire (SHSQ): A questionnaire
was designed for Scottish NHS clinical staff which measured hospital
safety climate and safety outcomes for both workers and patients. The
SHSQ constituted of four components: the 44 items of the Hospital
Survey on Patient Safety Culture (HSOPSC), plus 10 workers safety
behavior items, two items measuring self-reported worker and
patient injuries, and seven demographic questions (see additional
le 1: SHSQ).
2.3.1. Safety climate
The HSOPSC, developed in the USA (Sorra & Nieva, 2006), was
selected as it covers 12 dimensions of safety climate (e.g. hospital
management's commitment to safety, supervisory practices), two of
which are labeled as safety outcome measures (Overall Perceptions of
Safety and Incident reporting). It also contains two single items
labeled as safety outcome measures (Patient safety grade; Number
of incidents reported). Conrmed as 12 factors for this Scottish sample
(Sarac, Flin, Mearns, & Jackson, 2011), each dimension of safety climate
was assessed by three or four items measured on a 5 point Likert scale,
ranging from strongly disagree to strongly agree, and for Incident
reporting, the scale ranged from never to always. The Patient safety
grade item was measured on a 6 point Likert scale ranging from excellent to failing and the Incident reporting item was assessed on a 6
point Likert scale ranging from no incident reports to 21 or more
incident reports. This climate scale was chosen as it has been used
extensively in northern Europe, (Blegen, Gearhart, O'Brien, Sehgal, &
Alldredge, 2009; Hellings, Schrooten, Klazinga, & Vleugels, 2007;
Mardon, Khanna, Sorra, Dyer, & Famolaro, 2010; Olsen, 2010; Pfeiffer
& Manser, 2010; Smits, Dingelhoff, Wagner, van der Wal, &
Groenewegen, 2008) and was recommended by the European Society
Table 1
Overall response rates per Board (one hospital).
Boards
Hospital size
N returned
A
B
C
D
E
H
L
M
L
L
L
S
380
219
250
398
526
93
20
22
12
27
26
31
for Quality in Healthcare (2010). It was decided not to include the four
HSOPSC variables identied as safety outcome measures in the regression analysis, as these variables did not appear to be very robust measures of outcome. Instead, safety behavior and injury items derived
from industrial safety research were incorporated, as explained below.
2.3.2. Worker safety behaviors
Health care workers safety behaviors can be accessed through
observational techniques (Zohar et al., 2007). However, since observations can be difcult to gather in hospitals, self report measures
are frequently used to assess safety behaviors such as workers safety
compliance and safety participation behaviors (Neal et al., 2000) Positive associations between safety climate and self-reported measure
of desirable workers safety behaviors were shown from a sample of
525 employees in an Australian hospital (Neal et al., 2000). Later,
improvements in these behaviors at the group level were linked to
a reduction in future accident rates (Neal & Grifn, 2006).
The HSOPSC did not contain a safety behavior scale. For this
reason, a ten item scale was included to measure self reports of
workers safety compliance and participation behaviors (rated on a
ve-point scale ranging from strongly disagree to strongly
agree). Safety participation (Cronbach's = .77) was assessed by 4
items from Neal and Grifn's scale (Neal et al., 2000); an example
item is; I put in extra effort to improve the safety of the workplace.
For safety compliance (Cronbach's = .82), 6 items were incorporated from safety research on offshore oil installations (Mearns et al.,
2003) and reworded for healthcare workers; an example of a negatively scored item is I get the job done better by ignoring some
rules. The rules for healthcare staff include behaviors such as hand
washing and reporting incidents (Flin, 2007).
2.3.3. Patient and worker injuries
As it was not possible to obtain hospital-recorded patient or worker injury data for this study, in order to measure injuries experienced
both by the workers and patients, two self-report items were used.
The rst item (based on the Offshore Safety Questionnaire (Mearns
et al., 2001)) asked how often the individual had experienced a
work-related injury in this hospital, in the last 12 months. This was
rated on a scale from 0 / None (1) to 5 or more (4). A second question
asked about the number of witnessed errors that had harmed a patient in the last 12 months, rated on a scale from 0 / None (1) to 15
or more (5), with six options for indicating the reason for the last incident witnessed (based on the question used in the UK NHS Staff
Survey (Aston Business School, 2007).
The rst section asked for biodata: experience within the current
occupation, organization and work area, unit, and the nal section
provided an open space for comments.
2.4. Statistical Analyses
Data analysis was performed using SPSS (version 18). Sample
characteristics, composite mean scores and the average percentage
of positive responses were calculated following the reverse coding
the negatively worded items. Next, in order to test the hypotheses
that the higher scores on individual safety climate components
(HSOPSC) were associated with increased workers safety behavior
(i.e. higher safety compliance and participation), and lower rates of
self-reported experienced worker and witnessed patient injury
rates, Pearson correlation coefcients were calculated and stepwise
regression analyses (stepwise selection method) were conducted by
entering 10 safety climate dimensions (HSOPSC) as predictors since
it provides the most parsimonious model (Field, 2009) The same
procedure was completed for each of the four outcome measures.
Finally, to assess the impact of climate perceptions related to patient
care separately from more general safety climate on patient and
worker-related safety outcomes, we calculated two composite scores.
97
The rst composite score consisted of climate items (n = 15) that explicitly mentioned aspects of patient care e.g. After we make changes
to improve patient safety, we evaluate their effectiveness. For the
second composite score, we used the generic items (n = 20) that
had no specic focus on patient care, e.g. Whenever pressure builds
up, my supervisor wants us to work faster, even if it means taking
shortcuts. We then explored the relationships between workers
safety behaviors, worker and patient related outcomes with these
two safety climate composite scores (targeting patient care versus
generic safety climate items) using hierarchical regression models.
3. Results
3.1. Sample characteristics
Nurses constituted the majority of the sample (53%) followed by
Allied Health Professionals (22%), Nursing or Healthcare Assistants
(13%), and Medical and Dental consultants (12%). Regarding the
work area/unit of the respondents, the majority (22%) was from surgical units, followed by medicine (17%). A total of 37% of the participants had worked more than ten years within their current hospital,
and 32% worked between 15 years. Regarding their current profession, 30% had more than 21 years of experience. The majority of the
staff (74%) worked 2039 hours, 17% 4059 hours, and 1.7% worked
more than 60 hours per week. A total of 93% of the respondents
reported having direct contact with patients.
3.2. Descriptive ndings
The composite mean scores and the average percentage of positive
responses were computed for each HSOPSC dimension and outcome
variables (a higher score indicates a more positive response), see
Table 2. Results indicated that Teamwork within units, Supervisors
expectations and Organizational learning dimensions were rated
favorably. The highest agreement (73%) was reported for Teamwork
within the units, but Teamwork across units had only 39% positive response rate. Less favorable opinions on stafng levels (45% positive),
and feedback about error within their work unit suggest possible
areas for improvement in relation to patient safety. The scores are
generally comparable with those reported for hospitals in other
northern European countries (cited above).
Regarding the safety outcomes, 81% of the staff reported
complying with the safety rules (M = 4.02, SD = 0.66) and 75% indicated participating in safety activities (M = 3.85, SD = 0.59) 0.66).
For worker injuries, 75% of the participants reported no injuries in
the last 12 months, while 21% reported 1 to 2, 4% reported 3 or
Table 2
Means and standard deviations for HSOPSC scores.
HSOPSC SCALES / (Number of items
SD
3.6 (65%)
3.6 (64%)
3.7 (73%)
3.5 (54%)
3.3 (45%)
3.2 (44%)
3.2 (45%)
0.78
0.64
0.76
0.75
0.89
0.85
0.72
3.0 (38%)
3.0 (39%)
3.2 (32%)
0.83
0.70
0.64
3.6 (56%)
3.4 (56%)
0.94
0.76
98
3.4. The inuence of patient safety climate and generic safety climate on
safety-related outcomes
3.3. The associations between the HSOPSC dimensions and the safety outcome measures
In order to examine the associations between HSOPSC climate
dimensions and safety outcome measures, Pearson inter-correlation
coefcients were calculated between the 12 dimensions, for the two
HSOPSC outcome measures (patient safety grade and number of incidents reported), the two safety behavior measures, and the two items
measuring the frequency of worker and patient-related injuries in the
last 12 months (see additional le 2: Inter-correlation coefcients between HSOPSC scales and the outcome measures). The signicant
correlation coefcients between the 12 climate dimensions ranged
between r = .19 and .77 (p b .001), most showing a moderate effect.
For the correlations between the 12 climate dimensions and the safety behaviors: safety compliance and safety participation, the coefcients ranged between r=.07 and .44 (pb .001). Although signicant,
the effect sizes were smaller (ranged between r=.04 and -.32) when
examined in relation to self-reported worker and patient injuries.
To test the relative inuence of safety climate dimensions on
self-reports of workers safety behaviors and worker and patient injuries, as well as to determine the effect size, stepwise regression analyses
were performed. (Examination of variance ination factors and tolerance statistics indicated that multi-collinearity was not an issue). A
total of 10 predictor (climate dimensions) and 4 criterion variables
were included in the analysis. As can be seen in Table 3, stafng levels
and hospital management's support consistently predicted all the outcome variables. Every criterion measure was predicted by both the
unit and hospital level safety climate dimensions. The safety climate
scores were positively related to self-reported behavioral measures
with two exceptions; non-punitive response to error and stafng. The
strongest predictor of safety compliance behavior was the stafng levels
dimension (accounting for 15% of variance). For safety participation
behavior, the dimension of organizational learning was strongest (accounting for 10% of the variance). Only three of the climate dimensions
were negatively related to self-reports of worker and patient injury rates.
The climate dimensions of stafng, communication openness, and
management support explained 6% of variance in worker injuries. For
patient injuries, management support, stafng and teamwork across
units explained 13% of the variance, (the most signicant was Management support (Adj R 2 = .10).
4. Discussion
This study is rst to explore the clinical staff's perceptions of safety
within a sample of Scottish acute hospitals. It used a specially designed
questionnaire, employing the HSOPSC as the main component of the instrument. Although the HSOPSC has been used widely, to date, very few
Table 3
Stepwise Regression Analyses: HSOPSC dimensions as Predictors.
Dependent variables
Predictors
Adjusted R2
Std
Safety compliance
Stafng
Supervisors Expectations
Management Support
Handovers
Communication Openness
Organizational learning
Feedback & Communication
Stafng
Communication Openness
Non-punitive response
Teamwork Within Units
Management Support
Stafng
Communication Openness
Management Support
Management Support
Stafng
Teamwork Across Units
.383
.438
.465
.478
.481
.312
.327
.345
.349
.353
.356
.359
.225
.237
.243
.321
.350
.360
.146
.191
.215
.227
.229
.097
.106
.118
.120
.122
.124
.125
.051
.056
.059
.103
.121
.128
0.17
0.13
0.12
0.12
0.05
0.23
0.06
0.11
0.06
0.06
0.05
0.04
0.14
0.05
0.04
0.13
0.16
0.13
.19
.15
.15
.12
.06
.25
.10
-.13
.07
-.08
.06
.05
-.18
-.07
-.06
-.15
-.16
-.12
Safety participation
Worker injuries
Patient injuries
99
Table 4
Hierarchical Regression Results for Safety compliance, safety participation, worker and patient injuries.
Safety compliance
Safety participation
Worker injuries
Predictor variables
Std
Std
Std
Std
-.01
4.03
-.001
-.01
3.87
0.003
-.02
1.33
0.01
.03
1.50
0.01
2.95
-.004
0.001
.17**
0.20
.07
0.08
2
R = .01 for Step 1
R2 = .05** for Step 2
Patient injuries
2.19
-.03
0.01
-.08*
0.09
-.15**
0.17
2
R = .00 for Step 1
R2 = .05** for Step 2
2.99
.01
0.004
-.28**
0.40
-.03
0.05
2
R = .00 for Step 1
R2 = .09** for Step 2
100
also important to note the cross-sectional study design and the use of
retrospective self-reported injury data. For this reason, it was not possible to ascertain the direction of the causality.
4.3. Conclusions and impact on industry
The Scottish Hospital Safety Questionnaire with the combination of
climate, behavioral and outcome measures was found to produce an informative data set on the level and components of the hospitals safety
culture. The resulting prole revealed areas of strength but also of concern, relating to stafng levels and hospital management's commitment
to safety, factors which were associated with poorer safety outcomes.
Based on these ndings, it is suggested that healthcare organizations
need to ensure strong managerial commitment to safety and address
stafng decits in order to achieve the desired level of safety. Finally,
by demonstrating the impact of staff perceptions of patient safety climate on the safety of both patients and workers, this study illustrates
that a safety climate supporting patient care should also help to ensure
the safety of clinical workers.
Competing interest
None declared.
Author contributions
CA, RF & KM were responsible for the study conception and
design. CA performed the data collection, the data analysis and was
responsible for the drafting of the manuscript. CA, RF & KM made
critical revisions to the paper for important intellectual content. RF
obtained funding.
Funding
This research was funded by a Scottish Funding Council Strategic
Research Development Grant to the Scottish Patient Safety Research
Network.
Acknowledgements
We would like to thank all the NHS Scotland staff who gave their
time to complete our questionnaire.
References
Aston Business School (2007). NHS Staff Survey. Retrieved from www.nhsstaffsurveys.com
Blegen, B. A., Gearhart, S., O'Brien, R., Sehgal, N., & Alldredge, B. (2009). AHRQ's hospital
survey on patient safety culture: Psychometric analyses. Journal of Patient Safety, 5,
139144.
Blegen, M. A., Goode, C. J., Spetz, J., Vaughn, T., & Park, S. H. (2011). Nurse stafng effects on
patient outcomes: safety-net and non-safety-net hospitals. Medical Care, 49, 406414.
Christian, M. S., Bradley, J. S., Wallace, J. C., & Burke, M. J. (2009). Workplace safety: a
meta-analysis of the roles of person and situation factors. Journal of Applied
Psychology, 94, 11031127.
Clarke, S. (2006). The relationship between safety climate and safety performance: A
meta-analytic review. Journal of Occupational Health Psychology, 11, 315327.
Colla, J., Bracken, Kinney, L., & Weeks, W. (2005). Measuring patient safety: a review of
surveys. Quality & Safety in Health Care, 14, 364366.
European Society for Quality in Healthcare (2010). Use of patient safety culture instruments and recommendations. EUNetPas Project Report, Aarhus, Denmark. Ref Type:
Report.
Field, A. P. (2009). Discovering statistics using SPSS (3rd ed.). London: Sage.
Flin, R. (2007). Measuring safety culture in healthcare: A case for accurate diagnosis.
Safety Science, 45, 653667.
Flin, R., Burns, C., Mearns, K., Yule, S., & Robertson, E. M. (2006). Measuring safety
climate in health care. Quality & Safety in Health Care, 15, 109115.
Flin, R., Mearns, K., O'Connor, & Bryden, R. (2000). Measuring safety climate: Identifying the common features. Safety Science, 34, 177192.
Francis, R. (2013). Report of the Mid Staffordshire NHS Foundation Trust Public Inquiry.
London: The Stationery Ofce.
Gershon, R. R. M., Karkashian, C. D., Grosch, J. W., Murphy, L., Escamilla-Cejudo, A.,
Flanagan, P., et al. (2000). Hospital safety climate and its relationship with safe
101
Appendix 1
Inter-correlation coefcients between HSOPSC scales and the outcome measures.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Supervisors Expectations
Organizational Learning
Teamwork Within Units
Communication Openness
Feedback & Communication
Non-punitive Response
Stafng
Management Support
Teamwork Across Units
Hospital Handovers
Incident Reporting
Overall Perceptions of Safety
Incidents Reported
Patient Safety Grade
Safety Compliance
Safety Participation
Worker Injuries
Patient Injuries
10
11
12
13
14
15
16
17
.51
.47
.53
.51
.44
.44
.40
.32
.30
.29
.53
-.05*
.48
.36
.19
-.15
-.18
.52
.45
.52
.36
.35
.36
.32
.22
.33
.49
.07
.48
.23
.30
-.09
-.14
.50
.44
.43
.40
.32
.40
.28
.22
.48
-.05
.47
.25
.19
-.13
-.15
.61
.46
.37
.32
.31
.29
.31
.45
-.02
.46
.30
.21
-.16
-.15
.38
.37
.42
.36
.26
.38
.45
-.04
.48
.29
.24
-.12
-.19
.47
.30
.30
.28
.19
.44
-.03
.35
.25
.07
-.17
-.13
.43
.35
.36
.20
.77
-.13
.51
.39
.03ns
-.23
-.26
.51
.34
.24
.48
-.06
.47
.35
.14
-.16
-.27
.43
.19
.42
-.12
.39
.26
.10
-.14
-.25
.22
.36
-.06*
.36
.30
.11
-.14
-.25
.31
.12
.33
.27
.23
-.04
-.11
-.10
.63
.44
.11
-.17
-.32
-.08
-.09
.15
.19
.28
.41
.16
-.18
-.30
.22
-.14
-.25*
.06*
.01ns
.19
Cakil Agnew was a Research Fellow at the University of Aberdeen, where she received
her doctoral degree in Applied Psychology. Her research interests include safety culture and leadership behaviours in healthcare.
Rhona Flin is Professor of Applied Psychology and Director of the Industrial
Psychology Research Centre at the University of Aberdeen. Her research