Upload
sutanto-map-msc
View
216
Download
0
Embed Size (px)
Citation preview
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
1/7
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 b s t r a c ta 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
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 staf ng 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 2010–2011, 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-
nizations’ (SPSP, n.d.). Safety culture can be dened as “the product of
individual and group values, attitudes, perceptions, competencies and
patterns of behavior that determine the commitment to, and the style
and prociency of, an organization's health and safety management”
(Schein, 2004) with safety climate as a surface manifestation (Flin,
Mearns, O'Connor, & Bryden, 2000), a ‘snapshot’ of the prevailing safety
culture, typically measured by questionnaires.
1.2. Patient safety climate and safety outcomes
Using questionnaire measures, safety climate has been shown to
be related to safety outcomes in a number of industrial settings
(Clarke, 2006; Christian, Bradley, Wallace, & Burke, 2009; Neal &
Grif n, 2006; Neal, Grif n, & Hart, 2000). For example, a meta-
analysis (Christian et al., 2009) demonstrated a link between better
safety climate scores and lower archival worker accident data and
self-reported accident/injuries.
Several safety climate instruments have been developed to
assess hospital staff's perceptions of workplace safety (for reviews see
Colla, Bracken, Kinney, & Weeks, 2005; Flin, Burns, Mearns, Yule, &
Robertson, 2006; Jackson, Sarac, & Flin, 2010). Management commit-
ment to safety has been identied as a dominant theme in safety
climate measurement within the industrial safety literature (Flin et al.,
2000). Healthcare organizationsalso includethis factor in safetyclimate
assessment, using items focusing on management's support for
the safety of workers (Gershon, Karkashian, Grosch, 2000; Gershon,
Journal of Safety Research 45 (2013) 95–101
⁎ Corresponding author. Tel.: +44 1224 273212.
E-mail address: [email protected] (R. Flin).
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
Contents lists available at SciVerse ScienceDirect
Journal of Safety Research
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j s r
http://dx.doi.org/10.1016/j.jsr.2013.01.008http://dx.doi.org/10.1016/j.jsr.2013.01.008http://dx.doi.org/10.1016/j.jsr.2013.01.008mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.jsr.2013.01.008http://www.sciencedirect.com/science/journal/00224375http://www.sciencedirect.com/science/journal/00224375http://dx.doi.org/10.1016/j.jsr.2013.01.008mailto:[email protected]://dx.doi.org/10.1016/j.jsr.2013.01.008
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
2/7
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 pa-
tient 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 tech-
niques, and showed both unit and hospital level safety climates (using
bothgenericand patientfocused items) were predictors of workers’ safe-
ty behaviors. Similarly, a study of 789 hospital workers in the USA, found
thatwhen senior managementsupport for worker safety, safety feedback
and training were perceived favorably, workers experienced fewer blood
andbodyuid exposureincidents (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 work-
er 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 staf ng
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-reportsof 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 safetyaspects, separately for both patient and worker-related safety out-
comes. We expected that a more favorable safety climate focusing on
patient care would be associatedwith, 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 ques-
tionnaire survey was conducted with six hospitals from different
regions of Scotland (during 2009). Paper questionnaires (plus cover-
ing 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 obtainedfromthe UK National Research Ethics Service (NRES)
was that this study was a Service Evaluation and therefore would notrequire 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 ques-
tionnaires 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 excel-
lent 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 Response rate (%)
A L 380 20
B M 219 22
C L 250 12
D L 398 27
E L 526 26
H S 93 31
* L ≥500 beds, M=250 –
499 beds, S=50 –
249 beds.
96 C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
3/7
for Quality in Healthcare (2010). It was decided not to include the four
HSOPSC variables identiedas ‘safety outcome’ measures in the regres-
sion analysis, as these variables did not appear to be very robust mea-
sures 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 obser-vations can be dif cult 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) Pos-
itive 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 & Grif n, 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 α=.77) was assessed by 4
items from Neal and Grif n'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 α=.82), 6 items were incorporat-
ed from safety research on offshore oil installations (Mearns et al.,
2003) and reworded for healthcare workers; an example of a nega-
tively 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 work-
er 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 wasrated 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 pa-
tient 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 in-
cident 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 percentageof 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 coef cients 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.
The rst composite score consisted of climate items (n=15) that ex-
plicitly 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 versusgeneric 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 sur-
gical units, followed by medicine (17%). A total of 37% of the partici-
pants had worked more than ten years within their current hospital,
and 32% worked between 1–5 years. Regarding their current profes-
sion, 30% had more than 21 years of experience. The majority of the
staff (74%) worked 20–39 hours, 17% 40–59 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 re-
sponse rate. Less favorable opinions on staf ng levels (45% positive),
and feedback about error within their work unit suggest possibleareas 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% indi-
cated 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 M SD
Safety Climate Dimensions (Unit level)Supervis ors' expectations and actions (4) 3.6 (65%) 0.78
Or ganizat io nal learning - impro vement ( 3) 3.6 (64%) 0.64
Teamwork within hospital units (4) 3.7 (73%) 0.76
Communication openness (3) 3.5 (54%) 0.75
Feedback and communication about err or ( 3) 3.3 (45%) 0.89
Non-punitive response to error (3) 3.2 (44%) 0.85
Staf ng (4) 3.2 (45%) 0.72
Safety Climate Dimensions (Hospital level)
Hospital management support for patient safety (3) 3.0 (38%) 0.83
Teamwork across hospital units (4) 3.0 (39%) 0.70
Hospital handovers (4) 3.2 (32%) 0.64
HSOPSC Outcome Measures
Frequency of incident reporting (3) 3.6 (56%) 0.94
Overall perceptions of safety (4) 3.4 (56%) 0.76
*The average percentages of positive responses are presented in parenthesis.
97C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
4/7
more incidents. For witnessing incidents harming patients, 40% of the
staff reported witnessing 1–5 incidents in the last year, 3% reported
6–10, 2% reported 11 or more incidents. More than half of the sample
(54%) had not witnessed any incidents harming patients.
3.3. The associations between the HSOPSC dimensions and the safety out-
come measures
In order to examine the associations between HSOPSC climatedimensions and safety outcome measures, Pearson inter-correlation
coef cients were calculated between the 12 dimensions, for the two
HSOPSC outcome measures (patient safety grade and number of inci-
dents 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 coef cients be-
tween HSOPSC scales and the outcome measures). The signicant
correlation coef cients between the 12 climate dimensions ranged
between r=.19 and .77 (pb .001), most showing a moderate effect.
For the correlations between the 12 climate dimensions and the safe-
ty behaviors: safety compliance and safety participation, the coef -
cients 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 inju-
ries, as well as to determine the effect size, stepwise regression analyses
were performed. (Examination of variance ination factors and toler-
ance 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, staf ng levels
and hospital management's support consistently predicted all the out-
come 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 staf ng. The
strongest predictor of safetycompliance behavior wasthe staf ng levels
dimension (accounting for 15% of variance). For safety participation
behavior, the dimension of organizational learning was strongest (ac-counting 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 staf ng, communication openness, and
management support explained 6% of variance in worker injuries. For
patient injuries, management support, staf ng and teamwork across
units explained 13% of the variance, (the most signicant was Man-
agement support (Adj R 2=.10).
3.4. The in uence of patient safety climate and generic safety climate on
safety-related outcomes
In order to examine the inuence of staff perceptions of climate
relating to patient care versus generic safety climate on safety related
outcomes for workers and patients, two composite scores were calcu-
lated. Items were selected in relation to their content. For the patient
safety climate perceptions (Cronbach's α=.82), we used 15 items fo-
cusing on patient care (Mean=3.34 SD=0.50) and for the genericitems (Cronbach's α =.89), we calculated the mean of the 20 items
with no specic focus on patient care (Mean= 3.28 SD= 0.54). Hier-
archical regression analyses (Table 4) were carried out to examine
the unique contribution of safety climate perceptions in predicting
worker and patient safety outcomes after controlling for the hospital
effect. Variance ination factors (VIF) and tolerance statistics were
examined among predictors and covariates, and were determined
not to be indicative of multicollinearity (The tolerance statistics
ranged from .361 to .997, and the maximum VIF was 2.77).
The results showed that the perceptions of patient safety climate
were signicantly related to both worker and patient injuries, and
to workers’ safety compliance and participations behaviors (see
Table 4). When more positive perceptions of patient safety climate
were reported, less patient and worker injuries were observed. On
the other hand, more positive patient safety climate was related to
workers’ increased safety compliance and participation behaviors. Re-
sults were mixed for the perceptions of safety climate with no specic
focus on patient care. Although, these generic safety climate scores
were signicantly and negatively related to worker injuries, no such
effect was found for patient injuries. Similarly, workers’ safety com-
pliance behaviors were found to be signicantly increased with more
positive generic safety climate scores, no such an inuence was found
on workers’ safety participation behaviors.
Overall, the results indicated that the two safety climate scores
explained a very small amount of the variance in safety participation
and self reported worker (R 2=.05) and patient injuries (R 2=.09),
and safety participation behaviors (R 2=.05). On the other hand,
both types of safety climate perception explained 21% of the total var-
iance for workers’ safety compliance.
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 componentof the in-
strument. Although the HSOPSC has been used widely, to date, very few
Table 3
Stepwise Regression Analyses: HSOPSC dimensions as Predictors.
Dependent variables Predictors R Adjusted R 2 B Std β
Safety compliance Staf ng .383 .146 0.17 .19
Supervisors’ Expectations .438 .191 0.13 .15Management Support .465 .215 0.12 .15
Handovers .478 .227 0.12 .12
Communication Openness .481 .229 0.05 .06
Safety participation Organizational learning .312 .097 0.23 .25
Feedback & Communication .327 .106 0.06 .10
Staf ng .345 .118 −0.11 -.13
Communication Openness .349 .120 0.06 .07
Non-punitive response .353 .122 −0.06 -.08
Teamwork Within Units .356 .124 0.05 .06
Management Support .359 .125 0.04 .05
Worker injuries Staf ng .225 .051 −0.14 -.18
Communication Openness .237 .056 −0.05 -.07
Management Support .243 .059 −0.04 -.06
Patient injuries Management Support .321 .103 −0.13 -.15
Staf ng .350 .121 −0.16 -.16
Teamwork Across Units .360 .128 −0.13 -.12
Note: All effects are signicant at pb
.05.
98 C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
5/7
studies (e.g. Olsen, 2010) examined its associations with additional
measures, such as self-reported safety behaviors and injury rates.
The descriptive data illustrated areas of strength and of concern re-
lating to safety climate within each participating hospital. For example,
the unit level dimensions were found to receive more favorable re-
sponses compared to hospital level safety climate scales. Specically,
scales concerned with supervisory practices, improvement efforts at
the unit level, and teamwork were signicantly higher than the rest of
the sub-scales. At the unit level, the staf ng dimension was a perceived
challenge by the respondents. Respondents seemed to be more cynical
about their hospital management but were more favorable about their
own work area, for example, handovers across units were perceived
less positively than teamwork within units. Similar to previous studies
showing respondents reporting more compliance with the safety rules
compared to voluntary safety-related activities (Neal & Grif n, 2006;
Neal et al., 2000), results showed that overall safety compliance re-
ceived higher scores compared to safety participation scale.
In respect to self-reported injury rates, more than half of the re-
spondents (54%) reported not witnessing any patient injuries. Most
respondents (75%) reported not experiencing any work-related inju-
ries. An earlier NHS survey (Healthcare Commission, 2007), found
79% of staff reported not seeing any incidents that could have hurtstaff and 75% had not seen any incidents or near- misses that could
have hurt patients, in the last 12 months.
4.1. Associations between HSOPSC scores and the self-reported outcome
measures
The associations between the HSOPSC dimensions and the
safety outcome variables were examined via regression analyses. The
results revealed that the climate dimensions of staf ng and hospital
management's support were signicantly related to every outcome
measure. The impact of staf ng adequacy in hospitals has been the sub-
ject of a major investigation where staf ng cuts andtheir impact on pa-
tient care in an English hospital have been highlighted (Francis, 2013).
Inadequate staf ng levels and staff workload have also been identiedas key variables determining outcomes such as hospital mortality
rates (Needleman,Buerhaus, & Pankratz, 2011), nurseburnout, dissatis-
faction (Holden et al., 2011) and prolonged length of stay (Blegen,
Goode, Spetz, Vaughn, & Park, 2011) We also expected that increased
safety participation would be reported in relation to better staf ng
levels. In fact, the results showed decreased ratherthan increased safety
participation when the staf ng levels were perceived favorably. Re-
spondents who are experiencing staff shortages and higher workload
maytake short cuts and comply less with the safety protocols.However,
in the short term, they mayengagein more voluntarysafetyactivities in
order to compensate forthe negative effects of staff shortageson patient
care. In this sense, a good teamwork climate might help to mitigate the
adverse consequences of perception of inadequate staf ng (Siassakos et
al., 2011). Favorable staf ng levels on the other hand, might contribute
to diffusion of responsibility withinthe workgroups; leading to workers
failing to take individual action and therefore reporting decreased vol-
untary safety activities.
The observed effects of the climate scales on worker and patient-
related injuries were smaller than the effects of climate on the behav-
ioral measures of safety compliance and safety participation. This
nding is supported by a meta-analysis (Christian et al., 2009) show-
ing safety climate perceptions were a more proximal measure of safe-
ty behaviors than of worker injuries. However, as mentioned above,
75% participants did not report any work-related injuries, a rate which
is not dissimilar to other industries (Mearns et al., 2001).
Finally, we found that patient-specic safety climate was related to
both worker and patient related outcomes, whereas generic safety cli-
mate scores only had an impact on worker safety compliance behaviors
and worker injury rates. The association of patient safety climate with
both worker and patient safety outcomes replicates the ndings of
Hofmann and Mark (2006) and Taylor et al. (2012), indicating a poten-
tial common causal effect. Future research should test facets of safety
climate against specic types of worker behaviors (as in this study)
and also against particular categories of worker and patient injuries,
extending the approach taken by Taylor et al. (2012).
4.2. Limitations
One of the limitations of the current study was the low response
rate. In order to maximize response rates, an online version of the
questionnaire was prepared and a condential feedback report was
offered to the participating hospitals. Although a reasonable sample
size was achieved, the overall estimated response rate remained
low (23%), and so there is a risk of selection bias in that the percep-
tions of safety culture reported might not represent the views of
non-respondents (Groves & Peytcheva, 2008). Rogelberg, Luong,
Sederburg, and Cristol (2000) found that workers who do not comply
to organizational survey requests show lower organizational commit-
ment, and less satisfaction with supervisors. In future, additional ef-
fort should be made to obtain a higher level of local support andmanagerial involvement prior to the data collection, as well as more
consultation with the government health department to reduce the
risk of scheduling conicts with other surveys.
For the comparison of generic and patient-specic safety climates,
the items were selected thematically and then the internal consistency
of the derived scales was established. This was an exploratory exercise
and we recognize that this is a preliminary analysis that would require
replication, ideally with independent generic (e.g. from industry or
worker safety research) and patient safety-specic climate scales.
Perceptions of safety climate factors and on self-reported safety out-
comes (safety behaviors and worker and patient injury rates) were col-
lected froma singlesourcewhere thereis a riskof commonmethod bias
(Podsakoff & Organ, 1986), that can inate the relationships between
the variables (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). It is
Table 4
Hierarchical Regression Results for Safety compliance, safety participation, worker and patient injuries.
Safety compliance Safety participation Worker injuries Patient injuries
Predictor variables Std β B Std β B Std β B Std β B
First step: control variables
Constant 4.03 3.87 1.33 1.50
Hospital -.01 -.001 -.01 −0.003 -.02 −0.01 .03 0.01
Second step: Independent variables
Constant 1.93 2.95 2.19 2.99
Hospital .01 0.004 -.004 −0.001 -.03 −0.01 .01 0.004
Patient safety climate .37** 0.49 .17** 0.20 -.08* −0.09 -.28** −0.40
Generic safety climate .13** 0.13 .07 0.08 -.15** −0.17 -.03 −0.05
R 2=.00 for Step 1
ΔR 2=.21** for Step 2
R 2=.01 for Step 1
ΔR 2=.05** for Step 2
R 2=.00 for Step 1
ΔR 2=.05** for Step 2
R 2=.00 for Step 1
ΔR 2=.09** for Step 2
**pb .01, *pb .05.
99C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
6/7
also important to note the cross-sectional study design and the use of
retrospective self-reported injury data. For this reason, it was not pos-
sible 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 in-
formative data set on the level and components of the hospitals’ safetyculture. The resulting prole revealed areas of strength but also of con-
cern, relating to staf ng levels and hospital management'scommitment
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
staf ng decits in order to achieve the desired level of safety. Finally,
by demonstrating the impact of staff perceptions of patient safety cli-
mate 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.comBlegen, 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, 5,139–144.
Blegen, M. A., Goode, C. J., Spetz, J., Vaughn, T., & Park, S. H. (2011). Nurse staf ng effects onpatient outcomes: safety-net and non-safety-net hospitals. Medical Care, 49, 406–414.
Christian, M. S., Bradley, J. S., Wallace, J. C., & Burke, M. J. (2009). Workplace safety: ameta-analysis of the roles of person and situation factors. Journal of Applied
Psychology, 94, 1103–
1127.Clarke, S. (2006). The relationship between safety climate and safety performance: A
meta-analytic review. Journal of Occupational Health Psychology, 11, 315–327.Colla, J., Bracken, Kinney, L., & Weeks, W. (2005). Measuring patient safety: a review of
surveys. Quality & Safety in Health Care, 14, 364–366.European Society for Quality in Healthcare (2010). Use of patient safety culture instru-
ments 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, 653–667.Flin, R., Burns, C., Mearns, K., Yule, S., & Robertson, E. M. (2006). Measuring safety
climate in health care. Quality & Safety in Health Care, 15, 109–115.Flin, R., Mearns, K., O'Connor, & Bryden, R. (2000). Measuring safety climate: Identify-
ing the common features. Safety Science, 34, 177–192.Francis, R. (2013). Report of the Mid Staffordshire NHS Foundation Trust Public Inquiry.
London: The Stationery Of ce.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
work practices and workplace exposure incidents. American Journal of InfectionControl, 28, 211–221.
Gershon, R., Vlahov, D., Felknor, S., et al. (1995). Compliance withuniversal precautionsamong health care workers at three regional hospitals. American Journal of InfectionControl, 23, 225–236.
Groves, R. M., Peytcheva, E., Vesley, D., Johnson, P., Delcios, G., et al. (2008). The impactof non-response rates on non-response bias: A meat-analysis. Public Opinion Quar-terly, 72, 167–189.
Healthcare Commission (2007). National NHS Staff survey, 2007. Retrieved from. http:// www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_ndings_200804183620.pdf
Hellings, J., Schrooten, W., Klazinga, N., & Vleugels, A. (2007). Challenging patientsafety culture: survey results. International Journal of Health Care Quality Assurance, 20, 620–632.
Hofmann, D., & Mark, B. (2006). An investigation of the relationship between safety cli-mate and medication errors as well as other nurse and patient outcomes. PersonnelPsychology, 59, 847–869.
Holden,R. J., Scanlon,M. C., Patel, N.R., Kaushal, R.,Escota,K. M., Brown,R. L.,et al. (2011).A human factors framework and study of the effect of nursing workload on patientsafety and employee quality of working life. Quality & Safety in Health Care, 20, 15–24.
Health and Safety Executive (HSE) (n.d.). Retrieved March 2012 from https://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15.
Huang, D. T., Clermont, G., Kong, L., Weissfeld, L. A., Sexton, J. B., Romn, K. M., et al.(2010). Intensive care unit safety culture and outcomes: a US multicenter study.International Journal of Quality in Health Care, 22, 151–161.
Jackson, J., Sarac, C., & Flin, R. (2010). Hospital safety climate surveys: measurementissues. Current Opinion in Critical Care , 16 , 632–638.
Mardon, R., Khanna, K., Sorra, J., Dyer, N., & Famolaro, T. (2010). Exploring relationshipsbetween hospital patient safety culture and adverse events. Journal of Patient Safety,6 , 226–232.
Mearns, K., Flin, R., Gordon, R., & Fleming, M. (2001). Human and organizational factorsin offshore safety. Work & Stress, 15, 144–160.
Mearns, K., Whitaker, S., & Flin, R. (2003). Safety climate, safety management practiceand safety performance in offshore environments. Safety Science, 41, 641–680.
Naveh, E., Katz-Navon, T., & Stern, Z. (2005). Treatment errors in healthcare: A safetyclimate approach. Management Science, 51, 948–960.
Neal, A., & Grif n, M. A. (2006). A study of the lagged relationships among safetyclimate, safety motivation, safety behavior, and accidents at the individual andgroup levels. Journal of Applied Psychology, 91, 946–953.
Neal, A., Grif n, M. A., & Hart, P. M. (2000). The impact of organizational climate onsafety climate and individual behaviour. Safety Science, 34, 99–109.
Needleman, J., Buerhaus, P., & Pankratz, S. (2011). Nurse Staf ng and Inpatient HospitalMortality. The New England Journal of Medicine, 364, 1037–1045.
Olsen, E. (2010). Exploring the possibility of a common structural model measuringassociations between safety climate factors and safety behaviour in healthcareand the petroleum sectors. Accident Analysis and Prevention, 42, 1507–1516.
Pfeiffer, Y., & Manser, T. (2010). Development of the German version of the HospitalSurvey on Patient Safety Culture: Dimensionality and psychometric properties.Safety Science, 48, 1452–1462.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common methodbiases in behavioral research: A critical review of the literature and recommendedremedies. Journal of Applied Psychology, 88, 879–903.
Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: prob-lems and prospects. Journal of Management , 12, 531–544.
Rogelberg, S., Luong, A., Sederburg, M., & Cristol, D. (2000). Employee attitude surveys:Examining the attributes of noncompliant employees. Journal of Applied Psycholo-
gy, 85, 284–293.Rosen, A. K., Singer, S., Zhao, S., Shokeen, P., Meterko, M., Gaba, D., et al. (2010). Hospi-
tal safety climate and safety outcomes: is there a relationship in the VA? MedicalCare Research and Review, 67 , 590–608.
Sarac, C.,Flin,R., Mearns, K.,& Jackson, J. (2011).HospitalSurveyon Patient SafetyCulture:Psychometric Analysis on a Scottish Sample. BMJ Quality & Safety, 20, 842–848.
Schein, E. H. (2004). Organizational Culture and Leadership (pp. 3–63) (3rd ed.). SanFrancisco, CA:Jossey-Bass.
Siassakos, D., Fox, R., Hunt, L., Farey, J., Laxton, C., Winter, C., et al. (2011). Attitudes tosafety and teamwork in a maternity unit with embedded team training. American
Journal of Medical Quality, 26 , 132–137.
Smith, D. R., Muto, T., Sairenchi, T., Ishikawa, Y., Sayama, S., Yoshida, A., et al. (2010).Hospital safety climate, psychosocial risk factors and needlestick injuries in
Japan. Industrial Health, 48, 85–95.Smits, M., Dingelhoff, I. C., Wagner, C., van der Wal, G., & Groenewegen, W. P. (2008).
The psychometric properties of the 'Hospital Survey on Patient Safety Culture' inDutch hospitals. BMJ Health Services Research, 8, 230.
Sorra, J., & Nieva, V. (2006). Reliability and validity of the Hospital Survey on Patient Safety. Rockville, MD: Westat.
Scottish Patient Safety Programme (n.d.). Retrieved December 2010, from http://www.patientsafetyalliance.scot.nhs.uk/programme/
Taylor, J., Dominici, F., Agnew, J., Gerin, D., Morlock, L., & Miller, M. (2012). Do nurseand patient injuries share common antecedents? An analysis of associations withsafety climate and working conditions. BMJ Quality and Safety, 21, 101–111.
Williams, D., Olsen, S., Crichton, W., Witte, K., Flin, R., Ingram, J., et al. (2008). Detectionof adverse events in a Scottish hospital using a consensus-based methodology.Scottish Medical Journal, 53, 29–33.
Zohar, D., Livne, Y., Tenne-Gazit, O., Admi, H., & Donchin, Y. (2007). Healthcare climate:A framework for measuring and improving patient safety. Critical Care Medicine,
35, 1312–1317.
100 C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101
http://www.nhsstaffsurveys.com/http://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttps://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15https://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15https://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15http://www.patientsafetyalliance.scot.nhs.uk/programme/http://www.patientsafetyalliance.scot.nhs.uk/programme/http://www.patientsafetyalliance.scot.nhs.uk/programme/http://www.patientsafetyalliance.scot.nhs.uk/programme/https://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15https://handson.hse.gov.uk/hse/public/tablesimple.aspx?RID15http://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.cqc.org.uk/_db/_documents/National_NHS_staff_survey_2007_summary_of_key_findings_200804183620.pdfhttp://www.nhsstaffsurveys.com/
8/16/2019 Patient Safety Climate and Worker Safety Behaviours
7/7
Cakil Agnew was a Research Fellow at the University of Aberdeen, where she receivedher doctoral degree in Applied Psychology. Her research interests include safety cul-ture and leadership behaviours in healthcare.
Rhona Flin is Professor of Applied Psychology and Director of the IndustrialPsychology Research Centre at the University of Aberdeen. Her research
interests include non-technical skills in safety critical occupations and safetyculture.
KathrynMearns was a Senior Lecturer at the University of Aberdeen. She has specic in-terests in risk management and safety culture (which she has studied in the oil industryand air traf c management) and is now a human factors inspector for a safety regulator.
Appendix 1
Inter-correlation coef cients between HSOPSC scales and the outcome measures.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1. Supervisors’ Expectations -
2. Organizational Learning .51 -
3. Teamwork Within Units .47 .52 -
4. Communication Openness .53 .45 .50 -
5. Feedback & Communication .51 .52 .44 .61 -
6. Non-punitive Response .44 .36 .43 .46 .38 -
7. Staf ng .44 .35 .40 .37 .37 .47 -8. Management Support .40 .36 .32 .32 .42 .30 .43 -
9. Teamwork Across Units .32 .32 .40 .31 .36 .30 .35 .51 -
10. Hospital Handovers .30 .22 .28 .29 .26 .28 .36 .34 .43 -
11. Incident Reporting .29 .33 .22 .31 .38 .19 .20 .24 .19 .22 -
12. Overall Perceptions of Safety .53 .49 .48 .45 .45 .44 .77 .48 .42 .36 .31 -
13. Incidents Reported -.05* .07 -.05 -.02 -.04 -.03 -.13 -.06 -.12 -.06* .12 -.10 -
14. Patient Safety Grade .48 .48 .47 .46 .48 .35 .51 .47 .39 .36 .33 .63 -.08 -
15. Safety Compliance .36 .23 .25 .30 .29 .25 .39 .35 .26 .30 .27 .44 -.09 .41 -
16. Safety Participation .19 .30 .19 .21 .24 .07 .03ns .14 .10 .11 .23 .11 .15 .16 .22 -
17. Worker Injuries -.15 -.09 -.13 -.16 -.12 -.17 -.23 -.16 -.14 -.14 -.04 -.17 .19 -.18 -.14 .06* -
18. Patient Injuries -.18 -.14 -.15 -.15 -.19 -.13 -.26 -.27 -.25 -.25 -.11 -.32 .28 -.30 -.25* .01ns .19
Note: All effects are signicant at pb .001, *pb .05, ns: Non-signicant.
101C. Agnew et al. / Journal of Safety Research 45 (2013) 95–101