Content uploaded by Dean Smith
Author content
All content in this area was uploaded by Dean Smith on Nov 11, 2019
Content may be subject to copyright.
Downloaded from https://journals.lww.com/hcmrjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD36i5Q5e8D5BvZwI9+0fntCI5gYU8ySQBaywsh+NvKQe4= on 11/11/2019
Downloadedfromhttps://journals.lww.com/hcmrjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD36i5Q5e8D5BvZwI9+0fntCI5gYU8ySQBaywsh+NvKQe4= on 11/11/2019
Hospital staffing patterns and safety
culture perceptions: The mediating
role of perceived teamwork and
perceived handoffs
Soumya Upadhyay
Robert Weech-Maldonado
Christy H. Lemak
Amber L. Stephenson
Dean G. Smith
Background: As hospitals are under increasing pressure to improve quality and safety, safety culture has become a
focal issue for high-risk organizations, including hospitals. Prior research has examined how structural characteristics
directly impact safety culture. However, and based on Donabedian’s structure–process–outcome quality model, there
is a need to understand the processes that intermediate the relationship between structural characteristics and safety
culture perceptions.
Purpose: The processes by which registered nurse (RN) and hospitalist staffing may affect safety culture perceptions
were examined in this study. Specifically, this study investigates the processes of perceived teamwork across units and
perceived handoffs.
Methodology: Data sources for this research included Hospital Survey on Patient Safety Culture from the Agency for
Healthcare Research and Quality, the American Hospital Association’s Annual Survey Data, the American Hospital
Association Information Technology supplement, and the Area Health Resource File. Two separate mediation models
Key words: handoffs, staffing, safety culture, teamwork
Soumya Upadhyay, PhD, MHA, is Assistant Professor, Department of Healthcare Administration and Policy, School of Public Health, University of
Nevada at Las Vegas. E-mail: Soumya.upadhyay@unlv.edu.
Robert Weech-Maldonado, PhD, MBA, is Professor, Department of Health Services Administration, School of Health Professions, University of
Alabama at Birmingham.
Christy H. Lemak, PhD, is Professor, Department of Health Services Administration, School of Health Professions, University of Alabama at
Birmingham.
Amber L. Stephenson, PhD, MPH, is Assistant Professor, The David D. Reh School of Business, Schenectady, New York.
Dean G. Smith, PhD, is Professor, School of Public Health, Louisiana State University Health Sciences Center, New Orleans.
This article was presented at the 2018 Academy of Management, Health Care Management Division Annual Conference, held in Chicago, IL, on
August 13, 2018.
The authors have disclosed that they have no significant relationship with, or financial interest in, any commercial companies pertaining to this article.
DOI: 10.1097/HMR.0000000000000264
Health Care Manage Rev, 2019 00(0), 00–00
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
Month–Month •2019 1
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
for each process were used. Propensity weights were assigned to each hospital in the sample (N=207)toadjustfor
potential nonresponse bias of hospitals that did not assess employee’s safety culture perceptions.
Results: Results suggest that RN staffing influences safety culture perceptions, but hospitalist staffing does not. In
addition, RN staffing has an indirect effect on safety culture perceptions through better processes.
Practice Implications: Our studysheds light on how staffing affects safety culture perceptions. Specifically, our findings
suggest that positive perceptions of teamwork across units and handoffs are integral in the relationship between
RN staffing and safety culture perceptions. Hospital managers should, therefore, invest resources in staff recruitment
and retention. In addition, a targeted focus on perceived teamwork and handoffs may allow hospital managers to
improve safety culture perceptions.
Safety culture is an important strategy for organiza-
tions and an imperative precursor to filling the wide-
spread gaps in patient safety outcomes in the health
care system (Pronovost & Sexton, 2005). The Joint
Commission (2017) defines patient safety culture as the
“product of individual and group beliefs, values, attitudes,
perceptions, competencies, and patterns of behavior that
determine the organization’s commitment to quality and
patient safety.”A key feature of safety culture is perceptions
among employees regarding the importance of patient
safety in their organizations and serves as a building block
for far reaching distal outcomes such as reduced mortalities,
hospital acquired infections, complications, and read-
missions (Fan et al., 2016). Hospitals are under pressure
to improve patient safety and safety culture, particularly
given the increasing attention to this issue by the Institute
of Medicine and the Affordable Care Act (The Patient
Protection and Affordable Care Act, 2010; Wachter, 2004).
In response to this challenge, hospitals should consider
changing specific structural characteristics, which would
improve processes, subsequently resulting in better out-
comes (Donabedian, 1988).
Although a growing body of literature has supported the
role of structural characteristics, registered nurse (RN) and
hospitalist staffing, as practical steps toward improving safety
and safety culture perceptions (Ford, Silvera, Kazley, Diana,
& Huerta, 2016; McHugh et al., 2016), there has been a
dearth of research exploring the intricacies of the relation-
ship between structural characteristics and safety culture
perceptions. The impact of structural characteristics such
as RN staffing and hospitalist staffing on safety culture per-
ceptions may be dependent, in part, on specific intermedi-
ary mechanisms that need to be affected to influence a
positive change on safety culture perceptions. Two of these
potential intermediate mechanisms, perceived teamwork
across units, and perceived handoffs are considered in this
study as process characteristics. Although there is research
on the effects of teamwork and handoffs on safety culture
perceptions (Choi & Staggs, 2014; Greenstein, Arora,
Staisiunas, Banerjee, & Farnan, 2013), there is a need to
further understand how these concepts may mediate the
relationship between hospital structures and desired quality
and safety outcomes. Therefore, the purpose of this study is
to examine how structural characteristics may enhance
processes—specifically perceived teamwork across units
and handoffs—which in turn may improve the outcome
of safety culture perceptions.
Conceptual Framework
This article uses Donabedian’s(1988)structure–process–
outcome quality framework to conduct a three-pronged
approach to assessment of quality. Structure refers to attri-
butes of the environment where care occurs, process refers
to the actions involved in the giving and receiving of care,
and outcome refers to the effects of care on the patients and
broader populations (Donabedian, 1988). This model as-
sumes that good structure makes it more likely to have ro-
bust processes, which in turn increase the chances of having
better care outcomes (Donabedian, 1988). In Figure 1, we
introduce perceived teamwork across units and perceived
handoffs as mediators (process) to explain the mechanisms
that connect RN and hospitalist staffing (structure) and
safety culture perceptions (outcome) more completely.
Staffing, a structural characteristic that includes attri-
butes of human resources, plays an essential role in the pro-
vision of high-quality and safe care (Alenius, Tishelman,
Runesdotter, & Lindqvist, 2014; Brennan, Daly, & Jones,
2013). RN staffing, an essential precursor to the quality of
patient care, is associated with patient outcomes like mor-
tality rates, falls, and pressure ulcers (Cho et al., 2015).
Similarly, hospitalist staffing influences patient outcomes
such as mortalities, readmission rates, complications, and
length of stay (Chin, Wilson, Bang, & Romano, 2014; Epane
& Weech-Maldonado, 2015).
Teamwork across units and handoffs are care coordinat-
ing mechanisms done to and for the patient during the de-
livery of care and were therefore chosen as processes that
provide mediating mechanisms (O'Leary, Sehgal, Terrell,
& Williams, 2012; O'Malley, Draper, Gourevitch, Cross,
& Scholle, 2015). According to the National Quality Forum
(2012), care coordination implies a need for meaningful
2Health Care Management Review Month–Month •2019
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
communication and cooperation among providers for
patients to receive efficient, safe, and high-quality care.
Teamwork across units refers to collaboration and coordi-
nation between two or more individuals across hospital
units (Kalisch & Lee, 2009). For teamwork across units
to be effective, nurses and physicians need to interact in-
terdependently and dynamically toward a common goal of
safety to avoid breakdowns in communication (Horwitz
et al., 2013; O'Leary et al., 2012). Previous studies have
shown that loopholes in communication due to lack of
collaboration has resulted in retained sponges, mismatched
blood transfusion, incorrect extremity nerve blocks and
surgery sites, and sentinel events resulting in catastrophic
consequences for patients and their families (Weaver, Callaghan,
Cooper, Brandman, & O'Leary, 2014).
A handoff is a process by which the responsibility for a
patient passes from one health care professional to another,
to exchange accurate information about patients’care, cur-
rent condition, expected changes, and treatment and ser-
vices needed (Cohen & Hilligoss, 2010). As a point of
transition, handoffs are known to introduce vulnerabilities
in communication and continuity of care (Greenstein et al.,
2013; Riesenberg, Leisch, & Cunningham, 2010). Incon-
sistency in handoff processes may present clinical errors
and potential adverse events (Riesenberg et al., 2010). In
expressing concerns for the risk of errors and adverse
events, the Joint Commission has highlighted the impor-
tance of standardized approaches to handoffs and informa-
tion transfer (Greenstein et al., 2015). Retaining important
patient information during handoffs is critical in meeting
patient safety goals (Cohen & Hilligoss, 2010; Riesenberg
et al., 2010). In the subsequent sections, we explain the
relationships of (a) structure and outcomes, (b) structure
and process, and (c) the mediating role of process between
structure and outcome.
Structure and Outcomes
Extant research suggests two key structural characteristics,
which may affect safety culture perceptions: RN staffing
and hospitalist staffing. RN staffing refers to having ade-
quate numbers of RNs scheduled to handle the workload
as well as ensuring shift schedules (or the number of hours
per shift) for RNs are appropriate toprovide the best care to
patients (Cho et al., 2015). Low or inadequate RN staffing
ratio has been associated with adverse patient safety out-
comes, such as in-hospital falls, medication errors, pressure
ulcers, and hospital-acquired infections (McHugh & Ma,
2014; Pettker et al., 2011). Furthermore, RN staffing ratio
has been positively associated with RN’sassessmentofpa-
tient safety culture at their hospital (Alenius et al., 2014;
Brennan et al., 2013; Cho et al., 2015). To examine the
direct association between RN staffing and safety culture
perceptions, we suggest the following hypothesis:
Hypothesis 1a: RN staffing ratio is positively related
to safety culture perceptions.
Hospitalists are specialists in hospital medicine that have
served a vital role in the growing patient safety needs amidst
theincreasingcomplexityofcare of hospitalized patients.
One advantage of having hospitalists is that they are readily
available to detect and treat inpatient problems and have
a role in preventing or reducing morbidities (West et al.,
2014). In addition, high hospitalist staffing intensity has
been associated with a lower length of stay (Epane &
Figure 1
Conceptual framework
Hospital Staffing and Safety Culture 3
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Weech-Maldonado, 2015), which minimizes the probabil-
ity of having infections and complications. To examine the
direct relationship between hospitalist staffing and safety
culture perceptions, we hypothesize:
Hypothesis 1b: Hospitalist staffing ratio is positively
related to safety culture perceptions.
Structure and Process
As noted above, we have considered two structural char-
acteristics, (a) RN staffing ratio and (b) hospitalists
staffing ratio, and two processes, (a) perceived teamwork
across units and (b) perceived handoffs. Perceived team-
work refers to the discernment of the quality of collabora-
tion and cooperation among providers within a designated
group (Weaver et al., 2014), whereas perceived hand-
offs refer to the assessment of the quality of the handoff
exchange.
Research on teamwork across intensive care units, oper-
ating rooms, and in general medical contexts has confirmed
that discrepancies exist in the perceptions of quality of col-
laboration and cooperation (e.g., teamwork), which may
associate with appropriate staffing levels (Weaver et al.,
2014). Higher RN staffing, or at least minimum nurse
staffing where shortages do not negatively affect care of
patients and increase burden of other employees, may im-
prove cooperation across team members, for example, pro-
viding support in terms of physical assistance (e.g., helping
another team member ambulate a patient) or through ver-
bal reminders (e.g., reminding a fellow team member to
wash their hands) by ensuring the availability of other team
members when need arises and preventing staff from being
overburdened (Kalisch & Lee, 2009). Low nurse staffing
groups when compared to high nurse staffing groups had
better processes (Alenius et al., 2014). Therefore, at least
a minimum or higher RN staffing ratio is needed so that
teams can effectively coordinate activities across units to
exchange critical information about patients (Riesenberg
et al., 2010; Tscholl et al., 2015).
Similarly, RN staffing ratios need to be adequate for
smoother handoffs. These ratios are especially important
during transfer of patients from physically and cognitively
demanding workspaces such as the Emergency Depart-
ment, where a balance between continuity of care and
restricting demands on health care professionals needs to
be maintained (McHugh et al., 2016). Ensuring at least
minimum nurse and patient ratio so that there are enough
RNs to overlap work shifts for seamless handoffs minimizes
the likelihood of RNs from the prior shift feeling over-
worked, thereby having ample time for rest and for the
handoff to the next RN, particularly in a time-constrained
environment (Choi & Staggs, 2014). Thus, we suggest the
following hypotheses:
Hypothesis 2a: RN staffing ratio is positively related
to perceived teamwork across units.
Hypothesis 2b: RN staffing ratio is positively related
to perceived handoffs.
Hospitalists have a role in improving teamwork across
units and handoffs given their broad clinical expertise
(Messler & Whitcomb, 2015). They are expected to play
a role in stimulating effective team coordination across
units by supporting an increase in the frequency of com-
munication, strengthening shared goals, and encouraging
mutual respect; these attributes are related to reduced post-
operative pain, improved postoperative functioning, and
decreased length of stay (Fan et al., 2016). In addition,
hospitalists provide verbal handoffs supported with docu-
mentation to other physicians for the exchange of impor-
tant information, such as a checklist of elements for an
ideal discharge of elderly and high-risk patients with multi-
ple comorbidities (Greenstein et al., 2013). These verbal
handoffs are particularly useful for retention of information
to facilitate better handoffs and minimize patient informa-
tion from falling through the cracks (Horwitz et al., 2013).
Thus, we suggest the following hypotheses:
Hypothesis 2c: Hospitalist staffing ratio is positively
related to perceived teamwork across units.
Hypothesis 2d: Hospitalist staffing ratio is positively
related to perceived handoffs.
Mediating Role of Process in the
Relationship Between Structure
and Outcome
Perceptions of teamwork across units and handoffs are in-
trinsic to the relationship between structural characteristics
and safety culture perceptions because they involve in-
terpersonal dynamics (Pronovost & Sexton, 2005; S. J.
Weaver et al., 2013). In hospitals, better care coordination
among staff leads to effective teamwork across units and
smoother handoffs, which can influence employees’safety
culture perceptions. This is because better care coordina-
tion minimizes confusion, prevents staff from disregarding
specific tasks to reduce their workload during busy periods,
and helps them prioritize important activities (Antonelli,
McAllister, & Popp, 2009). Staffing allows for better care
coordination so that workload distributes evenly and ex-
change of critical information about patients improves
knowledge about their health issues (Tscholl et al., 2015;
West et al., 2014). Therefore, we expect staffing to have
an indirect effect on safety cultural perceptions via perceived
4Health Care Management Review Month–Month •2019
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
teamwork across units and perceived handoffs. Thus, we sug-
gest the following hypotheses:
Hypothesis 3a: RN staffing ratio will have an indirect
positive effect on safety culture perceptions through
its positive effect on perceived teamwork across units.
Hypothesis 3b: Hospitalist staffing ratio will have an
indirect positive effect on safety culture percep-
tions through its positive effect on perceived team-
work across units.
Hypothesis 3c: RN staffing ratio will have an indi-
rect positive effect on safety culture perceptions
through its positive effect on perceived handoffs.
Hypothesis 3d: Hospitalist staffing ratio will have an
indirect positive effect on safety culture perceptions
through its positive effect on perceived handoffs.
Methods
Data and Sample
This study uses the following data sets: (a) the 2016 Hospi-
tal Survey on Patient Safety Culture (HSOPSC) data set
from the Agency for Healthcare Research and Quality
(AHRQ), (b) the American Hospital Association’s (AHA)
Annual Survey Data, (c) the AHA Information Technology
supplement, and (d) the Area Health Resource File. The
HSOPSC survey is a voluntary survey administered by hos-
pitals to the employees. Employees’perceptions on patient
safety culture are collected as part of these data. The 2016
HSOPSC database is a pooled cross-sectional data set that
contains survey data from years 2014 and 2015. Of the 680
hospitals that submitted their results to AHRQ, 207 pro-
vided identifiable information for research purposes. We
merged the HSOPSC data with the AHA data set, the
AHA Information Technology supplement, and the Area
Health Resource File data to obtain a final analytic sample
of 207 hospital observations. Each of these was a unique ob-
servation, and there were no repeated measures. Our study
obtained approval from the institutional review board.
Dependent Variable
The outcome represents the average positive score for “per-
ceptions of safety,”one of the safety culture dimensions
from the HSOPSC data. This measure has been psycho-
metrically validated in prior research (Blegen, Gearhart,
O'Brien, Sehgal, & Alldredge, 2009). The perceptions of
safety composite measure includes the following items: (a)
“It is just by chance that more serious mistakes don’thap-
pen around here.”(b) “Patient safety is never sacrificed to
get more work done.”(c) “We have patient safety problems
in this unit.”(d) “Our procedures and systems are good at
preventing errors from happening.”The average positive score
for the perceptions of safety culture composite measure
consists of the average of the percentages of positive re-
sponses to each item and is measured by aggregating unit-
level results to hospital level (Famolaro et al., 2016).
Independent Variables
To measure RN staffing ratio, we use the following formula
that normalizes the number of full-time employees (FTE):
(Number of registered nurse FTE / Total inpatient days) *
100. To measure hospitalist staffing ratio, we use the fol-
lowing formula that normalizes the number of FTE: (Num-
ber of hospitalist FTE / Total inpatient days) * 100.
Mediating Variables
We consider the average positive scores for two process
measures: (a) perceived teamwork across units and (b) per-
ceived handoffs and transitions. Psychometric analyses of
HSOPSC survey have validated these dimensions at the
hospital level (Blegen et al., 2009). The following four
questions are under the composite dimension “teamwork
across units’: (a) hospital units do not coordinate well with
each other, (b) there is good cooperation among hospital
units that need to work together, (c) it is often unpleasant
to work with staff from other hospital units, and (d) hospi-
tal units work well together to provide the best care for pa-
tients. “Hospital handoffs and transitions”is composed of
the following four questions: (a) things fell between the
cracks when transferring patients from one unit to another,
(b) important patient care information is often lost during
shift changes, (c) problems often occur in the exchange
of information across hospital units, and (d) shift changes
are problematic for patients in this hospital (Famolaro
et al., 2016). Individual employees were asked the above
questions at the unit level, and the results were aggregated
to obtain hospital-level composite scores.
Control Variables
The following control variables were used in this study: (a)
Ownership status (not-for-profit, for-profit, government),
(b) size (small, 0–99 beds; medium, 100–299 beds; and
large, 300 and above beds), (c) teaching status (hospitals
were coded as “teaching”if they were member of the Coun-
cil of Teaching Hospitals, or if they were affiliated to a med-
ical school, or if they provided a residency program), (d)
Hospital Staffing and Safety Culture 5
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
system membership (hospitals were coded as a system member
if they were a member or affiliate of a large health care sys-
tem), (e) proportion of Medicare patients (Hospital Medi-
care inpatient days / Hospital inpatient days * 100), (f )
proportion of Medicaid patients (Hospital Medicaid inpa-
tient days / Hospital inpatient days * 100), and (g) elec-
tronic health record (EHR) presence, categorized in three
ways: no EHR (no functionalities), basic EHR (10 function-
alities fully or partially implemented), and advanced EHR
(24 functionalities fully implemented).
We use the following as market-level control variables:
(a) market competition (Hirschman–Herfindahl index;
values ranged from 0 to 1, wherein1 indicates monopolistic
market and values close to 0 indicate highly competitive
markets). For Hirschman–Herfindahl index, market was
defined as the particular health services area to which the
hospital belongs to. To calculate market competition, we
used a two-step approach: (1) Market share = (Inpatient
days for hospital X/ Total inpatient days for all hospitals
in the market) and (2) Sum of square of market shares for
each market = (Σ(Market share)
2
); (b) location (metro
(a population of 250,000 to 1 million and more), urban
(a population of 2,500 to 20,000), and rural (less than a
population of 2,500); and (c) percent per capita income
(proxy for socioeconomic status of patients). For both loca-
tion and per capita income, market was defined as the
county in which the hospital is located.
Analytic Approach
Multivariable regression was used to test two mediation
models: (a) In Model 1, perceived teamwork across units
was used as a mediator, and (b) in Model 2, perceived hand-
offs were used as a mediator. Baron and Kenny’s (1986)
method of testing mediation was used, in which: First, the
dependent variable (outcome) was regressed on the inde-
pendent variable (structures); this regression was the direct
effect of structural characteristics on outcome. Second, the
mediators (processes) were regressed on the independent
variables (structures). Finally, the dependent variable (out-
come) was regressed on both the independent variable
(structure) and the mediator (process). This was the indi-
rect effect of structural characteristics on outcome. Equa-
tions for the first model were as follows:
Ysafety culture perception sðÞ
¼B0þB1RN staffing
þB2 hospitalist staffing þB3controls
Yperceived teamwork across unitsðÞ
¼B0þB1RN staffing
þB2 hospitalist staffing
þB3controls
Ysafety culture perceptionsðÞ
¼B0þB1RN staffing
þB2hospitalist staffing
þB3 perceived teamwork across units
þB4controls
A similar set of equations was used for the secondmodel.
We also ran the Sobel–Goodman test to assess what per-
centage of the total effect is being mediated. To adjust for
potential response bias of hospitals participating in the
HSOPSC, we included propensity score weights in the re-
gression analysis, which were calculated by calculating the
inverse of the propensity scores for hospitals that partici-
pated in the HSOPSC.
Results
Table 1 presents the descriptive statistics of all the variables
included in this study. Average RN staffing ratio per 100
hospital inpatient days (IPD) is higher at 1.3 than the aver-
age hospitalist staffing ratio per 100 hospital IPD at 0.04, as
expected. The average positive scores for perceived team-
work across units, perceived handoffs, and safety culture
perceptions were 60.7%, 47%, and 67.1%, respectively.
Table 2 presents the direct effect of staffing variables on
the outcome, safety culture perceptions. Hypothesis 1a was
supported, because there was a positive relationship be-
tween RN staffing and safety culture perceptions. Every ad-
ditional RN FTE per 100 hospital IPD is associated with a
3.1% increase in safety culture perceptions. Hypothesis 1b
was not supported. Hospitalist staffing ratio was not signif-
icantly associated with safety culture perceptions.
Table 3 presents regression results of processes (perceived
teamwork across units and perceived handoffs) on staffing
variables. For Model 1, Hypothesis 2a was supported. Re-
gression results show that RN staffing ratio is positively
associated with perceived teamwork across units. Every ad-
ditional RN FTE per 100 IPD is related to a 3.1% increase
in average positive score of perceived teamwork across units
(p< .001). On the other hand, Hypothesis 2b was not sup-
ported. Hospitalist staffing ratio was not significantly associ-
ated with perceived teamwork across units. With respect to
control variables, for-profit hospitals, as compared to non-
profit hospitals, tend to have a reduced perception of team-
work across units by 10.6%.
For Model 2, Hypothesis 2c was supported. Every addi-
tional RN FTE per 100 IPD is associated with a 2.6% in-
crease in perceived handoffs (p< .001). On the other hand,
Hypothesis 2d was not supported. Hospitalist staffing is not
significantly associated with perceived handoffs.
Finally, Table 4 presents the last step in mediation anal-
ysis, as suggested by Baron and Kenny (1986), the indirect
effect of staffing variables on safety culture perceptions via
6Health Care Management Review Month–Month •2019
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
process characteristics. In the first model, the outcome,
safety culture perceptions, is regressed on both RN staffing
and perceived teamwork across units. In this model, adding
perceived teamwork across units reduces the coefficient size
for RN staffing, but RN staffing remains significant. This
suggests that the relationship between RN staffing and
safety culture perceptions is partially mediated by perceived
teamwork across units, lending partial support to Hypothe-
sis 3a. Every additional RN FTE per 100 IPD is associated
with a 1.8% increase in safety culture perceptions when
the mediator-perceived teamwork across units is accounted
for (p< .01). Results from the Sobel–Goodman test of me-
diation show that the mediation effect of perceived team-
work across units was statistically significant and explained
approximately 68% of the total effect (of RN staffing ratio
on safety culture perceptions). Because there was no signif-
icant direct relationship between hospitalist staffing and
safety culture perceptions, the mediation model for hospi-
talist staffing was not considered. Therefore, Hypothesis
3b is not supported.
In Model 2, safety culture perceptions are regressed on
RN staffing and perceived handoffs. In this model, adding
perceived handoffs reduces the coefficient size for RN staff-
ing, but RN staffing remains significant, which shows par-
tial support for Hypothesis 3c. Every additional RN FTE
Table 1
Descriptive statistics of all variables in the
sample (N= 207)
Mean SD
Structural characteristics
RN staffing ratio (per 100 days) 1.37 1.12
Hospitalist staffing ratio (per
100 days)
0.04 0.05
Process measures
Perceived teamwork across units 60.7% 11.5%
Perceived handoffs 47.0% 10.3%
Outcome
Safety culture perceptions 67.1% 9.3%
Controls Frequency %
Organizational characteristics
Ownership
Not for profit (Ref ) 160 77.6%
For profit 6 2.9%
Government nonfederal 40 19.4%
Size
Small (Ref) 91 44.1%
Medium 54 26.2%
Large 61 29.6%
Teaching status
Teaching ( Ref ) 111 53.8%
Nonteaching 95 46.0%
System affiliation
System (Ref) 136 66.0%
Nonsystem 70 33.9%
EHR presence
No EHR (Ref) 22 10.6%
Basic EHR 103 50.2%
Advanced EHR 81 39.1%
Mean SD
Proportion Medicaid population 20.81 37.94
Proportion Medicare
population
51.42 42.73
Market characteristics Frequency %
Location
Metro (Ref) 144 69.9%
Urban 59 28.6%
Rural 3 1.4%
Mean SD
Market competition (HHI) 0.60 0.36
Per capita income (per 1,000) 44.08 11.59
Note. RN = registered nurse; Ref = reference; EHR = electronic hea lth
record; HHI = Hirschman–Herfindahl index.
Table 2
Regression results of the direct relationship
between RN and hospitalist staffing and safety
culture perceptions (N=207)
Safety culture perceptions
BSE
RN staffing ratio 3.17*** (0.62)
Hospitalists staffing ratio 7.52 (7.72)
EHR presence (Ref = No EHR)
Basic EHR 1.61 (3.19)
Advanced EHR 1.53 (3.18)
Ownership (Ref = Nonprofit)
Government nonfederal −2.79 (1.44)
Profit −4.22 (3.74)
Size (Ref = small)
Medium −1.5 (1.60)
Large −4.24 (2.54)
Teaching status (Ref = teaching)
Nonteaching 2.6 (1.63)
System membership (Ref = No)
Yes −2.90* (1.31)
Proportion Medicaid population −0.09*** (0.01)
Proportion Medicare population 0.05* (0.02)
Location (Ref = Metro)
Urban 0.81 (1.51)
Rural 1.33 (3.10)
Market competition (HHI) 0.3 (2.06)
Per capita income (per 1,000) 0.08 (0.05)
Note. RN = registered nurse; Ref = reference; EHR = electronic hea lth
record; HHI = Hirschman–Herfindahl index.
*p<.05.**p<.01.***p<.001.
Hospital Staffing and Safety Culture 7
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
per 100 IPD is related with a 2% increase in safety culture
perceptions when the mediator-perceived handoffs are in-
cluded (p< .01). Results from the Sobel–Goodman test
show that the mediation effect of handoff was statistically
significant and explained approximately 54% of the total
effect (of RN staffing ratio on safety culture perception).
Because the direct effect between hospitalist staffing and
safety culture perceptions was not significant, a mediation
model was not considered. This shows that Hypothesis 3d
is not supported. For both models, hospitals that belong
to a system are likely to have reduced safety culture percep-
tions as compared to hospitals that do not belong to a
system.
Discussion
The purpose of this study was to examine the association
between staffing and safety culture perceptions via pro-
cesses (perceived teamwork across units and perceived
handoffs). Findings suggest that RN staffing influences
safety culture perceptions, and RN staffing has an indirect
effect on safety culture perceptions through better pro-
cesses. However, hospitalist staffing was not associated with
safety culture perceptions.
Our observation regarding the positive linkage between
RN staffing and safety culture perceptions is consistent
with previous studies that have shown that a higher level
of RN staffing is related to better patient outcomes, includ-
ing lower hospital related mortality, decreased infections in
the intensive care unit, a shorter length of stay in the inten-
sive care units, and lower odds of several adverse patient
events (Choi & Staggs, 2014; Lee, Blegen, & Harrington,
2014). A higher level of RN staffing reflects a hospital’s
commitment to high-quality care, value for nurses, and an
effective nurse recruitment and retention strategy. A more
robust RN staffing ensures that nurses do not get over-
worked or exhausted, thereby avoiding potential medical
errors (McHugh & Ma, 2014). Likewise, given that daily
patient care activities require coordination, better RN
staffing may increase the likelihood of improved coordina-
tion across teams and proper handoffs. However, although
we have found that that there is a potential link between
high nurse staffing and better safety culture, there may be
Table 3
Regression results of the relationship between RN and hospitalist staffing and perceived teamwork
across units and perceived handoffs (N=207)
Perceived teamwork across units
(Model 1)
Perceived handoffs
(Model 2)
BSEBSE
RN staffing ratio 3.19*** (0.79) 2.62*** (0.73)
Hospitalists staffing ratio 10.85 (7.59) −2.16 (6.88)
EHR presence (Ref = No EHR)
Basic EHR −5.83 (3.72) −3.02 (4.64)
Advanced EHR −3.79 (3.89) −2.37 (4.7)
Ownership (Ref = Nonprofit)
Government nonfederal 0.042 (2.60) −0.85 (2.74)
Profit −10.67** (−3.31) −6.06 (3.18)
Size (Ref = Small)
Medium −0.77 (1.81) −0.60 (1.81)
Large −6.28 (3.01) −3.31 (2.90)
Teaching status (Ref = Teaching)
Nonteaching 2.73 (2.30) 2.41 (2.20)
System membership (Ref = No)
Yes 1.74 (1.92) 3.08 (1.90)
Proportion Medicaid population −0.04 (0.03) −0.07** (0.02)
Proportion Medicare population 0.04 (0.03) 0.09** (0.02)
Location (Ref = Metro)
Urban −2.21 (2.07) 1.00 (1.87)
Rural −3.36 (3.94) 3.90 (2.45)
Market competition (HHI) −1.60 (2.41) −1.71 (2.65)
Per capita income (per 1,000) 0.04 (0.09) 0.11 (0.08)
Note. RN = registered nurse; Ref = reference; EHR = electronic health record; HHI = Hirschman–Herfindahl index.
*p<.05.**p<.01.***p<.001.
8Health Care Management Review Month–Month •2019
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
apointwheremorenursesnolongerassociateswithsafety
culture. For example, a hospital that is constrained with fi-
nancial resources may impede access to patient’scarebydi-
verting funds from other beneficial uses into meeting high
RN ratio. This may not be in the best interest of patient
safety, especially in the case of safety net hospitals (Conway,
Konetzka, Zhu, Volpp, & Sochalski, 2008).
Our findings extend previous work by establishing desig-
nated processes as the intervening mechanisms between
staffing and safety culture perceptions. Partial mediation
by perceptions of teamwork across units and handoffs is
an indication that these processes provide a bridging link
between RN staffing and safety culture perceptions. Ade-
quate RN staffing is correlated to higher perceptions of
teamwork and handoffs, resulting in increased perceptions
of safety culture. With adequate staffing, RNs perform bet-
ter in teams (Kalisch & Lee, 2009). Alternatively, RN
staffing shortages leading to use of supplemental/temporary
nurses, who do not fully understand the policies and
procedures of the hospital, may cause safety concerns for
the patients (Alenius et al., 2014).
Another key finding was that hospitalist staffing ratio
does not have a direct or indirect relationship with safety
culture perceptions. One potential explanation for this is
that the question items in the AHRQ HSOPSC survey
may inquire about more nurse sensitive outcomes that
are outside the purview of hospitalists. For instance, given
the training and background of hospitalists, they may focus
on accreditation, technology, and malpractice systems as
important dimensions of safety. Another potential expla-
nation is that this survey usually has a greater participation
from RNs rather than physicians, including hospitalists.
The percentage of RN respondents in our sample is 20.9%,
whereas the percentage of hospitalist respondents is 2.5%.
The variation within hospitalist staffing ratio as compared
to that of RN staffing ratio was less, as shown in Table 1, which
may have accounted for lack of significant findings. Although
the role of hospitalists may be different than that of RNs, their
Table 4
Regression results of the indirect relationship between RN and hospitalist staffing and safety culture
perceptions through Model 1: Perceived teamwork across units and Model 2: Perceived
handoffs (N = 207)
DV = Safety culture perceptions Model 1 Model 2
B(SE)B(SE)
RN staffing ratio 1.83** (0.60) 2.04** (0.61)
Hospitalists staffing ratio 2.97 (6.03) 8.46 (6.78)
Perceived teamwork across units 0.42** (0.15)
Perceived handoffs 0.43*** (0.10)
EHR presence (Ref = No EHR)
Basic EHR 4.06 (2.35) 2.91 (3.26)
Advanced EHR 3.12 (2.34) 2.56 (3.12)
Ownership (Ref = Nonprofit)
Government nonfederal −2.80* (1.38) −2.42 (1.47)
Profit 0.26 (3.50) −1.61 (3.64)
Size (Ref = Small)
Medium −1.17 (1.30) −1.23 (1.31)
Large −1.60 (1.83) −2.81 (1.93)
Teaching status (Ref = Teaching)
Nonteaching 1.46 (1.37) 1.57 (1.32)
System membership (Ref = No)
Yes −3.64*** (1.07) −4.23*** (1.05)
Proportion Medicaid population −0.07*** (0.01) −0.06** (0.02)
Proportion Medicare population 0.03 (0.02) 0.01 (0.02)
Location (Ref = Metro)
Urban 1.73 (1.13) 0.38 (1.32)
Rural 2.74 (4.01) −0.37 (3.12)
Market competition (HHI) 0.97 (1.52) 1.03 (1.73)
Per capita income (per 1,000) 0.06 (0.05) 0.03 (0.05)
Note. RN = registered nurse; Ref = reference; EHR = electronic health record; HHI = Hirschman–Herfindahl index.
*p<.05.**p<.01.***p<.001.
Hospital Staffing and Safety Culture 9
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
work still requires interprofessional teamwork and handoffs
in an effort to serve their key role—coordination of patient
care within a hospital (Chesluk et al., 2012).
Our study has some limitations. First, because of the low
participation rate (30%) on the AHRQ HSOPSC survey
and low availability of identifiable data, our study sample
is small; this limits the generalizability of our findings to
the broader national hospital population. Second, self-
reported surveys may be predisposed to a methodological
issue, common method variance, which may alter associa-
tions among variables. We used composite scores based
on multiple items for patient safety culture to avoid bias
due to common method variance and deidentified data to
keep the anonymity of respondents.
Future studies should further investigate whether hos-
pitalists’perspectives about patient safety and safety culture
may be different as compared to that of other employees in-
cluding RNs and, if so, what effect it may have on overall
safety culture perception scores. In addition, performing
these analyses at the unit level to capture various nuances
of different units would be an interesting projection of this
hospital-level study. For example, some units may require
high-intensity work and may be very demanding, such as
intensive care unit or trauma as compared to low-intensity
units such as OB/GYN. An in-depth fine-grained explora-
tion of the above analyses at the unit level is a topic that
future studies should consider.
Practical Implications
Our study highlights the importance of processes that con-
nect the link between structural characteristics (staffing)
and proximal quality outcomes (safety culture perceptions).
Specifically, our study begins to illuminate “whether”
and “how”structural characteristics affect safety culture
perceptions.
There are implications of adequate staffing levels on
certain other areas such as workforce, systems, and health
care delivery. Staffing levels may affect stress, burnout,
and exhaustion among the workforce. Adequate staffing
levels may reduce stress associated with over exhaustion
and keeps motivation levels high among the workforce,
further affording better teamwork and handoffs (West
et al., 2014). Improvement in quality and safety requires
a focus on both individuals and systems; therefore, man-
agers should invest in time and resources toward staff re-
cruitment and retention as well as to train employees for
better care coordination to achieve a safety culture. High
turnover rates of the health care workforce impact pro-
cesses that are needed to execute patient care and pa-
tients’safety. Finally, from a policy standpoint, given the
economic and human costs associated with adverse medi-
cal events, health care systems need to be improved. Some
errors may not lead to harm, but they provide a learning
opportunity to identify system improvements for safety
to be designed within the health care delivery system. Al-
though the workforce needs to be vigilant and careful, im-
proving systems for better patient safety would provide
promising results in the years to come.
Acknowledgments
The authors would like to thank the numerous reviewers
who have provided valuable suggestions to improve this
article.
References
Alenius, L. S., Tishelman, C., Runesdotter, S., & Lindqvist, R.
(2014). Staffing and resource adequacy strongly related to
RNs’assessment of patient safety: A national study of RNs
working in acute-care hospitals in Sweden. BMJ Quality Safety,
23(3), 242–249.
Antonelli, R. C., McAllister, J. W., & Popp, J. (2009). Making
care coordination a critical component of the pediatric health
system: A multidisciplinary framework. The Commonwealth
Fund Pub. No. 1277.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator
variable distinction in social psychological research: Concep-
tual, strategic, and statistical considerations. Journal of Person-
ality and Social Psychology,51(6), 1173.
Blegen, M. A., Gearhart, S., O'Brien, R., Sehgal, N. L., &
Alldredge, B. K. (2009). AHRQ's Hospital Survey on Patient
Safety Culture: Psychometric analyses. Journal of Patient Safety,
5(3), 139–144.
Brennan, C. W., Daly, B. J., & Jones, K. R. (2013). State of the
science: The relationship between nurse staffing and patient
outcomes. Western Journal of Nursing Research,35(6), 760–794.
Chesluk, B. J., Bernabeo, E., Hess, B., Lynn, L. A., Reddy, S., &
Holmboe, E. S. (2012). A new tool to give hospitalists feed-
back to improve interprofessional teamwork and advance pa-
tient care. Health Affairs,31(11), 2485–2492.
Chin, D. L., Wilson, M. H., Bang, H., & Romano, P. S. (2014).
Comparing patient outcomes of academician-preceptors, hospitalist-
preceptors, and hospitalists on internal medicine services in
an academic medical center. Journal of General Internal Medi-
cine,29(12), 1672–1678.
Cho, E., Sloane, D. M., Kim, E. Y., Kim, S., Choi, M., Yoo, I. Y.,
…Aiken, L. H. (2015). Effects of nurse staffing, work envi-
ronments, and education on patient mortality: An observa-
tional study. International Journal of Nursing Studies,52(2),
535–542.
Choi, J., & Staggs, V. S. (2014). Comparability of nurse staffing
measures in examining the relationship between RN staffing
and unit-acquired pressure ulcers: A unit-level descriptive, cor-
relational study. International Journal of Nursing Studies,51(10),
1344–1352.
Cohen, M. D., & Hilligoss, P. B. (2010). The published literature
on handoffs in hospitals: Deficiencies identified in an extensive
review. Quality and Safety in Health Care,19(6), 493–497.
Conway, P. H., Konetzka, T., Zhu, J., Volpp, K. G., & Sochalski,
J. (2008). Nurse staffing ratios: Trends and policy implications
for hospitalists and the safety net. Journal of Hospital Medicine,
3(3), 193–199.
10 Health Care Management Review Month–Month •2019
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Donabedian, A. (1988). The quality of care: How can it be
assessed? Journal of American Medical Association,260(12),
1743–1748.
Epané, J. P., & Weech-Maldonado, R. (2015). Hospitalists as a
staffing innovation: Does it impact hospital efficiency? Innova-
tion and Entrepreneurship in Health,2,1–8.
Famolaro, T., Yount, N. D., Burns, W., Flashner, E., Liu, H., &
Joann Sorra, J. (2016, March). Hospital Survey on Patient
Safety Culture 2016 user comparative database report. (pre-
pared by Westat, Rockville, MD, under Contract No.
HHSA290201300003C). Rockville, MD: Agency for Health-
care Research and Quality.
Fan, C. J., Pawlik, T. M., Daniels, T., Vernon, N., Banks, K.,
Westby, P., …Makary, M. A. (2016). Association of safety
culture with surgical site infection outcomes. Journal of the
American College of Surgeons,222(2), 122–128.
Ford, E. W., Silvera, G. A., Kazley, A. S., Diana, M. L., & Huerta,
T. R. (2016). Assessing the relationship between patient
safety culture and EHR strategy. International Journal of Health
Care Quality Assurance,29(6), 614–627.
Greenstein, E. A., Arora, V. M., Staisiunas, P. G., Banerjee, S. S.,
& Farnan, J. M. (2013). Characterising physician listening be-
haviour during hospitalist handoffs using the HEAR checklist.
BMJ Quality & Safety,22(3), 203–209.
Horwitz, L. I., Rand, D., Staisiunas, P., Van Ness, P. H., Araujo,
K. L., Banerjee, S. S., …Arora, V. M. (2013). Development
of a handoff evaluation tool for shift-to-shift physician handoffs:
The handoff CEX. Journal of Hospital Medicine,8(4), 191–200.
Joint Commission. (2017). 11 tenets of a safety culture. Retrieved
from https://www.jointcommission.org/assets/1/6/Sentinel_
Events_11_tenets_of_a_safety_culture_infographic_2018.pdf
Kalisch, B. J., & Lee, H. (2009). Nursing teamwork, staff charac-
teristics, work schedules, and staffing. Health Care Management
Review,34(4), 323–333.
Lee, H. Y., Blegen, M. A., & Harrington, C. (2014). The effects
of RN staffing hours on nursing home quality: A two-stage
model. International Journal of Nursing Studies,51(3), 409–417.
McHugh, M. D., & Ma, C. (2014). Wage, work environment, and
staffing: Effects on nurse outcomes. Policy, Politics, & Nursing
Practice,15(3–4), 72–80.
McHugh, M. D., Rochman, M. F., Sloane, D. M., et al. American
Heart Association’s Get With the Guidelines–Resuscitation
Investigators (2016). Better nurse staffing and nurse work en-
vironments associated with increased survival of in-hospital
cardiac arrest patients. Medical Care,54(1), 74–80.
Messler, J., & Whitcomb, W. F. (2015). A history of the
hospitalist movement. Obstetrics and Gynecology Clinics of
North America,42(3), 419–432.
National Quality Forum. (2012). Endorsement summaries. Retrieved
from http://www.qualityforum.org/News_And_Resources/
Endorsement_Summaries/Endorsement_Summaries.aspx
O'Leary, K. J., Sehgal, N. L., Terrell, G., & Williams, M. V. (2012).
Interdisciplinary teamwork in hospitals: A review and practi-
cal recommendations for improvement. Journal of Hospital
Medicine,7(1), 48–54.
O'Malley, A. S., Draper, K., Gourevitch, R., Cross, D. A., &
Scholle, S. H. (2015). Electronic health records and support
for primary care teamwork. Journal of the American Medical
Informatics Association,22(2), 426–434.
Pettker, C.M., Thung, S. F., Raab, C. A., Donohue, K. P., Copel,
J. A., Lockwood, C. J., & Funai, E. F. (2011). A comprehen-
sive obstetrics patient safety program improves safety climate
and culture. American Journal of Obstetrics and Gynecology,
204(3), 216-e1–216-e6.
Pronovost, P., & Sexton, B. (2005). Assessing safety culture: Guide-
lines and recommendations. Quality and Safety in Health Care,
14(4), 231–233.
Riesenberg, L. A., Leisch, J., & Cunningham, J.M. (2010). Nurs-
ing handoffs: A systematic review of the literature. The American
Journal of Nursing,110(4), 24–34.
The Patient Protection and Affordable Care Act. (2010). Pub. L.
No. 111-148, 124 Stat. 119.
Tscholl, D. W., Weiss, M., Kolbe, M., Staender, S., Seifert, B.,
Landert, D., …Noethiger, C. B. (2015). An anesthesia prein-
duction checklist to improve information exchange, knowl-
edge of critical information, perception of safety, and possibly
perception of teamwork in anesthesia teams. Anesthesia &
Analgesia,121(4), 948–956.
Wachter, R. M., The end of the beginning: Patient safety five years
after “to err is human.”(2004). Health Affairs,23,W4.
Weaver, A. C., Callaghan, M., Cooper, A. L., Brandman, J., &
O'Leary, K. J. (2014). Assessing interprofessional teamwork
in inpatient medical oncology units. Journal of Oncology Prac-
tice,11(1), 19–22.
Weaver, S. J., Lubomksi, L. H., Wilson, R. F., Pfoh, E. R., Martinez,
K. A., & Dy, S. M. (2013). Promoting a culture of safety as a
patient safety strategy: A systematic review. Annals of Internal
Medicine,158(5_Part_2), 369–374.
West, E., Barron, D. N., Harrison, D., Rafferty, A.M., Rowan, K.,
& Sanderson, C. (2014). Nurse staffing, medical staffing and
mortality in intensive care: An observational study. International
Journal of Nursing Studies,51(5), 781–794.
Hospital Staffing and Safety Culture 11
Copyright © 2019 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.