Content uploaded by Brendan Martin
Author content
All content in this area was uploaded by Brendan Martin on Dec 16, 2020
Content may be subject to copyright.
Major Article
Evaluation of hospital nurse-to-patient staffing ratios and sepsis bundles
on patient outcomes
Karen B. Lasater PhD, RN
a,b,
*, Douglas M. Sloane PhD
a
, Matthew D. McHugh PhD, RN, FAAN
a,b
,
Jeannie P. Cimiotti PhD, RN, FAAN
c
, Kathryn A. Riman BSN, RN
a,b
, Brendan Martin PhD
d
,
Maryann Alexander PhD, RN, FAAN
d
, Linda H. Aiken PhD, RN, FAAN
a,b
a
Center for Health Outcomes and Policy Research, School of Nursing, University of Pennsylvania, Philadelphia, PA
b
Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, PA
c
Nell Hodgson Woodruff School of Nursing, Emory University, Atlanta, GA
d
National Council of State Boards of Nursing, Chicago, IL
Background: Despite nurses’responsibilities in recognition and treatment of sepsis, little evidence docu-
ments whether patient-to-nurse staffing ratios are associated with clinical outcomes for patients with sepsis.
Methods: Using linked data sources from 2017 including MEDPAR patient claims, Hospital Compare, Ameri-
can Hospital Association, and a large survey of nurses, we estimate the effect of hospital patient-to-nurse
staffing ratios and adherence to the Early Management Bundle for patients with Severe Sepsis/Septic Shock
SEP-1 sepsis bundles on patients’odds of in-hospital and 60-day mortality, readmission, and length of stay.
Logistic regression is used to estimate mortality and readmission, while zero-truncated negative binomial
models are used for length of stay.
Results: Each additional patient per nurse is associated with 12% higher odds of in-hospital mortality, 7%
higher odds of 60-day mortality, 7% higher odds of 60-day readmission, and longer lengths of stay, even after
accounting for patient and hospital covariates including hospital adherence to SEP-1 bundles. Adherence to
SEP-1 bundles is associated with lower in-hospital mortality and shorter lengths of stay; however, the effects
are markedly smaller than those observed for staffing.
Discussion: Improving hospital nurse staffing over and above implementing sepsis bundles holds promise for
significant improvements in sepsis patient outcomes.
© 2020 Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
Key Words:
Nursing
Health services research
Acute care
BACKGROUND
For nearly a decade, New York state has been a leader in efforts to
reduce high rates of mortality for people with sepsis. In 2013, after
the highly publicized death of a 12-year-old boy with sepsis, New
York enacted Rory’s Regulations,
1
which required hospitals to
implement evidence-based protocols for the screening, early diagno-
sis, and timely treatment of patients with severe sepsis and/or septic
shock. Included in Rory’s Regulations was the public reporting of hos-
pital adherence to protocols. More recently, hospitals in all states fol-
lowed suit when the Centers for Medicare and Medicaid Services
(CMS) implemented a similar public reporting measure based on evi-
dence-based guidelines from the Surviving Sepsis Campaign.
2,3
Today, Hospital Compare publicly reports hospitals’adherence scores
on the Early Management Bundle for patients with Severe Sepsis/
Septic Shock (SEP-1).
2
Despite the national adoption of evidence-
based protocols for the care of patients with sepsis, few hospitals are
consistently delivering the requisite care for sepsis patients,
4,5
lead-
ing to potentially preventable deaths.
6-8
Another evidence-based intervention that has received little
attention in the context of caring for patients with sepsis but has
been associated with better clinical outcomes for patients with
* Address correspondence to Karen B. Lasater, PhD, RN, Center for Health Outcomes
and Policy Research, University of Pennsylvania School of Nursing, 418 Curie Boule-
vard, Fagin Hall, Philadelphia, PA 19104
E-mail address: karenbl@nursing.upenn.edu (K.B. Lasater).
Funding: Funding for this work was provided by the National Council of State
Boards of Nursing (NCSBN) (Lasater, PI); National Institute of Nursing Research,
National Institutes of Health (R01NR014855, Aiken, PI; T32NR007104, Aiken, Lake,
McHugh, MPIs); and Agency for Healthcare Research and Quality (R01HS026232
Cimiotti, PI).
Conflicts of interest: None to report.
https://doi.org/10.1016/j.ajic.2020.12.002
0196-6553/© 2020 Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
ARTICLE IN PRESS
American Journal of Infection Control 000 (2020) 1−6
Contents lists available at ScienceDirect
American Journal of Infection Control
journal homepage: www.ajicjournal.org
various medical and surgical conditions is patient-to-nurse staffing
ratios.
9-11
Some previous research has shown nurse staffing to be
associated with the incidence of hospital acquired infections.
12
Less
is known about the associations between patient-to-nurse staffing
ratios and clinical outcomes for patients with sepsis; however, some
recent research suggests that sepsis patients admitted to hospitals
with better nursing resources, including better staffing ratios, have
better clinical outcomes including lower odds of mortality, readmis-
sion, intensive care unit utilization, shorter lengths of stay, and lower
costs of care.
13
No research prior to this study has considered the
association of patient-to-nurse staffing ratios and recommended evi-
dence-based sepsis care bundles on outcomes for sepsis patients.
In this study, we directly evaluate whether patient-to-nurse staff-
ing ratios are associated with clinical outcomes for patients admitted
with sepsis in 116 New York state hospitals. We simultaneously eval-
uate the effects of hospital adherence to the SEP-1 evidence-based
care bundle on patient outcomes to determine whether and to what
extent improving patient-to-nurse staffing ratios might benefit
patients. This research question is timely and policy relevant since
New York state requires sepsis bundles and is currently considering
the Safe Staffing for Quality Care Act (A2954/S1032),
14
which would
require hospitals to comply with safe nurse staffing ratios.
METHODS
Design and data sources
A cross-sectional analysis of multiple linked data sources was
undertaken. Data about hospitals were provided from several sources
including a large survey of registered nurses licensed in New York
state, the 2017 American Hospital Association Annual Survey, and
publicly available 2017 Hospital Compare data from the CMS.
5
Infor-
mation about patient characteristics and outcomes was derived from
CMS MEDPAR data of Medicare patients hospitalized during 2017.
Nurses practicing in hospitals were used as informants about staff-
ing levels and other features of their work environments. The survey of
registered nurses was conducted between December 2019 and Febru-
ary 2020. Email addresses of all actively licensed nurses were obtained
from the New York state licensure list. All nurses, not a sample, were
contacted by email to complete the survey and responses were
returned anonymously. Nurses who did not respond to the initial sur-
vey invitation received up to 10 follow-up invitations during the study
period. Once nurses completed the survey, they no longer received
these follow-up invitations; and nurses could opt-out at any time.
Nurses were asked to report the name of their hospital employer,
which allowed us to aggregate responses from nurses working in the
same hospital and create hospital-level measures of nursing resources,
such as patient-to-nurse staffing ratios. Additional details of the survey
methodology have been reported elsewhere, including results of a
nonresponse second survey revealing no response bias in the variables
of interest.
15
The nurse-level response rate was 17% yielding 13,000
responses, an average of 24 registered nurses per hospital working in
adult medical surgical units, thus providing reliable estimates of staff-
ing in most acute care general hospitals in New York state.
16
Study sample of hospitals and patients
The analytic sample of hospitals included acute care hospitals in
New York state. Hospitals were included in the sample if they had at
least 5 registered nurses who responded to the survey and reported
working on a medical-surgical unit as a direct care staff nurse. Among
the final sample of 116 study hospitals, the average number of nurse
respondents was 24 and ranged from 5 to 139 nurses per hospital.
The final patient sample consisted of 52,177 Medicare beneficia-
ries between the ages of 65 and 99 years old who were discharged
from one of the 116 study hospitals between January 1, 2017 and
December 31, 2017. To be included in the study sample, patients
were required to have a principal diagnosis of sepsis present on
admission. ICD-10 codes used to identify sepsis are provided in
Appendix 1.
Outcome variables
The patient outcome variables of interest were in-hospital mortal-
ity, 60-day mortality, 60-day readmission, and hospital length of stay.
In-hospital mortality was defined as a death occurring during the
index admission for sepsis; 60-day mortality was defined as a death
occurring either in or outside of the hospital within 60 days of the
index admission date. A readmission was identified if a patient was
readmitted to a hospital (either the index hospital or some other hos-
pital in our study sample) within 60 days of discharge. Our readmis-
sion measure excludes patients who died during the index admission
(n = 7,773) or who were transferred out to another hospital (n = 962).
Hospital length of stay was calculated during the index hospitaliza-
tion as the number of days the patient was hospitalized. Patients
with lengths of stay longer than 60 days (n = 165) and patients who
died during the index admission or who were transferred out to
another hospital were excluded.
Predictor variables
The predictor variables of interest included patient-to-nurse staff-
ing ratios and hospital performance on the sepsis bundle for timely
and effective sepsis care (SEP-1). Patient-to-nurse staffing ratios
were derived from the survey responses of direct care registered
nurses working on medical-surgical units. Nurses were asked to
report the number of patients they were assigned during their last
shift worked. Responses were averaged among nurses working in the
same hospital to create a hospital-level measure of medical-surgical
patient-to-nurse staffing.
Hospital performance on timely and effective sepsis care was
obtained from CMS Hospital Compare data collected between January
1, 2017 and December 31, 2017. The SEP-1 score is a National Inpatient
Quality Measure that began in October 2015 as part of CMS’quality
reporting program.
17
Chart abstraction is used to identify the percent-
age of patients who received appropriate care for severe sepsis and
septic shock. Appropriate care includes interventions such as obtaining
lactate measurements, blood cultures, and delivering a broad-spec-
trum antibiotic within 3 hours of sepsis onset for individuals with
severe sepsis. Additionally, patients with septic shock require intrave-
nous fluids within 3 hours of onset, vasopressors within 5 hours, and
repeat volume assessments within 6 hours. Hospital SEP-1 scores can
range between 0% and 100% indicated the percentage of patients who
received appropriate care for severe sepsis and septic shock. Although
it is not within the clinical scope of bedside nurses to order and initiate
the sepsis care bundle, nurses are directly responsible for ensuring
timely completion of the relevant diagnostic testing and administra-
tion of treatments. Thus, the direct care nurse is a key contributor to a
hospital’s performance on the SEP-1 bundle.
Covariates
Hospital covariations were included in the modeling to control for
potentially confounding relationships. The American Hospital Associa-
tion survey provided data on hospital size, teaching status, and technol-
ogy capabilities. Size was defined by the number of inpatient beds and
categorized as small (≤100 beds), medium (101-250 beds), and large
(>250 beds). Teaching status was categorized as nonteaching (no medi-
cal trainees), minor teaching (0-4 medical trainees per bed), and major
teaching (≥4 medical trainees per bed). Hospitals with the capabilities
ARTICLE IN PRESS
2K.B. Lasater et al. / American Journal of Infection Control 00 (2020) 1
−
6
to perform major organ transplantation and/or open-heart surgery were
defined as having a high-technology status. A measure of patient-to-
nurse staffing ratios in intensive care units was derived from the survey
of nurses and included as a control variable in the models.
Patient covariates were obtained from MEDPAR data and
included: age, sex, 32 Elixhauser comorbidities, a dummy indicator
for whether or not the patient was a transfer-in from another hospi-
tal, and 5 dummy variables for diagnostic-related groups (DRGs),
which represented over 98% of the study patients.
Data analysis
Descriptive statistics including mean, median, standard deviation,
and ranges are used to report the number of direct care medical-surgical
nurse respondents per hospital and the number of Medicare patients
with sepsis within the 116 study hospitals. Patient-to-nurse ratios and
SEP-1 scores are reported for the 116 study hospitals and the hospitals
are described by their distribution of size, teaching status, and technol-
ogy capabilities. Multilevel random intercept logistic regression models
were estimated using MLWin to compute odds ratios for mortality and
readmission outcomes. Zero-truncated negative binomial models with
clusteredrobuststandarderrorswereusedtocomputeincidentrate
ratios (IRRs) for length of stay. Models are first presented unadjusted
and then adjusted for hospital and patient covariates. Staffing was mod-
eled as a one-unit change in the number of patients per nurse. SEP-1
score was modeled as a 10-point change in the hospital score.
RESULTS
The numbers of medical-surgical nurse respondents and sepsis
patients in the 116 study hospitals are described in Table 1. Among
the 116 New York hospitals in our sample, we obtained data about
patient-to-nurse staffing ratios from 2,747 registered nurses, with an
average of 23.7 nurses per hospital. Data from 52,177 Medicare
patients hospitalized with a principal diagnosis of sepsis were
included in our analysis, with an average of roughly 450 patients per
hospital.
Characteristics of the study hospitals and the patient-to-nurse
staffing ratios and performance on the CMS sepsis bundle (SEP-1) are
described in Table 2. The majority of hospitals in our sample had
greater than 250 beds (58.6%) and did not have a high technology sta-
tus (53.5%). Hospitals were relatively evenly distributed by their
teaching status: nonteaching (26.7%), minor teaching (37.1%), major
teaching (31.0%). The average medical-surgical patient-to-nurse staff-
ing ratio among the 116 hospitals was 6.3 patients per nurse (SD:
1.0). Larger, major teaching hospitals tended to have higher staffing
ratios as compared to smaller nonteaching hospitals. The average
SEP-1 score was 47.0% (SD 17.5%). SEP-1 scores were higher (indicat-
ing better performance on sepsis bundle adherence) in medium-sized
hospitals (101-250 beds) and hospitals without high technology
capabilities.
Among the 52,177 sepsis patients in our sample, 14.9% died dur-
ing the index admission and 28.6% of patients died within 60 days of
admission (Table 3). Excluding individuals who died during the index
hospitalization and those who were transferred out, 23.5% were read-
mitted within 60 days of discharge. The average length of stay during
the index hospitalization was 8.5 days (SD 7.3 days). The distribution
of patients’age and sex are described in Table 3. A slightly larger per-
centage of patients were female (52.3%) as compared to male (47.7%).
The most common DRGs were for severe sepsis without mechanical
ventilation with (DRG 872) and without (DRG 871) major complica-
tion/comorbidity. Common comorbidities included hypertension,
fluid and electrolyte disorders, congestive heart failure, and chronic
pulmonary disease.
Table 1
Numbers of medical-surgical nurse respondents in the 116 study hospitals, and the numbers of sepsis patients used in the different analyses
Nurse respondents per hospital
Number of nurse respondents Mean SD Median Minimum Maximum
Medical-surgical nurses 2,747 23.7 24.3 14 5 139
Sepsis patients per hospital
Sepsis patients used in analyses of - Number of patients Mean SD Median Minimum Maximum
In-hospital and 60-day mortality 52,177 449.8 403.5 289 11 1,943
Readmissions 43,442 374.5 348.6 232 10 1,699
Length of stay 43,227 373.1 347.3 229 10 1,696
Table 2
Selected characteristics of the 116 hospitals in the study sample, and medical-surgical staffing and SEP-1 scores in hospitals with different characteristics
Medical-surgical staffing
(patients per nurse)
Severe sepsis and septic shock
management bundle (SEP-1)
Number of hospitals Percent of hospitals Mean SD Median Mean SD Median
Hospital size
<=100 beds 12 10.3% 5.8 0.5 5.7 50.9 17.5 49.0
101-250 Beds 36 31.0% 6.3 1.2 6.1 52.2 18.3 53.0
>250 68 58.6% 6.4 0.9 6.2 43.6 16.6 41.0
Total 116 100.0% 6.3 1.0 6.1 47.0 17.5 46.0
Teaching status
Nonteaching 31 26.7% 5.9 0.7 5.8 46.8 17.1 48.0
Minor Teaching 43 37.1% 6.3 1.1 6.1 49.6 16.7 48.0
Major teaching 36 31.0% 6.5 1.1 6.5 43.4 17.9 42.0
Missing 6 5.2% 6.5 1.2 6.2 52.0 23.7 57.5
Total 116 100.0% 6.3 1.0 6.1 47.0 17.5 46.0
Technology
Nonhigh technology 62 53.5% 6.2 0.9 6.1 47.4 17.2 46.5
High technology 32 27.6% 6.0 0.8 5.9 41.6 15.3 42.5
Missing 22 19.0% 6.9 1.4 6.7 53.9 19.5 58.5
Total 116 100.0% 6.3 1.0 6.1 47.0 17.5 46.0
ARTICLE IN PRESS
K.B. Lasater et al. / American Journal of Infection Control 00 (2020) 1
−
63
Table 4 presents the unadjusted and adjusted effects of nurse
staffing and SEP-1 scores on clinical outcomes of sepsis patients. In
the unadjusted models, we find large and significant effects of staff-
ing on all 4 clinical outcomes of interest. Each additional patient per
nurse is associated with 19% higher odds of in-hospital mortality
(odds ratio [OR] 1.19, 95% confidence interval [CI] 1.10-1.29, P<
.0001), 13% higher odds of 60-day mortality (OR 1.13, 95% CI 1.07-
1.19, P<.0001), 6% higher odds of readmission (OR 1.06, 95% CI 1.02-
1.10, P.004), and longer lengths of stay by a factor of 6% (IRR 1.06,
95% CI 1.02-1.11, P= .009). In the unadjusted models, a 10-point
increase in a hospital’s performance on the SEP-1 bundle was signifi-
cantly associated with shorter lengths of stay by a factor of 3% (IRR
0.97, 95% CI 0.95-1.00, P= .017), but was not significantly associated
with lower mortality or readmission.
The fully adjusted models jointly modeled both the staffing and
SEP-1 score effects, in addition to hospital-level and patient-level
covariates. The effects of each additional patient per nurse in a
nurse’s workload remained large and significant for all of the clinical
outcomes studied. A 10-point improvement in SEP-1 scores was sig-
nificantly associated with shorter lengths of stay and 5% lower odds
of in-hospital mortality (OR 0.95, 95% CI 0.91-0.99, P= .018).
DISCUSSION
This analysis demonstrates that patient-to-nurse staffing ratios on
medical-surgical units vary across hospitals in New York—a state cur-
rently considering hospital nurse staffing regulation—with the aver-
age medical-surgical nurse caring for 2.3 more patients than is
recommended under the proposed legislation of 4 patients per
nurse.
14
In this study, we found each additional patient in a nurse’s
workload to be strongly and significantly associated with a higher
probability of in-hospital and 60-day mortality, readmission, as well
as longer lengths of stay, even after accounting for hospital and
patient characteristics.
There have been substantial policy efforts over the last decade to
reduce mortality among sepsis patients—both through Rory’s Regu-
lations in the case of New York state and through CMS sepsis bundles
that apply to hospitals nationally. This study finds that patients in
hospitals with greater adherence to the SEP-1 sepsis bundle have
lower in-hospital mortality and somewhat shorter lengths of stay. No
significant relationships were found between hospital adherence to
the SEP-1 bundle and 60-day mortality or readmissions.
Notably, the effects of nurse staffing on patient outcomes are
more pronounced than is hospital adherence to the SEP-1 bundle. For
example, each additional patient per nurse is associated with 12%
higher odds of in-hospital mortality compared with a 10% change in
SEP-1 adherence associated with only a 5% change in in-hospital
mortality. Higher SEP-1 scores were also associated with shorter
lengths of stay, but staffing had more than twice as large an effect on
shorter lengths of stay, even when accounting for hospitals’SEP-1
scores. Moreover, the effect of staffing was large and significant in
terms of 60-day mortality and readmissions, while the SEP-1 scores
revealed no association.
Nurse staffing levels have not been previously studied in relation
to evaluations of the impact of the SEP-1 bundle. However, it is not
entirely surprising to find nurse staffing workloads are associated
with sepsis outcomes, above and beyond hospitals’adherence to the
SEP-1 bundles. The interventions comprising the SEP-1 bundle and
the overall care of a septic patient are heavily reliant on nurses with
the adequate time and resources to surveil for signs and progression
of sepsis, to obtain blood samples in a timely manner, and to adminis-
ter antibiotic and vasopressor medication and fluid resuscitation,
which requires close monitoring and titration.
The findings suggest that sepsis patient outcomes would likely be
substantially improved by establishing a minimum safe hospital
nurse staffing standard, like the one currently under consideration in
New York state, in addition to the policies such as Rory’s Regulations
to promote adherence to the SEP-1 bundles. Attention to nurse staff-
ing ratios may not only reduce mortality and readmission among sep-
sis patients as we show here but is likely to impact patients with a
wide range of medical and surgical conditions, as previous research
has suggested.
18-22
Table 3
Outcomes and selected characteristics of sepsis patients in the study hospitals
Patients
Patient outcomes Number Percent
In-hospital mortality/cases 7,773/52,177 14.9%
60-day mortality/cases 14,898/52,177 28.6%
60-day readmissions/cases 10,206/43,442 23.5%
Number Mean (SD)
Length of stay 43,277 8.5 (7.3)
Patient characteristics Number Percent
Age
65-69 8,743 16.8%
70-74 8,343 16.0%
75-79 8,628 16.5%
80-84 8,820 16.9%
85-89 8,804 16.9%
90-99 8,839 16.9%
Total 52,177 100.0%
Sex
Female 27,294 52.3%
Male 24,883 47.7%
Total 52,177 100.0%
Transfer status
Not transferred in 50,421 96.6%
Transferred in 1,756 3.4%
Total 52,177 100.0%
Diagnostic-related group
853: Infectious disease with MCC 4,065 7.8%
854: Infectious disease with CC 805 1.5%
870: Severe sepsis with MV >96 h 2,316 4.4%
871: Severe sepsis without MV >96 h
without MCC M:CC
34,030 65.2%
872: Severe sepsis without
MV >96 h with MCC
10,191 19.5%
Other DRG 770 1.5%
Total 52,177 100%
Common comorbidities
Hypertension 40,071 76.8%
Fluid and electrolyte disorders 31,960 61.3%
Congestive heart failure 16,640 31.9%
Chronic pulmonary disease 15,718 30.1%
Renal failure 14,746 28.3%
Deficiency anemias 14,471 27.7%
Diabetes with chronic complications 12,012 23.0%
Other neurological disorders 10,389 19.9%
Hypothyroidism 9,613 18.4%
Diabetes wo chronic complications 7,978 15.3%
Weight loss 7,455 14.3%
Coagulopathy 6,797 13.0%
Depression 6,447 12.4%
Valvular disease 6,273 12.0%
Obesity 6,205 11.9%
Peripheral vascular disease 4,712 9.0%
Paralysis 4,410 8.5%
Metastatic cancer 2,934 5.6%
Solid tumor without metastasis 2,744 5.3%
MCC, major complication or comorbidity; CC, complication or comorbidity; MV,
mechanical ventilation.
Readmissions are based on cases that exclude cases that died in the hospital or were
transferred to another hospital. Cases used to calculate length of stay exclude cases
involving in-hospital deaths, patients transferred to another acute care facility, and
lengths of stay longer than 60 days. Comorbidities shown are those that involved at
least 5% of the patients in either patient group, ordered according to their prevalence.
The percentages of cases with different comorbidities do not sum to 100% due to
patients with multicomorbidities.
ARTICLE IN PRESS
4K.B. Lasater et al. / American Journal of Infection Control 00 (2020) 1
−
6
Limitations
The study findings should be considered in the context of both its
strengths and limitations. While this was a cross-sectional design,
and can therefore not claim a causal relationship between nurse staff-
ing and patient outcomes, other studies using multiple waves of
panel data have shown nurse staffing to be associated with patient
outcomes over time.
23
Our study uses a measure of nurse staffing
derived from staff nurses providing direct clinical care on medical-
surgical units in a large and representative sample of New York hospi-
tals. Other studies of nurse staffingoftenrelyonmeasuresofstaffing
created by administrators and include nurse positions in direct as well
as indirect patient care roles and in ambulatory as well as inpatient
care, which creates a less precise measure of the workload for nurses at
the bedside. Finally, we rely on publicly available data from Hospital
Compare of hospital-level adherence to the SEP-1 bundle to understand
whether and to what extent sepsis patients receive appropriate and
timely care. The SEP-1 bundle uses an “all or nothing”approach, such
that for a hospital to receive credit for administering appropriate and
timely sepsis care, they need to have performed all the interventions
within the bundle. Hospitals that provide some of the interventions are
not given credit for those interventions.
CONCLUSIONS
Despite public attention to high and potentially avoidable deaths
resulting from a diagnosis of sepsis as well as resulting policies pro-
moting adherence to evidence-based sepsis care measures, the aver-
age hospital provides appropriate sepsis care to only a little more
than half of patients. Nurses are central to the recognition, manage-
ment, and treatment of sepsis; and thus, the resultant clinical out-
comes for older adults with sepsis are associated with hospital
nurses’workloads. In this study, we find that every additional patient
in a nurses’workload is associated with higher odds of death, as well
as higher odds of readmission and longer lengths of stay, which sug-
gests attention to patient-to-nurse staffing ratios may be critical to
increasing adherence to sepsis care and improving patient outcomes.
SUPPLEMENTARY MATERIALS
Supplementary material associated with this article can be found
in the online version at https://doi.org/10.1016/j.ajic.2020.12.002.
References
Tagged P 1. New York State Department of Health. Rory’s regulations. Available at: https://
www.health.ny.gov/facilities/public_health_and_health_planning_ council/meet
ings/2013-02-07/docs/13-01.pdf. Accessed December 14, 2020.
2. Centers for Medicare & Medicaid. Severe sepsis and septic shock: management
bundle (composite measure). Available at: https://cmit.cms.gov/CMIT_public/View
Measure?MeasureId=1017. Accessed December 14, 2020.
3. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus
definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315:801–810.
4. Barbash IJ, Davis B, Kahn JM. National performance on the Medicare SEP-1 sepsis
quality measure. Crit Care Med. 2019;47:1026–1032.
5. Centers for Medicare & Medicaid. Hospital compare. Available at: https://www.
medicare.gov/hospitalcompare/search.html. Accessed December 14, 2020.
6. Levy MM, Dellinger RP, Townsend SR, et al. The surviving sepsis campaign: results
of an international guideline-based performance improvement program targeting
severe sepsis. Intensive Care Med. 2010;36:222–231.
7. Levy MM, Rhodes A, Phillips GC, et al. Surviving sepsis campaign: association
between performance metrics and outcomes in a 7.5-year study. Intensive Care
Med. 2014;40:1623–1633.
8. Rhodes A, Phillips G, Beale R, et al. The surviving sepsis campaign bundles and out-
come: results from the international multicentre prevalence study on sepsis (the
IMPreSS study). Intensive Care Med. 2015;41:1620–1628.
9. McHugh MD, Aiken LH, Windsor C, Douglas C, Yates P. Case for hospital nurse-to-
patient ratio legislation in Queensland, Australia, hospitals: an observational
study. BMJ Open. 2020;10: e036264.
10. Shekelle PG. Nurse−patient ratios as a patient safety strategy: a systematic review.
Ann Intern Med. 2013;158:404–409.
11. Griffiths P, Ball J, Drennan J, et al. Nurse staffing and patient outcomes: strengths
and limitations of the evidence to inform policy and practice. A review and discus-
sion paper based on evidence reviewed for the National Institute for Health and
Care Excellence Safe Staffing guideline development. Int J Nurs Stud. 2016;63:213–
225.
12. Cimiotti JP, Aiken LH, Sloane DM, Wu ES. Nurse staffing, burnout, and health care-
associated infection. Am J Infect Control. 2012;40:486–490.
13. Lasater KB, McHugh MD, Rosenbaum PR, et al. Evaluating the costs and outcomes
of hospital nursing resources: a matched cohort study of patients with common
medical conditions. J Gen Intern Med. 2020:1–8.
14. New York Senate and Assembly. Safe Staffing for Quality Care Act (A2954/S1032).
15. Lasater KB, Jarrin OF, Aiken LH, et al. A methodology for studying organizational
performance. Med Care. 2019;57:742–749.
16. National Research Council. Nonresponse in Social Science Surveys: A Research
Agenda. National Academies Press; 2013.
17. Schorr C. Nurses can help improve outcomes in severe sepsis. Am Nurse Today.
2016;11:20–25.
18. Kane RL, Shamliyan TA, Mueller C, Duval S, Wilt TJ. The association of registered
nurse staffing levels and patient outcomes: systematic review and meta-analysis.
Med Care. 2007;45:1195–1204.
19. Mark BA, Harless DW, McCue M, Xu Y. A longitudinal examination of hospital reg-
istered nurse staffing and quality of care. Health Serv Res. 2004;39:279–300.
20. Aiken LH, Clarke SP, Sloane DM, Sochalski J, Silber JH. Hospital nurse staffing and
patientmortality, nurseburnout, and job dissatisfaction. JAMA. 2002;288:1987–1993.
21. Aiken LH, Cimiotti JP, Sloane DM, et al. Effects of nurse staffing and nurse educa-
tion on patient deaths in hospitals with different nurse work environments. J Nurs
Adm. 2012;42:S10.
Table 4
Unadjusted and adjusted effects of medical surgical staffing (patient-to-nurse ratio) and SEP-1 scores on sepsis patient outcomes
Unadjusted models Fully adjusted models
Patient outcome Staffing effect SEP-1 effect Staffing effect SEP-1 effect
In-hospital mortality Odds ratio 1.19
z
0.96 1.12
y
0.95*
95% CI (1.10, 1.29) (0.91,1.00) (1.03, 1.21) (0.91, 0.99)
P>|z| <0.0001 0.051 0.008 0.018
60-day mortality Odds ratio 1.13
z
0.99 1.07*0.97
95% CI (1.07, 1.19) (0.96, 1.02) (1.01, 1.14) (0.94, 1.00)
P>|z| <0.0001 0.40 0.028 0.056
60-day readmission Odds ratio 1.06
y
0.99 1.07
y
0.99
95% CI (1.02, 1.10) (0.96, 1.01) (1.03, 1.12) (0.97, 1.02)
P>|z| 0.004 0.179 0.001 0.613
Length of stay IRR 1.06
y
0.97*1.05
y
0.98*
95% CI (1.02, 1.11) (0.95, 1.00) (1.02, 1.09) (0.97, 1.00)
P>|z| 0.009 0.017 0.002 0.024
Odds ratios for mortality and readmission models are from random intercept models estimated using MLWin. Incident rate ratios (IRR) for length of stay models are from zero trun-
cated negative binomial models with clustered standard errors. In the adjusted models for all outcomes, hospital-level controls include hospital size, technology status, teaching sta-
tus, and ICU staffing, and patient controls include age, sex, transfer status, 32 Elixhauser comorbidities, and dummy variables for the different DRGs.
*P<.05.
y
P<.01.
z
P<.001.
ARTICLE IN PRESS
K.B. Lasater et al. / American Journal of Infection Control 00 (2020) 1
−
65
22. Aiken LH, Sloane DM, Bruyneel L, et al. Nurse staffing and education and hospital
mortality in nine European countries: a retrospective observational study. Lancet.
2014;383:1824–1830.
23. Aiken LH, Sloane DM, Barnes H, et al. Nurses’and patients’appraisals show
patient safety in hospitals remains a concern. Health Aff (Millwood). 2018;37:
1744–1751.
ARTICLE IN PRESS
6K.B. Lasater et al. / American Journal of Infection Control 00 (2020) 1
−
6
