Relationship between nurse staffing level and adult nursing-sensitive outcomes in tertiary hospitals of Korea: Retrospective observational study

Accepted Manuscript Title: Relationship Between Nurse Staffing Level and Adult Nursing-Sensitive Outcomes in Tertiary Hospitals of Korea: Retrospective Observational Study Authors: Chul-Gyu Kim, Kyun-Seop Bae PII: DOI: Reference:

S0020-7489(18)30001-4 https://doi.org/10.1016/j.ijnurstu.2018.01.001 NS 3079

To appear in: Received date: Revised date: Accepted date:

1-3-2017 28-12-2017 1-1-2018

Please cite this article as: Kim, Chul-Gyu, Bae, Kyun-Seop, Relationship Between Nurse Staffing Level and Adult Nursing-Sensitive Outcomes in Tertiary Hospitals of Korea: Retrospective Observational Study.International Journal of Nursing Studies https://doi.org/10.1016/j.ijnurstu.2018.01.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Running Head: Relationship Between Nurse Staffing Level and Nursing-Sensitive Outcomes

SC R

Tertiary Hospitals of Korea: Retrospective Observational Study

The first author:

Department of Nursing, Chungbuk National University

N

Chungdae-ro 1, Seowon-Gu, Cheongju,

A M

[email protected]

PT

ED

Tel: 82-43-249-1860 Fax: 82-43-266-1710

U

Chul-Gyu Kim, Associate Professor, Ph.D., RN

Chungbuk, Korea 28644

IP T

Relationship Between Nurse Staffing Level and Adult Nursing-Sensitive Outcomes in

The corresponding author:

CC E

Kyun-Seop Bae, Professor, M.D., Ph.D. Asan Medical Center, University of Ulsan College of Medicine 88, Olympic-Ro 43-gil, Songpa-Gu, Seoul 05505, Korea

A

[email protected]

Tel: 82-2-3010-4611 Fax: 82-2-3010-4623

1

This work was supported by the research grant of the Chungbuk National University in 2015 (2015100244)

IP T

Background: Nurse staffing level is an important factor on nursing sensitive outcome. The relationships of nurse staffing level with nursing sensitive outcomes such as mortality, upper

SC R

gastrointestinal bleeding and pressure ulcer have been explored in the United States, Canada, Australia, and New Zealand. Lower level of hospital nurse staffing seems associated with more adverse outcomes, especially mortality. However, there is insufficient evidence of the nurse staffing

U

level-outcome relationship in other indicators.

N

Objectives: This study was conducted to describe the status and prove the relationships of nurse

A

staffing level with nursing sensitive outcome indicators for adult medical and surgical inpatients in

M

Korea. Patient and hospital characteristics as covariates on nurse sensitive outcome were also explored.

ED

Design: This was a retrospective observational study. Setting: The study setting was all 46 tertiary hospitals in Korea.

PT

Participants: We selected all anonymized patients aged 19 years or older and admitted at tertiary hospitals for two years (2013-2014) using electronic reimbursement claims data.

CC E

Method: Multiple logistic regression was used to examine relationships of nurse staffing level

(accounted for full-time registered nurses in general ward only) with Nursing-sensitive outcomes (NSOs) adjusted for patient and hospital characteristics. NSOs included urinary tract infection, upper

A

gastrointestinal tract bleeding, deep vein thrombosis, hospital-acquired pneumonia, pressure ulcer, sepsis,

shock/cardiac

arrest,

CNS

complication,

in-hospital

death,

wound

infection,

physiologic/metabolic derangement and pulmonary failure. Results: The total number of patients in 46 tertiary hospitals in Korea for two years was 3,665,307. Among these, number of patients who had at least one nursing-sensitive outcome was 338,369 2

(9.23%). The significant relationships of nurse staffing level with six nursing-sensitive outcome rates (urinary tract infection, upper gastrointestinal tract bleeding, hospital-acquired pneumonia, shock/cardiac arrest, in-hospital death, and wound infection) were shown. These six nursing-sensitive outcomes showed an increasing trend as nurse staffing level degraded even after adjusting for patient and hospital characteristics. When the nursing-sensitive outcomes between those of group 1 (bed-to-

IP T

nurse ratio < 2:1) and group 3 (between 2.5:1 and 3:1) were compared, the adjusted incidence rate of shock/cardiac arrest showed the highest difference (1.06%).

SC R

Conclusion: We demonstrated strong evidence for the relationships of nurse staffing level with six nursing-sensitive outcomes. We can use this study to improve nursing quality and to inform patients

of the nursing quality of hospitals so they can choose hospitals with better nursing quality. The nurse

N

U

staffing level should be optimized for better outcomes.

A

What is already known about the topic?

Higher levels of nurse staffing could have positive impact on the quality of care.



Lower levels of hospital nurse staffing are associated with more adverse outcomes, especially mortality.

PT

What this paper adds

ED

M



It identifies nursing-sensitive outcomes using large administrative databases in Korea.



It provides additional evidence supporting the relationships of nurse staffing level with patient

CC E



outcomes in urinary tract infection, upper gastrointestinal tract bleeding, hospital-acquired

A

pneumonia, shock or cardiac arrest, in-hospital death, and wound infection.

3

IP T

Introduction

SC R

Purchasers, payers, and policymakers have demanded better healthcare quality for decades to avoid spending money for little value on healthcare. To meet these needs in the healthcare system, two basic strategies have been applied: one is public reporting of each healthcare institution’s quality,

U

which aims to bring forth more informed consumers who will make wiser choices; the second is ‘pay

N

for performance’, which offers incentives to healthcare institutions for providing better results (Kane

A

& Radosevich, 2011). In Korea, the government strives to improve the quality and cost-effectiveness

M

of healthcare services, while medical service providers aim for higher competitiveness and quality improvement, and service consumers demand more rights and information on health services (Health

ED

Insurance Review and Assessment Service [HIRA], 2016). As a continuing effort on this issue, accreditation for healthcare organizations by the Korea Institute for Healthcare Accreditation,

PT

assessment of healthcare quality, public reporting of the results, and value incentive programs in accordance with performance of healthcare institutions by HIRA have been implemented in Korea

CC E

(Kim et al., 2012).

Objective evaluation of nursing outcomes in the health services is a crucial subject in this

issue (Aiken et al. 2002). However, a multi-disciplinary team approach is necessary for modern

A

healthcare service which requires health professionals from all fields (Fowler, 2008), and we cannot measure the outcomes separately. In other words, the result of patient treatment could not be described as an outcome of a single specific discipline (Douglas and Robb, 1995). Therefore, an important issue in recent decades has been on how to measure the contribution and influence of nursing on patient outcomes. Those in the area of nursing service need to prove that their service 4

ultimately contributes to better outcomes and improved efficiency. These efforts are essential for establishing the status of nursing within a health provision system (Park, 2003). In this context, researchers tried to quantify the relationships between nursing structural factors and patient outcomes (White et al., 2011). Nursing structural variables include nurse staffing level, academic degree, and application of protocols (Needleman et al., 2002; Tourangeau et al., 2007).

IP T

In their review, Griffiths et al., (2016) stated that higher nurse staffing correlates with better nursing service in terms of these aforementioned variables. If nurse staffing level is low, part of nursing

SC R

service is delegated to patient’s caregivers and quality of service degrades (HIRA, 2016). If the nurse staffing level is high, direct nursing time is increased (Park, 2003) and quality of nursing is also increased (Rochefort & Clarke, 2010). Measures of nurse staffing include RN full-time equivalent per

U

patient day (nursing hours per patient day, NHPPD), the number of patients per nurse/shift, patient to

N

RN ratio, staff/skill mix (proportion of total hours of care provided by RN) (Kane et al., 2007; Wilson

A

et al., 2011), and bed-to-nurse ratio (Cho & Yun, 2009). Patient outcomes related to nursing quality

M

include clinical outcomes, satisfaction, and quality of life (Brennan et al., 2013). The patient-to-nurse relationship under the condition of reimbursement systems is also important (Romeyke et al., 2012).

ED

Pressure ulcer, failure to rescue, infections such as hospital acquired pneumonia and urinary tract infection (UTIs), medication error and inpatient mortality are the most frequently used indicators for

PT

assessing nursing quality (Wilson et al., 2011). The relationships of nurse staffing level with mortality, fall, and pressure ulcer have been explored in the United States, Canada, Australia, and New Zealand

CC E

(White et al., 2011; Schreuders et al., 2014). Specifically, the fact that nurse staffing level is related to mortality has been reported in various studies from the United States, Europe, the United Kingdom, Thailand, and Australia (Aiken et al., 2002, 2014; Rafferty et al., 2007; Sasichay-Akkadechanunt et

A

al., 2003; Twigg et al., 2011). However, there is insufficient amount of data on the nurse staffing level–outcome relationship on other indicators such as pressure ulcer, falls, and urinary tract infection (Griffiths et al., 2016; Kane et al, 2007). Nursing-sensitive outcomes (NSOs) are indicators of patient health status, changes of which are caused directly by nursing activities. Those for adults include central nervous system (CNS) 5

complication, wound infection, pulmonary failure, urinary tract infection, pressure ulcer, pneumonia, deep vein thrombosis, upper gastrointestinal tract bleeding, sepsis, physiologic/metabolic derangement, shock or cardiac arrest, and in-hospital death (Needleman et al., 2002; McCloskey and Diers, 2005; Schreuders et al., 2014; Twigg et al., 2011, 2015). These indicators are useful for monitoring nursing quality but need to be adjusted with patient acuity and hospital characteristics for

IP T

the comparison of differences in nursing quality among hospitals (White et al., 2011; Wilson et al., 2013). System, process, patient, and provider affected nurse staffing-outcome relationship (Brennan et

SC R

al., 2013). Patient characteristics that affect the relationship are age, sex, socioeconomic status, surgical operation, comorbidities, acuity status such as the use of emergency room or ICU, while those of hospital characteristics are number of beds, location and teaching hospital, patient volume

U

(Brennan et al., 2013; Needleman et al., 2002; Schreuders et al., 2014). Additionally, because the

N

incidence of adverse outcomes is not high, previous studies on NSOs were conducted using big data

A

such as health service administration data (Wilson et al., 2012, 2013; Schreuders et al., 2014).

M

In Korea, however, the relationships of nurse staffing level with patient outcomes have yet to be extensively studied. Three previous studies in Korea have assessed the relationship between nurse

ED

staffing level and outcomes, but their subjects were limited to surgical patients (Kim et al., 2012) and acute stroke patients in the ICU (Cho & Yun, 2009); also, the indicators only included factors such as

PT

mortality, readmission rate, and length of stay (Park, 2003). To expand on these findings and to examine the usefulness of NSOs for comparing nursing quality among hospitals, we derived 12 NSOs

CC E

from the accumulated health administrative data and investigated the relationships of nurse staffing level with patient outcomes.

A

Objective

The overall objective of this study was to explore the status of 12 NSOs in adult medical and

surgical inpatients of tertiary hospitals and their relationships with nurse staffing level using HIRA reimbursement data. More concrete objectives are as follows: 1. To derive the status of NSOs using Korean HIRA reimbursement data 6

2. To compare the NSOs incidence rate by nurse staffing grades of hospitals 3. To explore the relationships of nurse staffing level with NSOs

Materials and Methods

IP T

Database

We used data from the 2013 to 2014 HIRA databases, which contain electronic

SC R

reimbursement claims data of inpatient and outpatient practices in Korea. Most medical institutions (99.9%) in Korea are required submit medical care fee claims of all patients to HIRA in the form of

either electronic data interchange or electronic media (diskette, CD). The database has demographic

U

information of patients such as age, sex, up to 30 diagnoses, all kinds of operations and procedures

N

received, and the details of patient care such as use of emergency room and intensive care unit (ICU).

A

Diagnoses are coded using the International Classification of Diseases (ICD), 10th revision, and

Participants Inclusion criteria

ED

M

procedures are coded using the Korean claim code.

PT

We selected all anonymized patients aged 19 years or older and admitted to tertiary hospitals in a two-year period (2013–2014) using electronic reimbursement claims data. Indicators from tertiary

CC E

hospitals are more reliable because the accuracy of disease codes is higher in tertiary hospitals (Park et al., 2000).

A

Exclusion criteria We excluded patients who did not have nationwide individual identification numbers because

without the identification numbers and their past history of claim data, their comorbidity scores (Charlson’s comorbidity score) cannot be calculated using inpatient and outpatient claim data.

7

Patient Risk Factors We selected risk factors that were found in previous studies (Needleman et al., 2002; Schreuders et al., 2014) and were available in the HIRA database. The patient characteristics for the adjustment of NSO indicators were age (as a continuous variable), sex, and type of health security as a proxy index of socioeconomic status (health insurance for higher socioeconomic status or medical aid

IP T

for lower socioeconomic status).

Clinical practice characteristics included route of admission as a relative urgent index

SC R

(outpatient or emergency room), surgical operation, use of ICU, and Charlson’s comorbidity score.

We adjusted comorbidity condition using Charlson’s comorbidity score identified by ICD-10 code. We identified each patient's comorbidity condition that had occurred within the preceding six months

U

using the data from the outpatient and inpatient HIRA database, in order to adjust for the comorbidity

N

that was presently being treated. Additionally, we did not include new comorbidities developed during

A

admission because we could not distinguish comorbidity from complications. After identifying

M

patients’ comorbidity conditions, we computed Charlson’s comorbidity score developed by Charlson et al. (1987).

ED

Hospital characteristics included number of beds and location with reference to previous studies (American Nurses Association, 2000; Needleman et al., 2002). The number of beds is closely

PT

related to the number of patients, and more beds indicate larger patient volume. Although patient volume is significantly related to patient outcome (Dudley et al., 2000), we did not adjust volume for

CC E

each disease because we did not consider nursing quality for each disease. We classified hospitals reflecting patient volume into three categories: fewer than 1,000 beds, 1,000~1,499 beds, and 1,500 beds or more. Hospital location was classified into three categories: large metropolitan area (Seoul,

A

the capital of Korea), other metropolitan areas, and city areas. All tertiary hospitals in Korea are teaching hospitals, and this could not be included as a hospital characteristic.

Measures Nurse Staffing Level 8

Nurse staffing level refers to the level of staff for nursing in a hospital. The nurse staffing grades of 2013 and 2014 by HIRA were used as the nurse staffing levels in this study. These consist of seven grades, but all tertiary hospitals had one of the higher three grades. Grades are determined by the ratio of beds to registered nurses (bed-to-nurse ratio) in the general wards: grade 1 requires a ratio of lower than 2:1, grade 2 one ranging from 2:1 to less than 2.5:1, and grade 3 one ranging from 2.5:1 to less

IP T

than 3:1 (HIRA, 2015). The bed-to-nurse ratio was calculated by dividing the total number of beds by the total number of full-time equivalent registered nurses (RNs) working in the general ward (Cho &

SC R

Yun, 2009). Beds and RNs at ICU, ER, delivery room, operating and recovery room, isolated room,

aseptic clean room, haemodialysis room, and day care centre were not included in determining the nurse staffing level. Only the bedside registered nurses are included. RNs of managers, non-unit based

U

nurses including wound care, infection prevention, and QI departments were not included in

N

determining the nurse staffing level. Also, assistive personnel and vocational/practical nursing license

Nursing-Sensitive Outcomes

M

A

nurses were excluded (HIRA, 2015).

ED

We defined the numerator, denominator, and exclusion criteria using the ICD-10 codes to calculate incidences of 12 NSOs for each indicator. Criteria for the calculation of each NSO were

PT

defined by the mapping of the Korean ICD-10 code and Korean diagnosis-related group (DRG) consulting a medical recorder and researchers in HIRA according to the definitions of the ICD-9

CC E

disease codes of Needleman et al. (2002) and the mapping of ICD-10 and Australian DRG by Wilson et al. (2013). NSOs were derived from the inpatient reimbursement claim data of HIRA using these criteria, which have been mapped with disease codes and the Korean DRG. Patients who underwent

A

surgical operations were identified using the Korean DRG for surgery. The incidences of nine NSOs (urinary tract infection, upper gastrointestinal tract bleeding,

deep vein thrombosis, hospital-acquired pneumonia, pressure ulcer, sepsis, shock/cardiac arrest, CNS complication, in-hospital death) of adult patients (aged 19 or older) were calculated using the following equation: 9

The remaining three NSOs for adult patients who underwent surgical operations (incidences of wound infection, physiologic/metabolic derangement, pulmonary failure) were calculated using the

IP T

following equation:

SC R

Ethical consideration

Ethical approval for this study was granted by the institutional review board (IRB) of Asan Medical Center (IRB Approval No. 2015-0788). Informed consent from each patient was waived

N

U

because the data was anonymized administrative data.

A

Statistical Analysis

M

The software used for the data analysis was SAS® 9.4 (Cary, NC). Descriptive statistics such as mean, standard deviation (SD), frequency, and proportion were calculated to show the status of

ED

NSOs as well as patient and hospital characteristics. The unit of analysis was admission case. We conducted a chi-squared test, Student’s t-test, and ANOVA to test the differences of the proportions or

PT

means according to patient and hospital characteristics and nurse staffing grade. Multiple logistic

CC E

regression was used to examine the relationships of nurse staffing level with NSOs adjusted for patient and hospital characteristics, including route of admission as a relative urgent index (outpatient or emergency room), surgical operation, use of ICU, Charlson’s comorbidity score, number of beds,

A

and location of hospital. Two-tailed tests were used in all relevant analysis, and statistical significance was determined at p < 0.05.

Calculation of Adjusted NSO We adjusted patient risk factors and hospital characteristics to examine the effect of nurse 10

staffing level using multiple logistic regression. We calculated the predicted NSO rate as the summation of NSO probability of all patients according to nurse staffing grade from the logistic model. Each adjusted NSO rate was calculated using the following equation:

IP T

Our patient risk adjusted models for the 12 NSOs had intermediate or good measures of

SC R

discrimination, with C statistics between 0.68 and 0.88.

PT

ED

M

A

N

U

The flow of data analysis is shown in Figure 1.

CC E

Figure 1. Main study flow

A

Results

Distribution of patient and hospital characteristics The total number of adult medical and surgical inpatient cases treated in 46 tertiary hospitals in Korea in a two-year span (from 2013 to 2014) was 3,665,307. Among these, the number of cases with at least one NSO was 338,369 (9.23%). Compared to cases without any NSO indicator, those 11

with at least one NSO indicator had statistically different characteristics of age, sex, medical security type, admission route, ICU experience, surgical operation, Charlson’s comorbidity score, nurse staffing level, hospital bed number, and hospital location (p < 0.001). Length of hospital stay was 23.0 days on average for patients with at least one NSO, while the same was 15.9 days for patients who did not have any NSO (p < 0.001).

IP T

The number of cases that involved surgical operation was 1,291,898 (35.3%), and 20,925 (1.62%) of these had at least one of the three NSOs. Compared to cases that did not have any NSO

SC R

indicator, surgical cases with at least one of the three NSOs had different characteristics of age, sex,

medical security type, admission route, ICU experience, Charlson’s comorbidity score, nurse staffing level, hospital bed number, and hospital location (p < 0.001). Length of stay for cases with NSOs was

A

CC E

PT

ED

M

A

N

U

32.0 days, which is longer than the 13.7 days for cases without NSOs (p < 0.001, Table 1).

12

I N U SC R

Table 1. Distribution of patient and hospital characteristics by case mix Medical and surgical inpatient cases Total* (n = 3665307)

Cases with any Cases without any † † NSO (n = NSO P value 3326938) (n = 338369)

Cases without any Cases with any Total* p surgical NSO† surgical NSO† (n = 1291898) value (n = 1270973) (n = 20925)

Mean age, y (SD) 19–29 30–39 40–49 50–59 60–69 Over 70 Sex (%)

57.0 (16.1) 230976 (6.3) 361521 (9.9) 520346 (14.2) 831464 (22.7) 782174 (21.3) 938826 (25.6)

56.2 (16.0) 223406 (96.7) 348439 (96.4) 492462 (94.6) 773925 (93.1) 707227 (90.4) 781479 (83.2)

65.4 (14.7) 7570 (3.3) 13082 (3.6) 27884 (5.4) 57539 (6.9) 74947 (9.6) 157347 (16.8)

<.001 <.001

54.8 (16.5) 101481 (7.9) 164291 (12.7) 202480 (15.7) 280895 (21.7) 256461 (19.9) 286290 (22.2)

54.7 (16.5) 100781 (99.3) 163079 (99.3) 200366 (99.0) 276680 (98.5) 251453 (98.1) 278614 (97.3)

61.9 (15.0) 700 (0.7) 1212 (0.7) 2114 (1.0) 4215 (1.5) 5008 (1.9) 7676 (2.7)

<.001 <.001

1647416 (89.7)

189380 (10.3)

<.001

614627 (47.6) 602704 (98.1)

11923 (1.9)

<.001

1828511 (49.9)

1679522 (91.9)

148989 (8.2)

677271 (52.4) 668269 (98.7)

9002 (1.3)

3536947 (96.5) 128360 (3.5)

3218019 (91.0) 108919 (84.9)

318928 (9.0) 19441 (15.1)

<.001

1257008 (97.3) 1237390 (98.4) 34890 (2.7) 33583 (96.3)

19618 (1.6) 1307 (3.7)

<.001

935351 (25.5) 2729956 (74.5)

782156 (83.6) 2544782 (93.2)

153195 (16.4) 185174 (6.8)

<.001

220033 (17.0) 213284 (96.9) 1071865 (83.0) 1057689 (98.7)

6749 (3.1) 14176 (1.3)

<.001 <.001

259956 (7.1) 3405351 (92.9)

169439 (65.2) 3157499 (92.7)

90517 (34.8) 247852 (7.3)

<.001

151106 (11.7) 144055 (95.3) 1140792 (88.3) 1126918 (98.8)

7051 (4.7) 13874 (1.2)

1291898 (35.3) 2373409 (64.7)

1202747 (93.1) 2124191 (89.5)

89151 (6.9) 249218 (10.5)

<.001

-

-

-

2912502 (79.5) 752805 (20.5)

2693449 (92.5) 633489 (84.1)

219053 (7.5) 119316 (15.9)

<.001

1191639 (92.2) 1175936 (98.7) 100259 (7.8) 95037 (94.8)

15703 (1.3) 5222 (5.2)

<.001

M

Male

1836796 (50.1)

PT

Female

CC E

Type of medical security Health insurance (%) Medical aid (%) Admission Emergency room (%) OPD (%) ICU Yes (%) No (%) Operation Yes (%) No (%) Charlson’s score <3 (%) ≥3 (%)

A

A

Category

ED

Surgical inpatient cases only‡

13

-

I N U SC R

Surgical inpatient cases only‡

Medical and surgical inpatient cases

Cases without any Cases with any Total* p surgical NSO† surgical NSO† (n = 1291898) value (n = 1270973) (n = 20925)

850725 (23.2) 1361550 (37.2) 1453032 (39.6)

797253 (93.7) 1236107 (90.8) 1293578 (89.0)

53472 (6.3) 125443 (9.2) 159454 (11.0)

<.001

327464 (25.3) 322911 (98.6) 481347 (37.3) 474410 (98.6) 483087 (37.4) 473652 (98.1)

4553 (1.4) 6937 (1.4) 9435 (1.9)

<.001

1821728 (89.6) 681863 (90.6) 823347 (93.5)

210598 (10.4) 70739 (9.4) 57032 (6.5)

<.001

676874 (52.4) 665172 (98.3) 270018 (20.9) 265404 (98.3) 345006 (26.7) 340397 (98.7)

11702 (1.7) 4614 (1.7) 4609 (1.3)

<.001

135844 (7.6) 103280 (10.9) 99245 (10.6) 23.0 (18.1)

<.001

657588 (50.9) 326888 (25.3) 307422 (23.8) 18.5 (14.2)

9258 (1.4) 6068 (1.9) 5599 (1.8) 32.0 (19.6)

<.001

A

Cases with any Cases without any † † NSO (n = NSO P value 3326938) (n = 338369)

2032326 (55.5) 752602 (20.5) 880379 (24.0)

ED

Nurse staffing grade 1 (%) 2 (%) 3 (%) Hospital bed number <1000 (%) 1000–1499 (%) Over 1500 (%) Hospital location Large metropolitan area (%) Other metropolitan area (%) City area Mean LOS, day (SD)

Total* (n = 3665307)

M

Category

CC E

PT

1782602 (48.6) 1446758 (92.4) 947954 (25.9) 844674 (89.1) 934751 (25.5) 835506 (89.4) 16.6 (14.7) 15.9 (14.2) Surgical patient only‡: patient who have done operation, * % in a column, † % in a row

A

OPD-outpatient department, ICU-Intensive Care Unit

14

<.001

648330 (98.6) 320820 (98.1) 301823 (98.2) 18.3 (14.0)

<.001

The status and incidence rates of NSOs by nurse staffing grade in Korea The incidences of nine NSOs (urinary tract infection, upper gastrointestinal tract bleeding, deep vein thrombosis, hospital-acquired pneumonia, pressure ulcer, sepsis, shock/cardiac arrest, CNS complication, in-hospital death) for all patients were 0.86 - 2.84%. The two lowest incidences were 0.86% and 0.96% for CNS complication and sepsis, respectively. The highest incidence was 2.84%

IP T

for hospital-acquired pneumonia, and that for other six NSOs ranged from 1.15 to 1.69% (Table 2).

Nine NSOs showed statistically significant differences according to nurse staffing level (p < 0.001).

SC R

Eight NSOs, except for pressure ulcer, showed an increasing trend as the nurse staffing grade increased from 1 to 3. Pressure ulcer incidence was highest in hospitals with nurse staffing grade 2.

Three NSOs for patients who underwent surgical operations—wound infection,

U

physiologic/metabolic derangement, and pulmonary failure—showed incidences of 0.52%, 1.29%,

N

and 0.28%, respectively (Table 2). Nurse staffing level was also a significant factor for the rate of

A

three NSOs (p < 0.001). Wound infection and pulmonary failure decreased with higher nurse staffing

M

grade. However, physiologic/metabolic derangement rate was lowest in hospitals with nurse staffing

A

CC E

PT

ED

grade 2.

15

I N U SC R

Table 2. Incidence rates of nursing-sensitive outcome by nurse staffing grade All hospitals Grade 1 a Grade 2 b Grade 3 c Nursing-sensitive outcome No. of Number Incidence No. ofNumber Incidence No. ofNumber Incidence No. ofNumber Incidence Patients of events rate (%) Patients of events rate (%) Patients of events rate (%) Patients of events rate (%) Urinary tract infection

3207323 36797

4765

0.60

1252736 20776

1.66

1327687 26573

2.00

<.001 (a
786208

0.76

1188613 13767

1.16

1232502 17062

1.38

<.001 (a
790109

A

tract

1.55

5968

Deep vein thrombosis

3665282 49884

1.36

850725

8719

1.02

1361543 17676

1.30

1453014 23489

1.62

<.001 (a
Hospital-acquired pneumonia

3432067 97301

2.84

810389

12660

1.56

1270505 36933

2.91

1351173 47708

3.53

ED

Upper gastrointestinal bleeding

3370532 52114

p value (Scheffe’s test)

<.001 (a
3643562 60675

1.67

849146

12190

1.44

1352221 25331

1.87

1442195 23154

1.61

<.001 (a
1350544 12915

0.96

267182

1793

0.67

531155

0.85

552207

1.19

<.001 (a
2909979 42353

1.46

734414

5143

0.70

1071183 14001

1.31

1104382 23209

2.10

<.001 (a
3335584 28571

0.86

806057

4605

0.57

1226496 11010

0.90

1303031 12956

0.99

<.001 (a
In-hospital death

3665307 61836

1.69

850725

10151

1.19

1361550 22336

1.64

1453032 29349

2.02

<.001 (a
Wound infection*

1291898 6775

0.52

327464

817

0.25

481347

2579

0.54

483087

3379

0.70

<.001 (a
Physiologic/ metabolic derangement*

932154

12034

1.29

260753

3318

1.27

346687

3420

0.99

324714

5296

1.63

<.001 (b
Pulmonary failure*

990445

2795

0.28

277303

499

0.18

362054

1152

0.32

351088

1144

0.33

<.001 (a
Sepsis Shock/Cardiac arrest

A

*

M

system

CC E

Central nervous complication

PT

Pressure ulcer

1.15

Nursing-sensitive outcomes for surgical patients. a, b, c: the results of Scheffe’s test

16

4538

6584

Relationship between nurse staffing level and NSOs after adjusting for patient and hospital characteristics Odds ratios for 12 NSOs are listed in Table 3. Odds ratios that increased as the nurse staffing grade increased after adjusting for patient and hospital characteristics were in urinary tract infection, upper gastrointestinal track bleeding, hospital-acquired pneumonia, shock/cardiac arrest, in-hospital

IP T

death, and wound infection (6 of 12 indicators). Odds ratios for CNS complication and pulmonary failure were lowest in grade 1 hospitals, but those of grade 2 hospitals were higher than those of grade

SC R

3 hospitals (2 of 12 indicators). The odds ratios for deep vein thrombosis, sepsis, and

physiologic/metabolic derangement were significantly lower in nurse staffing grade 2 hospitals than in

A

CC E

PT

ED

M

A

N

staffing grade 3 hospitals than in grade 1 and 2 hospitals.

U

grade 1 hospitals (3 of 12 indicators). Odds ratio for pressure ulcer were significantly lower in nurse

17

I N U SC R

Table 3. Odds ratio and 95% CI of nurse staffing level with adjusted patient and hospital characteristics by multiple logistic regression Upper HospitalCentral nervous Urinary tract Deep vein Pressure Shock/Cardiac In-hospital gastrointestinal acquired Sepsis system infection thrombosis ulcer arrest death tract bleeding pneumonia complication 3370532 3207323 3665282 3432067 3643562 1350544 2909979 3335584 3665307 Total umber 52114 36797 49884 97301 60675 12915 42353 28571 61836 Number of events 1.02* (1.02–1.03)

1.04* (1.04–1.05)

6775 12034 1.03* 1.01* 1.03* (1.02–1.03) (1.01–1.01) (1.02–1.03)

1 0.87* (0.86–0.89)

1 0.85* (0.83–0.87)

1 1 1 0.75* 0.77* 0.86* (0.74–0.77) (0.73–0.81) (0.83–0.89)

1 0.98 (0.90–1.05)

1 1 1 1 1.11* 1.49* 1.74* 1.43* (1.06–1.15) (1.44–1.53) (1.69–1.80) (1.33–1.55)

1 1.49* (1.43–1.55)

1 1.37* (1.30–1.44)

1 1 1 1.14* 1.67* 2.12* (1.09–1.18) (1.50–1.86) (1.97–2.28)

1 1.22¶ (1.04–1.45)

2.43* (2.38–2.49) 1

1.45* 1.62* 1.46* 2.03* (1.42–1.48) (1.60–1.64) (1.44–1.49) (1.95–2.11) 1 1 1 1

2.29* (2.24–2.34) 1

1.89* (1.84–1.94) 1

2.76* 1.04 3.38* (2.71–2.81) (0.98–1.10) (3.24–3.52) 1 1 1

3.22* (2.97–3.48) 1

3.11* (3.04–3.18) 1

3.56* (3.46–3.65) 1

2.40* 7.07* 7.91* 17.65* 14.65* (2.34–2.47) (6.95–7.19) (7.75–8.06) (16.97–18.35) (14.32–14.98) 1 1 1 1 1

3.90* (3.78–4.01) 1

9.98* 2.37* 3.63* (9.79–10.17) (2.24–2.52) (3.48–3.79) 1 1 1

16.00* (14.74–17.38) 1

0.57* (0.56–0.59) 1

0.55* (0.53–0.56) 1

0.67* 0.48* 0.61* 0.72* (0.65–0.68) (0.47–0.49) (0.60–0.63) (0.69–0.75) 1 1 1 1

1.09* (1.06–1.11) 1

0.71* (0.69–0.73) 1

0.39* (0.38–0.40) 1 -

1 1.94* (1.91–1.98)

1 1.76* (1.72–1.80)

1 1 1 1 2.45* 2.41* 2.90* 2.76* (2.40–2.49) (2.37–2.44) (2.84–2.95) (2.64–2.88)

1 1.91* (1.87–1.96)

1 1.92* (1.87–1.97)

1 1 1 3.87* 2.22* 3.06* (3.80–3.94) (2.08–2.37) (2.94–3.20)

1 2.96* (2.72–3.22)

1 2.17* (2.08–2.26)

1 1.60* (1.54–1.67)

1 1 1 1 0.93* 1.25* 1.10* 0.80* (0.90–0.96) (1.21–1.28) (1.07–1.13) (0.74–0.87)

1 1.77* (1.69–1.85)

1 1.57* (1.50–1.64)

1 1 1 1.08* 1.88* 0.76* (1.04–1.11) (1.70–2.08) (0.71–0.82)

1 1.52* (1.30–1.77)

1.03* (1.02–1.03)

1.02* (1.01–1.02)

1 1.76* (1.73–1.78)

1 0.69* (0.68–0.71)

1.03* 1.03* 1.04* 1.03* (1.03–1.04) (1.03–1.04) (1.04–1.05) (1.02–1.03)

A

Age

Medical aid Admission

OPD ICU Yes No Operation

A

Yes

No Charlson’s score <3 ≥3

Nurse staffing grade 1 2

1.70* (1.67–1.73) 1

CC E

Emergency room

1 1.43* (1.38–1.49)

1 1 1 1 1.30* 0.66* 0.83* 0.86† (1.27–1.32) (0.66–0.68) (0.82–0.85) (0.83–0.89)

ED

Type of medical security Health insurance

1 1.35* (1.29–1.41)

PT

Female

M

Sex Male

Physiologic/ Wound Pulmonary metabolic infection failure derangement 1291898 932154 990445

18

2795 1.02* (1.02–1.03)

-

-

-

-

I N U SC R

Upper HospitalUrinary tract Deep vein Pressure gastrointestinal acquired infection thrombosis ulcer tract bleeding pneumonia

Over 1,500

1.01 1.32* 0.80* 0.93 (0.96–1.05) (1.27–1.36) (0.77–0.83) (0.85–1.02)

2.44* (2.32–2.57)

1.50* (1.42–1.59)

1.17* 2.23* 1.26* (1.12–1.21) (1.99–2.50) (1.15–1.37)

1.22¶ (1.02–1.46)

1 0.74* (0.72–0.76) 0.70* (0.68–0.73)

1 0.99 (0.96–1.02) 1.14* (1.09–1.18)

1 0.87* (0.85–0.89) 0.82* (0.79–0.89)

1 1.06* (1.04–1.08) 0.65* (0.63–0.67)

1 0.86* (0.84–0.88) 1.03 (1.00–1.07)

1 0.65* (0.62–0.69) 0.76¶ (0.69–0.82)

1 0.94* (0.92–0.97) 1.67* (1.60–1.75)

1 0.83* (0.80–0.86) 1.26* (1.20–1.32)

1 1.12* (1.10–1.15) 0.93* (0.90–0.94)

1 1 1.18* 1.02 (1.11–1.25) (0.97–1.08) 0.76* 0.90‡ (0.69–0.84) (0.84–0.97)

1 0.73* (0.66–0.81) 0.96 (0.82–1.12)

1 1.16* (1.13–1.19) 1.33* (1.29–1.36) 0.73

1 1.14* (1.12–1.16) 1.13* (1.10–1.15) 0.80

1 1.29* (1.25–1.32) 1.11* (1.08–1.14) 0.82

1 1.32* (1.25–1.39) 1.26* (1.20–1.33) 0.88

1 2.26* (2.18–2.33) 2.06* (1.99–2.13) 0.84

1 1.34* (1.30–1.39) 0.95‡ (0.92–0.99) 0.78

1 1.23* (1.20–1.26) 0.98 (0.96–1.01) 0.87

1 1 1.31* 0.74* (1.22–1.40) (0.70–0.79) 0.76* 1.25* (0.70–0.82) (1.18–1.32) 0.68 0.81

1 2.11* (1.88–2.37) 1.18¶ (1.04–1.33) 0.88

Hospital location Large metropolitan area 1

1 0.99 (0.96–1.02) 0.84* (0.81–0.87) 0.75

PT

0.80* Other metropolitan area (0.78–0.82) 0.93* City area (0.91–0.95) 0.75 C-statistics

A

1,000–1,499

1.80* (1.72–1.89)

M

Hospital bed number <1,000

Central nervous Physiologic/ Shock/Cardiac In-hospital Wound Pulmonary system metabolic arrest death infection failure complication derangement

2.41* (2.30–2.52)

ED

3

Sepsis

A

CC E

* p < .0001, † < .001, ‡ < .01, ¶< .05

19

Adjusted NSO incidence rates by nurse staffing grade The pattern of adjusted 12 NSO incidence rates by nurse staffing grade differed according to indicators (Figure 2). Six NSOs—urinary tract infection, upper gastrointestinal tract bleeding, hospital-acquired pneumonia, shock/cardiac arrest, in-hospital death, and wound infection—showed increasing trends as nurse staffing level degraded, even after adjusting for patient and hospital

IP T

characteristics. Nurse staffing grade 2 hospitals showed the lowest adjusted incidence for deep vein thrombosis, sepsis, and physiologic/metabolic derangement, while they showed the highest adjusted

SC R

incidence for pressure ulcer, CNS complication, and pulmonary failure. The difference in adjusted

incidence rate by nurse staffing grade for the 12 NSOs ranged from 0.01% to 1.06%. The absolute difference in shock/cardiac arrest rate between nurse staffing grade 1 and 3 hospitals was highest at

A

CC E

PT

ED

M

A

N

U

1.06% (0.85% vs. 1.91%) (Figure 2).

20

I N U SC R A M ED PT CC E A Figure 2. Adjusted incidence rate of nursing-sensitive outcome by nurse staffing grade

21

Discussion Twelve NSOs were calculated using all tertiary hospital reimbursement claim data submitted to the Korean HIRA during a two-year span from 2013. The incidence rates of nine NSOs for all inpatients were 0.86 - 2.84%, and those for three NSOs for only patients who underwent surgical operations were 0.28 - 1.29%. All reimbursement claim data to HIRA from all Korean tertiary

IP T

hospitals are processed electronically; therefore, the data in this study can be considered complete and

not sampled. As Needleman et al. (2002) indicated, underreporting is likely to occur if the staff

SC R

number is not sufficient. However, health administrative data from only the tertiary hospitals that are

presumed to have adequate personnel including physicians, nurses, and medical recorders were included in this study and are thus considered more reliable. Therefore, it is meaningful that reliable

U

NSOs were derived from a complete dataset of all tertiary hospitals in Korea.

N

Nurse staffing grade as a nursing quality surrogate was an influential factor for NSO rate.

A

When not adjusted for patient and hospital characteristics, ten NSOs—except for pressure ulcer and

M

physiologic/metabolic derangement—showed increasing rates as nurse staffing grade decreased from 1 to 3. However, the adjusted odds ratios, that showed significantly increasing trends as nurse staffing

ED

grade increased, were those of six NSOs—urinary tract infection, upper gastrointestinal tract bleeding, hospital-acquired pneumonia, shock/cardiac arrest, in-hospital death, and wound infection. In other

PT

words, these six NSOs showed a clear relationship between nurse staffing level and nursing quality. Our findings in four indicators (urinary tract infection, upper gastrointestinal tract bleeding, hospital-

CC E

acquired pneumonia, shock/cardiac arrest) are consistent with the results of Needleman et al. (2002). Two indicators (in-hospital death, and wound infection) were statistically significant in this study, while those in Needleman et al.’s study (2002) were not. One possible reason for this difference might

A

be stem from the different definition of nurse staffing level—we defined nurse staffing as bed-tonurse ratio, while Needleman et al. (2002) defined it as number of RN-hours per patient-day. It is reassuring that nursing quality is reflected by not only the rate of surgical mortality (Griffiths et al., 2016), but also by the rates of urinary tract infection, upper gastrointestinal track bleeding, hospitalacquired pneumonia, shock/cardiac arrest, and wound infection. As the bed-to-nurse ratio increased 22

from grade 1 to 2 and from 1 to 3, the incidence of urinary tract infection increased by 2.17 times and 2.41 times, respectively; in the same manner, the incidence of upper gastrointestinal track bleeding increased by 1.60 times and 1.80 times, shock/cardiac arrest by 1.77 times and 2.44 times, and wound infection by 1.88 times and 2.23 times, respectively. Therefore, adverse outcomes are inversely correlated with nurse staffing level.

IP T

Five NSOs—deep vein thrombosis, physiologic/metabolic derangement, CNS complication,

pulmonary failure and sepsis—showed weak relationships with nurse staffing grade after adjusting for

SC R

patient and hospital characteristics. Therefore, we found statistically significant associations between

higher nurse staffing level and rates of eleven NSOs except only for pressure ulcer. These findings are in line with the studies by Twigg et al. (2011, 2015). Using data from public hospitals from a single

U

city, Twigg et al. (2011) found that increased nurse staffing level was significantly associated with

N

eight NSOs such as CNS complication, urinary tract infection, pressure ulcer, pneumonia, upper

A

gastrointestinal tract bleeding, sepsis, physiologic/metabolic derangement, and shock/cardiac arrest.

M

Also, Twigg et al. (2015) reported that understaffing was significantly associated with eight NSOs such as wound infection, urinary tract infection, pressure ulcer, pneumonia, deep vein thrombosis,

ED

upper gastrointestinal tract bleeding, physiologic/metabolic derangement, and sepsis in a metropolitan teaching hospital. Twigg et al. measured nurse staffing by nursing hours per patient-day (2011), and

PT

defined “understaffed” as staff hours less than 8 hours for morning or afternoon shift, or less than 10 hours for night (2015). Two indicators (in-hospital death, pulmonary failure) were statistically

CC E

significant in our study, while they were not significant in Twigg et al.’s studies (2011, 2015). One possible reason for the discrepancy might be due the difference of the definition of nurse staffing level, and another reason might be due to the adjustment of hospital and patient characteristics in our study,

A

which seems to have been not carried out in Twigg et al.’s studies (2011, 2015). Also, the sample sizes between our current study and those of Twigg et al.’s studies are very different. On the other hand, Kane et al. (2007) reported a significant association between higher nurse staffing level and lower rates of pulmonary failure in ICU through meta-analysis. Likewise, our study used a considerably larger and broader patient group to show that nurse staffing level was significant associated with 23

pulmonary failure for all inpatients. There is controversy about the effect of nurse staffing on patient outcomes among studies (Clarke, 2007). One of the reason that the relationship is not consistent among studies might be due to the different methodologies (Brennan et al., 2013). The significance of the relationship between nurse staffing and outcomes can be affected by the source of the data. A multivariate analysis suggested that the association was stronger when the nurse staffing data was

IP T

from the California Office for Statewide Health Planning and Hospitals database (Kane et al., 2007).

Thus, discrepancies between our results and those of earlier studies are considered to be from the

SC R

measure of nurse staffing, data source, and statistical analysis.

The rate of pressure ulcer was higher in nurse staffing grade 1 or 2 hospitals than in nurse staffing grade 3 hospitals, which is consistent with the result of Twigg et al. (2013). It may need to be

U

adjusted using pressure ulcer risk assessment tools such as the Braden and Norton scales, which

N

include more sensitive risk factors including sensory function, incontinence, activity level, mobility,

A

and nutritional status (National Pressure Ulcer Advisory Panel, European Pressure Ulcer Advisory

M

Panel and Pan Pacific Pressure Injury Alliance, 2014). Griffiths et al. (2016) reviewed the inconsistency of the relationship between nurse staffing and pressure ulcer among studies, and we also

ED

showed that pressure ulcer was the highest in grade 2 hospitals. Nevertheless, nursing still carries a crucial role in the management and prevention of pressure ulcer (White et al., 2011). He et al. (2016)

PT

also reported that pressure ulcer was associated with seasonality. Further studies are necessary for providing a clearer explanation on the relationship between pressure ulcer and nurse staffing.

CC E

One of the limitations in this study is that only the data from tertiary hospitals with the upper

three levels of all seven nurse staffing grades were used; this study design was employed in order to increase the reliability of the administrative data. If the reliability of data from hospitals with lower

A

nurse staffing levels is increased, we will conduct an additional study for studying the relationships of nurse staffing level with nursing outcomes in such environment. Also, our study is limited by its use of reimbursement data to identity the incidence rate of NSOs. Coding behaviour in an accounting system is also an important factor for precisely monitoring NSOs. There is no definite coding guideline for NSOs in HIRA dataset; therefore, underreporting may exist. To prevent this, a coding 24

guideline that incorporates inclusion and exclusion rules should be established. Another limitation is that administrative data are lacking in variables reflecting nursing quality such as direct nursing time, nurse education level, nursing-care-pathways, and physician-nurse relationship. Physician practice pattern, collaboration with nurse, nurse manager ability and support, collegial nurse-physician relations also affect patient outcomes (Kane et al., 2007). Therefore, nurse staffing should be viewed

IP T

as a factor for the quality rather than as a measure of the full nursing effects in hospitals (Needleman

et al., 2002). To fully understand the relationship between nursing quality and patient outcome, the

SC R

contents of the nursing structural and process variables in databases need to be improved in Korea.

Conclusion

U

The two important results of this study are the definition of algorithms for deriving NSOs

N

from the administrative data of Korean national health insurance system, and the significant

A

relationship between nurse staffing and nursing quality that was derived from the NSOs. If future

M

studies use the methods suggested in this study, the trend of nurse staffing and NSOs over time may be made evident. Furthermore, if HIRA and other investigators monitor nursing quality levels using

ED

the methods shown in this study, this will help enhance the nursing quality and to inform patients of the nursing quality of hospitals in Korea. Finally, we provided strong evidence for the relationships of

PT

nurse staffing level with six nursing-sensitive outcomes, and we suggest that optimal nurse staffing is

CC E

essential for delivering high-quality care with increased focus on value as well as cost-effectiveness.

A

Contribution of the Paper

What is already known about the topic? 

Higher levels of nurse staffing could have positive impact on the quality of care.



Lower levels of hospital nurse staffing are associated with more adverse outcomes, especially mortality. 25

What this paper adds 

It identifies nursing-sensitive outcomes using large administrative databases in Korea.



It provides additional evidence supporting the relationships of nurse staffing level with patient outcomes in urinary tract infection, upper gastrointestinal tract bleeding, hospital-acquired

A

CC E

PT

ED

M

A

N

U

SC R

IP T

pneumonia, shock or cardiac arrest, in-hospital death, and wound infection.

26

Reference Aiken, L.H., Clarke, S.P., Sloane, D.M., Sochalski, J., Silber, J.H., 2002. Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. J. Am. Med. Assoc. 288 (16), 1987– 1993.

IP T

Aiken, L.H., Sloane, D.M., Bruyneel, L., Van den Heede, K., Griffiths, P., Busse, R., Diomidous, M., Kinnunen, J., Kozka, M., Lesaffre, E., McHugh, M.D., Moreno-Casbas, M.T., Rafferty, A.M.,

SC R

Schwendimann, R., Scott, P.A., Tishelman, C., van Achterberg, T., Sermeus, W., (RN4CAST consortium), 2014. Nurse staffing and education and hospital mortality in nine European countries:

a

retrospective

observational

study.

Lancet.

(9931),

1824–1830.

U

doi:http://dx.doi.org/ 10.1016/S0140-6736(13)62631-8.

383

N

American Nurses Association, 2000. Nurse staffing and patient outcomes in the inpatient hospital

A

setting. American Nurses Publishing, Washington.

Brennan, C.W., Daly, B.J., Jones, K.R., 2013. State of the science: the relationship between nurse and

patient

outcomes.

West

M

staffing

J.

Nurs.

Res.

35

(6),

760-794.

ED

doi:http://dx.doi.org/10.1177/0193945913476577. Charlson, M.E., Pompei, P., Ales, K.L., MacKenzie, C.R., 1987. A new method of classifying

PT

prognostic comorbidity in longitudinal studies: Development and validation. J. Chronic. Dis. 40 (5), 373–383.

CC E

Cho, S.H., Yun, S.C., 2009. Bed-to-nurse ratios, provision of basic nursing care, and in-hospital and 30-day mortality among acute stroke patients admitted to an intensive care unit: crosssectional analysis of survey and administrative data. Int. J. Nurs. Stud. 46 (8), 1092-1101.

A

doi:http://dx.doi.org/ 10.1016/j.ijnurstu.2009.02.001.

Clarke, S.P., 2007. Registered nurse staffing and patient outcomes in acute care: looking back, pushing forward. Med. Care. 45 (12), 1126-1128. Douglas, D.I., Robb, A., 1995. Clarifying outcomes in clinical practice. Nurs. Stand. 9 (24), 29-30. Dudley, R.A., Johansen, K.L., Brand, R., Rennie, D.J., Milstein, A., 2000. Selective referral to high27

volume hospitals: Estimating potentially avoidable deaths. J. Am. Med. Assoc. 283 (9), 1159– 1166. Fowler, D.M., 2008. Guide to the Code of Ethics for Nurses: Interpretation and Application. Nursesbooks.org, Maryland. Griffiths, P., Ball, J., Drennan, J., Dall'Ora, C., Jones, J., Maruotti, A., Pope, C., Recio Saucedo, A.,

IP T

Simon, M., 2016. Nurse staffing and patient outcomes: Strengths and limitations of the

evidence to inform policy and practice. A review and discussion paper based on evidence

development.

Int.

J.

Nurs.

Stud.

63,

10.1016/j.ijnurstu.2016.03.012.

SC R

reviewed for the National Institute for Health and Care Excellence Safe Staffing guideline 213-225.

doi:http://dx.doi.org/

U

Health Insurance Review and Assessment Service, 2015. Details on the criteria and method of

N

application of medical claim. Health Insurance Review and Assessment Service. Seoul Korea.

A

Health Insurance Review and Assessment Service, 2016. Comprehensive Quality report of national

M

health insurance 2015. Health Insurance Review and Assessment Service, Wonju Korea. He J., Staggs V. S., Bergquist-Beringer S., Dunton N., 2016. Nurse staffing and patient outcomes: a

ED

longitudinal study on trend and seasonality. BMC Nursing, 1-10, doi:http://dx.doi.org/ 10.1186/s12912-016-0181-3.

Sudbury.

PT

Kane, R.L., Radosevich, D.M., 2011. Conducting health outcomes research. Jones & Bartlett Learning,

CC E

Kane, R.L., Shamliyan, T.A., Mueller, C., Duval, S., Wilt, T.J., 2007. The association of registered nurse staffing levels and patient outcomes: Systematic review and meta-analysis. Med. Care. 45 (12), 1195–1204, doi:http://dx.doi.org/ 1110.1097/MLR.1190b1013e3181468ca3181463.

A

McCloskey, B.A., Diers, D.K., 2005. Effects of New Zealand's health reengineering on nursing and patient outcomes. Med. Care. 43 (11), 1140-1146.

Kim, S.M., Jang, W.M., Ahn, H.A., Park, H.J., Ahn, H.S., 2012. Korean National Health Insurance value incentive program: achievements and future directions. J. Prev. Med. Public Health. 45 (3), 148-155. doi:http://dx.doi.org/10.3961/jpmph.2012.45.3.148. 28

Kim, Y., Cho, S.H., June, K.J., Shin, S.A., Kim, J., 2012. Effects of hospital nurse staffing on inhospital mortality, pneumonia, sepsis, and urinary tract infection in surgical patients. J. Korean Acad. Nurs. 42 (5), 719-729. doi:http://dx.doi.org/10.4040/jkan.2012.42.5.719. National Pressure Ulcer Advisory Panel, European Pressure Ulcer Advisory Panel and Pan Pacific Pressure Injury Alliance, 2014. Prevention and Treatment of Pressure Ulcers: Quick

IP T

Reference Guide. Emily Haesler (Ed.). Cambridge Media. Perth.

Needleman J., Buerhaus P., Mattke S., Stewart M., Zelevinsky K., 2002. Nurse-staffing levels and the

SC R

quality of care in hospitals. N. Engl. J. Med. 346 (22), 1715-1722.

Rafferty, A.M., Clarke, S.P., Coles, J., Ball, J., James, P., McKee, M., Aiken, L.H., 2007. Outcomes of variation in hospital nurse staffing in English hospitals: cross-sectional analysis of survey data

U

and discharge records. Int. J. Nurs. Stud. 44 (2), 175–182.

N

Park, J.K., Kim, K.S., Kim, C.B., Lee, T.Y., Lee, K.S., Lee, D.H., Lee, S., Jee S.H., Suh, I., Koh, K.,

A

Ryu, S.Y., Park, K., Park, W., Wang, S., Lee H., Chae, Y., Hong, H., Suh, J.S., 2000. The

Prev. Med. 33 (1), 76-82.

M

Accuracy of ICD codes for Cerebrovascular Diseases in Medical Insurance Claims. Korean J.

ED

Park, S. H., 2003. Analysis of direct nursing activity and patient outcomes related to graded fee of nursing management for inpatient. J. Korean Acad. Nurs. 33(1), 122-129.

quality

PT

Rochefort, C. M., Clarke, S. P., 2010. Nurses' work environments, care rationing, job outcomes, and of

care

on

neonatal

units.

J.

Adv.

Nurs.

66

(10),

2213-2224.

CC E

doi:http://dx.doi.org/10.1111/j.1365-2648.2010.05376.x.

Romeyke, T., Scheuer, H.C., Stummer, H., 2012. Introduction of the German case tariff fee system

A

and its effects on patient satisfaction in inpatient naturopathy. Eur. J. Integrative Med. 5 (2), 171-177. doi:https://doi.org/10.1016/j.eujim.2012.11.005.

Sasichay-Akkadechanunt, T., Scalzi, C.C., Jawad, A.F., 2003. The relationship between nurse staffing and patient outcomes. J. Nurs. Adm. 33 (9), 478-485. Schreuders, L.W., Bremner, A.P., Geelhoed, E., Finn, J., 2014. Using linked hospitalisation data to detect nursing-sensitive outcomes: a retrospective cohort study. Int. J. Nurs. Stud. 51 (3), 47029

478. doi:http://dx.doi.org/ 0.1016/j.ijnurstu.2013.06.006. Tourangeau, A.E., Doran, D.M., McGillis Hall, L., O'Brien Pallas, L., Pringle, D., Tu, J.V., Cranley, LA., 2007. Impact of hospital nursing care on 30-day mortality for acute medical patients. J. Adv. Nurs. 57 (1), 32-44. Twigg, D.E., Duffield, C.M., Bremner, A.P., Rapley, P., Finn, J., 2011. The impact of the nursing

of

patient

and

staffing

data.

Int.

J.

Nurs.

Stud.

48

(5),

540–548.

SC R

doi:http://dx.doi.org/10.1016/j.ijnurstu.2010.07.013.

IP T

hours per patient day (NHPPD) staffing method on patient outcomes: a retrospective analysis

Twigg, D.E., Geelhoed, E.A., Bremner, A.P., Duffield, C.M., 2013. The economic benefits of increased levels of nursing care in the hospital setting. J. Adv. Nurs. 69 (10), 2253–2261.

U

doi:http://dx.doi.org/10.1111/jan.12109.

N

Twigg D.E., Gelder L., Myers H., 2015 The impact of understaffed shifts on nurse-sensitive outcomes.

A

J. Adv. Nurs. 71 (7), 1564-72. doi:http://dx.doi.org/ 10.1111/jan.12616.

M

White, P., McGillis Hall, L., Lalonde, M., 2011. Adverse patient outcomes, in Doran M.D. (Eds.), Nursing outcomes: the state of the science. Jones & Bartlett Learning, Sudbury, pp. 241-284.

ED

Wilson, S., Bremner, A., Hauck, Y., Finn, J., 2011. The effect of nurse staffing on clinical outcomes of children in hospital: A systematic review. Int. J. Evidence-Based Healthcare. 9, 97-121.

PT

Wilson, S., Bremner, A.P., Hauck, Y., Finn, J., 2012. Identifying paediatric nursing-sensitive outcomes in linked administrative health data. BMC Health Serv. Res. 12 (209), 1-12.

CC E

doi:http://dx.doi.org/10.1186/1472-6963-12-209.

Wilson, S., Bremner, A.P., Hauck, Y., Finn, J., 2013. Evaluation of paediatric nursing-sensitive

A

outcomes in an Australian population using linked administrative hospital data. BMC Health Serv. Res. 13 (396), 1-14. doi:http://dx.doi.org/10.1186/1472-6963-13-396.

30