Post-intensive care unit depression among critical care survivors:

(1)

O R I G I N A L A R T I C L E

Post-intensive care unit depression among critical care survivors:

A nationwide population-based study

Jiyeon Kang¹ | Seonyoung Yun² | Young Shin Cho³ | Yeon Jin Jeong⁴

1College of Nursing, Dong-A University, Busan, Republic of Korea

2Department of Nursing, Youngsan University, Yangsan, Kyungnam, Republic of Korea

3Surgical Intensive Care Unit, Kosin University Gospel Hospital, Busan, Republic of Korea

4Department of Nursing, DongJu College, Busan, Republic of Korea

Correspondence

Seonyoung Yun, Department of Nursing, Youngsan University, 288 Junam-ro, Yangsan-si, Gyeongsangnam-do, 50510, Republic of Korea.

Email: [email protected]

Present address

Jiyeon Kang, College of Nursing, Dong-A University, 32, Daesingongwon-ro, Seo-gu, Busan, 49201, Republic of Korea.

Seonyoung Yun, Department of Nursing, Youngsan University, 288, Junam-ro, Yangsan-si, Gyeongsangnam-do, 50510, Republic of Korea.

Young Shin Cho, Surgical Intensive Care Unit, Kosin University Gospel Hospital, 262, Gamcheon-ro, Seo-gu, Busan, 49267, Republic of Korea.

Yeon Jin Jeong, Department of Nursing, DongJu College, 16, Sari-ro 55beon-gil, Saha-gu, Busan, 49318, Republic of Korea.

Funding information

Basic Science Research Program through the National Research Foundation of Korea, Grant/Award Number: NRF-

2016R1D1A1B03936044

Abstract

Aim:To investigate the incidence of post-intensive care unit (ICU) depression and its risk factors among critical care survivors.

Methods:The study data were extracted from the database of the National Health Insur- ance Service of Korea. We retrospectively analyzed data from 161,977 adult patients who were admitted to the ICU for more than 24 hr from January 1, 2012 to December 31, 2014 and survived for more than 1 year after discharge. Risk factors for newly diagnosed depression (Code F32) were analyzed using multiple logistic regression analysis.

Results:The incidence of post-ICU depression was 18.5%. The major risk factors were enteral nutrition (odds ratio [OR] = 2.28, 95% confidence interval [CI] = 2.19–2.36), cerebrovascular disease (OR = 1.59, 95% CI = 1.54–1.64), and hemi/paraplegia (OR = 1.48, 95% CI = 1.41–1.56). It was observed that cardiopulmonary resuscitation (OR = 0.55, 95% CI = 0.50–0.61) and myocardial infarction (OR = 0.75, 95% CI = 0.71–0.79) lowered depression.

Conclusions:The incidence of post-ICU depression was high and influenced by ICU treatment and physical impairments. Healthcare providers must pay attention to the psychological changes in survivors with major risk factors in the recovery process.

K E Y W O R D S

depression, enteral nutrition, intensive care units, risk factors, survivors

1 | I N T R O D U C T I O N

With population aging and the increase in chronic diseases, the number of patients in intensive care units (ICUs) has been increasing (Halpern & Stephen, 2010; Park et al., 2018). However, mortality in the ICU is declining, with

more than 80% of patients being discharged (Hill et al., 2016; Zimmerman, Kramer, & Knaus, 2013). The increase in ICU survivors has led to an interest in the long-term health problems they experience after discharge (Davydow, Gifford, Desai, Bienvenu, & Needham, 2009; Jackson et al., 2014; Needham et al., 2012).

DOI: 10.1111/jjns.12299

(2)

ICU survivors experience a variety of physical, mental, and cognitive impairments, collectively known as postintensive care syndrome (PICS) (Needham et al., 2012). In particular, PICS symptoms include muscle weakness, fatigue, sleep disturbances, decreased activities of daily liv- ing, anxiety, depression, posttraumatic stress disorder, and memory impairment (Davidson, Harvey, Schuller, & Black, 2013; Jackson et al., 2014; Needham et al., 2012). PICS symptoms persist for years after discharge and have a negative impact on survivors' quality of life (Dowdy et al., 2005;

Ringdal, Plos, Lundberg, Johansson, & Bergbom, 2009;

Steenbergen et al., 2015).

Among PICS symptoms, depression is a major factor in survivors' cognitive and functional impairments and the reduction in their quality of life (Davydow et al., 2009; Jutte, Erb, & Jackson, 2015). The prevalence of depressive symptoms at 12 months after ICU discharge is 29–33% higher than that of the general population (Jackson et al., 2014;

Rabiee et al., 2016). Jackson et al. (2014) reported that about 30% of depressed survivors did not have a history of pre- ICU depression. Wunsch et al. (2014) found that ICU survivors increased the risk of diagnosis of new psychiatric disorders during the first few months after discharge. In particular, the proportion of antidepressant prescriptions increased steadily during the first year after discharge from the ICU.

According to a systematic review (Rabiee et al., 2016), the demographic characteristics (e.g., age, sex), duration of hospital stay, disease severity, and use of sedatives and anal- gesics in the ICUs were not related to depressive symptoms after ICU discharge. The most relevant factor for the development of depressive symptoms in the survivors was pre- ICU depression history. However, previous studies on post- ICU depression included subjects who had depressive symptoms before entering ICUs. Among the 38 studies included in Rabiee et al.'s systematic review (2016), only four studies excluded subjects who had psychiatric problems prior to ICU admission. In such cases, it is unclear whether depression after ICU discharge is related to the ICU treatment or pre-existing depression. Therefore, it is necessary to investigate the incidence and risk factors for post-ICU depression, a newly developed depressive disorder after ICU discharge in survivors without pre-ICU depression.

There has been an increase in the number of critical care studies using long-term nationwide big data (Davydow, Hough, Langa, & Iwashyna, 2013; Park et al., 2018). Since Korea has achieved universal healthcare coverage, data related to the use of medical institutions and pharmacies by almost all citizens are stored and managed by the National Health Insurance Service (NHIS) (Lee, Kim, & Lee, 2012).

The NHIS has established the National Health Information Database (NHID) for all health insurance and medical

benefits entities, and currently has data from 2002 to 2015.

The NHIS provides NHID data in various forms to support policy development and academic research (National Health Insurance Sharing Service, 2017). Accordingly, we aimed to identify the incidence of post-ICU depression and its risk factors in critical care survivors using NHID data.

2 | M E T H O D S 2.1 | Study design

This was a retrospective case–control study using Korean NHID data to determine the incidence of post-ICU depression and its risk factors in critical care survivors.

2.2 | Data sources

The data for the current study, extracted from the NHID, were obtained after receiving formal approval from the NHIS (#REQ0000009365). To analyze the NHID data, we performed the following four steps. In the first stage, we identified all subjects with ICU admission records from January 1, 2011 to December 31, 2015. Then, the subjects who were admitted to the ICU before this period, those who had been diagnosed with depression, those who were under 18 years of age, those who were admitted for less than 24 hr, and those who died before discharge, were excluded. If a subject's ICU admission code was entered multiple times during this period, the data of the first ICU admission were extracted. The number of subjects extracted from this stage was 497,603. In the second stage, we performed washing out of the data. First, the subjects who were admitted to hospital after January 1, 2015 were excluded in order to obtain the death and depression diagnosis data for one year after discharge. Patients who were admitted between January 1 and December 31, 2011 were also excluded to confirm the experience of risk factors for depression before entering the ICU. The number of subjects selected through this stage was 300,233. In stage 3, we excluded those who died within one year of discharge, those who were diagnosed with depression during ICU admission, and those who had experienced possible risk factors for depression before ICU admission. The number of subjects selected through this step was 161,977. In stage 4, we classi- fied these subjects into a control group and a case group with new diagnosis of depression within one year of discharge.

The new depression was confirmed using the“F32”code of the Korean Standard Classification of Disease 7th edition (KCD-7) reflecting the International Classification of Dis- eases 10th revision. The F32.x code is a disease code for newly developed major depression (Korea Information Clas- sification of Diseases, 2016). In Korea, depressive disease codes must be entered at the time of prescribing anti-

(3)

depressants, so additional drug codes for depression were not identified. Therefore, among adult patients admitted to ICU for more than 24 hr during the period from January 1, 2012 to December 31, 2014, those who were diagnosed with depression for the first time within one year of discharge became the case group of the current study (Figure 1).

ICU admission and risk factors for depression were identified using the Health Insurance Review and Assessment billing code entered in the NHID (Health Insurance Review & Assess- ment Service, 2017a). The AJ100-AJ390 codes were used for ICU admission and codes for the neonatal and pediatric ICU and ICU in nursing homes were excluded. We directly extracted demographic characteristics such as age, sex, type of health insurance, income quintile, and disability from the NHID as risk factors for depression. The risk factors related to ICU treatment, including mechanical ventilator use (M5850, M5857, M5858, M5860), continuous renal replacement therapy (CRRT; O7051-O7054, O7031-O7034), extracorporeal membrane oxygenation (ECMO; O1901-O1904), cardiopulmonary resuscitation (CPR; M5877), defibrillation (M5880), intra-aortic balloon pumping (IABP; O1921, O1922), and enteral nutrition (Q2660, Q2661), were extracted. Comorbid disease and severity were confirmed using the Charlson Comorbidity Index (CCI) and KCD-7 codes (Charlson,

Pompei, Ales, & MacKenzie, 1987; Kim, 2010). Drugs such as muscle relaxants (e.g., vecuronium, cisatracurium, rocuronium, atracurium), sedatives (e.g., midazolam, loraze- pam, diazepam), and inotropes/vasopressors (e.g., dopamine, dobutamine, norepinephrine) were identified and analyzed using the corresponding active pharmaceutical ingredient codes (Health Insurance Review & Assessment Service, 2017b).

2.3 | Statistical analysis

Differences in demographic and ICU treatment characteristics between the case and control groups were analyzed using the Chi-square and t-tests. The Chi-square andt-tests were also used for additional analysis on enteral nutrition. Risk factors for depression were analyzed using multiple logistic regression analysis. The above analyses were performed using the SAS 9.4 program (SAS Institute Inc., Cary, NC, USA).

2.4 | Ethical considerations

This study was reviewed by the Institutional Review Board (IRB) (No: 2-104709-AB-N-01-201706-HR-028-02) of Dong-A University. We received the data from the NHIS with removed identifying information. Data analysis was

January 1, 2011-December 31, 2015 ICU admission record (NHID) Excluded:

- under 18 years old

- ICU admission in the past (2002-2010) - diagnosed with depression in the past (2002-2010) - ICU admission less than 24 hours

- death in ICU (death before discharge) n = 497,603

January 1, 2012 - December 31, 2014 Eligible subjects

n = 300,233

Case group (post-ICU depression

; F32 code) n = 30,023

Eligible subjects n = 161,977

Excluded:

- death within one year after discharge - depression occurred during ICU admission - those who experienced risk factors before

ICU admission

Control group (non-depression)

n = 131,954 Washout

1st: January 1-December 31, 2015 2nd: January 1-December 31, 2011

F I G U R E 1 Flowchart of subject selection. ICU, intensive care unit; NHID, National Health Information Database

(4)

T A B L E 1 Comparison of characteristics between depressive and control groups (N= 161,977)

Characteristics Categories

Total (N= 161,977) M ± SD orn(%)

Depressive group (n= 30,023) M ± SD orn(%)

Control

group (n= 131,954)

M ± SD orn(%) χ²ort(p) Age

18–29 30–39 40–49 50–59 60–69 70–79 80–89 90≤

60.5 ± 15.6 9,914 (6.1) 11,388 (7.0) 24,737 (15.3) 37,847 (23.4) 34,643 (21.4) 31,385 (19.4) 11,043 (6.8) 1,020 (0.6)

61.8 ± 14.8 1,338 (4.5) 1,801 (6.0) 4,339 (14.5) 6,926 (23.1) 6,781 (22.6) 6,590 (21.9) 2,116 (7.0) 132 (0.4)

60.3 ± 15.7 8,576 (6.5) 9,587 (7.3) 20,398 (15.5) 30,921 (23.4) 27,862 (21.1) 24,795 (18.8) 8,927 (6.7) 888 (0.7)

−15.48 (<.001) 414.89 (<.001)

Gender Male

Female

100,928 (62.3) 61,049 (37.7)

17,443 (58.1) 12,580 (41.9)

83,485 (63.3) 48,469 (36.7)

278.31 (<.001)

Type of insurance Local

Employer-provided Medicaid

59,685 (36.8) 92,363 (57.0) 9,929(6.2)

11,379 (37.9) 16,428 (54.7) 2,216 (7.4)

48,306 (36.6) 75,935 (57.5) 7,713 (5.9)

139.52 (<.001)

Income 0

1–4 5–10 11–15 16–20

37,640 (23.2) 24,101 (14.9) 42,404 (26.2) 8,078 (5.0) 49,754 (30.7)

7,203 (24.0) 4,495 (15.0) 7,740 (25.8) 1,463 (4.9) 9,122 (30.3)

30,437 (23.1) 19,606 (14.9) 34,664 (26.3) 6,615 (5.0) 40,632 (30.7)

13.76 (.01)

Disability registration Yes No

22,910 (14.1) 139,067 (85.9)

4,858 (16.2) 25,165 (83.8)

18,052 (13.7) 113,902 (86.3)

125.92 (<.001)

Length of hospital stay 15.3 ± 12.0 18.9 ± 13.9 14.5 ± 11.4 −58.55 (<.001)

Mechanical ventilation Yes No

35,291 (21.8) 126,686 (78.2)

8,012 (26.7) 22,011 (73.3)

27,279 (20.7) 104,675 (79.3)

518.96 (<.001)

CRRT Yes

No

2,763 (1.7) 159,214 (98.3)

609 (2.0) 29,414 (98.0)

2,154 (1.6) 129,800 (98.4)

22.88 (<.001)

ECMO Yes

No

635 (0.4) 161,342 (99.6)

114 (0.4) 29,909 (99.6)

521 (0.4) 131,433 (99.6)

0.14 (.71)

CPR Yes

No

3,958 (2.4) 158,019 (97.6)

519 (1.7) 29,504 (98.3)

3,439 (2.6) 128,515 (97.4)

79.01 (<.001)

Defibrillation Yes

No

4,543 (2.8) 157,434 (97.2)

603 (2.0) 29,420 (98.0)

3,940 (3.0) 128,014 (97.0)

85.72 (<.001)

IABP Yes

No

887 (0.5) 161,090 (99.5)

120 (0.4) 29,903 (99.6)

767 (0.6) 131,187 (99.4)

14.81 (.001)

Enteral nutrition Yes

No

23,469 (14.5) 138,508 (85.5)

8,538 (28.4) 21,485 (71.6)

14,931 (11.3) 117,023 (88.7)

5,787.82(<.001)

Sedatives Yes

No

84,170 (52.0) 77,807 (48.0)

17,575 (58.5) 12,448 (41.5)

66,595 (50.5) 65,359 (49.5)

638.14 (<.001)

Muscle relaxants Yes

No

53,774 (33.2) 108,203 (66.8)

10,268 (34.2) 19,755 (65.8)

43,506 (33.0) 88,448 (67.0)

16.68 (<.001)

Inotropes/vasopressors Yes No

46,627 (28.8) 115,350 (71.2)

8,489 (28.3) 21,534 (71.7)

38,138 (28.9) 93,816 (71.1)

4.70 (.03)

CCI 3.61 ± 2.63 4.00 ± 2.70 3.52 ± 2.60 −28.28 (<.001)

Cerebrovascular disease Yes No

51,358 (31.7) 110,619 (68.3)

14,094 (46.9) 15,929 (53.1)

37,264 (28.2) 94,690 (71.8)

3,951.44 (<.001)

Congestive heart failure Yes No

23,424 (14.5) 138,553 (85.5)

4,199 (14.0) 25,824 (86.0)

19,225 (14.6) 112,729 (85.4)

6.73 (.01)

(Continues)

(5)

performed at designated locations and computers. The resulting file was exported after approval from the NHIS.

The raw data could not be exported.

3 | R E S U L T S

3.1 | The incidence of post-ICU depression According to NHIS data, among the adult patients without a history of depression admitted to ICU for more than 24 hr between January 1, 2012 and December 31, 2014, 161,977 survived for more than one year after discharge. Of these, 30,023 were diagnosed with depression for the first time within one year of ICU discharge. The incidence of post- ICU depression in Korean critical care survivors was 18.5%.

3.2 | Comparison of characteristics between depressive and control groups

Table 1 shows the differences in demographic and ICU treatment characteristics between the two groups. Among the

demographic characteristics, the mean age (61.8

± 14.8 years vs. 60.3 ± 15.7 years) and the proportion of women (41.9% vs. 36.7%) and the registered disabled (16.2% vs. 13.7%) were higher in the depressive group.

Compared with the control group, the depressive group had lower rates of employee-insured type health insurance coverage (54.7% vs. 57.5%) and higher rates of medical aid bene- ficiaries (7.4% vs. 5.9%).

Among the ICU treatment characteristics, the length of stay (18.9 ± 13.9 days vs. 14.5 ± 11.4 days) and the use of mechanical ventilators (26.7% vs. 20.7%), CRRT (2.0%

vs. 1.6%), sedatives (58.5% vs. 50.5%), and muscle relaxants (34.2% vs. 33.0%) was higher in the depressive group than in the control group. However, CPR (1.7% vs. 2.6%), defibrillation (2.0% vs. 3.0%), IABP (0.4% vs. 0.6%), and the use of inotropes/vasopressors (28.3% vs. 28.9%) was lower in the depressive group.

The CCI score of the depressive group was higher than that of the control group (4.00 ± 2.70 vs. 3.52 ± 2.60).

Among comorbid diseases, the proportions of cerebrovascular disease (46.9% vs. 28.2%), hemi/paraplegia (11.3%

T A B L E 1 (Continued)

Total (N= 161,977) M ± SD orn(%)

Depressive group (n= 30,023) M ± SD orn(%)

Control

group (n= 131,954)

M ± SD orn(%) χ²ort(p) Chronic pulmonary disease Yes

No

70,327 (43.3) 91,650 (56.7)

13,669 (45.5) 16,354 (54.5)

56,658 (42.9) 75,296 (57.1)

66.82 (<.001)

Myocardial infarction Yes No

19,854 (12.3) 142,123 (87.7)

2,289 (7.6) 27,734 (92.4)

17,565 (13.3) 114,389 (86.7)

735.58 (<.001)

Diabetes Yes

No

63,102 (39.0) 98,875 (61.0)

12,373 (41.2) 17,650 (58.8)

50,729 (38.4) 81,225 (61.6)

78.76 (<.001)

Liver disease Yes

No

57,775 (35.7) 104,202 (64.3)

11,263 (37.5) 18,760 (62.5)

46,512 (35.2) 85,442 (64.8)

54.73 (<.001)

Peptic ulcer disease Yes No

64,497 (39.8) 97,480 (60.2)

12,339 (41.1) 17,684 (58.9)

52,158 (39.5) 79,796 (60.5)

25.19 (<.001)

Kidney disease Yes

No

7,774 (4.8) 154,203 (95.2)

1,669 (5.6) 28,354 (94.4)

6,105 (4.6) 125,849 (95.4)

46.54 (<.001)

Rheumatic disease Yes

No

8,675 (5.4) 153,302 (94.6)

1,895 (6.3) 28,128 (93.7)

6,780 (5.1) 125,174 (94.9)

66.47 (<.001)

Peripheral vascular disease Yes No

31,880 (19.7) 130,097 (80.3)

6,596 (22.0) 23,427 (78.0)

25,284 (19.2) 106,670 (80.8)

122.05 (<.001)

Cancer Yes

No

32,834 (20.3) 129,143 (79.7)

5,752 (19.2) 24,271 (80.8)

27,082 (20.5) 104,872 (79.5)

28.20 (<.001)

Dementia Yes

No

8,202 (5.1) 153,775 (94.9)

2,136 (7.1) 27,887 (92.9)

6,066 (4.6) 125,888 (95.4)

322.45 (<.001)

Hemi/paraplegia Yes

No

9,064 (5.6) 152,913 (94.4)

3,397 (11.3) 26,626 (88.7)

5,667 (4.3) 126,287 (95.7)

2,281.59 (<.001)

AIDS/HIV Yes

No

113 (0.1) 161,864 (99.9)

17 (0.1) 30,006 (99.9)

96 (0.1) 131,858 (99.9)

0.91 (.34)

Abbreviations: AIDS/HIV, acquired immune deficiency syndrome/human immunodeficiency virus; CCI, Charlson Comorbidity Index; CPR, cardiopulmonary resuscitation; CRRT, continuous renal replacement therapy; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pumping.

(6)

T A B L E 2 Logistic regression analysis of risk factors for post-ICU depression

Variables Categories B SE OR 95% CI p

Age ref = 18–29

30–39 0.22 0.04 1.24 1.15–1.34 <.001

40–49 0.29 0.04 1.34 1.25–1.43 <.001

50–59 0.30 0.03 1.36 1.27–1.45 <.001

60–69 0.31 0.04 1.37 1.28–1.46 <.001

70–79 0.27 0.04 1.31 1.22–1.40 <.001

80–89 0.06 0.04 1.06 0.98–1.15 .168

90≤ −0.35 0.10 0.70 0.58–0.86 .001

Gender Ref = male 0.15 0.01 1.16 1.13–1.20 <.001

Type of insurance Local (ref)

Employer-provided −0.09 0.01 0.92 0.89–0.94 <.001

Medicaid 0.03 0.03 1.03 0.97–1.10 .297

Income 0 (ref)

1–4 0.02 0.02 1.02 0.97–1.07 .413

5–10 −0.00 0.02 1.00 0.96–1.04 .933

11–15 −0.02 0.03 0.98 0.92–1.05 .552

16–20 0.00 0.02 1.00 0.96–1.05 .850

Disability registration Ref = no 0.03 0.02 1.03 1.00–1.07 .090

Length of hospital stay 0.02 0.00 1.02 1.02–1.02 <.001

Mechanical ventilation Ref = no 0.06 0.02 1.06 1.02–1.10 .003

CRRT Ref = no −0.05 0.05 0.96 0.87–1.05 .354

CPR Ref = no −0.59 0.05 0.55 0.50–0.61 <.001

Defibrillation Ref = no −0.08 0.05 0.92 0.84–1.01 .084

IABP Ref = no 0.08 0.10 1.08 0.88–1.32 .454

Enteral nutrition Ref = no 0.82 0.02 2.28 2.19–2.36 <.001

Sedatives Ref = no 0.13 0.02 1.14 1.10–1.17 <.001

Muscle relaxants Ref = no −0.10 0.02 0.91 0.88–0.94 <.001

Inotropes/vasopressors Ref = no −0.20 0.02 0.82 0.79–0.84 <.001

CCI 0.04 0.01 1.04 1.03–1.05 <.001

Cerebrovascular disease Ref = no 0.46 0.02 1.59 1.54–1.64 <.001

Congestive heart failure Ref = no −0.06 0.02 0.94 0.90–0.98 .003

Chronic pulmonary disease Ref = no 0.04 0.02 1.04 1.01–0.98 .011

Myocardial infarction Ref = no −0.29 0.03 0.75 0.71–0.79 <.001

Diabetes Ref = no 0.00 0.02 1.00 0.97–1.04 .838

Liver disease Ref = no 0.07 0.02 1.07 1.04–1.11 <.001

Peptic ulcer disease Ref = no 0.04 0.02 1.04 1.01–1.07 .008

Kidney disease Ref = no 0.08 0.03 1.08 1.02–1.15 .016

Rheumatic disease Ref = no 0.10 0.03 1.10 1.04–1.16 .001

Peripheral vascular disease Ref = no 0.05 0.02 1.05 1.01–1.09 .006

Cancer Ref = no −0.12 0.02 0.89 0.84–0.93 <.001

Dementia Ref = no 0.08 0.03 1.09 1.03–1.15 .004

Hemi/paraplegia Ref = no 0.39 0.03 1.48 1.41–1.56 <.001

Abbreviations: CCI, Charlson Comorbidity Index; CI, confidence interval; CPR, cardiopulmonary resuscitation; CRRT, continuous renal replacement therapy; IABP, intra-aortic balloon pumping; OR, odds ratio; ref, reference; SE, standard error.

(7)

vs. 4.3%), dementia (7.1% vs. 4.6%), chronic pulmonary disease (45.5% vs. 42.9%), diabetes (41.2% vs. 38.4%), liver disease (37.5% vs. 35.2%), peptic ulcer disease (41.1%

vs. 39.5%), kidney disease (5.6% vs. 4.6%), rheumatic disease (6.3% vs. 5.1%), and peripheral vascular disease (22.0%

vs. 19.2%) were higher in the depressive group than in the control group. Conversely, the rates of myocardial infarction (7.6% vs. 13.3%), heart failure (14.0% vs. 14.6%), and cancer (19.2% vs. 20.5%) were lower in the depressive group than in the control group.

3.3 | Risk factors for post-ICU depression Multivariate logistic regression analysis was performed to determine the risk factors for post-ICU depression, with the

characteristics that were significantly different in the univari- ate analysis as the independent variables and depression as the dependent variable. This regression model was statisti- cally significant (χ² = 9,881.75, p< .001), and the results are shown in Table 2. The risk of post-ICU depression was significantly higher in the group with enteral nutrition (OR = 2.28, 95% CI = 2.19–2.36), cerebrovascular disease (OR = 1.59, 95% CI = 1.54–1.64), and hemi/paraplegia (OR = 1.48, 95% CI = 1.41–1.56). Compared to the 18–29 age group, the risk of depression was higher in the groups aged between 30 and 70 and lower in the group aged 90 and over. Women had an increased risk of developing depression. In addition, the length of stay, use of the mechanical ventilators and sedatives, CCI score, chronic pulmonary disease, liver disease, peptic ulcer disease, kidney disease, T A B L E 3 Comparison of characteristics according to experience of enteral nutrition (N= 161,977)

Total (N= 161,977) M ± SD orn(%)

Enteral nutrition group (n= 23,469) M ± SD orn(%)

Non-enteral nutrition group (n= 138,508)

M ± SD orn(%) χ²or t (p)

Age 60.5 ± 15.6 63.4 ± 15.8 60.1 ± 15.5 −30.41 (<.001)

Gender Male 100,928 (62.3) 14,197 (60.5) 86,731 (62.6) 38.61 (<.001)

Female 61,049 (37.7) 9,272 (39.5) 51,777 (37.4)

Disability Yes 22,910 (14.1) 4,397 (18.7) 18,513 (13.4) 476.45 (<.001)

Registration No 139,067 (85.9) 19,072 (81.3) 119,995 (86.6)

Length of Hospital stay 15.3 ± 12.0 23.1 ± 15.3 14.0 ± 10.8 −111.11 (<.001)

Mechanical ventilation Yes 35,291 (21.8) 13,460 (57.4) 21,831 (15.8) 20,371.49 (<.001)

No 126,686 (78.2) 10,009 (42.6) 116,677 (84.2)

CRRT Yes 2,763 (1.7) 1,277 (5.4) 1,486 (1.1) 2,284.01 (<.001)

No 159,214 (98.3) 22,192 (94.6) 137,022 (98.9)

ECMO Yes 635 (0.4) 360 (1.5) 275 (0.2) 916.47 (<.001)

No 161,342 (99.6) 23,109 (98.5) 138,233 (99.8)

CPR Yes 3,958 (2.4) 1,635 (7.0) 2,323 (1.7) 2,355.39 (<.001)

No 158,019 (97.6) 21,834 (93.0) 136,185 (98.3)

Defibrillation Yes 4,543 (2.8) 1,058 (4.5) 3,485 (2.5) 292.11 (<.001)

No 157,434 (97.2) 22,411 (95.5) 135,023 (97.5)

IABP Yes 887 (0.5) 221 (0.9) 666 (0.5) 78.25 (<.001)

No 161,090 (99.5) 23,248 (99.1) 137,842 (99.5)

Sedatives Yes 84,170 (52.0) 17,836 (76.0) 66,334 (47.9) 6,351.18 (<.001)

No 77,807 (48.0) 5,633 (24.0) 72,174 (52.1)

Muscle relaxants Yes 53,774 (33.2) 10,130 (43.2) 43,644 (31.5) 1,228.86 (<.001)

No 108,203 (66.8) 13,339 (56.8) 94,864 (68.5)

Inotropes/

vasopressors

Yes 46,627 (28.8) 11,082 (47.2) 35,545 (25.7) 4,549.28 (<.001)

No 115,350 (71.2) 12,387 (52.8) 102,963 (74.3)

CCI 3.61 ± 2.63 3.77 ± 2.56 3.58 ± 2.64 −10.34 (<.001)

Abbreviations: CCI, Charlson Comorbidity Index; CPR, cardiopulmonary resuscitation; CRRT, continuous renal replacement therapy; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pumping.

(8)

rheumatic disease, peripheral vascular disease, and dementia were factors that increased the risk of depression. However, in subjects who underwent CPR (OR = 0.55, 95%

CI = 0.50–0.61) and those with myocardial infarction (OR = 0.75, 95% CI = 0.71–0.79), the risk of developing depression was reduced. Employee insurance coverage, muscle relaxants, inotropes/vasopressors, heart failure, and cancer also reduced the risk of developing depression.

3.4 | Additional analysis

In the current study, enteral nutrition was the risk factor with the highest odds for post-ICU depression among critical care survivors. We conducted additional analyses to characterize the subjects with enteral nutrition, and the results are shown in Table 3. The average age of the enteral nutrition group was higher than that of the non-enteral nutrition group (63.4

± 15.8 years vs. 60.1 ± 15.5 years). The proportion of women (39.5% vs. 37.4%) and persons with disabilities (18.7% vs. 13.4%) was also higher. The length of stay was longer in the enteral nutrition group (23.1 ± 15.3 vs. 14.0 ± 10.8 days). The use of mechanical ventilators (57.4% vs. 15.8%), CRRT (5.4% vs. 1.1%), ECMO (1.5%

vs. 0.2%), IABP (0.9% vs. 0.5%), sedatives (76.0%

vs. 47.9%) and muscle relaxants (43.2% vs. 31.5%), experience of CPR (7.0% vs. 1.7%), and CCI scores (3.77 ± 2.56 vs. 3.58 ± 2.64) were higher in the enteral nutrition group.

4 | D I S C U S S I O N

An analysis of Korea's nationwide data on critical care survivors showed that the incidence of post-ICU depression within one year of discharge was 18.5%. As of 2015, the worldwide prevalence of depression was estimated at 4.4%

(World Health Organization, 2017). In the 2014 Korea National Health and Nutrition Examination Survey, the prevalence of depression measured by the screening ques- tionnaire was 6.7%, and 2.9% of participants were diagnosed with depression by a physician (Ministry of Health and Wel- fare, 2015). It can, therefore, be seen that the incidence of depression in critical care survivors is considerably higher than that of the general population. Furthermore, considering the potential depressive patients who were not diagnosed with depression, the incidence of post-ICU depression in critical care survivors are expected to be higher.

The two advantages of the current study were that large- scale data were utilized and the incidence of post-ICU depression was verified by the diagnostic code. Most of the related studies on post-ICU depression (Battle, James, &

Temblett, 2015; Jacka, Mitchell, & Perez-Parada, 2016;

Rabiee et al., 2016) were confined to a relatively small number of subjects, and depression symptoms were measured

using self-report questionnaires, such as the Hospital Anxi- ety and Depression Scale-Depression subscale. In addition, while several studies (Battle et al., 2015; Jacka et al., 2016) have reported the incidence of post-ICU depression, their accuracy cannot be determined because pre-ICU depression history was not investigated. Contrarily, the results of the current study suggest that ICU treatment can increase the incidence of depression in critical care survivors regardless of previous depression history.

The highest risk factor for the development of post-ICU depression was enteral nutrition in the ICU. Enteral nutrition refers to the provision of nutritional support through naso- gastric or other enteral tubes. This is a major nutritional support measure for ICU patients who cannot feed orally owing to intentional or unintentional loss of consciousness. In the present study, 28.4% of the depressive group had experience of enteral nutrition, which was more than twice as high as in the control group. According to the additional analysis results, the subjects of the enteral nutrition group had more physical and mental disabilities than those of the non-enteral nutrition group. They were hospitalized for longer periods and had a two to seven times higher rate of advanced treatments such as use of the mechanical ventilator, CRRT, IABP, and drug use. In other words, it can be assumed that enteral nutrition reflects the severity of a patient's condition in the ICU. However, this hypothesis needs to be confirmed through further research.

In the current study, cerebrovascular disease and hemi/

paraplegia were also major risk factors for post-ICU depression. According to a study by Jackson et al. (2014) among depressive critical care survivors, only 4% had depressive symptoms related to cognitive problems, whereas 72% of survivors complained of depressive symptoms of physical problems. Their results suggest that post-ICU depression is mainly caused by physical impairments. Cerebrovascular disease and hemi/paraplegia are likely to progress to chronic disease and may cause physical impairments in survivors.

Moreover, depression after cerebrovascular disease is a well- known problem (Robinson & Jorge, 2016). Early rehabilitation in the ICU helps improve physical function and prevent PICS (Fuke et al., 2018). Considering that physical impairments can be closely related to post-ICU depression, it is necessary to try early in-ICU multidimensional rehabilitation that includes psychological aspects.

In addition to the above factors, the survivors' sex, type of insurance, disability registration, length of stay, and the use of mechanical ventilators, sedatives, muscle relaxants, and inotropes/vasopressors increased the risk of post-ICU depression. However, the odds of these factors were as small as 1.04–1.16. According to a previous study (Rabiee et al., 2016), the characteristics of survivors and ICU treatment were not related to depression. However, negative

(9)

experiences and memories in the ICU consistently correlated with post-ICU depression (Rabiee et al., 2016; Ringdal et al., 2009). Since the current study adopted a retrospective design using the national health insurance claim codes, we could not measure survivors' ICU-related negative experiences or memories. According to a study on the ICU experience (Ringdal et al., 2009), 26% of survivors had more than one delusional memory, such as hallucinations or nightmares of the ICU, and they were more likely to develop post-ICU depression. Particularly, those who underwent advanced treatment such as mechanical ventilation remained more negative about the ICU experience. The relationship between advanced treatments and subsequent negative experiences in the ICU and their impact on the occurrence of post-ICU depression in survivors should be further explored.

CPR was a factor that lowered the incidence of post-ICU depression, and this can be explained in several ways. First, CPR means that the patient had cardiac arrest. Cognitive impairment due to hypoxic brain injury is common in post- CPR patients (Lim, Verfaellie, Schnyer, Lafleche, & Alex- ander, 2014). It is difficult to identify mental disorders such as depression in a subject whose cognitive function is impaired (Elliott, Rodgers, & Brett, 2011). Therefore, depression in post-CPR survivors in our study might have been underestimated. Second, post-CPR survivors may actu- ally be less depressed. The quality of life of the patients who survived cardiac arrest and were subsequently discharged was not significantly different from that of the general population. There were also cases where the quality of life improved even before the cardiac arrest (Elliott et al., 2011;

Smith, Andrew, Lijovic, Nehme, & Bernard, 2015). In some studies (Elliott et al., 2011; Wilder Schaaf et al., 2013), depression after cardiac arrest was reported to vary according to survivors' functional status, which should also be taken into consideration. Third, the probability of post- CPR survival with some degree of neurological outcomes is as low as 2.6–11.0% (Jeong et al., 2016; Kim et al., 2013;

McNally et al., 2011). In the current study, only 2.4% of the survivors underwent CPR. Owing to this small proportion, the possibility that the statistical results might be exagger- ated should also be considered.

The findings of the current study confirm that the incidence of post-ICU depression in critical care survivors is quite high. Early identification and intervention of post-ICU depression is important for the recovery and quality of life of critical care survivors (Jackson et al., 2014; Myhren, Ekeberg, Tøien, Karlsson, & Stokland, 2010). The initial depression right after ICU discharge was persistent and did not improve significantly over time (Milton, Brück, Schandl, Bottai, & Sackey, 2017; Rabiee et al., 2016). Therefore, it is necessary to seek ways to identify depression early to reduce post-ICU depression in critical care survivors. The number

of interventional studies on post-ICU depression is still small and its effects are not clear (Rabiee et al., 2016). However, it is reported survivors who were provided psychological interventions at the time of admission were less depressed and had a significantly lower percentage of being on psychiatric medication (Milton et al., 2017). In the future, it is necessary to develop various interventions to reduce post-ICU depression in critical care survivors. In particular, it is necessary to develop and apply the early intervention that can be performed before ICU discharge.

The limitations of this study are as follows. First, the subjects who had depression before ICU admission were excluded from the analysis, but those who had psychiatric disorders other than depression were not excluded. Thus, it should be considered that other psychiatric disorders could affect post-ICU depression. Second, since the depressive group in the current study was selected by the diagnostic code, we could not include those who had not been clinically diagnosed. This might have reduced the incidence of post- ICU depression. Finally, it should be acknowledged that the risk factors for post-ICU depression were confined to the information provided by the NHIS because this was a retrospective study. Thus, the present study did not control factors such as the recovery or complications of critical illness after ICU discharge. These need to be complemented by future large-scale prospective studies.

5 | C O N C L U S I O N S

The incidence of post-ICU depression in critical care survivors within one year of discharge was 18.5%, which was very high compared with the general population. Advanced treatments in the ICU, as well as diseases and physical impairments in survivors, were the major risk factors for post-ICU depression. In particular, it is necessary to pay close attention to the possibility of post-ICU depression development in survivors who have experienced enteral nutrition in the ICU and those who have physical impairments. Various attempts are needed for the early detection or prevention of post-ICU depression, which can negatively affect the recovery and quality of life of critical care survivors.

A C K N O W L E D G M E N T S

This study was supported by the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education (Grant no. NRF- 2016R1D1A1B03936044).

(10)

C O N F L I C T O F I N T E R E S T

The authors declare no conflicts of interest.

A U T H O R C O N T R I B U T I O N S

All the authors contributed to the conception and design of this study; S. Y. performed the interpretation of data and dra- fted the manuscript; Y. S. C., and Y. J. J. critically reviewed the manuscript; J. K. supervised the whole study process.

All the authors read and approved the final manuscript.

O R C I D

Jiyeon Kang https://orcid.org/0000-0002-8938-7656 Seonyoung Yun https://orcid.org/0000-0003-2261-9325 Yeon Jin Jeong https://orcid.org/0000-0002-0015-921X

R E F E R E N C E S

Battle, C., James, K., & Temblett, P. (2015). Depression following critical illness: Analysis of incidence and risk factors.Journal of the Intensive Care Society, 16(2), 105–108. https://doi.org/10.1177/

1751143714559904

Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987).

A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation.Journal of Chronic Dis- eases, 40(5), 378–383.

Davidson, J. E., Harvey, M. A., Schuller, J., & Black, G. (2013). Post- intensive care syndrome: What it is and how to help prevent it.

American Nurse Today, 8(5), 32–36.

Davydow, D. S., Gifford, J. M., Desai, S. V., Bienvenu, O. J., &

Needham, D. M. (2009). Depression in general intensive care unit survivors: A systematic review. Intensive Care Medicine, 35(5), 796–809. https://doi.org/10.1007/s00134-009-1396-5

Davydow, D. S., Hough, C. L., Langa, K. M., & Iwashyna, T. J.

(2013). Symptoms of depression in survivors of severe sepsis: A prospective cohort study of older Americans.The American Jour- nal of Geriatric Psychiatry, 21(9), 1–18. https://doi.org/10.1016/j.

jagp.2013.01.017

Dowdy, D. W., Eid, M. P., Sedrakyan, A., Mendez-Tellez, P. A., Pronovost, P. J., Herridge, M. S., & Needham, D. M. (2005). Qual- ity of life in adult survivors of critical illness: A systematic review of the literature.Intensive Care Medicine, 31(5), 611–620. https://

doi.org/10.1007/s00134-005-2592-6

Elliott, V. J., Rodgers, D. L., & Brett, S. J. (2011). Systematic review of quality of life and other patient-centred outcomes after cardiac arrest survival. Resuscitation, 82(3), 247–256. https://doi.org/10.

1016/j.resuscitation.2010.10.030

Fuke, R., Hifumi, T., Kondo, Y., Hatakeyama, J., Takei, T., Yamakawa, K.,…Nishida, O. (2018). Early rehabilitation to prevent postintensive care syndrome in patients with critical illness: A systematic review and meta-analysis. British Medical Journal Open, 8(5), e019998. https://doi.org/10.1136/bmjopen-2017- 019998

Halpern, N. A., & Stephen, M. P. (2010). Critical care medicine in the United States 2000–2005: An analysis of bed numbers, occupancy

rates, payer mix, and costs.Critical Care Medicine, 38(1), 65–71.

https://doi.org/10.1097/CCM.0b013e3181b090d0

Health Insurance Review & Assessment Service. (2017a).Health insurance medical care costs(in Korean). Retrieved from https://www.

hira.or.kr/co/ebook/list.do?pgmid=HIRAA960001000000 Health Insurance Review & Assessment Service. (2017b).

National Health Insurance Scheme (in Korean). Retrieved from http://www.hira.or.kr/rd/insuadtcrtr/InsuAdtCrtrList.do?

pgmid=HIRAA030069000400

Hill, A. D., Fowler, R. A., Pinto, R., Herridge, M. S., Cuthbertson, B. H., & Scales, D. C. (2016). Long-term outcomes and healthcare utilization following critical illness–A population- based study.Critical Care, 20(1), 76–85. https://doi.org/10.1186/

s13054-016-1248-y

Jacka, M. J., Mitchell, N., & Perez-Parada, J. (2016). Incidence and prevalence of anxiety, depression, and post-traumatic stress disorder among critical care patients, families, and practitioners.Journal of Anesthesia & Intensive Care Medicine, 1(1), 1–4. https://doi.org/

10.19080/JAICM.2015.01.555555

Jackson, J. C., Pandharipande, P. P., Girard, T. D., Brummel, N. E., Thompson, J. L., Hughes, C. G.,…Ely, E. W. (2014). Depression, post-traumatic stress disorder, and functional disability in survivors of critical illness in the BRAIN-ICU study: A longitudinal cohort study.The Lancet Respiratory Medicine, 2(5), 369–379. https://doi.

org/10.1016/S2213-2600(14)70051-7

Jeong, S. Y., Kim, C. W., Yoon, T. H., Kim, Y. J., Hong, S. O., &

Choi, J. A. (2016). The factors influencing neurological outcome of out-of-hospital cardiac arrest with cardiac etiology.Journal of the Korean Society of Emergency Medicine, 27(2), 165–172 in Korean.

Jutte, J. E., Erb, C. T., & Jackson, J. C. (2015). Physical, cognitive, and psychological disability following critical illiness: What is the risk.

Seminars in Respiratory and Critical Care Medicine, 36(6), 943–958. https://doi.org/10.1055/s-0035-1566002

Kim, J. Y., Shin, S. D., Ro, Y. S., Song, K. J., Lee, E. J., Park, C. B., &

Hwang, S. S. (2013). Post-resuscitation care and outcomes of out- of-hospital cardiac arrest: A nationwide propensity score-matching analysis.Resuscitation, 84(8), 1068–1077. https://doi.org/10.1016/

j.resuscitation.2013.02.010

Kim, K. H. (2010). Comparative study on three algorithms of the ICD- 10 Charlson comorbidity index with myocardial infarction patients.

Journal of Preventive Medicine and Public Health, 43(1), 42–49 in Korean. https://doi.org/10.3961/jpmph.2010.43.1.42

Korea Information Classification of Diseases. (2016).Disease coding guide(2016 NEW) (in Korean). Retrieved from http://www.koicd.

kr/2016/coding/v7.do

Lee, J. Y., Kim, K. H., & Lee, J. S. (2012). Construction of an appro- priate sampling design and a sample database using the National Health Information Database.National Health Insurance Corpora- tion, 1–51 in Korean.

Lim, C., Verfaellie, M., Schnyer, D., Lafleche, G., & Alexander, M. P.

(2014). Recovery, long-term cognitive outcome and quality of life following out-of-hospital cardiac arrest. Journal of Rehabilitation Medicine, 46(7), 691–697. https://doi.org/10.2340/16501977-1816 McNally, B., Robb, R., Mehta, M., Vellano, K., Valderrama, A. L.,

Yoon, P. W.,… Kellermann, A. (2011).Out-of-hospital cardiac arrest surveillance: Cardiac Arrest Registry to Enhance Survival (CARES), United States, October 1, 2005-December 31, 2010. Mor- bidity and mortality weekly report. Surveillance Summaries, 60(8), 1–19.

(11)

Milton, A., Brück, E., Schandl, A., Bottai, M., & Sackey, P. (2017).

Early psychological screening of intensive care unit survivors: A prospective cohort study.Critical Care, 21(1), 273. https://doi.org/

10.1186/s13054-017-1813-z

Ministry of Health and Welfare. (2015). Korea Health Statics 2014:

Korea national health and nutrition examination survey (KNHANSE VI-2)(in Korean). Retrieved from https://knhanes.cdc.

go.kr/knhanes/sub04/sub04_03.do?classType=7

Myhren, H., Ekeberg, O., Tøien, K., Karlsson, S., & Stokland, O.

(2010). Posttraumatic stress, anxiety and depression symptoms in patients during the first year post intensive care unit discharge.Crit- ical Care, 14(1), R14. https://doi.org/10.1186/cc8870

National Health Insurance Sharing Service. (2017). Introduction of National Health Insurance Sharing Service(in Korean). Retrieved from https://nhiss.nhis.or.kr/bd/ab/bdaba018lv.do

Needham, D. M., Davidson, J., Cohen, H., Hopkins, R. O., Weinert, C., Wunsch, H., … Harvey, M. A. (2012). Improving long-term outcomes after discharge from intensive care unit: Report from a stakeholders' conference. Critical Care Medicine, 40(2), 502–509. https://doi.org/10.1097/CCM0b013e318232da75 Park, J., Jeon, K., Chung, C. R., Yang, J. H., Cho, Y. H., Cho, J.,…

Suh, G. Y. (2018). A nationwide analysis of intensive care unit admissions, 2009-2014- the Korean ICU National Data (KIND) study.Journal of Critical Care, 44, 24–30. https://doi.org/10.1016/

j.jcrc.2017.09.017

Rabiee, A., Nikayin, S., Hashem, M. D., Huang, M., Dinglas, V., Bienvenu, O. J.,…Needham, D. M. (2016). Depressive symptoms after critical illness: A systematic review and meta-analysis.Criti- cal Care Medicine, 44(9), 1744–1753. https://doi.org/10.1097/

CCM.0000000000001811

Ringdal, M., Plos, K., Lundberg, D., Johansson, L., & Bergbom, I.

(2009). Outcome after injury: Memories, health-related quality of life, anxiety, and symptoms of depression after intensive care.The Journal of Trauma, 66(4), 1226–1233. https://doi.org/10.1097/TA.

0b013e318181b8e3

Robinson, R. G., & Jorge, R. E. (2016). Post-stroke depression: A review. The American Journal of Psychiatry, 173(3), 221–231.

https://doi.org/10.1176/appi.ajp.2015.15030363

Smith, K., Andrew, E., Lijovic, M., Nehme, Z., & Bernard, S. (2015).

Quality of life and functional outcomes 12 months after out-of- hospital cardiac arrest. Circulation, 131(2), 174–181. https://doi.

org/10.1161/circulationaha.114.011200

Steenbergen, S., Rijkenberg, S., Adonis, T., Kroeze, G., van Stijn, I., &

Endeman, H. (2015). Long-term treated intensive care patients outcomes: The one-year mortality rate, quality of life, health care use and long-term complications as reported by general practitioners.

BioMed Central Anesthesiology, 15(142). https://doi.org/10.1186/

s12871-015-0121-x

Wilder Schaaf, K. P., Artman, L. K., Peberdy, M. A., Walker, W. C., Ornato, J. P., Gossip, M. R., & Kreutzer, J. S. (2013). Anxiety, depression, and PTSD following cardiac arrest: A systematic review of the literature.Resuscitation, 84(7), 873–877. https://doi.

org/10.1016/j.resuscitation.2012.11.021

World Health Organization. (2017). Depression and other common mental disorders: Global health estimates. Retrieved from http://

www.who.int/mental_health/management/depression/prevalence_

global_health_estimates/en/

Wunsch, H., Christiansen, C. F., Johansen, M. B., Olsen, M., Ali, N., Angus, D. C., & Sørensen, H. T. (2014). Psychiatric diagnoses and psychoactive medication use among nonsurgical critically ill patients receiving mechanical ventilation.The Journal of the Ameri- can Medical Association, 311(11), 1133–1142. https://doi.org/10.

1001/jama.2014.2137

Zimmerman, J. E., Kramer, A. A., & Knaus, W. A. (2013). Changes in hospital mortality for United States intensive care unit admissions from 1988 to 2012.Critical Care, 17(2), R81. https://doi.org/10.

1186/cc12695

How to cite this article: Kang J, Yun S, Cho YS, Jeong YJ. Post-intensive care unit depression among critical care survivors: A nationwide population-based study.Jpn J Nurs Sci. 2020;17:e12299.https://doi.

org/10.1111/jjns.12299