Article, Psychiatry

Safety of reassessment-and-release practice for mental health patients boarded in the emergency department

a b s t r a c t

Objectives: Among emergency department (ED) mental health and substance abuse (MHSA) patients, we sought to compare mortality and healthcare utilization by ED discharge disposition and inpatient bed request status. Methods: A retrospective cohort study of 492 patients was conducted at a single University ED. We reviewed three groups of MHSA patients including ED patients that were admitted, ED patients with a bed request that were discharged from the ED, and ED patients with no bed request that were discharged from the ED. We iden- tified main outcomes as ED return visit, re-hospitalization and mortality within 12 months based on chart review and reference from the National Death Index.

Results: The average age of patients presenting was 30.5 (SD16.4) years and 251 (51.0%) were female patients. Of these patients, 216 (43.9%) presented with mood disorder and 93 (18.9%) with self-harm. The most common rea- son for discharge from the ED after an admission request was placed was from stabilization of the patient (n = 138). An ED revisit within 12 months was significantly higher among patients discharged who had a bed request in place prior to departure (54.0%, p b 0.001), than those discharged from the ED (40.9%) or admitted to inpatient care (30.5%). The rate of suicide attempt and death did not show statistical significance (p = 0.55 and p = 0.88). Conclusion: MHSA patients who were discharged from ED after bed requests were placed were at greater risk for return visits to the ED. This implicates that these patients require outpatient planning to prevent further avoid- able healthcare utilization.

(C) 2018

Introduction

Background/rationale

About 420,000 individuals visit emergency departments (EDs) an- nually for attempted suicide and self-inflicted injury in the US; this number has been increasing in the last two decades [1]. A significant portion of these suicide attempters suffer from psychiatric conditions such as depression, anxiety, and substance abuse [2]. As EDs often func- tion as a last resort for Psychiatric patients, providing effective psychiat- ric services are important to ensure patient safety. However, in the ED, the patients are often subject to the complex processes of the mental health care system, including medical clearance, insurance verifications, regulatory and institutional requirements, and patient disposition [3]. The complex processes results in psychiatric boarding, defined as

* Corresponding author at: Department of Emergency Medicine, The University of Iowa Carver College of Medicine, 200 Harkins DR, Iowa City, IA 52246, USA.

E-mail address: [email protected] (S. Lee).

psychiatric patients’ waiting in hallways or other emergency depart- ment areas for inpatient beds [4]. Our ED started to see a rise of mental health patients waiting for inpatient psychiatry transfer in 2015, and this has negatively affected the quality of patient care.

In 2014, the Emergency Medicine Practice Committee of the American College of Emergency Physicians (ACEP) published a practice guideline that recommended reducing psychiatric inpatient admissions to avoid psychiatric boarding in the ED [3]. However, the pressure of re- ducing inpatient admissions in the context of bed shortage and psychi- atric boarding is leading to another growing issue of discharging mental health and substance abuse (MHSA) patients from the ED after a prolonged stay [5]. This discharge may be due to stabilization of the cri- sis by the time of subsequent assessments in the ED or to frustration with waiting N24 h to find an available bed. The practice pattern is pos- sibly comparable to discharge against medical advice (AMA), and the previous study reported the risk of increased suicide among discharged AMA from inpatient psychiatric ward [6]. No studies to our knowledge have examined the safety (e.g., subsequent suicide attempt and related injuries) and efficiency (e.g., ED revisit and hospitalization) of this

https://doi.org/10.1016/j.ajem.2018.02.026

0735-6757/(C) 2018

reassessment and release practice in MHSA patients. Particularly, MHSA patients who live in rural areas with limited access to psychiatric re- sources may differ in their Safety outcomes [7].

Objective

We wanted to determine whether a de-escalation and re- assessment approach for MHSA patients has an increased risk for an ad- verse event defined as suicide attempt or injury-related re- hospitalization and ED revisit within 12 months relative to patients who were released from ED without a bed request or those who were hospitalized in our institution. Our study hypothesis is that patients pre- senting with MHSA conditions who had a bed request on the day of ED arrival but were later discharged had a higher risk of adverse events (defined as suicide attempt or injury-related re-hospitalization and ED revisit within 12 months) than patients who were released from ED without a bed request or those who were hospitalized.

Methods

Study design

This is a retrospective cohort study using electronic health records from a single academic ED. The institutional review board approved this study with a waiver of informed consent. Our study adhered to the Strengthening the Reporting of Observational studies in Epidemiol- ogy (STROBE) statement [8].

Study setting

This study took place in the tertiary care center and the only aca- demic medical center in the State of Iowa with a 65,000-annual census.

Participants

Study participants were identified by retrospective query of the in- stitutional electronic medical record of patients that consecutively pre- sented to the ED in the specified time range from January 2015 to December 2016. Patients were included if any of the International Clas- sification of Disease, 9th Revision, Clinical Modification codes (ICD-9- CM) listed in the ED encounter diagnoses and procedures met the clin- ical classification criteria software (CCS) definition for MHSA. The Men- tal Health and Substance Abuse CCS (CCS-MHSA) groups mental health and substance abuse Diagnostic codes is commonly used to identify MHSA patients [9]. Patients of any age presenting to the ED with a diag- nosis meeting CCS-MHSA criteria were eligible to participate in the study. Our study included three cohorts. The first cohort was MHSA pa- tients discharged from the ED without a previous bed request, the sec- ond cohort was MHSA patients discharged from the ED after a bed request for an inpatient stay was requested, and the third cohort was MHSA patients admitted to the inpatient psychiatry unit from the ED.

Data collection

Demographic data, clinical data and medications were collected from electronic health records. Trained research assistants extracted these data. Initially, we reviewed 100 charts (20%) independently by the trainees and the principal investigator to measure the agreement of demographics, diagnosis and disposition by reporting weighted Cohen’s kappa. After the data were extracted from the hospital record, death status was requested from the National Death Index, a data source by the CDC to verify the status of each patient, date of death and the cause of death. Missing or unclear data was resolved by study team consensus.

Independent variables

We collected variables from our electronic health record and refer- enced the National Death Index. Variables collected from the electronic health record include patient name, date of birth, medical record num- ber, age, sex, gender (male, female and transgender), zip code of patient residence, insurance status (private, Tricare, Medicare, Medicaid, no in- surance and unknown), time and date of initial ED arrival, ED chief com- plaint(s), admitting MHSA diagnosis according to Clinical Conditioning Software (CCS), visit to primary care provider and psychiatrist visit within the last 12 months, family or friend support during ED evaluation based on chart review, mode of ED presentation (ambulance or other), the utilization of crisis support determined by the utilization of mobile crisis unit, and daily psychiatric medications. Daily medications were categorized into classes: antidepressant/antianxiety agent, tricyclic an- tidepressants, opioids, benzodiazepines, antipsychotics and others. We also measured the Severity of disease/condition by using the Clinical Global Impression Scale (CGIS), where a score of 1 is the mildest and 7 is the worst [10]. For those who had a bed request but were discharged from the ED, we elicited the reasons for change of disposition plans. Lastly, we defined frequent utilizer as those who had more than four visits to the ED in the year following the Initial ED encounter.

Outcomes variables

We defined outcome of interest as 1) ED revisit, 2) re- hospitalization, 3) nonfatal and fatal suicide attempt and 4) death with- in 12 months of ED visit. Thirty day outcomes were also examined for each outcome. Outcomes were determined by chart review and refer- ence to the National Death Index.

Sample size

Assuming an alpha of 0.05, power of 0.8 (Beta = 0.2) and a 9.8% dif- ference in the frequency of a suicide attempt within 12 months by those discharged from the ED versus discharged from an inpatient psychiatric unit, a sample size of 492 patient records (164 for each of the three groups) was needed. The estimated percentage difference was obtained from the observed difference of frequencies of 12 month suicide at- tempt of 16.3% after discharge from an ED setting and 6.5% in the inpa- tient setting [11].

Statistical analysis

We described demographic data by means, standard deviations of the mean, medians, and interquartile range values for continuous vari- ables; frequencies and proportions were listed for categorical variables. Wilcoxon-Mann-Whitney test, chi-square test, and Fisher’s exact tests were used to compare demographic data between groups, as appropri- ate. Kaplan-Meier survival curves were created to estimate the time-to- event for each of the four outcomes: ED revisit, rehospitalization, suicide attempt and death. Log-rank tests were used to compare probabilities of survival between groups. For analyses of 12 month outcomes, if no out- come event was found within the first 12 months of the initial ED visit, the subject was censored at 12 months. Thirty-day time-to-event out- comes were also analyzed for ED revisit, rehospitalization, and death.

Sensitivity analysis

We hypothesized a priori that we were likely to find a significant gap in the follow-up data available from patients who do not reside within the same county (Johnson County) as our institution. Although our elec- tronic health record has a record linkage to other institutions, the cover- age is not nearly complete. Therefore, we conducted an additional sensitivity analysis on the subgroup of patients who reside in Johnson County to assess the robustness of our study findings and test the

presence of ascertainment bias. County of residence was determined by patient’s zip code, using the county for the majority of the land area in the patient’s zip code.

Results

Since 2015, our ED has seen a rise in MHSA patients waiting for inpa- tient psychiatry transfer. As patients stabilized from the crisis situation, we found that a substantial number of them who had inpatient bed re- quests were released from our ED (Fig. 1). Many of the MHSA patients who were discharged directly from the ED after a bed request had ED length of stay times of over 24 h.

Participants

We identified a total of 537 patients consecutively presenting with MHSA related conditions from January 2015 (Fig. 2) through December 2016. After removing 45 charts due to duplicated entry, a total of 492 patients, 161 to 167 patients per cohort, remained for analysis. The kappa statistics for abstractor agreement were reported as 0.93 (95%CI 0.87-1.0) for sex, 0.86 (95%CI 0.77-0.95) for ED return visit, 0.66 (95% CI 0.51-0.81) for return hospitalization, and -0.13 (95%CI -0.26 to – 0.01) for CGI.

We reported demographic data in Table 1. The average age was

30.5 years (SD16.4), 251 (51.0%) patients were female, 184 (37.4%) pa- tients were insured by Medicaid, and 70 (14.2%) by Medicare. A total of 216 (43.9%) patients presented with mood disorder, 93 (18.9%) patients presented with suicide and self-inflicted injury, and 69 (14.0%) present- ed with Alcohol-related disorders.

Patient demographics differed by ED discharge and bed request sta- tus. The proportion of alcohol-related disorders was the largest for those who were discharged from the ED, followed by those who were discharged after bed requested (33.5%, 9.2% respectively, Table 1). Pa- tients discharged from the ED without a bed request and with a bed re- quest had similar frequencies of follow-up with a PCP (51.2%, 50.3% respectively) compared to those who were admitted (35.9%, p = 0.0007). Over half of patients who were admitted (60.5%) or were discharged after bed request (53.4%) presented with suicidal ideation, a significantly higher proportion than those who were discharged with- out a bed request (26.2%, p b 0.001). Among the patients who were discharged after bed requests, the most common reason for changing disposition was that patient condition was stabilized (137, 85.1%), followed by that patient became frustrated and provider did not think hospitalization were mandatory (19, 11.8%), and that patient left AMA and provider did not seek for involuntary hospitalization (4, 2.5%). No

Fig. 2. Flowchart of Study Enrollment.

differences were seen between groups on frequent ED utilization (p = 0.31).

Outcome data

The proportion of 12 month return ED visit and hospital readmis- sion differed significantly by group (Table 2). return ED visits within 12 months were highest, 54.0%, in ED patients discharged after a bed request (Group 2) compared to 30.5% for patients admitted after an ED bed request (Group 3) and 40.9% revisit rate in ED patients discharged with no bed request (Group 1) (Table 2 and Fig. 3). Twelve months return hospitalization after ED evaluation was also highest, 37.8% in Group 2 versus 18.3% and 25.6% in Groups 1 and 3 respectively (p b 0.001). Attempted suicide and death within 12 months were reported as highest in Group 2 at 6.8% and 2.5%, re- spectively, although this was not statistically different than the pro- portions found in Group 1 and 3 (P = 0.892, P = 0.880, respectively). Patients with an ED discharge following a bed request (Group 2), re- ported more frequently returning to the ED for psychiatric (24.8%) or suicidal (16.8%) conditions than those who were admitted (Group 3, psychiatric = 14.4%, suicidal = 6.0%) or discharged from the ED without a bed request (Group 1, psychiatric = 14.0%, suicidal =

Fig. 1. The trend of MHSA patient next-day discharge from the ED who had a previous bed request.

Table 1

Demographic information (attach table here for editing, and will move to the end for submission)

All (N = 164) no bed

request and ED discharge

(N = 161) bed request and ED discharge

(N = 167) Admitted from ED

p-value

Age (median (IQR), mean (SD))

25.5

30.5

26.4

30.6

25.5

30.9

24.1

29.8

0.445

(17.8-40.9)

(16.4)

(19.6-41.5)

(15.8)

(19.5-39.5)

(16.5)

(15.9-41.9)

(17.0)

Sex

0.726

Male

51.0%

251

52.4%

86

52.2%

84

48.5%

81

Female

49.0%

241

47.6%

78

47.8%

77

51.8%

86

Gender

0.461

Male

48.0%

236

47.0%

77

46.6%

75

50.3%

84

Female

50.8%

250

52.4%

86

50.9%

82

49.1%

82

Transgender

1.2%

6

0.6%

1

2.5%

4

0.6%

1

Insurance

0.075

Private

42.5%

209

51.2%

84

34.2%

55

41.9%

70

Tricare

0.8%

4

1.2%

2

0.6%

1

0.6%

1

Medicare

14.2%

70

9.2%

15

18.3%

30

15.0%

25

Medicaid

37.4%

184

32.3%

53

40.4%

65

39.5%

66

no insurance

4.9%

24

5.5%

9

6.2%

10

3.0%

5

Unknown

0.2%

1

0.6%

1

0.0%

0

0.0%

0

Homelessness

4.7%

23

3.7%

6

5.6%

9

4.8%

8

0.709

Admitting diagnosis (CCS)

Adjustment disorder

0.4%

2

0.0%

0

0.0%

0

1.2%

2

0.332a

Anxiety disorder

9.8%

48

13.4%

22

6.2%

10

9.6%

16

0.091

Attention-deficit, conduct, and disruptive behavior

9.4%

46

3.7%

6

6.8%

11

17.4%

29

b0.001

disorders

Delirium, dementia, and amnestic and other cognitive

1.0%

5

0.0%

0

1.8%

3

1.2%

2

0.252a

disorders

DevelopMental disorders

1.4%

7

0.6%

1

1.2%

2

2.4%

4

0.465a

Disorders usually diagnosed in infancy, childhood, or

0.4%

2

0.0%

0

0.6%

1

0.6%

1

0.773a

adolescence

Impulse control disorders, NEC

0.6%

3

0.0%

0

0.0%

0

1.8%

3

0.110a

Mood disorders

43.9%

216

29.9%

49

41.6%

67

59.9%

100

b0.001

Personality disorders

6.1%

30

1.8%

3

8.1%

13

8.4%

14

0.020

Schizophrenia and other psychotic disorders

10.0%

49

4.3%

7

11.2%

18

14.4%

24

0.007

Alcohol-related disorders

14.0%

69

25.0%

41

11.8%

19

5.4%

9

b0.001

Substance-related disorders

5.9%

29

5.5%

9

4.4%

7

7.8%

13

0.403

Suicide and intentional self-inflicted injury

18.9%

93

8.5%

14

19.9%

32

28.1%

47

b0.001

Screening and history of mental health and substance

0.4%

2

1.2%

2

0.0%

0

0.0%

0

0.217a

abuse codes

Miscellaneous mental health disorders

4.7%

23

3.1%

5

2.5%

4

8.4%

14

0.020

Clinical Global Impression Scale (CGI)

b0.001

Mild (1-3)

27.6%

136

57.9%

95

15.5%

25

9.6%

16

Moderate (4-5)

60.2%

296

40.2%

66

75.8%

122

64.7%

108

Severe (6-7)

12.2%

60

1.8%

3

8.7%

14

25.6%

43

Residential facility 7.7%

38

3.7%

6

9.8%

16

9.8%

16

0.058

Chief complaint

Depression

22.4%

110

18.3%

30

20.5%

33

28.1%

47

0.078

Bipolar

2.4%

12

0.6%

1

1.9%

3

4.8%

8

0.046a

Intoxication

15.2%

75

33.5%

55

9.2%

15

3.1%

5

b0.001

Hallucination/delusion

7.7%

38

4.3%

7

10.4%

17

8.5%

14

0.105

Substance abuse

4.9%

24

4.9%

8

5.5%

9

4.3%

7

0.877

Suicidal ideation

46.8%

230

26.2%

43

53.4%

86

60.5%

101

b0.001

Agitation/altered mental status

10.2%

50

3.1%

5

11.8%

19

15.6%

26

0.001

Overdose

6.9%

34

4.3%

7

9.3%

15

7.2%

12

0.197

Other

25.0%

123

29.9%

49

23.0%

37

22.2%

37

0.203

Injury

3.3%

16

1.2%

2

3.1%

5

5.5%

9

0.101

Regular follow-up with PCP

45.7%

225

51.2%

84

50.3%

81

35.9%

60

0.007

Regular follow-up with psychiatry provider

41.5%

204

32.3%

53

44.1%

71

47.9%

80

0.011

Daily medications

Antidepressant/antianxiety agent

49.6%

244

41.5%

68

53.4%

86

53.9%

90

0.039

Tri-cyclic antidepressant (TCA)

2.0%

10

1.8%

3

1.8%

3

2.4%

4

0.999a

Opioid

6.3%

31

8.5%

14

6.2%

10

4.2%

7

0.266

Benzodiazepines

13.0%

64

12.8%

21

13.4%

22

12.8%

21

0.982

Antipsychotics

24.4%

120

15.2%

25

26.1%

42

31.7%

53

0.002

Others

32.9%

162

26.8%

44

35.4%

57

36.5%

61

0.123

Family support

b0.001

Yes

53.9%

265

40.9%

67

62.7%

101

58.1%

97

No

8.7%

43

7.3%

12

8.1%

13

10.8%

18

Unknown

37.4%

184

51.8%

85

29.2%

47

31.1%

52

Crisis support

4.5%

22

1.2%

2

5.0%

8

7.2%

12

0.030

Accompanied to ED by:

0.081

Family/friend

40.5%

199

42.7%

70

38.5%

62

40.1%

67

law enforcement

14.3%

70

12.2%

20

19.3%

31

11.4%

19

Third party (Crisis unit, social worker, teacher, etc)

3.7%

18

1.8%

3

4.9%

8

4.3%

7

None

28.9%

142

31.1%

51

29.2%

47

26.4%

44

Unknown

12.8%

63

12.2%

20

7.9%

13

18.3%

30

Table 1 (continued)

All (N = 164) no bed

request and ED discharge

(N = 161) bed request and ED discharge

(N = 167) Admitted from ED

p-value

Ambulance

23.4%

115

26.2%

43

19.9%

32

24.0%

40

0.392

ED frequent userb

15.0%

74

15.2%

25

18.0%

29

12.0%

20

0.310

a p-value calculated with Fisher’s exact test.

b ED Frequent User Defined as N4 ED visits in 12 months.

6.7%). This relationship of increased return to ED for a psychiatric condition for those with an ED discharge following a bed request (Group 2), persisted through three ED re-visits.

Thirty day outcomes were similar to 12 month outcomes. ED pa- tients discharged after receiving bed requests had higher propor- tions of 30-day ED revisits (29.2% vs. 15.9% (Group 1) and 11.4%

(Group 3)) and hospital re-admission (16.2% vs. 7.9% (Group 1) and 6.6% (Group 3)). Two deaths occurred during the first 30 days following the initial encounter. One death occurred in a pa- tient discharged from the ED with a previous bed request, and one death occurred in a patient discharged from the ED without a previ- ous bed request.

Table 2

Outcomes.

Outcome 1. ED revisit

All Group 1: (N = 164) No

bed request and ED discharge

Group 2: (N = 161)

bed request and ED discharge

Group 3: (N = 167)

admitted from ED

p-value

ED Revisit 1

ED revisit within 12 mos.

41.7%

205

40.9%

67

54.0%

87

30.5%

51

b0.001

Reason for ED revisit

Psychiatric

17.7%

87

14.0%

23

24.8%

40

14.4%

24

0.02

Suicide

9.8%

48

6.7%

11

16.8%

27

6.0%

10

0.001

Medical

15.0%

74

15.9%

26

18.3%

30

11.0%

18

0.17

Overdose (intoxication)

6.7%

33

9.2%

15

6.8%

11

4.2%

7

0.2

Trauma (injury-related)

4.5%

22

5.5%

9

3.1%

5

4.9%

8

0.54

Surgical, not injury-related

0.2%

1

0.0%

0

0.6%

1

0.0%

0

0.33a

Unknown

0.2%

1

0.0%

0

0.6%

1

0.0%

0

0.33a

ED Revisit 2

23.0%

113

22.6%

37

29.2%

47

17.4%

29

0.04

ED Revisit 2

Reason for ED revisit

Psychiatric

8.3%

41

6.7%

11

13.4%

22

4.9%

8

0.01

Suicide

4.1%

20

3.7%

6

5.0%

8

4.0%

6

0.78

Medical

10.8%

53

10.4%

17

14.3%

23

7.8%

13

0.16

Overdose (intoxication)

3.5%

17

4.3%

7

3.1%

5

3.1%

5

0.78

Trauma (injury-related)

2.2%

11

3.7%

6

1.2%

2

1.8%

3

0.41a

Surgical, not injury-related

0.2%

1

0.6%

1

0.0%

0

0.0%

0

0.99a

Unknown

0.4%

2

0.6%

1

0.0%

0

0.6%

1

0.99a

ED revisit 3

23.0%

113

22.6%

37

29.2%

47

17.4%

29

0.04

ED Revisit 3

Reason for ED revisit

Psychiatric

6.1%

30

3.7%

6

9.3%

15

5.4%

9

0.09

Suicide

3.1%

15

0.6%

1

6.2%

10

2.4%

4

0.011

Medical

9.4%

46

11.6%

19

10.6%

17

6.0%

10

0.18

Overdose (intoxication)

2.6%

13

3.7%

6

2.4%

4

1.8%

3

0.69a

Trauma (injury-related)

1.4%

7

1.8%

3

1.2%

2

1.2%

2

0.99a

Surgical, not injury-related

0.0%

0

0.0%

0

0.0%

0

0.0%

0

Unknown

0.4%

2

0.6%

1

0.6%

1

0.0%

0

0.550a

Outcome 2. Hospital readmission

Readmission to hospital

27.2%

134

18.3%

30

37.8%

62

25.6%

42

b0.001

Reason for readmission

Psychiatric

18.3%

90

10.4%

17

27.4%

45

17.1%

28

b0.001

Suicide

10.2%

50

5.5%

9

17.1%

28

7.9%

13

0.001

Medical

6.7%

33

6.1%

10

9.8%

16

4.3%

7

0.13

Overdose (Intoxication)

3.1%

15

2.4%

4

4.3%

7

2.4%

4

0.54

Trauma (injury-related)

2.2%

11

1.8%

3

1.8%

3

3.1%

5

0.8a

Surgical, not injury-related

0.6%

3

1.2%

2

0.6%

1

0.0%

0

0.78a

Unknown

0.0%

0

0.0%

0

0.0%

0

0.0%

0

Outcome 3. Death

suicidal attempt

Death

5.3%

26

4.3%

7

6.8%

11

4.8%

8

0.55

0.85a

Yes

2.0%

10

1.8%

3

2.5%

4

1.8%

3

No

98.0%

482

98.2%

161

97.5%

157

98.2%

164

30-Day outcomes

ED Revisit within 30 days

18.7%

92

15.9%

26

29.2%

47

11.4%

19

b0.001

Readmission to hospital

Death

10.2%

50

7.9%

13

16.2%

26

6.6%

11

0.008

0.55a

Yes

0.4%

2

0.6%

1

0.6%

1

0.0%

0

No

99.6%

490

99.4%

163

99.4%

160

100.0%

167

a p-value calculated with Fisher’s exact test.

survival analysis“>Fig. 3. Kaplan-Meier 12-month Survival Curves. N = 492; log-rank test p-values are listed for each outcome. A. Survival curve for outcome of ED Revisits; B. Survival curve for Hospital Readmissions; C. Survival curve for outcome of death.

Survival analysis

In terms of death status, the data showed an overall 12-month Death rate of 2.0% with the main causes of death reported as: medical [5], over- dose [4], and unspecified [1]. Log-rank tests demonstrate that there are differences in the probability of ED revisit between ED patients discharged after receiving a bed request and ED patients that are admit- ted (p b 0.0001) or discharged with no bed request (p = 0.011) (Fig. 3). Hospital readmission also differed significantly between patients discharged after bed request versus those who were discharged without a bed request (p b 0.0001). No differences in mortality were found.

Sensitivity analysis

A sensitivity analysis was performed on the population of subjects residing in the same county as the ED facility. Of all study participants, 46.8% (n = 230) resided in the same county as the ED facility. Of the 230 MHSA patients included in the sensitivity analysis, 98.3% of patients lived within zip codes where the entire zip code area is in Johnson County. The local (in-county residents) were more likely to be discharged after bed request, when compared to out-of-county residents (39.6% local vs. 26.7% non-local, p b 0.001). Further, local residents were less likely to be admitted after bed request (25.2% local vs. 41.6% non-local, p b 0.001). Of patients discharged after bed request, local subjects experi- enced shorter time to ED revisit compared to non-local subjects (17 days local vs. 49 days non-local for 25% revisit times, p = 0.028) (Fig. 4). Hospital readmission times for this group did not differ based on patient residence (p = 0.819). Similar trends were noted when com- paring non-local and local subjects admitted after bed request (Fig. 4).

Discussion

Main findings

Our study demonstrated the increasing number of any MHSA patients who were discharged after bed request was placed since 2015, and that about half of these patients returned to ED within 12 months. This finding of increased healthcare utilization such as ED return visit and hospital re- admission was disproportionally higher in this patient group than other group of patients with MHSA conditions, and raises an important ques- tion of placing a safe and reliable outpatient plan in place from ED when timely inpatient psychiatry admission is not a viable option.

Interpretation of study results

We found that the number of boarding MHSA patients are increasing in our institution. This trend is similar to the national data, as it is

reported by previous studies [4,12]. The most common reason that pa- tients were released after overnight stay in the ED was due to stabiliza- tion of Acute conditions. The impact of overnight observation and de- escalation is reported by Zeller et al. who found a reduction of inpatient psychiatric hospitalization by 75% after opening a mental health boarding unit [13]. Our study added insight that the next day assess- ment can be valuable to reassess the patients’ condition when length of stay is prolonged.

However, the rate of healthcare utilization was higher among those who had a bed request and were subsequently released. This group had more revisits to the ED than the other two groups, and this remained consistent both at the 30-day and 12 month follow-ups. Interestingly, this group also reported a higher percentage of psychiatric or suicidal conditions as reasons for returning to the ED than the other two groups. This is important data as it points to the areas of ED procedures which we can improve. Our previous study from private insurer’s data demon- strated that the rate of 30-day ED return visit was 14.2%, increased risk of return among mood disorders, and our current data revealed a slight- ly higher rate of returns, possibly due to the mixed payer status [14]. The potential solution could be establishing protocols for disposition, setting up a protocol for the next day assessment for boarding psychiatric pa- tients in the ED, and the last would be to provide outpatient resources to accommodate patients with ongoing psychiatric needs in a rural state. The previous study of follow-up care for psychiatric patients seen in the ED noted that the odds of follow up care were lower among males, African Americans, those patients with mood disorders and those who lived in rural areas [15]. Our study did not look at wheth- er or not psychiatric patients discharged from the ED had an outpatient appointment and this may be helpful in determining discharge planning in the future.

Another interesting point is that 75.8% of patients who were reassessed and discharged had CGI scores in the moderately severe range as opposed to 64.7% of those patients who were admitted. How- ever, those patients who were hospitalized to inpatient psychiatry did have a higher percentage of severe CGI scores (25.6% versus 8.7%). The previous studies used CGI for drug therapy and Treatment response, so its utility is somewhat limited in our study design [16,17]. We speculate that if more psychiatric beds had been available it would have been pru- dent to admit the patients in this group.

Although healthcare utilization was increased, the rate of suicide at- tempt and death were not significantly different between each group. This study was not powered to detect the statistically significant rate of suicide attempt or death, yet the study finding is valid since we refer- enced the National Death Index to verify the status of each patient in our database. The validity of this innovative record-linkage system was re- ported in the previous surveillance studies [18,19]. Another study used the Social Security Database to verify the 15-day mortality after ED discharge and although infrequent, a short-term post-ED discharge

Fig. 4. Sensitivity Analysis Comparing Local and Non-Local Residents by bed request group. A, B. Group 1: MHSA subjects discharged with no bed request. C,D. Group 2: MHSA subjects discharged after bed request. E,F. Group 3: MHSA subjects admitted.

mortality rate was significantly higher for those with psychiatric history than ED patients without [20]. A previous study reported that about one in five suicide victims had contact with mental health services within a month before their suicide, which implies the role of emergency psychi- atry would be imperative to screen and intervene [21]. Further research would be necessary to elucidate any modifiable factors associated with suicide attempt and death.

Study implications

This study findings implied that there are increasing needs for cur- rent disposition assessment and suicide prevention plans for those who were released. The rate of suicide attempt and suicide related death after discharge was infrequent and was not powered to demon- strate statistical significance. Study results showed some insights into

overall suicidal behavior, healthcare utilization patters, and the need for an alternative outpatient follow-up strategy for MHSA patients consid- ered at-risk who presented to the ED and were discharged due to lack of inpatient beds. Alakeson et al. described several potential solutions to mitigate psychiatric ED boarding issues in the past [4]. The recent ED based study utilizing universal screening plus an intervention consisting of an expanded suicide screening and provision of a self- administered safety plan in the ED followed by a telephone-based inter- vention holds promise and could be utilized in our study population [22]. Among them, our data support a community-based, collaborative behavioral health crisis assessment such as active involvement of crisis intervention and the establishment of post-acute care crisis manage- ment center locally. Another question is the cost regarding recurrent ad- missions and Negative outcomes associated with discharge from ED despite initial bed request, which is best addressed by a separate cost- effective analysis.

Limitations

Our study has several limitations. First, we did not have an adequate number of patients to examine suicide attempts and related death as an outcome. We were aware that rare events such as completed suicide might not reach statistical significance using our study design. In addi- tion to the rare but important outcomes, we also found that the healthcare utilization was increased in this cohort and emphasized the importance of allocating the right outpatient resources. Second, there was concern of differential collection of follow-up data on patients who may not return to our ED for care, particularly among non-local pa- tients. To address this ascertainment bias, we completed a sensitivity analysis which found that local patients return to our ED in a shorter length of time than non-local patients but no differences were found in hospital readmission. One might hypothesize that non-local patients may have received MHSA treatment at an ED closer to home rather than returning to our ED that went uncaptured. If this were true, you would expect this to reduce our ability to find a significant association for re- turn to ED within 12 months and we were still able to detect a signifi- cant difference. Third, the use of CGI was somewhat limited due to significant discordance between chart reviewers. Fourth, the retrospec- tive data had some missing values that could impact Data quality. Fifth, this was a single center study and generalizability to other practice en- vironments requires a caution.

Conclusion

This study concludes that the patients with MHSA conditions who were released from emergency department following an inpatient bed request continued to utilize healthcare resource, and it is likely that these patients would benefit from intensive outpatient follow up to pre- vent the return visit, return hospitalization and suicide attempt when it is avoidable. This study allows us to look further into how we can im- prove our Disposition decision making and to develop close outpatient

follow up. These results also pose the overriding question of how best to meet the needs of ED patients with mental health issues that will im- prove patient outcome, save costs to the patient and be a step in the di- rection of improved treatment for this deserving population.

References

  1. Turecki G, Brent DA. Suicide and suicidal behaviour. Lancet 2016;387:1227-39.
  2. Ting SA, Sullivan AF, Boudreaux ED, Miller I, Camargo Jr CA. Trends in US emergency department visits for attempted suicide and self-inflicted injury, 1993-2008. Gen Hosp Psychiatry 2012;34:557-65.
  3. Grabowitz L, Kukuy OL, Lidar M, et al. P01-029-microscopic hematuria in FMF. Pediatr Rheumatol 2013;11:A33.
  4. Alakeson V, Pande N, Ludwig M. A plan to reduce emergency room ‘boarding’ of psy- chiatric patients. Health Aff (Millwood) 2010;29:1637-42.
  5. Fuller DA, Sinclair E, Geller J, Quanbeck C, Snook J. Going, Going, Gone: Trends and Consequences of Eliminating State Psychiatric Beds. Treatment Advocacy Center; 2016.
  6. Kuo CJ, Tsai SY, Liao YT, Lee WC, Sung XW, Chen CC. Psychiatric discharge against medical advice is a risk factor for suicide but not for other causes of death. J Clin Psy- chiatry 2010;71:808-9.
  7. Druss BG, Bornemann T, Fry-Johnson YW, McCombs HG, Politzer RM, Rust G. Trends in mental health and Substance abuse services at the nation’s community health centers: 1998-2003. Am J Public Health 2008;98:S126-31.
  8. von Elm E, Altman DG, Egger M, et al. The strengthening the reporting of observa- tional studies in epidemiology (STROBE) statement: guidelines for reporting obser- vational studies. J Clin Epidemiol 2008;61:344-9.
  9. Elixhauser A SC SC. Hospital inpatient statistics. Healthcare Cost and Utilization Pro- ject (HCUP) Research Note. Rockville, MD: Agency for Health Care Policy and Re- search; 1999. AHCPR Pub. No. 99-0034; 1996.
  10. Busner J, Targum SD. The clinical global impressions scale: applying a research tool in clinical practice. Psychiatry (Edgmont) 2007;4:28-37.
  11. Gunnell D, Hawton K, Ho D, et al. Hospital admissions for self harm after discharge from psychiatric inpatient care: cohort study. BMJ 2008;337:a2278.
  12. McCaig LFNE. National Hospital Ambulatory Care Survey: 2004 emergency depart- ment summary (centers for disease control and prevention: advanced data). Vital Health Stat 2006;372:1-30.
  13. Zeller S, Calma N, Stone A. Effects of a dedicated regional psychiatric emergency ser- vice on boarding of psychiatric patients in area emergency departments. West J Emerg Med 2014;15:1-6.
  14. HJ Lee S, Bobo WV, Johnson R, Sangaralingham LR, Campbell RC. Predictors of return visits among insured emergency department mental health and substance abuse pa- tients, 2005-2013. West J Emerg Med 2017;18(5):884-93 Aud.
  15. Croake S, Brown JD, Miller D, et al. Follow-up care after emergency department visits for mental and Substance use disorders among Medicaid beneficiaries. Psychiatr Serv 2017;68:566-72.
  16. Bandelow B, Baldwin DS, Dolberg OT, Andersen HF, Stein DJ. What is the threshold for symptomatic response and remission for major Depressive disorder, panic disor- der, social anxiety disorder, and generalized anxiety disorder? J Clin Psychiatry 2006;67:1428-34.
  17. Leucht S, Kane JM, Kissling W, Hamann J, Etschel E, Engel R. Clinical implications of brief psychiatric rating scale scores. Br J Psychiatry 2005;187:366-71.
  18. Wohler B, Qiao B, Weir HK, MacKinnon JA, Schymura MJ. Using the National Death Index to identify duplicate cancer incident cases in Florida and New York, 1996- 2005. Prev Chronic Dis 2014;11:E167.
  19. Hanna DB, Pfeiffer MR, Sackoff JE, Selik RM, Begier EM, Torian LV. Comparing the na- tional death index and the social security administration’s death master file to ascer- tain death in HIV surveillance. Public Health Rep 2009;124:850-60.
  20. Chang BP, Pany MJ, Obermeyer Z. early death after emergency department discharge in patients with psychiatric illness. Am J Emerg Med 2017;35:784-6.
  21. Luoma JB, Martin CE, Pearson JL. Contact with mental health and primary care pro- viders before suicide: a review of the evidence. Am J Psychiatry 2002;159:909-16.
  22. Miller IW, Camargo Jr CA, Arias SA, et al. Suicide prevention in an emergency depart- ment population: the ED-SAFE study. JAMA Psychiat 2017;74:563-70.

Leave a Reply

Your email address will not be published. Required fields are marked *