©20

08 T

hom

son

Reu

ters

Psychiatric Care in General Hospitals With and Without Psychiatric Units: How Much

and for Whom? 

Tami L. Mark, Thomson Reuters

Elizabeth Stranges, Thomson Reuters

Rita Vandivort-Warren, SAMHSA

Carol Stocks, AHRQ

Pam Owens, Consultant AHRQ

2009 AHRQ Annual Conference

September 14, 2009

©20

08 T

hom

son

Reu

ters

2

Background• Despite the shift from long term inpatient stays to community

treatment, hospitalization remains a key component of mental health care today, primarily for people in crisis

• Most inpatient psychiatric treatment occurs in general acute care hospitals rather than specialty psychiatric hospitals

• General hospital psychiatric care can be provided in two distinct ways– Psychiatric Units:

• Set up and staffed specifically for psychiatric treatment• Separate, often locked, space within hospital

– “Scatter Beds”• General medical care beds located throughout the hospital

©20

08 T

hom

son

Reu

ters

3

Public Policies Affecting Psychiatric Unit Supply and Demand

• Decline in beds in public psychiatric hospitals and more recently private psychiatric hospitals

©20

08 T

hom

son

Reu

ters

Change in Psychiatric Beds in U.S.

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

450,000

500,000

1970 1976 1980 1986 1990 1992 1995 1998 2000 2002

State and county mental hospitals

Private psychiatric hospitals

Non-Federal general hospitals with separate psychiatric services

©20

08 T

hom

son

Reu

ters

5

Public Policies Affecting Psychiatric Unit Supply and Demand

• PPS Exemption of Psychiatric Units, October 1983

• Managed Care Limits on Inpatient Care in 1990s

• PPS implementation, January 2005

• IMD Exclusion encourages use of psych units in community hospitals– Medicaid will not pay for inpatient treatment for persons

age 21-64 who receive care in an “institution for mental disease”, defined as an institution of more than 16 beds that primarily treats people with mental illness

©20

08 T

hom

son

Reu

ters

6

Shortage of Psychiatric Beds?

• Overcrowding in emergency rooms due to psychiatric patients

• Many hospitals report “ED boarding” of patients with psychiatric illness

• Survey of state mental health authorities revealed that more than 80 percent of states reported a shortage of psychiatric beds

©20

08 T

hom

son

Reu

ters

7

Research Questions

• How much psychiatric care in general hospitals is occurring in psychiatric units and how much in scatter beds?

• What types of patients are being treated in psychiatric units and what types in scatter beds?

©20

08 T

hom

son

Reu

ters

Motivation

• To what extent are patients treated in community hospitals receiving the specialized services that psychiatric units offer?

• Are scatter beds being used more in regions where there are not specialized units to supplement psychiatric beds?

• Are scatter beds being used primarily to treat medical comorbid conditions or do patients being treated there primarily have psychiatric conditions?

©20

08 T

hom

son

Reu

ters

9

Outline

• Data Sources

• Prior Research

• Part 1: Number of community hospital psychiatric patients treated in psychiatric units vs scatter beds

• Part 2: Characteristics of patients treated in psychiatric units and scatter beds

• Conclusions

©20

08 T

hom

son

Reu

ters

10

Data Sources on Care in Psychiatric Units

• American Hospital Association Survey of Hospitals

• Medicare Cost Reports

• SAMHSA Survey of Mental Health Specialty Facilities (IMHO, SMHO)

• HCUP-SID Revenue Codes

©20

08 T

hom

son

Reu

ters

11

Prior Research

• Kiesler & Simpkins: The Unnoticed Majority in Psychiatric Inpatient Care, 1995

• Methods – 1980 Hospital Discharge Survey by NCHS

– Identified psychiatric unit using NIMH survey of psychiatric units (now carried out by SAMHSA) and AHA

• Findings – In 1980, 38% of psychiatric inpatient episodes in

community hospitals occurred in scatter beds

©20

08 T

hom

son

Reu

ters

Part I: Estimating the Percent of Psychiatric Discharges from General Hospitals in Psychiatric Units and Scatter Beds

©20

08 T

hom

son

Reu

ters

13

Methods

• HCUP-SID discharges (2000 – 2006)– Total number of discharges from community hospitals in

participating states

– Examined those with principal psychiatric diagnoses (excluding substance abuse)

• Linked to Medicare Cost Report through AHA ID– Information on whether have PPS exempt psychiatric unit

• Checked information on psychiatric unit against volume of MH discharges– With additional web searching for verification

©20

08 T

hom

son

Reu

ters

14

Study Sample (2006)

U.S. Community Hospitals

4,927

HCUP-SID Hospitals

4,309

(38 States)

AHA-SID-MCR Linked Hospitals

4,220

86% of Community Hospitals

©20

08 T

hom

son

Reu

ters

15

Of Community Hospitals, 27% Have Psychiatric Units, Down from 36% in 2002

Percent of Community Hospitals with Psychiatric Units

3436

3129 28 27

0

5

10

15

20

25

30

35

40

2001 2002 2003 2004 2005 2006

©20

08 T

hom

son

Reu

ters

16

About 20% of Discharges are from Hospitals without Psychiatric Units Based on MCR and SID

Percent of Discharges from Hospitals with and without Psychiatric Units: Based on MCR and SID

15 17 16 17 18 19

85 83 84 83 82 81

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2001 2002 2003 2004 2005 2006

Without Psychiatric Units With Psychiatric Units

©20

08 T

hom

son

Reu

ters

Psychiatric Discharges Across States

Rate of Psychiatric Discharges per 10,000 State Population from Psych Units and Scatterbeds, 2003 (Based on Presence of MCR Indicator)

-

10

20

30

40

50

60

70

80

IA NV WA TX UT GA AZ CO CA SC HI WI

OR NH VT MI IN W

V NC NJ VA RI KS NE FL ME TN MA OH MN KY NY CT PA IL SD MO

Psychiatric Unit Discharges Scatterbed Discharges

©20

08 T

hom

son

Reu

ters

18

Summary of Analysis Thus Far

• Summary: About 20% of discharges from hospitals without psychiatric units

• Maybe over-estimate scatterbeds: Assumes no under-reporting of psychiatric units by hospitals

• Maybe under-estimate scatterbeds: Assumes that all discharges from hospitals with psychiatric units are from psychiatric units

©20

08 T

hom

son

Reu

ters

19

Under Reporting Analysis

• 94% of community hospitals without an MCR psychiatric unit indicator had less than 100 MH discharges (based on HCUP-SID counts).

• 6% of hospitals without an MCR psychiatric unit indicator had 100 or more MH discharges

• 39 of the 50 hospitals (78%) with >100 MH discharges but no MCR indicator had a psychiatric unit indicated on their website

• Conclusion: Discharge volume can be used to impute missing MCR psychiatric unit status

©20

08 T

hom

son

Reu

ters

20

About 2% of Discharges from Hospitals without Units (If > 100 MH discharges is used as a proxy for a unit)

Percent of Discharges from Hospitals with and without Units based on 100+ MH Discharges

2 2 2 2 2 2

98 98 98 98 98 98

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2001 2002 2003 2004 2005 2006

Without Psychiatric Units With Psychiatric Units

©20

08 T

hom

son

Reu

ters

21

Over Estimation Analysis

• Used revenue codes for room & board charges for 12 states to examine whether discharges had revenue codes indicating psychiatric unit room and board charge

• Found 3.6% of discharges from hospitals with psychiatric units were from scatter beds

©20

08 T

hom

son

Reu

ters

22

About 6% of Discharges are from Scatter Beds after Correcting for Under and Over Estimation

Percent of Discharges from Psychiatric Units and Scatter Beds: Best Estimate

6 5 6 6 6 6

94 95 94 94 94 94

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2001 2002 2003 2004 2005 2006

Psychiatric Unit Discharges Scatter Bed Discharges

©20

08 T

hom

son

Reu

ters

Part II: Characteristics of Patients in Psychiatric Units and Scatter Beds

©20

08 T

hom

son

Reu

ters

24

Methods

• Used states that had revenue codes that accurately captured room and board

• Examined discharges that had a psychiatric room & board revenue code as compared to those from medical surgical rooms

©20

08 T

hom

son

Reu

ters

25

Data

• 12 HCUP-SID States

• Kentucky, Maine, Massachusetts, Nebraska, Nevada, New York, North Carolina, Pennsylvania, Tennessee, Texas, Washington, and West Virginia

©20

08 T

hom

son

Reu

ters

26

Characteristics Examined

• Age

• Gender

• Length of stay

• ICD-9-CM mental health diagnoses

• Existence of any secondary mental health, substance abuse, or non-mental health substance abuse ICD-9-CM diagnoses

• ICD-9-CM Procedures

• Total charges

• Admission source

• Discharge type

©20

08 T

hom

son

Reu

ters

27

Scatter Bed Discharges are More Female

Percent of Females in Scatter Beds and Units

62.859.3

54

0

10

20

30

40

50

60

70

80

90

100

Scatter Beds -Hospitals without Units

Scatter Beds -Hospitals with Units

Psychiatric Units

Perc

en

t o

f D

isch

arg

es

©20

08 T

hom

son

Reu

ters

28

Scatter Beds Discharges Are Older

Percent Over Age 65 in Scatter Beds and Units

30.3

20.5

9.2

0

10

20

30

40

50

60

70

80

90

100

Scatter Beds - Hospitalswithout Units

(Mean Age = 51)

Scatter Beds - Hospitalswith Units

(Mean Age = 43)

Psychiatric Units (Mean Age = 39)

Per

cen

t o

f D

ish

carg

es

©20

08 T

hom

son

Reu

ters

29

Scatter Bed Discharges are More Medicare and Less Medicaid

Percent of Discharges by Payer in Scatter Beds and Units

0

5

10

15

20

25

30

35

40

45

Scatter Beds -Hospitals without Units

Scatter Beds -Hospitals with Units

Psychiatric Units

Per

cen

t o

f D

isch

arg

es

Medicare

Medicaid

Private Payer

Uninsured

©20

08 T

hom

son

Reu

ters

30

Scatter Beds Have Lower Lengths of Stay

Length of Stay in Scatter Beds and Units

4

6

10

0

2

4

6

8

10

12

Scatter Beds - Hospitalswithout Units

Scatter Beds - Hospitals withUnits

Psychiatric Units

Ave

rag

e N

um

ber

of

Day

s

©20

08 T

hom

son

Reu

ters

31

Scatter Beds have More Emergency Room Admissions

Emergency Room Admissions to Scatter Beds and Units

68.1

60.2

54.1

0

10

20

30

40

50

60

70

80

Scatter Beds - Hospitalswithout Units

Scatter Beds - Hospitals withUnits

Psychiatric Units

Per

cen

t o

f D

isch

arg

es

©20

08 T

hom

son

Reu

ters

32

Scatter Bed Discharges are More Likely to be Transferred

Percent of Discharges Transfered to Another Facility

20.6

17.6

10.1

0

5

10

15

20

25

Scatter Beds - Hospitalswithout Units

Scatter Beds - Hospitals withUnits

Psychiatric Units

Per

cen

t o

f D

isch

arg

es

©20

08 T

hom

son

Reu

ters

33

Units have more schizophrenia and episodic mood disorders, scatter beds have more anxiety and other nonorganic psychosis

8

23

28

139

29

23

10

22

54

25

0

10

20

30

40

50

60

Schizophrenia Episodic MoodDisorders

Anxiety Disorders Other NonorganicPsychoses

Per

cen

t o

f D

isch

arg

es

Scatter Beds - Hospitals without Units Scatter Beds - Hospitals with Units Psychiatric Units

©20

08 T

hom

son

Reu

ters

34

Scatter beds have more secondary non-MHSA diagnoses

Percent with Secondary non-MHSA diagnosis

89

79

70

0

10

20

30

40

50

60

70

80

90

100

Scatter Beds - Hospitalswithout Units

Scatter Beds - Hospitals withUnits

Psychiatric Units

Per

cen

t o

f D

isch

arg

es

©20

08 T

hom

son

Reu

ters

35

Most Discharges Across Settings do not have Procedures Coded

Percent of Discharges without Procedure Codes

8277 80

0

10

20

30

40

50

60

70

80

90

100

Scatter Beds - Hospitalswithout Units

Scatter Beds - Hospitals withUnits

Psychiatric Units

Per

cen

t o

f D

isch

arg

es

©20

08 T

hom

son

Reu

ters

36

Limitations

• Data on psychiatric unit status is imperfect

• Data on details of clinical treatment being provided to patients in scatter beds is limited

©20

08 T

hom

son

Reu

ters

37

Conclusions

• Psychiatric units may play a more critical role than previously appreciated in ensuring an adequate supply of inpatient psychiatric care

• Scatter beds tend to be used for a short amount of time (4 days on average) and 20% of patients are transferred. More likely to be used for older Medicare patients with anxiety although 1/3 have schizophrenia or mood disorders

©20

08 T

hom

son

Reu

ters

38

Policy Implications

• There are no U.S. policies to regulate, monitor, or create incentives for adequate access to psychiatric beds across the country

• This may need to be addressed to ensure adequate access to inpatient care

• Additionally, need to consider how and whether psychiatric units can be supplemented with good quality psychiatric care provided in hospitals without psychiatric units

©20

08 T

hom

son

Reu

ters

THANK YOU

Tami.Mark@ThomsonReuters.Com