1. Analyzing MHSA Hospital Treatment: Results from an AHRQ-SAMHSA Collaboration Carol Stocks, RN, MHSA Sam Schildhaus, PhD Katharine Levit Pat Santora, PhD AHRQ  September 19, 2011

2. Overview of Session HCUP Data Overview HCUP Data Overview Carol Stocks Carol Stocks Emergency Departments Emergency Departments – MHSA Visits to Emergency Departments Carol Stocks Carol Stocks – SA Visits to Emergency Departments for the Uninsured Sam Schildhaus Sam Schildhaus Inpatient Stays Inpatient Stays – MHSA Inpatient Stays in Community Hospitals Katharine Levit Katharine Levit 2

3. HCUP Data Overview Carol Stocks 3

4. Healthcare Cost and Utilization Project (HCUP) What is HCUP? What is HCUP? – Hospital-based administrative data – Large collection going back many years – Encounter-level with all “payers” including the uninsured – Includes inpatient, emergency department and ambulatory surgery data 4

5. 5 Demographic Data DiagnosesProceduresCharges The Foundation of HCUP Data is Hospital Billing Data

6. 6 Patient enters hospital Hospital sends billing data and any additional data elements to Data Organizations States store data in varying formats Billing record created AHRQ standardizes data to create uniform HCUP databases The Making of HCUP Data

7. What Are Community Hospitals? AHA definition of community hospitals: Non-Federal, short- term, general, and other specialty hospitals, excluding hospital units of other institutions (e.g., prisons) Include these hospitals: Multi-specialty general hospitals Multi-specialty general hospitals OB-GYN OB-GYN ENT ENT Orthopedic Orthopedic Pediatric Pediatric Public Public Academic medical centers Academic medical centers Exclude these hospitals: Long-term care Psychiatric Alcoholism/chemical dependency Rehabilitation DoD / VA / IHS 7

8. HCUP Databases Research Publications User Support Research Products SASDSEDD NIS SID KID NEDS 8 HCUP is a Family of Databases, Tools, and Products

9. Key : Participating Non-participating AZ CA UT CT FL GA IA IL KS MA MD MO NJ NY OR PA SC TN CO WA WI VA ME MN MI NC TX KY WV RI NE VT NV OH SD AR IN NH MT ID WY ND NM OK LA MS AL DE HI AK 9 HCUP Partners Providing 2010 Inpatient Data

10. HI AK Key : Participating Non-participating RI AZ CA UT CT FL GA IA IL KS MA MD MO NJ NY OR PA SC TN CO WA WI VA ME MN MI NC TX KY WV NE VT NV OH SD AR IN NH MT ID WY ND NM OK LA MS AL DE 10 HCUP Partners Providing 2010 Emergency Department Data

11. HCUP National Databases are Sampled from State Databases 11 State Inpatient Databases NIS State Emergency Department Databases KID NEDS

12. What is HCUP and What Is It Not? HCUP is... A collection of electronic discharge records from health care encounters All payer, including the uninsured Hospital, ambulatory surgery, emergency department data All hospital discharges from participating states (currently 44) Accessible multiple ways: raw data, reports, on-line aggregate statistics HCUP is NOT... A survey Specific to a single payer, e.g. Medicare Office visits, pharmacy, laboratory, radiology Only a sample Inaccessible 12

13. Recap: Use of HCUP Databases Benefits Large sample size Large sample size Uniformity of coding Uniformity of coding Routine, regular collection Routine, regular collection Ease of access Ease of access All-payer All-payer Available at local, state, regional, national level Available at local, state, regional, national level Supplemental files available Supplemental files available Limitations Differences in coding across hospitals Differences in coding across hospitals No data on individuals outside of hospital system No data on individuals outside of hospital system May not show complete episode of care May not show complete episode of care May not include all hospitals May not include all hospitals Lack revenue information Lack revenue information Limited clinical details Limited clinical details ED data do not contain information on time to triage, time to treatment, time to disposition, etc. ED data do not contain information on time to triage, time to treatment, time to disposition, etc. 13

14. Mental Health and Substance Abuse (MHSA) Emergency Department (ED) Visits, 2007 Carol Stocks 14

15. Characteristics of MHSA- related Adult ED Visits 12.5 percent of all ED visits (12 million visits) were MHSA- related: 12.5 percent of all ED visits (12 million visits) were MHSA- related: 41 percent of visits resulted in hospital admission – over 2.5 times the rate of admission for other conditions 41 percent of visits resulted in hospital admission – over 2.5 times the rate of admission for other conditions 54 percent of MHSA ED visits were for women 54 percent of MHSA ED visits were for women 18-44 year olds comprised the largest share (47 percent) of adult ED visits 18-44 year olds comprised the largest share (47 percent) of adult ED visits Medicare was the most frequently billed payer (30 percent of visits) Medicare was the most frequently billed payer (30 percent of visits) 64 percent of visits involved MH conditions, 24 percent SA conditions, and 12 percent co-occurring MHSA conditions 64 percent of visits involved MH conditions, 24 percent SA conditions, and 12 percent co-occurring MHSA conditions 15

16. Most Common Reasons for MHSA-related Adult ED Visits Five all-listed MHSA conditions accounted for 96 percent of documented MHSA conditions during ED visits: Five all-listed MHSA conditions accounted for 96 percent of documented MHSA conditions during ED visits: Mood disorders (43 percent of visits) Mood disorders (43 percent of visits) Anxiety disorders (26 percent of visits) Anxiety disorders (26 percent of visits) Alcohol disorders (23 percent of visits) Alcohol disorders (23 percent of visits) Drug disorders (18 percent of visits) Drug disorders (18 percent of visits) Schizophrenia and other psychoses (10 percent of visits) Schizophrenia and other psychoses (10 percent of visits) 16

17. Percentage of Hospital Admissions for Adult ED Visits with MHSA Conditions, 2007 17

18. Payers for MHSA Adult Care in Community Hospitals, 2007 18

19. Adult ED Visits with MHSA Conditions by Age Groups, 2007 19

20. Expected Payer for ED Visits with MHSA Conditions, 2007 20

21. ED Visits and MHSA-related Conditions MHSA conditions were documented for 12.5 percent of the 122.3 million total ED visits for all conditions. MHSA conditions were documented for 12.5 percent of the 122.3 million total ED visits for all conditions. Mental health diagnoses were involved in 8 percent of all ED visits (9.9 million visits). Mental health diagnoses were involved in 8 percent of all ED visits (9.9 million visits). Alcohol-related disorders were involved in 2.3 percent of ED visits (2.8 million visits). Alcohol-related disorders were involved in 2.3 percent of ED visits (2.8 million visits). Drug-related disorders were involved in 1.8 percent of visits (2.2 million visits). Drug-related disorders were involved in 1.8 percent of visits (2.2 million visits). 21

22. MHSA Discharge Status from the ED 22

23. Most Frequent Types of MHSA – related ED Visits 23

24. Substance Use Disorder (SUD) Emergency Department Visits for the Uninsured, 2009 Sam Schildhaus 24

25. Emergency Department Major portal for entry into hospital and institutional care. Major portal for entry into hospital and institutional care. Emergency Department (ED) source of admission to hospital of 50% of all non- obstetric admissions in 2006, up from 36% in 1996. Emergency Department (ED) source of admission to hospital of 50% of all non- obstetric admissions in 2006, up from 36% in 1996. Legal mandate under Emergency Medical Treatment and Labor Act (EMTALA) – those who come to ED must receive medical screening and be stabilized regardless of insurance status or ability to pay Legal mandate under Emergency Medical Treatment and Labor Act (EMTALA) – those who come to ED must receive medical screening and be stabilized regardless of insurance status or ability to pay 25

26. Increase in ED Visits Between 1997 and 2007, ED visits increased by 23% from 95 million to 117 million* Between 1997 and 2007, ED visits increased by 23% from 95 million to 117 million* ED is crucial to patients with substance use disorders (SUD), saving the lives of those with drug/alcohol overdoses and treating the consequences of SUD ED is crucial to patients with substance use disorders (SUD), saving the lives of those with drug/alcohol overdoses and treating the consequences of SUD * National Hospital Ambulatory Medical Care Survey: 1997 Emergency Department Summary, Vital and Health Statistics, Centers for Disease Control and Prevention, National Center for Health Statistics, number 304, May 6, 1999, Table 1, page 4; National Hospital Ambulatory Medical Care Survey: 2007 Emergency Department Summary, Centers for Disease Control and Prevention, National Center for Health Statistics, number 26, August 6, 2010,Table 1, page 7 26

27. Prior Related Research Owens and Mutter: HCUP-NEDS (2006) – Owens and Mutter: HCUP-NEDS (2006) – – Treat-and-Release (routine discharge) 1.4 times higher among the uninsured than the insured – Admission among insured 2.1 times higher among insured than uninsured Owens, Mutter, and Stocks: HCUP-NEDS (2007) – Owens, Mutter, and Stocks: HCUP-NEDS (2007) – – Uninsured mental health and substance use related ED visits were two to four times less likely to result in hospitalization than patient visits with insurance coverage 27

28. Study Purpose Analyze ED visits with principal or secondary SUD diagnosis (Dx) to examine the following: Analyze ED visits with principal or secondary SUD diagnosis (Dx) to examine the following: – Does payer status differ among types (e.g., alcohol only, drug only, both) of SUD patients? – When the relationships are statistically controlled, is discharge of SUD patients to hospital or institution associated with patient, payer, and hospital characteristics? 28

29. Operational Definition: SUD – Any SUD diagnosis (Dx), both principal and secondary Dx – ICD-9-CM Alcohol Abuse: 291.0-291.9 303.00-303.92, 305.00-305.02, but excluding remission code of 303.03. Alcohol Abuse: 291.0-291.9 303.00-303.92, 305.00-305.02, but excluding remission code of 303.03. Drug Abuse: 292.0-292.9, 304.00-304.92, 305.20- 305.92,648.30-648.34, 965.00-965.02, but excluding medication error and remission codes 292.81,304.03, 304.13, 304.23, 304.33, 304.43, 304.63, 304.73, 304.83, 305.43, 305.53, 305.63, 305.73 Drug Abuse: 292.0-292.9, 304.00-304.92, 305.20- 305.92,648.30-648.34, 965.00-965.02, but excluding medication error and remission codes 292.81,304.03, 304.13, 304.23, 304.33, 304.43, 304.63, 304.73, 304.83, 305.43, 305.53, 305.63, 305.73 29

30. Findings Approximately 19 million of 77 million (25%) emergency department visits were by the uninsured ages 18-64 years Approximately 19 million of 77 million (25%) emergency department visits were by the uninsured ages 18-64 years Approximately 1.4 million of the 19 million (7%) had a diagnosed substance use disorder Approximately 1.4 million of the 19 million (7%) had a diagnosed substance use disorder 30

31. Findings Payer status of ED visits by those 18-64 Payer status of ED visits by those 18-64 – Uninsured: 25% – Private insurance: 39% – Medicaid: 20% – Medicare: 9% – Other payers: 6% SUD discharges more likely than non-SUD discharges to be uninsured (35% vs. 25%) SUD discharges more likely than non-SUD discharges to be uninsured (35% vs. 25%) 31

32. Multivariate Analysis Would the substantial difference in discharge disposition between the SUD and non-SUD patients be associated with many patient and facility characteristics? Would the substantial difference in discharge disposition between the SUD and non-SUD patients be associated with many patient and facility characteristics? To test the relationship among the characteristics, we used a multivariate model that statistically controls for patients’ socio- demographic characteristics, chronic conditions, self harm, insurance, and hospital characteristics To test the relationship among the characteristics, we used a multivariate model that statistically controls for patients’ socio- demographic characteristics, chronic conditions, self harm, insurance, and hospital characteristics 32

33. Likelihood of discharge to hospital/institution after ED visit Older patients (45-64) 9% less likely than younger (18-44) patients Older patients (45-64) 9% less likely than younger (18-44) patients Women 21% less likely than men Women 21% less likely than men Patients residing in poorest zip codes 17% less likely than patients residing in wealthier zip codes Patients residing in poorest zip codes 17% less likely than patients residing in wealthier zip codes Patients with Medicare 15% more likely than uninsured Patients with Medicare 15% more likely than uninsured Patients with private insurance 41% more likely than uninsured Patients with private insurance 41% more likely than uninsured 33

34. Likelihood of discharge to hospital/institution after ED visit Patients with other insurance 57% more likely than uninsured Patients with other insurance 57% more likely than uninsured Patients with higher number of Dx 42% more likely than with lower number of Dx Patients with higher number of Dx 42% more likely than with lower number of Dx Visits by patients with higher number of chronic conditions 31% more likely than with lower number of chronic conditions Visits by patients with higher number of chronic conditions 31% more likely than with lower number of chronic conditions Visits by patients who intended to hurt self 3.9 times more likely than others Visits by patients who intended to hurt self 3.9 times more likely than others Visits at teaching hospital 31% more likely than visits at nonteaching hospital Visits at teaching hospital 31% more likely than visits at nonteaching hospital 34

35. Issues Over one third (35%) of MHSA visits treated in community hospital EDs are uninsured Over one third (35%) of MHSA visits treated in community hospital EDs are uninsured Lack of insurance is associated with decreased post-ED care in community hospitals even after demographic, diagnostic, and hospital characteristics are statistically controlled Lack of insurance is associated with decreased post-ED care in community hospitals even after demographic, diagnostic, and hospital characteristics are statistically controlled Important to monitor this relationship under expanded insurance coverage through the Affordable Care Act Important to monitor this relationship under expanded insurance coverage through the Affordable Care Act 35

36. Mental Health and Substance Abuse (MHSA) Community Hospital Inpatient Visits, 2008 Katharine Levit 36

37. MHSA Conditions Accounted for 5% of Hospital Stays 39.9 million inpatient stays in 2008, 1.8 million (about 5%) for MHSA 39.9 million inpatient stays in 2008, 1.8 million (about 5%) for MHSA 6 MHSA stays per 1,000 population 6 MHSA stays per 1,000 population MHSA stays averaged 7.1 days compared to 4.6 days for all stays MHSA stays averaged 7.1 days compared to 4.6 days for all stays – MH stays: 10.8 days per stay – SA stays: 4.7 days per stay MHSA stays cost $5,500 per stay compared to $9,100 for all stays MHSA stays cost $5,500 per stay compared to $9,100 for all stays 37

38. Mood Disorders were the Single Largest Reason for an MHSA Stays 38

39. ALOS 2.5 Days Longer for MHSA Stays than for All Diagnoses 39

40. ALOS Varied Considerably by MHSA Diagnosis 40

41. MHSA Stays Accounted for 21% of All Discharges Leaving the Hospital Against Medical Advice (AMA) 41

42. MHSA Diagnoses had a Higher Rate of Discharges AMA than All Other Diagnoses 42

43. Non-elderly Adults had a Disproportionate Share of All MHSA Stays Relative to their Share of the Total Population and All Hospital Stays 43

44. There were 60 MHSA Hospital Stays per 10,000 Population 44

45. Most Frequent Principal MHSA Diagnoses by Age Mood disorders was the most frequent principal MHSA diagnosis across all age groups in 1997 and 2008 Mood disorders was the most frequent principal MHSA diagnosis across all age groups in 1997 and 2008 Alcohol-related disorders accounted for 12 percent of MHSA stays among 18-44 year olds, 21 percent of MHSA stays among 45-64 year olds, and 12 percent of MHSA stays for 65- 84 year olds Alcohol-related disorders accounted for 12 percent of MHSA stays among 18-44 year olds, 21 percent of MHSA stays among 45-64 year olds, and 12 percent of MHSA stays for 65- 84 year olds The number of hospital stays for drug-related conditions rose rapidly for all age groups over 45 years old (87-117-percent increase from 1997-2008), while remaining relatively stable (11- percent decline) among 18-44 year olds The number of hospital stays for drug-related conditions rose rapidly for all age groups over 45 years old (87-117-percent increase from 1997-2008), while remaining relatively stable (11- percent decline) among 18-44 year olds The underlying causes of this increase were rapid growth in drug-induced delirium and in poisonings by opiate-based pain medications The underlying causes of this increase were rapid growth in drug-induced delirium and in poisonings by opiate-based pain medications 45

46. Rise in Drug-induced Delirium and Poisonings by Opiate-based Pain Medications Fueled Increase in Drug-related Hospitalizations for Patients 85 and Older Drug-induced delirium and poisonings by opiate-based pain medications accounted for 78 percent of the drug-related stays and 89 percent of the increase in drug-related stays for patients 85 and older Drug-induced delirium and poisonings by opiate-based pain medications accounted for 78 percent of the drug-related stays and 89 percent of the increase in drug-related stays for patients 85 and older Drug-induced delirium can result from side-effects of medications and occurs often in elderly hospitalized patients Drug-induced delirium can result from side-effects of medications and occurs often in elderly hospitalized patients Drug-induced delirium and poisonings by opiate-based pain medications were also responsible for a large number of drug-related discharges in 45-64 year olds (19 percent) and 65-84 year olds (60 percent) Drug-induced delirium and poisonings by opiate-based pain medications were also responsible for a large number of drug-related discharges in 45-64 year olds (19 percent) and 65-84 year olds (60 percent) 46 NUMBER OF DRUG- RELATED DISCHARGES IN 2008 CUMULATIVE GROWTH IN DRUG-RELATED DISCHARGES 1997-2008 PERCENT CONTRIBUTION TO GROWTH IN DRUG- RELATED DISCHARGES 1997-20008 PRINCIPAL ICD-9-CM DIAGNOSIS45-64 Years 65-84 Years 85+ Years 45-64 Years 65-84 Years 85+ Years 45-64 Years 65-84 Years 85+ Years All drug-related discharges65,40016, 0003,200117%96%87%100.0% Drug withdrawal (ICD-9-CM 292.0)20,3002,0001002701077141.913.53.9 Drug-induced delirium (ICD-9-CM 292.81) 4,2006,4002,10014356987.029.069.8 Poisonings by codeine (methylmorphine), meperdine (pethidine), morphine (ICD-9-CM 965.09) 8,3003,30040069338124520.632.919.1 All other drug related conditions*32,6004,30060049802430.624.67.3

47. Adults 18-44 Accounted for Large Shares of Stays for the Most Frequent MHSA Conditions 47

48. The Gender Split for MHSA Stays Varied by Diagnosis 48

49. 14% of All Discharges had a Secondary MH Diagnosis 49

50. 5% of All Discharges had a Secondary SA Diagnosis 50

51. MHSA Stays were More Commonly Uninsured or Insured by Medicaid than All Stays 51

52. The Uninsured and Medicaid Covered a Disproportionate Share of Costs for MHSA Stays 52

53. Schizophrenia was the Most Costly MHSA Diagnosis 53

54. The Average MHSA Hospital Stay Cost $1,200 Less than Stays without a Major OR Procedure 54

55. MH Hospitalization Rates were Higher in Poorest Communities than in All Other Communities 55

56. SA Hospitalization Rates in Poorest Communities were Similar to All Other Communities 56

57. Summary 26% of treatment spending for MHSA conditions went for hospital care in 2005, making stays key parts of treatment 5% of inpatient stays are for MHSA conditions 5% of inpatient stays are for MHSA conditions MHSA stays are longer on average but less costly MHSA stays are longer on average but less costly MHSA conditions vary by age and gender and are often secondary conditions for a stay MHSA conditions vary by age and gender and are often secondary conditions for a stay MHSA stays were 2 to 5 times more likely to be uninsured, depending on the condition MHSA stays were 2 to 5 times more likely to be uninsured, depending on the condition Hospitalized patients with schizophrenia, depression, or bipolar disorder were more likely to reside in the poorest communities Hospitalized patients with schizophrenia, depression, or bipolar disorder were more likely to reside in the poorest communities On the Web at http://www.hcup- us.ahrq.gov/reports/factsandfigures/2008/TOC_2008.jsp On the Web at http://www.hcup- us.ahrq.gov/reports/factsandfigures/2008/TOC_2008.jsp 57

58. 58 Questions and Discussion

59. For Further Information HCUP Facts and Figures: HCUP Facts and Figures: http://www.hcupus.ahrq.gov/reports/factsandfigures/2008/ TOC_2008.jsp HCUP Topical Reports: HCUP Topical Reports: http://www.hcup-us.ahrq.gov/reports/mhsa.jsp HCUP Statistical Briefs: HCUP Statistical Briefs: http://www.hcup-us.ahrq.gov/reports/statbriefs/sbtopic.jsp 59