Super-Utilizers and 30-day Readmissions at CCRMC

USING THE DATA TO UNDERSTAND THE PATIENT POPULATION AND DESIGN THE INTERVENTIONS

Sara Levin, MD February 10, 2014

Safety Net Systems Need to Examine Their Data on Readmissions

May not be an optimal quality indicator for a safety net

population

The burdens of mental illness, poor social support, poverty, marginalization impact the unplanned readmissions

Readmission rates as a quality indicator may penalize systems that care for poor or vulnerable populations

Readmits are driven by factors unrelated to “medical care” interventions.

Why Readmissions?

If our population is different, how do our interventions and measures need to be different?

Need to understand what our readmissions represent

Improved discharge planning and care coordination is a good goal

What Is Our Readmission Rate?

11.2 Includes all-cause medical, surgical, critical care, gynecology admissions;

excludes psych, OB, duplications

Two metrics are presented: Rates, calculated by visit (~16000 visits) and percentages of patients who readmitted during study period (~11000 patients)

For July 2008-June 2011:

11.2 per 100 discharges at risk for readmission

10% of patients have 30-day readmit 1 or more times during study period.

The Safety-Net Readmit Population: YOUNGER

2.44

7.74 9.14 9.68

12.23 13.16 11.10 11.16

Under 18 19-25 26-35 36-45 46-65 66-75 Over 75 All Ages

Re

adm

issi

on

Rat

e

Age Ranges of Patients at Visit (in Years)

30 Day Readmission Rate by Age Ranges of Patients at Visit

Overall Readmit Rate

The Safety-Net Readmit Population: SICKER

6.84

10.36

16.53 17.32

19.89

12.93 11.16

1 Day 2 Days 3-4 Days 5-7 Days 8-14 Days 15-30 Days All Durations

Re

adm

issi

on

Rat

e

Duration of Previous Length of Stay

30 Day Readmission Rate by Patient Length of Stay at Previous Visit

Overall Readmit Rate

The Safety-Net Readmit Population: SICKER

7.7% 11.5%

48.8%

9.9%

No PQI Dx 1 PQI Dx 2 or More PQI Dx All Dx

Quantity of Ambulatory Sensitive Dx per MRN

Percent of Patients with Readmission within 30 Days

Overall percent of patients that readmit

The Safety-Net Readmit Population: SMALLER

Conclusion: The majority of patients do not readmit. In our safety population 1% of patients readmitted more than 2 times in the study period. 2.5% of patients account for 50% of readmission visits.

10164

844 160 46 31 16 5 7 2 2 1 1 1 1

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Nu

m b

er

of

Pat

ien

ts

Number of Readmissions

On Cost - Superutilizers

Patient Distribution Percent of Total Charges

Top 0.1 Percentile = 124 Patients

12 %

Top 1 Percentile = 1235 Patients

36 %

Top 2.5 Percentile = 3126 Patients

50 %

Top 5 Percentile = 6177 Patients

61 %

Top 10 Percentile = 12354 Patients

73 %

*Data from 2010 Calendar Year Claims Data Across CCHP/CCRMC/HC Delivery

Top 2.5% utilizers responsible for 83% of Readmissions (941 out of total 1133 Readmissions were accounted for by top 2.5% of patients)

Superutilizers – On Readmissions

Top 2.5%

Utilizers, 941,

83%

Next 2.5%

Utilizers, 125,

11%

Remaining 95%

Utilizers,67, 6%

Readmission Distribution

*Data from 2010 Calendar Year Claims Data Across CCHP/CCRMC/HC Delivery System

Superutilizers - Where Is The Utilization?

5.3

0.8 0.1 0.4

20.3

3.8

1.5

10.7

0.0

5.0

10.0

15.0

20.0

25.0

Avg. Outpatient

Visit

Avg. ED Visit Avg. Inpatient

Admission

Avg. Inpatient

Days

All Patients

Top 2.5% Utilizers

*Data from 2010 Calendar Year Claims Data Across CCHP/CCRMC/HC Delivery System

Superutilizers - On Burden of Disease

6.1

4.5

3.7

3.2

2.7

2.2 1.7

1.4 1.0 0.9

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Top 10

Percentile

10-20

Percentile

20-30

Percentile

30-40

Percentile

40-50

Percentile

50-60

Percentile

60-70

Percentile

70-80

Percentile

80-90

Percentile

90-100

Percentile

Number of Comorbidities

*Data from 2010 Calendar Year Claims Data Across CCHP/CCRMC/HC Delivery System

Two physicians and one epidemiologist asked:

Can we explain the drivers of readmission rate in the safety-net population based on what we see day-to-day in this small population?

And Hypothesized:

We can define “conditions of vulnerability” that are driving the readmission rate on this population based on clusters of ICD-9 codes that communicate vulnerable states.

Conditions of Vulnerability

We defined 5 “conditions of vulnerability” by clusters of ICD-9 codes:

End-of-life

Frailty

Substance Use

Mental Illness

Chronic Pain/Immobility

Conditions of Vulnerability

• Alcohol/drug-induced psychosis

• Opioid, Methamphetamine, Cocaine, other drug abuse

• Alcohol Abuse

• Alcohol or Drug Withdrawal • Esophageal Varices

• Ascites

• Hepatitis (alcoholic)

• Cirrhosis (alcoholic)

• Toxic ingestions

Conditions of Vulnerability: Substance Use

38% of Patients with Substance Abuse Diagnosis

13.6%

7.7%

Substance Abuse No Substance Abuse

Percent of patients with one or more readmission

• Severe Mood Disorders

• Anxiety Disorders

• Psychoses

• Somataform

Conditions of Vulnerability: Mental Illness

18% of patients have at least one diagnosis associated with Mental Illness

17.0%

8.3%

Mental Health No Mental Health

Percent of patients with one or more readmission

• Chronic pain syndromes

• Pressure ulcers in non-paraplegic

• Osteomyelitis, chronic

• Fistulas, intestinal, entero, vesicular

• Chronic, Non-Healing Wound

• Morbid Obesity

Conditions of Vulnerability: Chronic Pain, Immobility, Obesity

24% of patients have Chronic Pain/Immobility/Obesity

19.3%

7.0%

Immobility, Chronic Pain No Immobility, Chronic Pain

Percent of patients with one or more readmission

Poisoning by Rx Meds

Dehydration Metabolic disorders Electrolyte disorders Delirium/Dementia Senility Failure to thrive Cachexia

Conditions of Vulnerability: Frailty

19.8%

6.0%

Frailty No Frailty

Percent of patients with one or more readmission

28 % of patients have at least one diagnosis associated with Frailty

• Malignant Pleural Effusion

• Malignant Pain

• Malignant ascites

• Neutropenia (Chemo-induced)

• Aspiration Pneumonia

• Malignancies of digestive systems and Respiratory systems

• Ovarian Malignancies

• Hematologic Malignancy • Not in remission

Conditions of Vulnerability: End-of-Life

27.3%

7.9%

End of Life No End of Life

Percent of patients with one or more readmission

10% of patients have at least one diagnosis associated with End of Life

Conclusion: Among patients with ambulatory sensitive admissions, a presence of a condition of vulnerability MARKEDLY increases risk of readmission for patient

Conditions of Vulnerability Drive Readmissions in Patients Admitted with PQI Diagnoses

15%

25%

11%

29%

16%

30%

13%

31%

15%

40%

WithoutSubstance

Abuse

SubstanceAbuse

WithoutFrailty

Frailty WithoutMentalHealth

MentalHealth

WithoutImmobility,

Chronic Pain

Immobility,Chronic Pain

Without Endof Life

End of Life

Pe

rce

nt

of

pat

ien

ts w

ith

on

e o

r m

ore

30

day

re

adm

issi

on

s

Readmissions among patients with ambulatory sensitive admissions

72% 171% 83% 138%

160%

Conditions of vulnerability more significant than any other single disease or

demographic entity

Using ICD-9 codes indicators of basic needs instability exacerbated by

illness

Superutilizers Understanding and defining the populations at-risk for

becoming superutilizers can help us to design effective and compassionate interventions at the front end

Conclusions

Ambulatory palliative care

Superutilizer program – High Intensity Health Home (Camden Model?)

Align and Integrate Mental Health, Behavioral Health and Physical Health services

Coordinate with our Public Health, Health Plan and Community partners and integrate social services across the system and into health care delivery sites

Conclusions

PROACT Team: Patient Touchstone PROACT TEAM: Inpatient teams; Home Visit MDs; PROACT RN; Clinical Pharmacist; Case Mgmt Coaching (CTI); SNF FNP

Based on risk level: Phone Visit<72hrs PROACT Clinical Pharmacist

Coaching/Case mgmt. Home Health referral Home Visit MD High Intensity Health Home(?)

Warm Hand-offs/ active management

Primary Care Health Home

Connect to services: mental health specialty care addiction treatment hospice community support

PROACT

PROACT Outputs

Risk stratification tool applicable to our safety net population Assess every hospitalized patient for readmission risk Deliver a variable intensity care intervention upon discharge

Every patient discharged from CCRMC will receive a post-discharge contact within 72 hours of discharge to assess obstacles with discharge plan and assist with follow-up plan

Projected Savings from Readmissions Reductions

$196K/Mo now! (12% reduction 30-day readmits) $327K/Mo when we reach goal of 20% $3.9 Mil/Yr

Contra Costa County Care Transitions (future state?)