A Population-Based Perspective on the Care of Complex Patients: - - PowerPoint PPT Presentation

UNC Cancer Outcomes Research Seminar April 14, 2015

• C. Annette DuBard, MD, MPH

SVP, Informatics and Evaluation

A Population-Based Perspective on the Care of Complex Patients: Knowing When to Intervene

CCNC Population Health Management Approach

Key Ingredients:

• Connecting patients to a primary care medical home; providing those

medical homes with support

• Care management model that uses analytics to target highest need

beneficiaries in the appropriate settings

• Practice support model that provides resources to medical homes to

provide higher value care

• Assisting DHHS/DMA in effectively deploying programs (e.g. pharmacy

initiatives, clinical policy changes)

• Shared informatics platform for data integration, analytics and

reporting, and clinical applications

CCNC Footprint Statewide

• 5,000 primary care providers
• 1,800 Practices
• 90% of PCPs in NC
• 1.3 million Medicaid Patients
• 300,000 Aged, Blind, Disabled
• 150,000 Dually Eligible
• 22,000 uninsured in HealthNet

programs; >20,000 privately insured

All 100 NC Counties 14 Networks

Each network averages:

• 1.4 Medical Directors, 1.0 Psychiatrist
• 42.8 Local Care Managers
• 1.8 Pharmacists
• Multiple disciplines: RN, LCSW, RD, …

4

Informatics Center Data Sources and Applications

• Weekly payer claims, monthly

eligibility/enrollment data

• Real-time Admission/Discharge/

Transfer data from >60 hospitals

• Pharmacy claims via ESI,

Surescripts transactions, Medicare Part D

• Lab results (Labcorps, Solstas,

UNCH)

• Behavioral Health-MCO encounter

data

• Other DHHS data: birth certificate,

immunizations

• PCP Electronic Health Records
• Patient data recorded in user

applications Care Management Information System  2,300 active users, embedded locally in primary care practices, hospitals, local health departments, and other settings  >140,000 patients/month with recorded care management activities PharmaceHome  1,100 active users of shared medication management platform,  >13,000 patients/month with medication reconciliation activities Provider Portal  2,500 active users; 40,000 patient records accessed per month Reports and Pop Health Management Tools  secure web-based reports distribution  member, cost, and utlizaiton dashboards; predictive modeling and opportunity analysis; disease registries; geomapping; quality and performance measurement

Population Needs System Resources

CCNC Care Management Evolution

Disease Management Care of Complex Patients

Focus on High Cost/High Risk

Focus on Most Impactable

One Size Fits All

Right sizing of intervention to maximize ROI

Among Medicaid Beneficiaries with Chronic Disease, Multiple Chronic Conditions is the Norm

Data Brief | Clarifying Multimorbidity Patterns to Improve Targeting and Delivery of Clinical Services for Medicaid

• Populations. Boyd et al. CHCS, December 2010

“The near universality of comorbidity across these index conditions highlights the need for care management strategies that explicitly acknowledge this clinical complexity.”

Beneficiaries with 6 or more Chronic Conditions account for 70% of ALL Medicare readmissions

Source: Centers for Medicare and Medicaid Services. Chronic Conditions among Medicare Beneficiaries, Chart book: 2012 Edition Baltimore, MD. 2012.

Targeting Patients At Risk For Readmission

• 190,000 NC Medicaid

recipients are admitted to the hospital every year, and 31,000 have multiple hospital admissions. – Nearly one in ten admissions represents a readmission within 30 days

• f a prior discharge

– Cross-hospital traffic is common: 23% of 30-day readmissions occur in a different facility.

Cross-state traffic: Complex Medicaid patients admitted to UNC from 80 counties

10

11

Real-time notification of hospital admission. Priority flagging based on overall risk profile using historical claims.

Typical patient identified as high priority for CCNC Transitional Care

58 year old man with severe diabetes, kidney disease and Hepatitis C  Earlier in the year: Two ED visits at Duke and Durham Regional; Two UNC hospitalizations with uncontrolled DM and hyperosmolarity coma  Recently hospitalized at Duke with hepatic encephalophathy and aspiration pneumonitis/ acute respiratory failure  Re-hospitalized at UNC with c diff colitis and hepatic coma  Primary care provider is in a Duke-affiliated practice

Medication Review

20 medicines in patient’s possession based on prescription fill history. Additional 10 (unmatched) medicines listed on hospital discharge summary.

Transitional Care Team in Action

• RN care manager and health educator visited patient’s home 2 days after discharge

– Noted chaotic household; patient was “completely confused” about hospital events; unaware blood sugar had been >1000 at admission; “absent-minded” – CM worked with patient & family to develop a person-centered plan of care

• Follow-up PCP visit

– CM accompanied patient to medical home

• Team-based care

– Follow-up home visit by health educator and registered dietician – Patient/family education on “red flags” and use of glucometer – Nutritional assessment – baseline habits and knowledge – Provided bus pass to endocrinology appointment

• Network pharmacist consultation

– Clarified active med list – Corresponded with patient’s endocrinologist to simplify insulin regimen for better manageability, and switch to pen due to visual impairment

Peer-reviewed research

Transitional Care

• 20% reduction in readmissions for patients with multiple

chronic conditions

• Benefit persists far beyond the

first 30 days

• One hospital readmission

avoided for every 6 interventions – strong ROI

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12

Transitional Care (N=1,966) Usual Care (N=1,035)

Months since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to First Readmission for Patients Receiving Transitional Care Vs. Usual Care

Lighter shaded lines represent time from initial discharge to second and third readmissions

(Significant Chronic Disease in Multiple Organ Systems, Levels 5 & 6; ACRG3 = 65-66)

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12

Transitional Care (N=1,747) Usual Care (N=2,451)

Months since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to First Readmission for Patients Receiving Transitional Care Versus Usual Care

Lighter shaded lines represent time from initial discharge to second and third readmissions

(History of Significant Acute Disease, all severity levels ; ACRG3 = 20-25)

Example of an ACRG with a LOW risk of readmission that didn’t benefit from transitional care.

• 20%
• 10%

0% 10% 20% 30% 40% 50%

Reduction in Readmission Risk When Managed

*Size of bubble reflects the number of hospital discharges

Low Risk Medium Risk High Risk

Higher is better Which Patients Benefit the Most from Transitional Care Management?

Most Cancer Patients Have Complex Care Needs

0% 10% 20% 30% 40% 50% 60% 70% 80% Hypertension Mental Health Condition Diabetes Chronic Gastrointestinal Disease Chronic Obstructive Pulmonary Disease Cardiovascular/Ischemic Vascular Chronic Neurologic Chronic Kidney Disease Severe and Persistent Mental Illness 3 or More Chronic Conditions

Comorbidities Among Current NC Medicaid Recipients with Cancer Diagnosis (n=26,827)

NC Medicaid Cancer Patients Discharged from UNC Hospital

20

Home addresses of cancer patients discharged from UNC Hospital Apr 2012 to Mar 2013

NC Medicaid Cancer Patients Discharged from Duke Medical Center

21

Home addresses of cancer patients discharged from Duke Medical Center Apr 2012 to Mar 2013

NC Medicaid Cancer Patients Discharged from Five Major Medical Centers

22

Home addresses of cancer patients discharged from all five major medical centers from Apr 2012 to Mar 2013

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 Yes (N=445) No (N=248) Months since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to Readmission for Patients Receiving and Not Receiving Any Transitional Care

(Adults with “Dominant, Metastatic, and Complicated Malignancies”)

*Includes non-dually enrolled Medicaid patients with an initial discharge in SFY2011 (excluding obstetrics, newborns, malignancies, burns and traumas) who were CCNC enrolled at time of discharge or month after. Significant group differences: Wilcoxon (Gehan) statistic = 25.68, p<.0001

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 Yes (N=74) No (N=91) Months since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to Readmission for Patients Receiving and Not Receiving Any Transitional Care

(Children with “Dominant, Metastatic, and Complicated Malignancies”)

*Includes non-dually enrolled Medicaid patients with an initial discharge in SFY2011 (excluding obstetrics, newborns, malignancies, burns and traumas) who were CCNC enrolled at time of discharge or month after. Significant group differences: Wilcoxon (Gehan) statistic = 7.73, p<.01

Patient Education Timely Follow- up with Outpatient Providers Medication Management Face-to-Face Patient Encounters

Components of Transitional Care: Who Needs What?

26

Answer: It matters more for some than

• thers!

Question: How important is an early follow-up appointment after hospital discharge?

.0 30-Day 90-Day DIALYSIS WITH DIABETES LEVEL - 4 2941 43.3 71.3 CHRONIC RENAL FAILURE - DIABETES - OTH DOM CHRON DIS LEVEL - 6 4170 42.9 66.6 CONGESTIVE HEART FAILURE - DIABETES - COPD LEVEL - 6 3934 35.9 64.5 HIV DISEASE LEVEL - 4 2568 31.9 54.5 TWO OTHER DOMINANT CHRONIC DISEASES LEVEL - 6 2651 28.8 51.6 DIABETES - ADVANCED CAD - OTH DOM CHRON DIS LEVEL - 6 2625 28.2 50.6 DIABETES AND OTH DOM CHRON DIS LEVEL - 6 2254 29.9 50.3 SCHIZOPHRENIA AND OTH MOD CHRON DIS LEVEL - 6 3254 23.9 47.5 CONGENITAL QUADRIPLEGIA, DIPLEGIA OR HEMIPLEGIA LEVEL - 4 2134 23.9 41.6 COPD AND OTH DOM CHRON DIS LEVEL - 6 2052 21.1 40.0 DIABETES AND OTH MOD CHRON DIS LEVEL - 6 2450 20.6 38.0 SCHIZOPHRENIA AND OTH MOD CHRON DIS LEVEL - 5 2771 18.9 37.8 HIV DISEASE LEVEL - 3 2678 19.5 37.5 ONE OTH DOM CHRON DIS AND ONE OR MORE MOD CHRON DIS LEVEL - 6 2142 20.2 37.5 ONE OTH DOM CHRON DIS AND ONE OR MORE MOD CHRON DIS LEVEL - 5 1893 13.5 25.8 SCHIZOPHRENIA AND OTH MOD CHRON DIS LEVEL - 4 3207 12.2 21.4 SCHIZOPHRENIA AND OTH MOD CHRON DIS LEVEL - 3 4665 9.0 17.3 TWO OTHER MODERATE CHRONIC DISEASES LEVEL - 4 1894 7.1 14.1 SCHIZOPHRENIA AND OTH MOD CHRON DIS LEVEL - 2 3321 7.1 11.9 ASTHMA AND OTH MOD CHRON DIS LEVEL - 2 2241 4.1 8.4 Top 20 Largest CRG's N Readmission Rates

Even within the multiple chronic population, the readmission rates are vastly different across CRG’s; with highest risk CRG’s having a 4-6 times greater risk of readmission.

Baseline Readmission Rates by Clinical Risk Group

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 YES (N=3,690) NO (N=6,337)

Days since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to Readmission for Patients Receiving Outpatient Follow-up Within 7 Days of Discharge

(Patients with single dominant or moderate chronic condition; ACRG3 = 51-56) All CCNC enrolled at discharge; inpatient discharges during the period 4/1/12-3/31/13, excluding deliveries, newborns, discharges to another facility and members dually enrolled at discharge.

Analysis of Time to 30-day readmission for patients who did vs. did not have an

• utpatient follow-up visit

(With repeated analyses for each clinical risk group, testing different time intervals to the follow-up appointment)

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 YES (N=581) NO (N=1,304)

Days since discharge from the hospital

Proportion still out of the hospital

Survival Function

Time to Readmission for Patients Receiving Outpatient Follow-up Within 7 Days of Discharge

(Patients with multiple chronic conditions and 40-50% expected risk of readmission) All CCNC enrolled at discharge; inpatient discharges during the period 4/1/12-3/31/13, excluding deliveries, newborns, discharges to another facility and members dually enrolled at discharge.

As patients’ clinical risk increased, they were more likely to benefit from earlier outpatient follow-up.

Opportunity Analysis for Patients Receiving 7-day Follow-up

Recommended Follow-up Period Did the patient receive follow-up within 7 days of discharge? NO YES Total Risk Strata Grouping 0 30 days 16,082 10,242 26,324 1 21 days 9,834 4,237 14,071 2 14 days 9,099 4,151 13,250 3 7 days 11,515 5,510 17,025 Total 46,530 24,140 70,670

For every patient getting a 7-day follow-up who doesn’t need it, there is a patient who would have benefitted from 7-day follow-up who did not get it.

Key Insight: Current Outpatient Visit Resources are Mis-matched

Question: Are Home Visits worth the additional cost? Answer: Yes! If targeted appropriately

What is the incremental savings benefit of a home visit, compared to less intensive TC management activities?

Incremental Savings Achieved From Home Visits by Clinical Risk Strata

Non-MCC 0-9%

• $539

10-19% $490 20-29% $395 30-39% $2,080 40-49% $3,736 50%+ $5,818

• $2,000
• $1,000

$0 $1,000 $2,000 $3,000 $4,000 $5,000 $6,000 $7,000

Difference in Total Cost of Care During 6 Month Period After Index Discharge

*Percentages reflect the relative clinical risk for patients in that strata with Multiple Chronic Conditions (MCC), based upon their expected risk of a 90-day readmission. ‘Non-MCC’ reflects the number of non-delivery/newborn discharges incurred by all other CCNC enrolled patients without MCC.

For patients with >30% readmission risk, savings far exceed the cost of the home visit

Putting it into Action

Other Flags to Inform Next Steps:

Home Visit Priority

Readmission risk >30%

Palliative Care Priority

High risk of mortality and preventable end-of-life spend

Chronic Pain Priority

Pattern of frequent narcotic fills and ED visits

Risk of Drug Therapy Problem

Risk of drug interaction, duplication, or adherence problems based on real- time medication data from multiple sources

Real-time notification of hospital admissions with care management priorities

Admission and Readmission Rate Trends

35

10.5% 10.2% NC Medicaid Beneficiaries with Multiple Chronic Conditions, 2008-2012

460 480 500 520 540 560 580 600

CY2008 CY2009 CY2010 CY2011 CY2012 CY2013 SFY2014 Inpatient Admissions per 1,000 Beneficiaries Inpatient Admissions Per 1,000 MCC Beneficiaries per Year

Inpatient Admission Trends among NC Medicaid Beneficiaries with Multiple Chronic Conditions, 2008-FY2014

This means >8,000 fewer inpatient admissions in SFY2014 compared to 2008 performance

90 95 100 105 110 115 120 125

CY2008 CY2009 CY2010 CY2011 CY2012 CY2013 SFY2014 30-day readmissions per 1,000 beneficiaries

30-day readmissions Per 1,000 MCC Beneficiaries per Year

Readmission Trends among NC Medicaid Beneficiaries with Multiple Chronic Conditions, 2008-FY2014 This means >2,200 fewer readmissions in SFY2014 compared to 2008 performance

Where are we going from here?

Impactability Scores as opposed to Risk Scores

• Risk Scores predict the likelihood of a given event. They are

designed to predict events/outcomes as part of usual care (i.e., if we didn’t intervene, what might be expected to happen). The dependent variable in the predictive models are typically events (e.g., hospital utilization) or costs.

• Impactability Scores are designed to identify members who will

benefit the most from a given intervention. The dependent variable in the predictive models are the estimated savings from care management interventions, based on rigorous, controlled real-world evaluations.

Care Management (CM) Impactability Score:

General Overview of Methodology

Time Cost

]

Impactability scores predict how much change can be expected when intervened.

Care Manager Intervenes

Typical “risk scores predict where a person is expected to be in the future.

Impactability vs. Risk: Impact on ROI

Received Care Management Intervention? N Costs PMPM PRE Costs PMPM POST Difference Net Effect

Highest Impactable Patients

(Patients with a CM Impactability Score of 300+)

NO 1219 $2,276 $2,294 $18 YES 2865 $2,930 $2,163

• $767
• $785

Highest Risk Patients

(Patients with a 40% risk of an inpatient admission in the next 12 months)

NO 1369 $2,627 $2,516

• $111

YES 3019 $3,387 $2,786

• $601
• $489

Targeting the most impactable patients generates almost twice as much savings as targeting the highest risk patients.

Where are we going from here?

Populati

• n

Needs System Resourc es Populati

• n

Needs System Resourc es Large Population Who to Touch How to Touch Them

Intensive Transitional Care for Highest Risk Discharges

Target Interventions to maximize ROI Efficient use of limited care team resources

Face to Face Outreach for High Risk/ Telephonic for low Targeted QI Efforts to change Practice Patterns

Logistics: Using Data to Actively Guide Intervention Strategy

NC Medicaid oncology costs trend 13% increase in per patient cost since 2010

NC Medicaid vs Commercial oncology population comparison ED utilization

North Carolina Medicaid Claims July 1, 2012 – June 30, 2013. Incudes 7,244 Medicaid recipients with ICD9 code of cancer (140 – 239) who also received chemotherapy and/or radiation treatment.

NC Medicaid vs Commercial oncology population comparison Inpatient utilization

North Carolina Medicaid Claims July 1, 2012 – June 30, 2013. Incudes 7,244 Medicaid recipients with ICD9 code of cancer (140 – 239) who also received chemotherapy and/or radiation treatment.

Average providers per beneficiary (excluding radiology) 81% of beneficiaries visited 6 or more providers each year

Treatment side effect profile may be a useful component of risk segmentation

Treatment therapy risk category Chemotherapy combinations with high probability of severe side effects Risk Category High Probability

• f Emesis

High Probability

• f Myelo-

suppression High Probability

• f Both

Moderate Probability of Emesis None % Of Patients 7% 1% 11% 19% 62% % of Patients in Risk Category with ED Visit 73% 76% 63% 68% 63% % of Patients in Risk Category with Inpatient Visit 74% 67% 65% 69% 60%

Treatment therapy risk category Average allowed per beneficiary receiving severe side-effect chemotherapy

Wake/Johnston Pilot Program

Patient Case Study

55 year-old male with stage IV carcinoid of the colon, high blood pressure and diabetes

• Issue: Patient struggling with ongoing nausea, pain and vomiting due to oral

chemotherapy; ED and inpatient visits monthly due to intense side effects. (Patient would call 911 when the nausea began). Extensive difficulty with appointment adherence due to transportation constraints.

• Intervention: Direct patient engagement by oncology care manager; coordination with

provider.

• Education around side effect management (utilize antiemetic on 6-8 hour schedule).
• Referral to nutritionist and palliative care specialist
• Taxi voucher
• Connection to community resources for a new mattress and blood pressure machine
• Outcome: Greatly improved side effect management. To date, the patient has not been

admitted to the hospital, emergency department, or missed any appointments since enrollment in the program.

Wake/Johnston Pilot Program Patient case study

55 year-old male with stage IV carcinoid of the colon, high blood pressure and diabetes

• Issue: Patient struggling with ongoing nausea, pain and vomiting due to oral

chemotherapy; ED and inpatient visits monthly due to intense side effects. (Patient would call 911 when the nausea began). Extensive difficulty with appointment adherence due to transportation constraints.

• Intervention: Direct patient engagement by oncology care manager; coordination with

provider.

• Education around side effect management (utilize antiemetic on 6-8 hour schedule).
• Referral to nutritionist and palliative care specialist
• Taxi voucher
• Connection to community resources for a new mattress and blood pressure machine
• Outcome: Greatly improved side effect management. To date, the patient has not been

admitted to the hospital, emergency department, or missed any appointments since enrollment in the program.

49

Source: Biologics demonstration oncology care management program results from July 1, 2014 through November 14, 2014

Contact for more information: Annette DuBard, MD, MPH SVP for Informatics and Evaluation adubard@n3cn.org www.communitycarenc.org

Acknowledgements

Carlos Jackson, PhD, CCNC Director of Program Evaluation CCNC Clinical Program Leadership: Jennifer Cockerham, RN; Tom Wroth, MD; Troy Trygstad PharmD 800+ nurse care managers, clinical pharmacists, and other care team members out there doing the good work!

50