Targeting Health System Improvement to High Cost Patients CAHSPR May 2011 Walter Wodchis, PhD X.Camacho, I. Dhalla, A. Guttman, E.Lin, G.Anderson 1
Older adults with complex conditions Three research studies: HSPRN: 1. Target populations for health system improvement W.Wodchis, X.Camacho, I. Dhalla, A. Guttman, E.Lin, G.Anderson 2. Caring for older adults with multi-morbidity A. Bierman, G. Mery, J. Hamilton, E. Adler, N. Nanwa, W.Wodchis ICES: 3. Seniors Atlas: Identifying risk in older adults S.Bronskill, X. Camacho, S.Gill, A.Grunier, J.Poss, W.Wodchis 2
Overview 1. Why target? What populations? 2. Epidemiology and economic burden in Ontario 3. Identifying target populations for intervention - predicting risk 3
Why Target? What Populations? Populations with high health utilization rates who move from one sector of the health care system (e.g. acute) to another (e.g. community) may represent opportunities to improve quality and reduce costs primarily by reducing adverse events and preventing acute and LTC admission. While quality of care within providers is being enhanced by performance measurement and reporting, payment incentives and quality improvement programs Care transitions between providers are fraught with lack of coordination, poor communication, safety issues related to medication management...etc,etc. 4
Why Target? What Populations? Focus is on multi-morbidity Source: The Chief Public Health Officer s Report on the State of Public Health in Canada. 2010 : Growing Older Adding Life to Years 5
Why Target? What Populations? There are many older adults with multi-morbidity in Canada: Seniors with three or more reported chronic conditions accounted for 40% of reported health care use among seniors Gaps exist in preventive and collaborative care Though most seniors have access to PHC: fewer than half (48%) reported talking at least some of the time to a health professional about their treatment goals. Source: Canadian Institute for Health Information: Seniors and the Health Care System: What Is the Impact of Multiple Chronic Conditions? July 2001. Based on data from the Statistics Canada Canadian Survey of Experiences With Primary Health Care, 2008. Canadian Institute for Health Information 6
Epidemiology and Economic Burden in ON What we ve done: 1. Identify community-based cohort of clients aged 66+ admitted and discharged from Acute care between April 2007-March 2008 in Ontario with : 1. 2 or more ACSC conditions (Angina, Asthma, COPD, Diabetes, Grand Mal Seizure, Heart Failure, Hypertension) or any one of the following tracer chronic conditions: Stroke, Cardiac Arrhythmia, Hip Fracture, Spinal Stenosis, PVD, DVT/PE Follow for 365 days (until March 2009) 2. Link all administrative clinical databases and incorporate costs to understand system utilization and costs 3. Subset patients admitted to acute who were receiving home care prior to acute admission to identify risk groups for acute and LTC admissions after discharge. 7
Epidemiology and Economic Burden in ON Acute Diagnosis Prevalence Cardiac Arrhythmia 14,976 38.4% Stroke 8,707 22.3% ACSC (>1 diagnosis) 7,351 18.9% Hip Fracture 5,749 14.7% DVT/PE 1,887 4.8% PVD 1,634 4.2% Spinal Stenosis 1,418 3.6% Total 38,978 8
Epidemiology and Economic Burden in ON Summarize Utilization and Costs in 365 days following index acute care admission: Total Population 38,978 (0.3% population) Average Annual Cost $35,935 System Cost $1,400,689,862 (3% system cost) 9
Epidemiology and Economic Burden in ON Total health system cost 1 year following index Average cost = $35,935; Total System Cost: $1,400,689,862 Pharma cost Physician cost (3.5%): (5.2%): $1,454.29 HC cost (6.1%): $1,909.62 (82.9% users) $3,732.60 (94.3% users) (56.9% users) ED cost (0.3%) : $201.49 (55.1% users) LTC cost (7.1%): $19,700.03 (12.4% users) CCC cost (10.3%): $33,296.85 (10.7% users) Rehab cost (10.5%): $21,230.81 (17.2% users) Index hospitalization AC cost (36.1%) : $12,517.29 (100% users) Acute care cost (20.9%) : $17,961.13 (40.3% users) 10
Epidemiology and Economic Burden in ON $45,000 $40,000 $35,000 $30,000 $25,000 $20,000 $15,000 $10,000 $5,000 $- Average System Cost in 365 Days Following Acute Discharge (2008 $) Pharma MD HC LTC CCC Rehab ED Acute care 11
Epidemiology and Economic Burden in ON $450 $400 $350 $300 $250 $200 $150 $100 $50 $- Total System Cost in 365 Days Following Acute Discharge (2008 $1,000,000's) Pharma MD HC LTC CCC Rehab ED Acute care 12
Epidemiology and Economic Burden in ON ACSC Arrhythmia Hip Fracture Stroke ED VISITS 30 d 28% 25% 14% 18% 90 d 46% 41% 24% 31% 1 yr 66% 61% 45% 44% ACUTE READMITS 30 d 16% 13% 7% 6% 90 d 28% 22% 13% 13% 1 yr 53% 45% 31% 30% POST ACUTE 1 YEAR SYSTEM COST $24,202 ± 29,419 $20,670 ± 29,466 $31,907 ± 33,499 $27,021 ± 37,036 TOTAL POPULATION COST (000 s) $177,034 $307,531 $183,658 $235,138 13
Epidemiology and Economic Burden in ON Number of Different Pharmacies Used Number of different non-institutional physicians Total all non-institutional provider visits (physician, pharmacy, home care) ACSC Arrhythmia Hip Fracture Stroke 0-1 42% 41% 38% 44% 2 30% 30% 32% 30% 3+ 28% 29% 30% 26% 0 12% 12% 12% 24% 1-5 44% 43% 58% 47% 6-12 36% 38% 27% 26% 13+ 8% 7% 3% 4% 67±105 IQR [13-74] 56±87 IQR [15-57] 69±115 IQR [13-73] 49±96 IQR [5-44] 14
Epidemiology and Economic Burden in ON Individuals with complex conditions are costly: System burden is a combination of prevalence and cost. Some cohorts (ACSC, Arrhythmias) use more acute, primary care and pharmacy. Some cohorts (Stroke, Hip Fracture) use Rehabilitation and Complex Continuing Care and are at higher risk for LTC admission. 15
Target Populations for System Improvement (Among 16,605 discharged to community) Examine likelihood of readmission to acute care within 7-30 days and 7-90 days associated with: 1. Home care nursing visit (show 1 day vs 3 days) 2. Primary care visit (show <7 days vs >7 days) (controlling for host of risk factors using logistic regression - 51 covariates). 16
Target Populations for System Improvement Population Discharged to Community n = 16,605 Risk of Readmission to Inpatient Acute Care Independent Variable Home Nursing Visit within 1 day (vs 2-3 days) Primary Care Visit within 7 days New Filled Prescription 7-30 days Adj. Odds Ratio* (95% Confidence Interval) 0.72 ł (0.53, 0.98) 0.91 (0.81, 1.03) 1.07 ł (1.04, 1.10) 7-90 days Adj. Odds Ratio* (95% Confidence Interval) 0.70 ł (0.55, 0.90) 0.85 ł (0.78, 0.93) 1.04 ł (1.01, 1.06) * Adjusted for 51 measures of patient characteristics, prior medical treatment, diagnoses and geography Ł significant at the 5% level 17
Measurement that follows patients Rehab / CCC / Sub-acute Care Specialist Care Acute (ED, IP, SDS) CCAC Home Care Pharmacy LTC Primary Care Patient Flow Patient Rebound 18
Identifying target populations for intervention - predicting risk Risk for LTC MAPLe [5 levels: Low-Very High] (Method for Assigning Priority Levels) Activities of Daily Living Cognitive Performance Behaviour Wandering Decision-making decline Environment or medication mgmt Ulcers Self-reliance (Geriatric screen) Meal preparation assistance Few meals or swallowing problem Falls Risk for Acute LACE [0-18] (Length of stay, Acuity, Charlson comorbidity, Emergency Use) Acute length of stay Acuity on admission (admit via ED) Charlson comorbidity (AMI, CVA, PVD, diabetes, CHF, COPD, liver, tumor, renal, AIDS) Number of emergency visits in 6 months prior to admission 19
Identifying target populations for intervention - predicting risk Acute and LTC Admission for Homecare Clients Discharged Home from Acute 40% 35% 30% 25% 20% 15% 10% 5% 0% LACE Low LACE High MAPLe Low LACE Low LACE High MAPLe High LTC 365 admit rate 30 day readmit rate 90 day readmit rate M-Low is MAPLe Low, Mild & Moderate; L-low is LACE < 10 LTC Admission within 365 days after acute discharge 20
Summative comments Risk profiles differ for acute and LTC. Acute readmission more related to medical care. LTC admission more related to functional and social status. What is the role for medical and social care interventions? What portion of the risk is modifiable? 21
Older adults with complex conditions 22
Summative comments Many good intervention ideas How to identify service package for different clients Targeting may be key: Who is at risk for what outcome? What is the best intervention to avoid that outcome? For example: risk of acute readmission and LTC placement among home care clients who were admitted to acute 23
Ideas for innovative strategies Chad Boult: 1. Interdisciplinary primary care 2. Care/case management 3. Disease management 4. Preventative home visits 5. Comprehensive geriatric assessment, geriatric evaluation and management 6. Pharmaceutical care 7. Chronic disease management 8. Proactive rehabilitation 9. Caregiver support and education 10. Transitional care 11. Substitutive hospital at home 12. Care in nursing home 13. Prevention and management of delerium 14. Comprehensive inpatient care 24
Strategies that balance safety and patient-centeredness Goals of care: Avoidance of adverse events including stroke, falls and fractures, acute admissions and death. Patient-centered care involves patient preferences and involvement of caregivers Maintenance of independence / function Goals of care for progressively older persons may focus more on function (and less on secondary prevention?) 25