Uploaded by Mohab Fawzy

07angusmccann-161005191448

advertisement
Population Health
Management
- one person at a time
Angus McCann
IBM Global Healthcare Team
angus.mccann@uk.ibm.com
@eHealthAngus
Spot the ‘patient’…
3
Source: Bipartisan Policy Center,
“F” as in Fat: How Obesity Threatens
America’s Future (TFAH/RWJF, Aug.
2013)
4
Focusing on sickest does not bend the cost trend
5
One ‘patient’ at a time?
Current View
30 Patients Per Day
14 have Chronic Conditions
Unknown Health Risks
Visits Too Short for Coaching
Volume-Based/Episodic
New Population View
2500 Patient Population
900 have Chronic Conditions
1100-1250 have Mod-High Health Risk
Care enhanced through IT & data
Value-Based/Continuous
6
Key facets of population health management
• Comprehensive view of ‘health’ – physical, mental, whole person
• Early Intervention, Health Promotion and Prevention
• Wider determinants of health considered – eg Income maximisation,
legal advice, housing, education
• Addressing lifestyle behaviours
• Use of data
• Population stratification / risk prediction
• Care pathways defined and used
• Self-management
• Integration across agencies
Well
At Risk
Acute
Self-Limiting
Chronic
Illness
Complex Care
Comprehensive view of ‘health’ / wider determinants
8
Population definition / stratification
1. Diabetes stands out
with a low overall
compliance rate
of 38%
2. Significant percent
of diabetic patients
with A1c rates >9
9
Bring people into the system (appropriately)
11
Automate
Team based care – integrated across agencies
Patient engagement / self management
Identify variance in care by practice
Identify variances by
practice to target
improvement
strategies
16
Identify variance by clinician
17
New delivery models require integrated data…
19
20
Medical text analytics
Diseases
Medications
Symptoms
Modifiers
21
What really causes heart failure readmissions at Seton
The Data We Thought
Would Be Useful … Wasn’t
•
•
•
113 candidate predictors
from structured and
unstructured data sources
Structured data was less
reliable then unstructured
data – increased the
reliance on unstructured
data
New Unexpected
Indicators Emerged …
Highly Predictive Model
Predictor Analysis
% Encounters
Structured Data
% Encounters
Unstructured
Data
Ejection Fraction
(LVEF)
2%
74%
Smoking Indicator
35%
(65% Accurate)
81%
(95% Accurate)
Living Arrangements
<1%
73%
(100% Accurate)
Drug and Alcohol
Abuse
16%
81%
Assisted Living
0%
13%
22
Helping the population be healthy
Data Driven
Every Person
has a Plan
Team Based
Automation to Manage
a Population Down to
the Individual
23
BCS Health Scotland Conference
• Strathclyde University
• 11/12 Oct
24
Population Health
Management
- one person at a time
Angus McCann
IBM Global Healthcare Team
angus.mccann@uk.ibm.com
@eHealthAngus
Download