The document discusses predictive case modelling in social care and health. It describes how predictive models can identify high-risk patients using their medical history to help avoid costly hospital admissions. The document outlines how predictive models are developed using years of patient data and validated on separate data. It also discusses how predictive risk scores can be used to target intensive case management programs at those most likely to benefit.
3. Case Finding
• NHS predictive models
• Models for social care
Evaluation
Remuneration
4. Why Predictive Modelling?
• BMJ in paper* in 2002 showed Kaiser Permanente in
California seemed to provide higher quality healthcare
than the NHS at a lower cost
*Getting more for their dollar: a comparison of the NHS with California's Kaiser Permanente BMJ 2002;324:135-143
• Kaiser identify high risk people in their population and
manage them intensively to avoid admissions
• Inaccurate Approaches:
– Clinician referrals
– Threshold approach (e.g. all patients aged >65 with 2+
admissions)
5. Frequently-admitted patients
50
Average number of emergency bed days
45
40
35
30
25
20
15
10
5
0
-5 -4 -3 -2 -1 Intense +1 +2 +3 +4
year
6. Regression to the mean
Average number of emergency bed days
50
45
40
35
30
25
20
15
10
5
0
Intense
-5 -4 -3 -2 -1 year +1 +2 +3 +4
7. Emerging Risk
50
45
Average number of emergency bed days
40
35
30
25
20
15
10
5
0
-5 -4 -3 -2 -1 +1 +2 +3 +4
Intense
year
8. Kaiser Pyramid
Small numbers of
people at very high
risk
The Pyramid
represents the
distribution of
risk across the
population
Large numbers
of people at
low risk
[Size of shape is proportional to number of patients]
9. Patterns in routine data
Inpatient
Inpatient A&E data
A&E data GP Practice
GP Practice
data
data data
data
Outpatient
Outpatient
data
data PARR
Combined
Model
Census
Census
data
data
10. Scotland Wales
• SPARRA • PRISM model
• SPARRA-MD • Welsh Predictive Risk
Service
11. Name, Address, DOB 131178 J7KA42
Encrypted,
linked data
Inpatient
Inpatient
Name, Address, DOB 131178 J7KA42
Outpatient
Outpatient
J7KA42
A&E
A&E
Name, Address, DOB 131178 J7KA42
GP
GP
Name, Address, DOB 131178 J7KA42
J7KA42 76.4
131178 76.4
Decrypted data
with risk score
attached
12. 10 Million Patient-Years
10 Million Patient-Years
of Data
of Data
Development Validation
5 Million Patient-Years
5 Million Patient-Years 5 Million Patient-Years
5 Million Patient-Years
of Data
of Data of Data
of Data
13. Inpatient
Inpatient
Outpatient
Outpatient Development
A&E
A&E
GP
Sample
GP
J7KA42 J7KA42 J7KA42
YH8TPP YH8TPP YH8TPP
G8HE9F G8HE9F G8HE9F
3LWZ67 3LWZ67 3LWZ67
2NX632 2NX632 2NX632
LG5DSD LG5DSD LG5DSD
3V9D54R 3V9D54R 3V9D54R
Year 1 Year 2 Year 3
14. Inpatient
Inpatient
Outpatient
Outpatient Development
A&E
A&E
GP
Sample
GP
J7KA42 J7KA42 J7KA42
YH8TPP YH8TPP YH8TPP
G8HE9F G8HE9F G8HE9F
3LWZ67 3LWZ67 3LWZ67
2NX632 2NX632 2NX632
LG5DSD LG5DSD LG5DSD
3V9D54R 3V9D54R 3V9D54R
Year 1 Year 2 Year 3
15. Inpatient
Inpatient
Outpatient
Development
Outpatient
A&E
A&E Sample
GP
GP
J7KA42 J7KA42 J7KA42
YH8TPP YH8TPP YH8TPP
G8HE9F G8HE9F G8HE9F
3LWZ67 3LWZ67 3LWZ67
2NX632 2NX632 2NX632
LG5DSD LG5DSD LG5DSD
3V9D54R 3V9D54R 3V9D54R
Year 1 Year 2 Year 3
16. Inpatient
Inpatient
Outpatient
Validation
Outpatient
A&E
A&E Sample True
False
Positive
Negative
GP
GP
A89KP5 A89KP5 A89KP5
833TY6 833TY6 833TY6
I9QA44 I9QA44 I9QA44
85H3D 85H3D 85H3D
6445JX 6445JX 6445JX
233UMB 233UMB 233UMB
False
Positive
RF02UH RF02UH RF02UH
True
Negative
Year 1 Year 2 Year 3
17. Inpatient
Inpatient
Outpatient
Outpatient
Using the Model
A&E
A&E
GP
GP
A89KP5 A89KP5
833TY6 833TY6
I9QA44 I9QA44
85H3D 85H3D
6445JX 6445JX
233UMB 233UMB
RF02UH RF02UH
Last Year This Year Next Year
18. Distribution of Future Utilisation
£4,500
Actual Average cost per patient
£4,000
£3,500
£3,000
£2,500
£2,000
£1,500
£1,000
£500
£0
0 10 20 30 40 50 60 70 80 90
Predicted Risk (centile rank)
23. How the output of predictive
models are used
• Case Management
• Intensive Disease Management
• Less Intensive Disease
Management
• Wellness Programmes
Potential Misuses
Dumping
Cream-skimming
Skimping
24. Health Needs Social Care Needs
• Diagnoses • Client group
• Prescriptions • Disabilities
• Record of Health • Record of care
Contacts history
PAST Predictive
Model
FUTURE
Health Service Use Social Care Use
• GP visits • Residential care
• Community care • Intensive home
• Hospital care care
• Direct payments
32. Person-Based Resource Allocation
• Historically, GP practice budgets set on area-
based variables
• New approach is person-based
• Exclude certain variables to avoid perverse
incentives
– Procedures
– Disease severity
35. Trend
Model
Cost
predicts:
Details Model predicts
which patients
will become
high-cost over
next 6 or 12
months
Examples Low-cost
patient this
year will
become high-
cost next year
36. Trend
Model
Cost Event
predicts:
Details Model predicts Model predicts
which patients which patients
will become will have an
high-cost over event that can
next 6 or 12 be avoided
months
Examples Low-cost Patient will be
patient this hospitalized
year will
become high- Patient will
cost next year have diabetic
ketoacidosis
37. Trend
Model
Cost Event Actionability
predicts:
Details Model predicts Model predicts Model predicts
which patients which patients which patients
will become will have an have features
high-cost over event that can that can readily
next 6 or 12 be avoided be changed
months
Examples Low-cost Patient will be Patient has
patient this hospitalized angina but is
year will not taking
become high- Patient will aspirin
cost next year have diabetic Patient does
ketoacidosis not have
pancreatic
cancer
(Ambulatory
Care Sensitive)
38. Trend
Model
Cost Event Actionability Readiness to
predicts: engage
Details Model predicts Model predicts Model predicts Model predicts
which patients which patients which patients which patients
will become will have an have features are most likely
high-cost over event that can that can readily to engage in
next 6 or 12 be avoided be changed upstream care
months
Examples Low-cost Patient will be Patient has Patient does
patient this hospitalized angina but is not abuse
year will not taking alcohol
become high- Patient will aspirin
cost next year have diabetic Patient does Patient has no
ketoacidosis not have mental illness
pancreatic
cancer
(Ambulatory Patient
Care Sensitive) previously
compliant
39. Trend
Model
Cost Event Actionability Readiness to Receptivity
predicts: engage
Details Model predicts Model predicts Model predicts Model predicts Model predicts
which patients which patients which patients which patients what mode and
will become will have an have features are most likely form of
high-cost over event that can that can readily to engage in intervention
next 6 or 12 be avoided be changed upstream care will be most
months successful for
each patient
Examples Low-cost Patient will be Patient has Patient does Patient prefers
patient this hospitalized angina but is not abuse email rather
year will not taking alcohol than telephone
become high- Patient will aspirin
cost next year have diabetic Patient does Patient has no Patient prefers
ketoacidosis not have mental illness male voice
pancreatic rather than
cancer female
(Ambulatory Patient
Care Sensitive) previously
compliant Readiness to
change