Predictive case modelling in social care and health www.nuffieldtrust.org.uk Geraint Lewis FRCP FFPH
t : 020 7631 8450 e : info@nuffieldtrust.org.uk www.nuffieldtrust.org.uk The Nuffield Trust
<ul><li>Case Finding </li></ul><ul><ul><li>NHS predictive models </li></ul></ul><ul><ul><li>Models for social care </li></...
Why Predictive Modelling? <ul><ul><li>BMJ in paper* in 2002 showed  Kaiser   Permanente  in California seemed to provide  ...
Frequently-admitted patients 0 5 10 15 20 25 30 35 40 45 50 - 5 - 4 - 3 - 2 - 1 Intense year + 1 + 2 + 3 + 4 Average numbe...
Regression to the mean 0 5 10 15 20 25 30 35 40 45 50 Average number of emergency bed days - 5 - 4 - 3 - 2 - 1 Intense yea...
Emerging Risk 0 5 10 15 20 25 30 35 40 45 50 Average number of emergency bed days - 5 - 4 - 3 - 2 - 1 Intense year + 1 + 2...
Kaiser Pyramid The Pyramid represents the distribution of risk across the population Small numbers of people at very high ...
Inpatient data A&E data GP Practice data Outpatient data PARR Patterns in  routine  data Combined Model Census data
Scotland <ul><ul><li>SPARRA </li></ul></ul><ul><ul><li>SPARRA-MD </li></ul></ul>Wales <ul><ul><li>PRISM model </li></ul></...
J7KA42 J7KA42 J7KA42 J7KA42 J7KA42 J7KA42 76.4 131178 76.4 Encrypted, linked data Decrypted data with risk score attached ...
10 Million Patient-Years  of Data 5 Million Patient-Years  of Data 5 Million Patient-Years  of Data Development Validation
J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3...
J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3...
J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3...
A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D...
A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH Using the Model Last Year ...
Distribution of Future Utilisation £0 £500 £1,000 £1,500 £2,000 £2,500 £3,000 £3,500 £4,000 £4,500 0 10 20 30 40 50 60 70 ...
NHS Combined Model
Clinical   Profiles
Tackling the  Inverse Care Law
Developing Business Cases
How the output of predictive models are used <ul><ul><li>Case Management </li></ul></ul><ul><ul><li>Intensive Disease Mana...
 
 
 
Evaluation of Preventive Care 5
Start of intervention Overcoming regression to the mean using a control group (1)
Overcoming regression to the mean using a control group (2) Start of intervention
Overcoming regression to the mean using a control group (3) Start of intervention
Start of intervention Overcoming regression to the mean using a control group (4)
Person-Based Resource Allocation <ul><ul><li>Historically, GP practice budgets set on area-based variables </li></ul></ul>...
[email_address] t:  020 7631 8450 e:  [email_address]
Trend Model predicts: Details Examples
Trend Model predicts: Cost Details Model predicts which patients will  become  high-cost over next 6 or 12 months Examples...
Trend Model predicts: Cost Event Details Model predicts which patients will  become  high-cost over next 6 or 12 months Mo...
Trend Model predicts: Cost Event Actionability Details Model predicts which patients will  become  high-cost over next 6 o...
Trend Model predicts: Cost Event Actionability Readiness to engage Details Model predicts which patients will  become  hig...
Trend Model predicts: Cost Event Actionability Readiness to engage Receptivity Details Model predicts which patients will ...
Upcoming SlideShare
Loading in …5
×

Geraint_London_Slides25Jan 2011

253 views

Published on

Published in: Health & Medicine
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
253
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
5
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Geraint_London_Slides25Jan 2011

  1. 1. Predictive case modelling in social care and health www.nuffieldtrust.org.uk Geraint Lewis FRCP FFPH
  2. 2. t : 020 7631 8450 e : info@nuffieldtrust.org.uk www.nuffieldtrust.org.uk The Nuffield Trust
  3. 3. <ul><li>Case Finding </li></ul><ul><ul><li>NHS predictive models </li></ul></ul><ul><ul><li>Models for social care </li></ul></ul><ul><li>Evaluation </li></ul><ul><li>Remuneration </li></ul>
  4. 4. Why Predictive Modelling? <ul><ul><li>BMJ in paper* in 2002 showed Kaiser Permanente in California seemed to provide higher quality healthcare than the NHS at a lower cost </li></ul></ul><ul><li>*Getting more for their dollar: a comparison of the NHS with California's Kaiser Permanente BMJ 2002;324:135-143 </li></ul><ul><ul><li>Kaiser identify high risk people in their population and manage them intensively to avoid admissions </li></ul></ul><ul><ul><li>Inaccurate Approaches: </li></ul></ul><ul><ul><ul><li>Clinician referrals </li></ul></ul></ul><ul><ul><ul><li>Threshold approach (e.g. all patients aged >65 with 2+ admissions) </li></ul></ul></ul>
  5. 5. Frequently-admitted patients 0 5 10 15 20 25 30 35 40 45 50 - 5 - 4 - 3 - 2 - 1 Intense year + 1 + 2 + 3 + 4 Average number of emergency bed days
  6. 6. Regression to the mean 0 5 10 15 20 25 30 35 40 45 50 Average number of emergency bed days - 5 - 4 - 3 - 2 - 1 Intense year + 1 + 2 + 3 + 4
  7. 7. Emerging Risk 0 5 10 15 20 25 30 35 40 45 50 Average number of emergency bed days - 5 - 4 - 3 - 2 - 1 Intense year + 1 + 2 + 3 + 4
  8. 8. Kaiser Pyramid The Pyramid represents the distribution of risk across the population Small numbers of people at very high risk Large numbers of people at low risk [Size of shape is proportional to number of patients]
  9. 9. Inpatient data A&E data GP Practice data Outpatient data PARR Patterns in routine data Combined Model Census data
  10. 10. Scotland <ul><ul><li>SPARRA </li></ul></ul><ul><ul><li>SPARRA-MD </li></ul></ul>Wales <ul><ul><li>PRISM model </li></ul></ul><ul><ul><li>Welsh Predictive Risk Service </li></ul></ul>
  11. 11. J7KA42 J7KA42 J7KA42 J7KA42 J7KA42 J7KA42 76.4 131178 76.4 Encrypted, linked data Decrypted data with risk score attached 131178 131178 131178 131178  Inpatient  Outpatient  A&E  GP Name, Address, DOB Name, Address, DOB Name, Address, DOB Name, Address, DOB
  12. 12. 10 Million Patient-Years of Data 5 Million Patient-Years of Data 5 Million Patient-Years of Data Development Validation
  13. 13. J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R Year 1 Year 2 Year 3 Development Sample  Inpatient  Outpatient  A&E  GP
  14. 14. J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R Development Sample Year 1 Year 2 Year 3  Inpatient  Outpatient  A&E  GP
  15. 15. J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R J7KA42 YH8TPP G8HE9F 3LWZ67 2NX632 LG5DSD 3V9D54R Development Sample Year 1 Year 2 Year 3  Inpatient  Outpatient  A&E  GP
  16. 16. A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH Validation Sample True Positive False Positive False Negative True Negative Year 1 Year 2 Year 3  Inpatient  Outpatient  A&E  GP
  17. 17. A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH A89KP5 833TY6 I9QA44 85H3D 6445JX 233UMB RF02UH Using the Model Last Year This Year Next Year  Inpatient  Outpatient  A&E  GP
  18. 18. Distribution of Future Utilisation £0 £500 £1,000 £1,500 £2,000 £2,500 £3,000 £3,500 £4,000 £4,500 0 10 20 30 40 50 60 70 80 90 Predicted Risk (centile rank) Actual Average cost per patient
  19. 19. NHS Combined Model
  20. 20. Clinical Profiles
  21. 21. Tackling the Inverse Care Law
  22. 22. Developing Business Cases
  23. 23. How the output of predictive models are used <ul><ul><li>Case Management </li></ul></ul><ul><ul><li>Intensive Disease Management </li></ul></ul><ul><ul><li>Less Intensive Disease Management </li></ul></ul><ul><ul><li>Wellness Programmes </li></ul></ul><ul><li>Potential Misuses </li></ul><ul><ul><ul><ul><li>Dumping </li></ul></ul></ul></ul><ul><ul><ul><ul><li>Cream-skimming </li></ul></ul></ul></ul><ul><ul><ul><ul><li>Skimping </li></ul></ul></ul></ul>
  24. 27. Evaluation of Preventive Care 5
  25. 28. Start of intervention Overcoming regression to the mean using a control group (1)
  26. 29. Overcoming regression to the mean using a control group (2) Start of intervention
  27. 30. Overcoming regression to the mean using a control group (3) Start of intervention
  28. 31. Start of intervention Overcoming regression to the mean using a control group (4)
  29. 32. Person-Based Resource Allocation <ul><ul><li>Historically, GP practice budgets set on area-based variables </li></ul></ul><ul><ul><li>New approach is person-based </li></ul></ul><ul><ul><li>Exclude certain variables to avoid perverse incentives </li></ul></ul><ul><ul><ul><li>Procedures </li></ul></ul></ul><ul><ul><ul><li>Disease severity </li></ul></ul></ul>
  30. 33. [email_address] t: 020 7631 8450 e: [email_address]
  31. 34. Trend Model predicts: Details Examples
  32. 35. Trend Model predicts: Cost 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
  33. 36. Trend Model predicts: Cost Event Details Model predicts which patients will become high-cost over next 6 or 12 months Model predicts which patients will have an event that can be avoided Examples Low-cost patient this year will become high-cost next year Patient will be hospitalized Patient will have diabetic ketoacidosis
  34. 37. Trend Model predicts: Cost Event Actionability Details Model predicts which patients will become high-cost over next 6 or 12 months Model predicts which patients will have an event that can be avoided Model predicts which patients have features that can readily be changed Examples Low-cost patient this year will become high-cost next year Patient will be hospitalized Patient will have diabetic ketoacidosis Patient has angina but is not taking aspirin Patient does not have pancreatic cancer (Ambulatory Care Sensitive)
  35. 38. Trend Model predicts: Cost Event Actionability Readiness to engage Details Model predicts which patients will become high-cost over next 6 or 12 months Model predicts which patients will have an event that can be avoided Model predicts which patients have features that can readily be changed Model predicts which patients are most likely to engage in upstream care Examples Low-cost patient this year will become high-cost next year Patient will be hospitalized Patient will have diabetic ketoacidosis Patient has angina but is not taking aspirin Patient does not have pancreatic cancer (Ambulatory Care Sensitive) Patient does not abuse alcohol Patient has no mental illness Patient previously compliant
  36. 39. Trend Model predicts: Cost Event Actionability Readiness to engage Receptivity Details Model predicts which patients will become high-cost over next 6 or 12 months Model predicts which patients will have an event that can be avoided Model predicts which patients have features that can readily be changed Model predicts which patients are most likely to engage in upstream care Model predicts what mode and form of intervention will be most successful for each patient Examples Low-cost patient this year will become high-cost next year Patient will be hospitalized Patient will have diabetic ketoacidosis Patient has angina but is not taking aspirin Patient does not have pancreatic cancer (Ambulatory Care Sensitive) Patient does not abuse alcohol Patient has no mental illness Patient previously compliant Patient prefers email rather than telephone Patient prefers male voice rather than female Readiness to change

×