Studying mortality trends: The IMPACT CHD Policy Model

1
Studying mortality trendsThe IMPACT CHD Policy
Model
Charity No 1110067

• Prof Simon Capewell
• Chair of Clinical Epidemiology
• DIVISION OF PUBLIC HEALTH
• LIVERPOOL UNIVERSITY UK
• 14th January 2008
• Particular thanks to Julia Critchley, Kath
  Bennett
• Martin OFlaherty, Robin Ireland, Ann Capewell

2
International mortality trends 1968-2003 men,
coronary heart disease CHD
SourceBHF Heartstats (WHO statistics Men aged
35 - 74, Standardised)
3
International mortality trends 1968-2003 men,
coronary heart disease CHD
Why have CHD mortality rates halved?
SourceBHF Heartstats (WHO statistics Men aged
35 - 74, Standardised)
4
Explaining the fall in coronary heart disease
deaths in England Wales 1981-2000
?
Why did CHD mortality halve in spite of
population ageing??
68,230 fewer deaths in 2000 ?
2000
1981
  Unal, Critchley Capewell Circulation 2004
109(9) 1101
5
Explaining the fall in coronary heart disease
deaths in England Wales 1981-2000
Risk Factors worse 13 Risk Factors better
-71    Treatments -42
68,230 fewer deaths in 2000 ?
2000
1981
  Unal, Critchley Capewell Circulation 2004
109(9) 1101
6
Explaining the fall in coronary heart disease
deaths in England Wales 1981-2000
Risk Factors worse 13 Obesity (increase)
3.5 Diabetes (increase) 4.8 Physical
activity (less) 4.4 Risk Factors better
-71 Smoking -41 Cholesterol
-9 Population BP fall -9 Deprivation
-3 Other factors -8   Treatments
-42 AMI treatments -8 Secondary
prevention -11 Heart failure
-12 AnginaCABG PTCA -4 Angina Aspirin
etc -5 Hypertension therapies -3
68,230 fewer deaths in 2000 ?
2000
1981
  Unal, Critchley Capewell Circulation 2004
109(9) 1101
7
?Risk Factors obviously powerful but was it
tablets or lifestyles??
8
CHD prevention in England Wales 1981-2000
Population v. High Risk StrategiesDeaths
prevented or postponed (Sensitivity analysis )
Population secular trends
C h o l e s t e r o l
Population diet
Blood Pressure
Diet in CHD patients
Treating High Risk
Statins CHD patients
Secular trends CHD patients
High Risk Statins
Unal et al BMJ 2005 331 614
9
CHD prevention in England Wales 1981-2000
Population v. High Risk StrategiesDeaths
prevented or postponed (Sensitivity analysis )
Population secular trends
C h o l e s t e r o l
Population diet
Blood Pressure
Diet in CHD patients
Treating High Risk
Statins CHD patients
Secular trends CHD patients
High Risk Statins
Unal et al BMJ 2005 331 614
10
Explaining the fall in CHD deaths in USA
1980-2000 RESULTS
NEJM 2007 356 2388.
Risk Factors worse 17 Obesity (increase)
7 Diabetes (increase) 10 Risk Factors
better -65 Population BP fall -20 Smoking
-12 Cholesterol (diet) -24 Physical activity
-5  Treatments -47 AMI treatments
-10 Secondary prevention -11 Heart
failure -9 AnginaCABG PTCA
-5 Hypertension therapies -7 Statins (primary
prevention) -5   Unexplained -9
341,745 fewer deaths in 2000 ?
2000
1980
11
Comparisons with other studies CHD mortality
falls attributed to
NEJM 2007 356 2388.
12
EXPLOITING THE IMPACT MODEL

• Replication in other populations
• Populations with RISING CHD
• Calculating life-years gained
• Cost effectiveness
• WHAT IF treatment uptakes increased?
• WHAT IF risk factors reduced further?

13
WHAT IF Treatment Uptakes in England Wales
Increased? Actual Uptakes ? 50 25,805
Deaths prevented or postponed (DPPs) IF
Uptakes at least 80 20,910 additional DPPs
Capewell et al Heart 2006 92 521 Capewell,
Pell et al et al Eur Heart J 1999 20 1836
14
IF Treatment Uptakes Increased in England
Wales Actual Uptakes ? 50 25,805 Deaths
prevented/ postponed (DPPs) IF Uptakes
at least 80 20,910 additional DPPs
Capewell et al Heart 2006 92 521
15
EXPLOITING THE IMPACT MODEL

• Replication in other populations
• Populations with RISING CHD
• Calculating life-years gained
• Cost effectiveness
• WHAT IF treatment uptakes increased?
• WHAT IF risk factors reduced further?

16
Estimating the potential changes in CHD mortality
in England Wales between 2000 and 2010

• WHAT IF risk factors
• a) continue recent trends ?
• b) undergo additional reductions ?
• (as already achieved in
• Australia, USA, Sweden, Finland etc)

  Unal et al J Clin Epid 2005 58 733 Kelly
Capewell HDA 2004 www
17
Potential changes in CHD mortality in England
Wales between 2000 and 2010 IF risk factors
a) continue recent trends b) additional
reductions already achieved elsewhere
  Unal et al J Clin Epid 2005 58 733
18
The IMPACT Model Thus, to reduce CHD mortality
in UK

• Modest additional risk factor reductions already
  achieved in USA Scandinavia
• could prevent or postpone over 50,000 deaths by
  2010
• halving current CHD deaths (100,000) in UK

  Unal et al J Clin Epid 2005 58 733
19
hence Heart of Mersey CHD primary prevention
programme

• Key Targets
• Healthier food
• Smoking reduction

20
Is the party over?
21
US Trends in age-adjusted CHD mortality rates
men women aged 35 years
Ford Capewell JACC 2007 50 2128
22
US Trends in AGE-SPECIFIC CHD mortality rates
men women aged 35 years
Ford Capewell JACC 2007 50 2128
23
Trends in age-specific CHD mortality rates
England Wales 1984-2004 lines indicate 5
year moving averages
Heart July 2007 doi 10.1136/hrt.2007.118323
24
Using IMPACT to explain CHD trends and examine
future policy options Conclusions

• CHD mortality big falls in UK elsewhere
• 25-50 due to evidence-based therapies
• 50 -75 due to risk factor reductions
• (especially smoking cholesterol)

25
Using IMPACT to explain CHD trends and examine
future policy options Conclusions

• CHD mortality big falls in UK elsewhere
• 25-50 due to evidence-based therapies
• 50 -75 due to risk factor reductions
• (especially smoking cholesterol)
• small reductions in UK risk factors
• could Halve CHD deaths
• Healthy diet Tobacco control
• remain top policy priorities

26
Reserve slides
27
IMPACT Model Main Components
RISK
F FACTORS Patient Groups
TREATMENTS OUTCOMES
Cholesterol
BMI Diabetes
Smoking
Physical Activity
Blood Pressure
Age Sex
CABG/PTCA surgery
Medical
Death
Survival
Ford et al NEJM 2007 356 2388
28
Modelling UK trends to 2010
2010 predicted
  Unal, Critchley Capewell Circulation 2004
109 (9) 1101 J Clin Epid 2005 58 733 Heart
2006 92 521
29
Modelling UK trends to 2010 predicted IF
additional treatmentsper 100,000
If more treatments for more patients ? 21,000
fewer deaths
  Unal, Critchley Capewell Circulation 2004
109 (9) 1101 Heart 2006 92 521 J Clin Epid
2005 58 733
30
Modelling UK trends to 2010 predicted IF
additional treatmentsper 100,000 IF modest risk
factor reductions
Additional risk factor reductions ? 50,000 fewer
deaths
If more treatments for more patients ? 21,000
fewer deaths
  Unal, Critchley Capewell Circulation 2004
109 (9) 1101 Heart 2006 92 521 J Clin Epid
2005 58 733
31
The Natural History of CHD
Natural Course of CHD
? Atheroma
Atheroma Thrombosis
Hanlon, Capewell et al 1997
32
CHD Prevention options
Natural Course of CHD
Hanlon, Capewell et al 1997
33
US Trends in AGE-SPECIFIC CHD mortality rates
men women aged 35 years
Ford Capewell JACC 2007 50 2128
34
IMPACT2 CVD Policy Model
Biological Risk Factors
Population Policies Behaviours
Combined CVD Risk
CVD Patient Groups
OUTPUTS
35
Population Policies Behaviours
Biological Risk Factors
Combined CVD Risk
CVD Patient Groups
OUTPUTS
SUDS
NON-SUDS
Chronic Angina
Unstable Angina
CHD Death
Combined CVD Risk
First MI
Early Heart Failure
From any State
Severe Heart Failure
Recurrent MI
Non-CHD Death
MI survivors
Stroke Other CVD
Populations UKgtEWgtRegionsgtPCTs
Outputs Population-based incidence, prevalence
Deaths prevented Life-Years Life expectancy
Costs Cost-effectiveness ratios
36
Population Policies Behaviours
Biological Risk Factors
Combined CVD Risk
CVD Patient Groups
OUTPUTS
Diabetes or IGT
SUDS
NON-SUDS
Physical Activity
Unstable Angina
Chronic Angina
CHD Death
Combined CVD Risk
Obesity (BMI)
Diet
Cholesterol LDL ( HDL)
Early Heart Failure
Acute MI
From any State
Smoking
Blood Pressure
Severe Heart Failure
Recurrent MI
Non-CHD Death
MI survivors
Deprivation
Additional CVD Risk Factors
Stroke Other CVD
Populations UKgtEWgtRegionsgtPCTs
Outputs Population-based incidence, prevalence
Deaths prevented Life-Years Life expectancy
Costs Cost-effectiveness ratios
37
IMPACT2 structure
38
Additional risk factor reductions achieved
elsewhere
UK 2000 level TARGET in
2010 SMOKING 26 17 (USA 2002)
CHOLESTEROL 5.6 5.2 (mmol/l)
Gothenberg (Sweden), Stanford (USA)
Perth (Australia) BLOOD PRESSURE (Diastolic BP
mmHg) 77 73 (4 mmHg fall) New Zealand
(4.4 ) Finland (5.2 ) France (6.0
) OBESITY USA 15 prevalence reduction by 2010
(??) PHYSICAL ACTIVITY increase prevalence by
5 (?) best DIABETES No successful community
reductions Assume 5 decrease in prevalence
(??)
  Unal et al J Clin Epid 2005 58 733