1

Commissioning to reduce health inequalities:Supporting analysis

Ester RomeriDilwyn SheersHealth Inequalities National Support TeamDepartment of Health

2

What needs to be done to improve health and reduce inequalities?

• Set and quantify ambitions based on peers’ best performance (e.g. Number of deaths to prevent)

Where does the problem lie? • Identify ‘excess’ deaths by cause and age (where gains can be made, e.g. CHD

mortality in the over 75s?)

How to understand health needs in local population?• Identify variation in primary care performance and impact on mortality (e.g. Number

of patients with pre-existing CHD but untreated)

Way forward ‘modelling the numbers’• Quantify the contribution of systematic and selected evidence based interventions

needed to achieve the challenge (e.g. Oldham’s ambition)

Outline

3

Setting Ambitions: Best in Peer Group (Males)

Male All Age All Cause Mortality (AAACM) Rates by GP Consortia, 2007-09

*Peer group = Former Spearhead PCTs in ‘Centres with Industry’ ONS area classification

500

550

600

650

700

750

800

850

900

950

GP Consortia in Peer Group*

Mo

rta

lity

Ra

te p

er

10

0,0

00

SouthBirmingham

Oldham

4

Estimating the scale of the challenge (Males)

Oldham Male AAACM rate 2001-2009, forecast and trajectory to 2013-15 ambition

500

550

600

650

700

750

800

850

900

950

1000

01-03 02-04 03-05 04-06 05-07 06-08 07-09 08-10 09-11 10-12 11-13 12-14 13-15

3 year average

Mal

e A

AA

CM

per

100

,000

Target

Actual

Forecast

Equivalent to 417 (13%) fewer male deaths in 2013-15

270 fewer deaths in 2013-15 expected based on current trend

5

Setting Ambitions: Best in Peer Group (Females)

Female All Age All Cause Mortality (AAACM) Rates by GP Consortia, 2007-09

*Peer group = Former Spearhead PCTs in ‘Centres with Industry’ ONS area classification

300

350

400

450

500

550

600

650

GP Consortia in Peer Group*

Mo

rta

lity

Ra

te p

er

10

0,0

00 Walsall

Oldham

6

Estimating the scale of the challenge (Females)

Oldham Female AAACM rate 2001-2009, forecast and trajectory to 2013-15 ambition

300

350

400

450

500

550

600

650

700

01-03 02-04 03-05 04-06 05-07 06-08 07-09 08-10 09-11 10-12 11-13 12-14 13-15

3 year average

Fe

ma

le A

AA

CM

pe

r 1

00

,00

0

Target

Actual

Forecast

Equivalent to 636 (17%) fewer female

deaths in 2013-15

297 fewer deaths in 2013-15 expected based on current trend

7

Estimating the scale of the challenge : Summary

2007-09AAACM

(rate)

2013-15 ambition

(rate)

2007-09 deaths

(number)

2013-15 ambition(number)

Required reduction(number)

Expected reduction (number)

Additional reductionRequired

Males 833.6 721.2 3100 2683 417 270 147

Females 597.7 497.8 3410 2774 636 297 339

Persons - - 6510 5457 1053 567 486

Reductions in mortality numbers necessary to meet 2013-15 targets

8

Identifying ‘excess’ mortality by age group

Number of excess deaths by age group in Oldham compared to England average, 2006-08

0

50

100

150

200

250

<1yr

1-4 5-9 10-14

15 -19

20 -24

25 -29

30 -34

35 -39

40 -44

45 -49

50 -54

55 -59

60 -64

65 -69

70 -74

75 -79

80 -84

85+

Age group

Exc

ess

deat

hs 2

006-

08(p

erso

ns)

Source: Derived from London Public Health Observatory Health Inequalities Intervention Tool data

9

Identifying ‘excess’ mortality by cause

Number of excess deaths in Oldham by cause, gender and broad age groupcompared to England average, 2006-08

0

50

100

150

200

250

CH

D

Str

oke

All

canc

ers

CO

PD

Pne

umon

ia

Live

r di

seas

e

Dia

bete

s

Oth

er c

ause

s

CH

D

Str

oke

All

canc

ers

CO

PD

Pne

umon

ia

Live

r di

seas

e

Dia

bete

s

Oth

er c

ause

s

Exc

ess

num

ber

of d

eath

s 20

06-0

8

75+ years

<75 years

Source: Derived from NCHOD standardised mortality ratios (SMR) and mortality numbers by age and cause.Excess mortality = ‘observed’ minus ‘expected’ deaths

Males Females

10

Identifying Primary Care performance to outcomes

QOF registered prevalence and CHD Mortality(<75) in Oldham (MSOAs)

11

Index of Multiple Deprivation (2007)

Most deprived areas in Oldham (darker areas)

12

Identifying the untreated patients (GP practice)

CHD: Expected vs QOF Registered Prevalence (Percentage)

A quarter of patients with a history of CHD are estimated undiagnosed (untreated)

CHD 8 - % patients whose last measured cholesterol <= 5mmol/l (measured in last 15 months)

0%

20%

40%

60%

80%

100%

Practice code

Target Met Target Missed Exception Coded Undiagnosed based on Expected Prevalence

13

Identifying the untreated patients (GP practice)

CHD: Expected vs QOF Registered Prevalence (Number)

CHD 8- Number of patients whose last measured cholesterol <= 5 mmol/l (measured in last 15 months)

0

100

200

300

400

500

600

700

800

Practice code

Target Met Target Miissed Exception code Undiagnosed based on Expected Prevalence

14

Identifying the untreated patients (GP practice)

Hypertension: Expected vs QOF Registered Prevalence (Percentage)

More than half patients with Hypertension are estimated undiagnosed (untreated)

15

Identifying the untreated patients (GP practice)

Hypertension: Expected vs QOF Registered Prevalence (Number)

BP5- Number of patients whose BP <= 150/90 (measured in last 9 months)

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

Practice code

Target Met Target Miissed Exception code Undiagnosed based on Expected Prevalence

16

Identifying the untreated patients (GP practice)

COPD: Expected vs QOF Registered Prevalence (Percentage)

On average half of patients with COPD are estimated undiagnosed (untreated)

17

Identifying the untreated patients (GP practice)

COPD: Expected vs QOF Registered Prevalence (Number)

COPD 10- Number of patients with COPD with FeV1 recorded (in last 15 months)

0

100

200

300

400

500

600

Practice code

Target Met Target Miissed Exception code Undiagnosed based on Expected Prevalence

18

Potential impact of evidence-based interventions on reducing mortality numbers

InterventionDeaths

postponedTreatment population

NNT to postpone one death

Secondary prevention following CVD eventFour treatments (beta blocker, aspirin, ACE inhibitor, statin)

Currently untreated: CVD deaths averted 31 4,335 136

Currently partially treated: CVD deaths averted 61 15,335 253

Additional treatment for hypertensives

Additional hypertensive therapy 62Statin treatment for hypertensives with high CVD risk 27

Warfarin for atrial fibrillation >65 years

Stroke deaths averted 17 609 35

Improving diabetes management

Reducing blood sugars (HbA1c) over 7.5 by one unit 13 3,092 232

Treating CVD risk among COPD patients

Statins for eligible mild & moderate COPD patients 45 1,833 40

Total 258 - -

38,053 425

NNT = Number Needed to Treat to postpone one death

19

Aim: Deliver a short-term plan to place the PCT on a target AAACM trajectory for males

The Plan: Focus on six evidence based interventions:1. Full implementation of evidence based treatments for patients with CVD who are

currently untreated2. Full implementation of evidence based treatments for patients with CVD who are

currently partially treated3. Finding and treating undiagnosed hypertensives4. Moving patients on Atrial Fibrillation registers from aspirin to warfarin 5. Statins to address CVD risk among COPD patients. 6. Reducing blood sugar in diabetic patients

Expected Outcomes• Improved management of primary and secondary prevention of CVD• Postponement of up to 257 deaths from CVD if the interventions are fully

implemented, although this would depend on pace of incremental delivery • Achieving 38% of full implementation of all interventions would deliver the AAACM

target although again this depends on pace of incremental delivery

Using the model: a worked example (1)

Source: Rochdale PCT AAACM Recovery Plan, Nov 2010

20

Using the model: a worked example (2)

• Intervention:Moving patients over 65 years on AF registers from aspirin to warfarin

• Evidence Base: Evidence shows Warfarin reduces stroke among AF patients by 64%, compared with 22% for aspirin

• Treatment population: Local audit identified 887 (45%) of AF patients eligible (46% already on Warfarin, 9% contraindicated). Compared with estimate of 587 in the model

• Outcomes: Estimated 24 stroke deaths prevented. (By applying model’s NNT ratio of 37 to 887 patients). Compared with estimate of 16 deaths prevented in the model

• Costs: Recurrent costs of £348,000 (Additional cost of prescribing plus additional referrals to the anticoagulation Service)

21

Using the model: a worked example (3)

• Intervention:Statins to address CVD risk among patients with mild or moderate COPD

• Evidence Base: Observational studies show CVD is the leading cause of mortality among patients with mild and moderate COPD, yet CVD risk is often untreated among this patient group

• Treatment population: Aim to increase coverage from 26% to 66% of all COPD patients. (Current treatment coverage of 26% estimated from local audit of COPD registers plus estimate of undiagnosed COPD from APHO prevalence estimate.) Equates to an additional 2,450 COPD patients on a statin

• Outcomes: Estimated 55 deaths prevented (consistent with model which shows effect of additional 40% COPD patients on a statin)

• Costs: Recurrent costs of £95,000 (includes finding additional patients)

Prescribing costs versus outcomes

•Fig 7a - Prescribing Costs per Diabetic Pt (Apr 06-Mar 07) v Percentage of diabetic patients whose HbA1C

•has been 7.4 or less in the last 15 months (Apr 06-Mar 07)

•16

•1

•25

•15

•26•22

•8

•21

•27

•11

•20

•14•30•6

•4

•17

•19

•5

•10

•33

•7

•3

•2

•18

•32

•28

•29

•24

•31

•13

•12

•9

•23

•30

•35

•40

•45

•50

•55

•60

•65

•70

•75

•80

•£150 •£200 •£250 •£300 •£350 •£400 •£450

•NIC (£) per diabetic patient

• % t

arg

et m

et

•Higher % pts at target - Low prescribing

•Lower % pts at target - Low prescribing

•Higher % pts at target - High prescribing

•Lower % pts at target - High prescribing

Prescribing costs versus outcomes (see HINST ‘How-to’ Guide @ www.dh.gov.uk/HINST )