CHRONIC HEART FAILURE

National clinical guideline for diagnosis and management

in primary and secondary care

Appendix G

August 2010
Chronic heart failure update (Appendix G)

Appendix G Health Economics Evidence Tables

Diagnostics questions

J. Mant, J. A. Doust, A. K. Roalfe, and P. Barton. Systematic Review and Individual Patient Data meta-Analysis of Diagnosis of Heart Failure, with Modelling of Implications
of Different Diagnostic strategies in Primary Care. 2009.
Study details Population & Health outcomes Costs Cost effectiveness
interventions
Economic analysis: Population: Health outcomes Cost components ICER calculation (all scenarios):
Cost-effectiveness analysis Patients with incorporated: incorporated: ICERs were calculated in term of cost per
suspected chronic Sensitivity and specificity of Cost for echo: £100 per additional case found.
Study design: heart failure in BNP at specific cut-points investigation ICERs were calculated by subgroups of patients
Decision-analytic model (based primary care (Tables 20 to 24 in (according to MICE score).
on a meta-analysis) publication) Cost for BNP: £15 per 3 ICERs were calculated per subgroups (‘do
Patients were investigation nothing’ versus ‘BNP’, ‘BNP’ versus ‘echo’, and ‘do
WTP threshold calculation stratified in sub- Sensibility of echo + specialist nothing’ versus ‘echo’).
Calculation of WTP for group by MICE assessment: 100% Costs used to calculate WTP
interpretation of results from a score for misdiagnosed cases: Base case analysis:
UK perspective Survival raise as result of early • Hospitalisation cost Cost-effectiveness analysis
Intervention 1: diagnosis (additional for late Reported cost per additional case found
(1) WTP calculation not ‘Do nothing’ (no • 0.163 (3 years) diagnosis) 6-month time horizon
considering QALY gained more test post sign • 0.247 (5 years) • Drug cost (additional for WTP of £270 (not considering QALY gained)
• Considered the cost and symptom • 0.390 (10 years) early diagnosis)
associated to misdiagnosed evaluation using Results (Table 20 in publication):
cases (diagnosed 6 months MICE scoring QALY gained as result of early Cost used for WTP calculation • MICE score 0,2,3: No test
later): system) diagnosis (1) Avoided cost on treatment • MICE score 5,6,7,8: BNP then echo
o additional hospitalisation • 0.106 (3 years) for late diagnosis (6-month • MICE score 9,10,11: Echo
cost Intervention 2: • 0.247 (5 years) delay):
o drug cost avoided ‘BNP’ (post sign • 0.390 (10 years) • Beta-blocker (carvedilol) BNP threshold for referring to echo
• Calculated WTP value = £270 and symptom, £12.30 per month (for 34% • MICE score 5 – refer if BNP > 149pg/ml
per misdiagnosed case perform BNP and Probabilities within the of patients) = £25 per • MICE score 6 – refer if BNP > 117pg/ml

2
Chronic heart failure update (Appendix G)
(represents cost incurred for
a 6-month delay in diagnosis)
(2) WTP calculation considering
QALY gained (using NICE
threshold of £20k per QALY
gained)
• QALY gain * £20k (NICE
threshold) + £270 (see
above)
o £2,370 to detect a case (3-
year time horizon)
o £3,470 (5 years)
o £5,370 (10 years)
Perspective:
UK NHS
Primary care
Time horizon:
6 months (base-case analysis)
3 years (secondary analysis)
5, 10 years (sensitivity analysis)
Discounting:
No discounting for cost (there is
no cost incurred after 1 year).
QALYs after 1 year were not
discounted.
then echo analysis patient
depending upon • 6-month probability of • ACE inhibitor (lisinopril)
the result of the hospital admission if £2.41 per month (69% of
BNP test, using patient treated (0.413); if patients) = £10 per patient
decision cut-points patient untreated (0.645) • ARB (losartan) £18.09 per
for BNP) • Probability for a patient of month (16% of patients) =
having CHF according to £17 per patient
Intervention 3: MICE score (Tables 20 to
‘Echo’ (post sign 24 in publication) (2) Additional hospitalisation
and symptom, cost (for late diagnosis)
proceed straight to • 23% extra admissions
echo) within 6 months
• £1400 per admission (NHS
reference cost [18 Sept
2007])
• £320 (‘cost per patient’ –
0.23*£1400)
Sensitivity analysis
• Cost of echo (£50 - £150)
• Cost of BNP testing (£10 -
£20)
Currency & cost year:
GBP, year not specified
(probably 2007)
• MICE score 7 – refer if BNP > 92pg/ml
• MICE score 8 – refer if BNP > 72pg/ml
Secondary analysis
Cost-effectiveness analysis
Reported cost per additional case found
3-year time horizon
WTP of £2,370 (considering QALY gained)
Results (Table 21 in publication):
• MICE score 0: BNP then echo (refer if BNP >
38pg/ml
• MICE score 5,6,7,8, 9,10,11: Echo
Sensitivity analysis
1- Base case analysis
a) Varying cost of investigation (‘extreme cases’ -
low cost echo and expensive BNP / high cost echo
and cheap BNP)
i) Cost of echo £50; cost of BNP £20
Results (Table 22 in publication):
• MICE score 0,2: No test
• MICE score 3,5,6,7,8, 9,10,11: Echo
ii) Cost of echo £150; cost of BNP £10
Results (Table 23 in publication):
• MICE score 0,2,3: No test
• MICE score 5,6,7,8, 9,10,11: BNP then echo
• MICE score 13,14,16: Echo
BNP thresholds for referring to echo
• MICE score 5 – refer if BNP > 313pg/ml
• MICE score 6 – refer if BNP > 247pg/ml
3
Chronic heart failure update (Appendix G)

• MICE score 7 – refer if BNP > 194pg/ml

• MICE score 8 – refer if BNP > 153pg/ml
• MICE score 9 – refer if BNP > 120pg/ml
• MICE score 10 – refer if BNP > 95pg/ml
• MICE score 11 – refer if BNP > 75pg/ml

2- Secondary analysis
a) Cost of echo £150; cost of BNP £10; 5-year time
horizon; WTP £2,370
Results (Table 24 in publication)
• MICE score 0,2,3: BNP then echo (cut-off 58,
36, 28 pg/ml)
• MICE score 5,6,...: Echo

b) Cost of echo £150; cost of BNP £10; 5-year time

horizon; WTP £3,470
Results (Table 24 in publication)
• MICE score 0,2: BNP then echo (cut-off 58, 36,
pg/ml)
• MICE score 3,5,6,...: Echo

c) ) Cost of echo £150; cost of BNP £10; 10-year

time horizon; WTP £5,370
Results (Table 24 in publication)
• MICE score 0: BNP then echo (cut-off 58 pg/ml)
• MICE score 2,3,5,6,...: Echo

3- Use NT-proBNP instead of BNP in the base-case

analysis
Do not affect results

Data sources
Health outcomes:

4
 Chronic heart failure update (Appendix G)

Specificity and sensibility estimates were taken from the meta-analysis presented in the HTA

Life-year gain:
1 2, 3
Survival curves were generated using the Framingham study ; risk ratios for effect on survival of ACEi and BB from meta-analyses .
1
Ho KKL, Anderson KM, Kannel WB et al. Survival after the onset of congestive heart failure in Framingham Heart Study Subjects. Circulation 1993; 88:.107-15
2
Shibata MC, Flather MD, Wang D. Systematic review of the impact of beta blockers on mortality and hospital admissions in heart failure. European Journal of Heart Failure
2001; 3: 351-357
3
Flather MD, Yusuf S, Kober L et al for the ACE-Inhibitor Myocardial Infarction Collaborative Group. Long term ACE-inhibitor therapy in patients with heart failure or left
ventricular dysfunction: a systematic overview of data from individual patients. Lancet 2000; 355:1575-81

Quality-of-life weights:
Calvert 2005 = EQ-5D score of 0.6 for patients with significant heart failure; assume 0.65 for all patients with CHF for use in the analysis

Cost sources:
Drug cost: BNF online accessed 18 September 2007
Intervention cost (echo and BNP): Not referred
Hospitalisation cost: NHS reference cost, assessed 18 September 2007

Comments
Source of funding:
Not reported

Limitations:
It was assumed that if a patient with CHF was misdiagnosed, then the diagnosis would be made after an average delay of 6 months. There is no data on which to base this
assumption.

Waiting lists for echo were not considered (the model assumed there was sufficient capacity). A limited supply of echo and a delay for the investigation could affect the
results of the analysis.

Life-year gain estimates used to calculate QALYs (using a fixed utility score) have not been discounted after one year.

Overall quality*: Minor limitations Overall applicability**: Directly applicable

5
Chronic heart failure update (Appendix G)

Abbreviations: ICER = incremental cost-effectiveness ratio; Echo = echocardiography; WTP = willingness to pay; MICE = Male (2pts) / Infarction (6pts) / Crepitations (5pts) / Edema (3pts);
QALY = cost per quality-adjusted life-years; NHS = National Health Service.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Beta-blockers

G. Yao, N. Freemantle, M. Flather, P. Tharmanathan, A. Coats, Wilson PA Poole, and Investigators SENIORS. Long-term cost-effectiveness analysis of nebivolol compared with
standard care in elderly patients with heart failure: an individual patient-based simulation model. Pharmacoeconomics 26 (10):879-889, 2008.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes: Cost components incorporated: Cost-effectiveness result (ICER):
Cost-effectiveness - Elderly patients (≥ 70 - Drugs - €3066 per life-year gained
analysis, reporting cost years) with heart failure (1) Probabilities o Nebivolol (based on maximum - €3926 per QALY gained
per QALY gained and reduced LVEF incorporated to the model: dosage that each patient
- N total = 2128; 1067 in - Hospitalisation for maintained during SENIOR trial; Sensitivity analysis:
Study design: nebivolol group; 1061 in cardiovascular event 1.25mg/day in 7.2%; 2.5 in 7.6%; 5 One-way sensitivity analysis:
- Individual patient- standard care group. (SENIORS) in 13.3%; 10 in 71.9%) - Age at start treatment (60, 65, 75, 80,
simulation model - Mean age of 76.1 years - Cardiac death (SENIORS) o Other relevant cardiac drugs used instead of 70 for the base-case analysis)
based on a Markov (SD 4.8) in nebivolol - Sudden death (SENIORS) (based on SENIOR trial) - Discount rate: 0% and 5% (instead of 3.5%)
modelling group; 76.1 years (SD - Death due to other - GP visits (assumptions) - Number of outpatient visits after
framework 4.6) in standard care causes (derived from UK o Nebivolol group: each month for 3 cardiovascular hospitalisations: 3 instead of
- Based on the group population based on months, then every 3 months 2.
SENIORS trial - 82.1% on ACEi age- and sex-specific o Standard care group: every 3
- 6.6% on ARB mortality excluding months Probabilistic sensitivity analysis:
Perspective: - 27.6% on aldosterone cardiac related deaths; - Outpatient specialist visits - Key values were varied using their respective
UK NHS antagonist sex based on SENIORS; (assumption) distribution
- 39.3 on glycosides assume every patient o Cardiovascular hospitalisations
Time horizon: - 42.2% on aspirin started at 70 years old) were followed by 2 outpatient Sensitivity analysis results (one-way):
Lifetime horizon - 82.1% on diuretics attendances - Age at start of treatment: ICERs (per QALY
(2) Health-Utility scores - Cardiovascular hospitalisations gained) varied from €3348 for 60 y.o. to
Discounting: Intervention 1: - For each NYHA class (SENIOR) €5291 for 80 y.o.
Future costs and - Nebivolol + standard - Based on reported - Subgroups for cardiovascular - Discount rate: €4187 and €3657 per QALY

6
Chronic heart failure update (Appendix G)

benefits were care results of CARE-HF trial hospitalisations for gained for 0% and 5% respectively
discounted at 3.5% per - Nebivolol was up- - NYHA class I, II, III, IV: o Worsening heart failure (captured - Number of outpatient visits after
annum. titrated during a period respectively 0.815, 0.72, the cost of treatment for severe cardiovascular hospitalisations (3 instead of
of 16 weeks: initial 0.59, 0.508 adverse events) 2 in the base-case analysis): €3923 cost per
dosage of 1.25mg once - When a patient was o MI QALY gained
daily, if tolerated hospitalised, a disutility o Stroke
increase to 2.5 and then score of -0.1 was applied o Rates of these events from SENIOR
5mg every 1-2 weeks, - Cardiovascular deaths (involved
reaching a target of hospital stay cost and intravenous
10mg once daily over a drug cost)
maximum of 16 weeks.
Currency & cost year:
Intervention 2: - Euro 2006
Placebo + standard care - Exchange rate used: £1 = €1.478
Data sources
Health outcomes:
See ‘Health Outcomes’ section

Cost sources:
- Drug cost: BNF No. 52
- GP and outpatient visit cost: Costs and Health Social Care 2006
- Cardiovascular hospitalisation cost: National schedule of reference cost 2006

Comments
Source of funding:
Menarini Research SpA

Limitations:
- Several potentially important economic outcomes were not collected in SENIOR and assumptions where necessary (GP and outpatient attendances)
- Quality of life estimates were not collected in SENIOR and data used were from another source (CARE-HF trial)
- Same cost of hospitalisation used regardless of severity of heart failure

Overall quality*: Minor limitations Overall applicability**: Directly applicable

7
Chronic heart failure update (Appendix G)

Abbreviations: ICER = incremental cost-effectiveness ratio; NHS = National Health Service; BB = Beta-blockers; ACEi = Angiotensin-converting enzyme inhibitors; MI = Myocardial Infarction; SD
= Standard Deviation; NYHA = New York Heart Association Classification; SENIORS = Study of Effects of Nebivolol Intervention on Outcomes and Rehospitalisation in Seniors with Heart Failure;
LVEF = Left Ventricular Ejection Fraction; CARE-HF = Cardiac Resynchronisation – Heart Failure trial.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Aldosterone antagonists

M. Duerden and M. Tabberer. A budget impact model for a drug in heart failure: Eplerenone. British Journal of Cardiology 15 (2):101-105, 2008.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes Cost components Cost-effectiveness result (ICER):
- Cost impact analysis Patients with acute incorporated: incorporated: £6,730 per life year saved
- Cost effectiveness analysis myocardial infarction (MI) (1) Base-case cohort - Hospitalisation cost
complicated by left - Incidence of hospitalised (£1,492 per bed day; Cost-impact analysis result:
Study design: ventricular dysfunction and MI: 433/250,000 in from England 2005- In a primary care trust with a population of 250,000,
- Base-case cohort (no heart failure England; 524/250,000 in 2006 NHS reference eplerenone treatment results in a reduction of:
eplerenone): resource use Scotland (from population cost database) - 46 bed days for re-hospitalisations due to heart
+ cost and survival from Intervention 1: data from the ‘Office of - Eplerenone cost (£555 failure (£1,469 per bed day) = £67,574 (as reported
UK-based statistical data Eplerenone + optimal National statistic and NHS per annum; from in the abstract)
- Treatment cohort medical treatment data’ in England [2005- Haymarket Publishing, - 47 bed days for re-hospitalisations due to heart
(eplerenone): difference in (including ACEi and BB) 2006] and Scotland [2006]) eMIMs) failure (£1,492 per bed day) = £70,124 (as reported
resource and survival - Cost and hospital length of in the text)
based on the EPHESUS trial stay for MI and heart Currency & cost year:
Intervention 2: failure patients (from 2006 GBP Sensitivity analysis:
Perspective: Placebo + optimal medical England 2005-2006 NHS Base-case cohort: increase mortality rate by 10%, 15%,
UK NHS treatment reference cost database) and 20%
- 20% of hospitalised MI
Time horizon: patient shows signs of Sensitivity analysis results :
3 years heart failure (NICE); 87
patients in England (7 per Sensitivity analysis for year 3
Discounting: month) Reduction in 10% 15% 20%
Future costs and benefits - Survival probabilities were mortality over

8
Chronic heart failure update (Appendix G)

were discounted at 3.5% per derived from an baseline

annum. epidemiologic study of all- Cost per life- £2,771 £2,180 £1,812
cause heart failure carried years saved
out in West London up to Bed days 32 25 18
18 months (Cowie 2000); avoided with
extrapolated to 36 months. eplerenone
The extrapolation Cost per bed £2,359 £3,177 £4,634
predicted a 48% 3-year day avoided (£75,488) (£79,425) (£83,412)
survival. (total cost
avoided)
(2) Treatment cohort:
- Reduction in overall
mortality taken from the
EPHESUS study (up to 16
months)
- Post EPHESUS study
mortality: survival curve
runs parallel to base-case
cohort mortality curve up
to 36 months.
- Rate of hospitalisation
from the EPHESUS study
(rates of re-admission and
length of stay)
Data sources
Health outcomes:
Base-case cohort
- Incidence of hospitalised MI: Population data from the ‘Office of National statistic and NHS data’ in England [2005-2006] and Scotland [2006].
- Probability of hospitalised MI patients with heart failure: NICE. Costing report NICE clinical guideline 48. MI: secondary prevention. London, May 2007.
- Survival probabilities: Cowie MR, Wood DA, Coasts AJ et al. Survival of patients with a new diagnosis of heart failure: a population based study. Heart 2000; 83:505-10.
Treatment cohort
- EPHESUS study

Cost sources:

9
Chronic heart failure update (Appendix G)

- Hospitalisation cost: England 2005-2006 NHS reference cost database;

- Eplerenone cost: Haymarket Publishing, eMIMs.
- Difference in resource use and survival for the treatment cohort based on the EPHESUS trial.

Comments
Source of funding:
Pfizer Ltd

Limitations:
- The analysis assumed a 100% adherence and compliance for patients on eplerenone treatment.
- Extrapolation of survival
- Short time horizon (3 years)
- Difference in survival and resource use from the EPHESUS trial, a multi-centre and multi-national study. Result applied to a UK context.

Overall quality*: Potentially serious limitations Overall applicability**: Directly applicable

Abbreviations: ICER = incremental cost-effectiveness ratio; NHS = National Health Service; BB = Beta-blockers; ACEi = Angiotensin-converting enzyme inhibitors; EPHESUS = Eplerenone Post-
Acute Myocardial Infarction Heart Failure Efficacy and Survival Study; MI = Myocardial Infarction.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

L. Tilson, B. McGowan, M. Ryan, and M. Barry. Cost-effectiveness of spironolactone in patients with severe heart failure. Irish Journal of Medical Science 172 (2):70-72, 2003.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes Cost components Cost-effectiveness result:
- Cost effectiveness analysis - Patients with severe incorporated: incorporated: Analysis Result
- Reporting cost per life-year chronic heart failure - Probabilities of death and - Spironolactone Base-case analysis €466/LYG
gained - NYHA class III & IV and hospitalisation for the treatment cost (Irish (pDeath = 0.18; pHosp
LVEF ≤35 placebo cohort were taken Monthly Index of = 0.25; 1 additional
Study design: - Mean age of 65 years from a cohort of patients Medical Specialties, outpatient visit for
- Based on the RALES study followed over 12 months in July 2002) spironolactone cohort;
- Developed using a Markov Intervention 1: an Irish teaching hospital - Hospitalisation cost hosp cost = €3,019)
Model Spironolactone added to - The difference in for severe heart failure
o 3 health states: (1) severe optimal medical probabilities of death and (McGowan B. et al.

10
Chronic heart failure update (Appendix G)
CHF; (2) severe CHF +
hospitalisation; (3) death
o 1-year period before
possible transition from
one state to another.
Perspective:
Irish public healthcare system
Time horizon:
10 years
Discounting:
Future costs and outcomes
were discounted at 5% and
1.5% respectively.
Data sources
Health outcomes:
See above
Cost sources:
See above
Comments
Source of funding:
management
Intervention 2:
Optimal medical
management (might include
diuretics, ACEi, digoxin, BB,
or a combination of these)
hospitalisation for the Cost of treating heart Two-way sensitivity from € 309/LYG to
treatment cohort were failure in an Irish analysis – variation of €624/LYG
taken from RALES teaching hospital. Ir probabilities of death
- Assumed no difference in Med Sci 2001; (0.16, 0.21) and
death and hospitalisation 169:241-44) hospitalisation (0.21,
rates between cohorts - Hospital outpatient 0.29)
after the 2-year mean visit cost (McGowan B. One-way sensitivity from €466/LYG to
duration of RALES The clinical and analysis – additional €1,136/LYG
economic aspects of outpatient visit
the present required to initiate
management of heart medication for
failure in an Irish spironolactone group
teaching hospital. MSc (1, 2, 4)
thesis, Trinity College One-way sensitivity from €728/LYG to
Dublin 2001). analysis – cost of spironolactone cohort
hospitalisation varied dominates placebo
Currency & cost year: (€1,060; €9,319) cohort (adding
Euro 2002 spironolactone to
standard treatment is
more effective and less
costly)
11
Chronic heart failure update (Appendix G)

NR

Limitations:
- The study did not incorporated a quality of life measure
- The mean age of the population of patients in RALES study was lower than the Irish population of patient with chronic heart failure (65 vs 76 years)
- Some cost data were taken from published studies and not from Government sources, which can affect their relevance
Overall quality*: Potentially serious limitations Overall applicability**: Partially applicable
Abbreviations: RALES = Randomised Aldactone Evaluation Study; NYHA = New York Heart Association Classification; CHF = Chronic Heart Failure; BB = Beta-blockers; ACEi = Angiotensin-
converting enzyme inhibitors.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Isosorbide/hydralazine combination

D. C. Angus, Zwirble WT Linde, S. W. Tam, J. K. Ghali, M. L. Sabolinski, V. G. Villagra, W. C. Winkelmayer, M. Worcel, and American Heart Failure Trial African. Cost-
effectiveness of fixed-dose combination of isosorbide dinitrate and hydralazine therapy for blacks with heart failure. Circulation 112 (24):3745-3753, 2005.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes Cost components ICER:
Cost-effectiveness analysis - Black people with incorporated: incorporated: Reported cost per life-year gained.
moderate to severe Data from A-HeFT RCT (18- - Hospitalisation
Study design: heart failure month follow-up): (including physician Results reported using 2 total cost estimates:
Decision-analytic model - Mean age: 56.8 years - Mortality (treatment vs cost); - Heart failure-related cost;
(based on the A-HeFT study) - 100 % black people placebo): 6.2% vs 10.2%; - ER visits; - All healthcare-related cost
- 40% women p=0.016 - Unscheduled physician
Perspective: - 94.9% NYHA class III - Survival time: 403 vs 380 visits; Cost-effectiveness within the A-HeFT study
- US perspective heart failure days; p=0.01 - Scheduled physician period (18 months):
- Included direct healthcare - 93% were on ACEi or visits; - ISDN/HYD therapy is dominant (improved
costs and excluded indirect ARB Post trial survival: - ISDN/HYD therapy; survival and saved cost) using both heart
and non-healthcare costs - 87% were on BB - Use 5-year follow-up data - Concomitant failure-related cost ($9144-$8611=$553) and
- Treatment N=518; for NYHA class III patients medication; all healthcare-related cost ($19728-
Time horizon: control N=532 reported by Bardy 2005 - Other cares. $17998=$1730).
- 18 months (A-HeFT follow- - Baseline characteristics - Assumed that survival Bootstrap simulation sampling :

12
Chronic heart failure update (Appendix G)
up) were similar across arms
- Lifetime
Intervention 1:
Discounting: Standard therapy (BB, ACEi
Future costs and survival were or ARB, aldosterone
discounted at 3% per annum. antagonist, digoxin and
diuretics as appropriate)
Intervention 2:
Standard therapy +
ISDN/HYD therapy (20mg /
37.5mg); starting with 1
tablet 3 times daily and
titrating to 2 tablets 3 times
daily as tolerated.
Average dose: 4.2 tablets
per day.
68% took full doses of 6
tablets per day at some
time.
Failure of adherence
(treatment vs placebo):
3.2% vs 10.2%; p=0.016.
Mean follow-up: 12.8
months.
curves of treatment and Resource use as collected - Heart-failure related cost : 49% dominant;
control arms decayed at from the A-HeFT trial. 66% better than $10 000/life-year gained
the same rate. - All health costs : 71% dominant; 82% better
- Zero survival at 10 years. Currency & cost year: than $10 000/life-year gained
2004 US dollars
Lifetime horizon:
Resource use results - Using the assumption that there were no
from A-HeFT: additional benefits of ISDN/HYD therapy
- 43% of beyond the trial period (survival and
hospitalisations were resource use – except drug cost for
related to heart failure treatment arm); ICER = $41 800 per life-year
- 30% fewer heart gained (hearth failure-related costs); $44
failure-related 400 (all medical costs)
hospitalisations for - When considering that ISDN/HYD therapy
treatment group had benefit beyond the trial, considering one
- One-day reduction in additional year of effect, the lifetime ICER
the average LOS for was estimated to be $22 900 per life-year
each heart failure- gained (hearth failure-related cost); $32 900
related (all medical costs).
hospitalisations for the
treatment group Sensitivity analysis:
- All-cause Were varied:
hospitalisations, - Hospital cost and ISDN/HYD therapy cost ±
unscheduled office 50%;
visits, ER visits, and - Cost of concomitant medication,
use of concomitant unscheduled office visits, ER visits, and other
medications: not usual medical care from +100% to -50%.
significantly different - Other hospitalisation costs were used.
between groups
- All-cause Sensitivity analysis results :
hospitalisation LOS The incremental cost was most sensitive to
shorter for treatment hospital cost and treatment cost (ISDN/HYD
group therapy) variations.
13
 Chronic heart failure update (Appendix G)

Using an alternative hospitalisation cost

(Medicare heart failure-related hospitalisation
cost unadjusted for race, gender, and LOS
[adjusted for the main analysis]) :
- ISDN/HYD therapy no longer dominant
- Using hearth failure-related costs; ICER = $10
335 per life-year gained; 24% of simulations
dominant; 49% of simulations less than $10
000/life-year gain
- Using all medical costs; ICER = $1546 per life-
year gained; 46% of simulations dominant;
66% of simulations less than $10 000/life-
year gained

Data sources
Health outcomes:
- First 18-month survival from the A-HeFT study.
1
- Survival curves post-trial: was used 5-year follow-up data for NYHA class III patients reported by Bardy 2005 .
1
Gust H. Bardy, M.D., Kerry L. Lee, Ph.D., Daniel B. Mark, M.D., Jeanne E. Poole, M.D., Douglas L. Packer, M.D., Robin Boineau, M.D., Michael Domanski, M.D., Charles Troutman, R.N., Jill Anderson, R.N., George
Johnson, B.S.E.E., Steven E. McNulty, M.S., Nancy Clapp-Channing, R.N., M.P.H., Linda D. Davidson-Ray, M.A., Elizabeth S. Fraulo, R.N., Daniel P. Fishbein, M.D., Richard M. Luceri, M.D., John H. Ip, M.D., for the Sudden
Cardiac Death in Heart Failure Trial (SCD-HeFT) Investigators.Amiodarone or an Implantable Cardioverter–Defibrillator for Congestive Heart Failure. N Engl J Med 2005;352:225-237.

Cost sources:
- Hospitalisation cost: 2003 Medicare Hospital Discharge Database. Washington, DC. Data adjusted to years 2004. Increased this cost by 17% to account for physician fees
(published method).
- Cost for ER and unscheduled physician visits; cost for scheduled physician visits; and cost for other cares: Medicare Utilisation for Part B (supplementary Medical Insurance
SMI). Baltimore, Md: Centres for Medicare and Medicaid Services; 2004.
- Concomitant medication: 2004 Red book.
- ISDN/HYD therapy: Manufacturer’s (NitroMed Inc) announced price.

Comments

14
Chronic heart failure update (Appendix G)

Source of funding:
NitroMed Inc

Limitations:
This study did not use QALYs as health outcome (assessed in A-HeFT trial using the ‘Minnesota Living With Heart Failure’ survey, for which no validated approach to generate
utility scores exist). To note that patients in the treatment arm reported better quality of life.

Overall quality*: Minor limitations Overall applicability**: Partially applicable

Abbreviations: ICER = incremental cost-effectiveness ratio; NHS = National Health Service; A-HeFT = African-American Heart Failure Trial; ISDN/HYD = Fixed-dose combination of isosorbide
dinitrate and hydralazine; BB = Beta-blockers; ACEi = Angiotensin-converting enzyme inhibitors; ARB = Angiotensin receptor blockers; RCT = Randomised controlled trial; LOS = Length of stay;
NYHA = New York Heart Association.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Angiotensin-II receptor antagonist

J. J. McMurray, F. L. Andersson, S. Stewart, K. Svensson, A. C. Solal, R. Dietz, J. Vanhaecke, D. J. van Veldhuisen, J. Ostergren, C. B. Granger, S. Yusuf, M. A. Pfeffer, and K.
Swedberg. Resource utilization and costs in the Candesartan in Heart failure: Assessment of Reduction in Mortality and morbidity (CHARM) programme. European Heart Journal
27 (12):1447-1458, 2006.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes: Cost components Cost-effectiveness and cost-consequence
- Cost-consequence analysis - Patients with NYHA class II- - The primary outcome of incorporated: results:
for CHARM overall, IV HF the CHARM programme - Drug treatment (for the See tables below
alternative, added, and - CHARM-Added: patients was a composite of candesartan arm,
preserved; with LVEF < 0.40 on ACEi cardiovascular death or considering 4 GP visits and 4 Sensitivity analysis:
- Cost-effectiveness analysis - CHARM-Alternative: hospital admission for biochemistry tests for drug - Increased the length of non-cardiovascular
for patients with LVEF < patients with LVEF < 0.40 worsening HF (hazard initiation and up-titration) admissions of 30% - for the potential
0.40 (alternative, added, intolerant of ACEi ratio): CHARM-Added: - Hospital admission – for additional cost of certain adverse events
and the 2 pooled together); - CHARM-Preserved: 0.85 (95% CI 0.75-0.96, cardiovascular reason and (renal impairment) in the candesartan group
reporting cost per life-year patients with LVEF > 0.40 P=0.011); CHARM- not (number of admission, - Adding cost of 1 GP visit for treatment
gained. Alternative: 0.77 (0.67- length of stay, ward type) – related adverse events in the candesartan
Intervention 1: 0.89, P=0.0004; CHARM- Using 2 approaches: (1) DRG group (renal impairment and hypotension)

15
Chronic heart failure update (Appendix G)
Study design:
- Based on the CHARM
programme
- Within-trial cost-
consequence and cost-
effectiveness analyses of
CHARM
Perspective:
UK NHS; France; Germany
Time horizon:
Within-trial analysis.
Median follow-up of 41
months for CHARM-Added; 37
months for CHARM-preserved;
34 months for CHARM-
Alternative; and 38 months for
the overall CHARM
programme.
Discounting:
Future costs and benefits were
discounted at 3% per annum
(3 perspectives) – 3.5% for the
UK analysis in the sensitivity
analysis.
Data sources
Health outcomes:
CHARM programme
Cost sources:
Candesartan + optimal
medical treatment (diuretic,
digoxin, ACEi, ARB, BB
spironolactone)
Intervention 2:
Placebo + optimal medical
treatment
Mean daily dose of
candesartan was:
CHARM-Added: 16.9mg;
CHARM-Alternative: 16.8mg;
CHARM-Preserved: 19mg.
Preserved: 0.89 (0.77- cost; (2) Per diem cost for not leading to admission
1.03, P=0.118); Overall hospital bed days. - Length of hospital stay ± 20%
CHARM programme: - Cardiovascular procedures - Use 3.5% as discount rate for the UK analysis
0.84 (0.77-0.91),
P<0.0001. Currency & cost year:
- 2003-2004 local currency for
- All-cause mortality UK, France, Germany.
(hazard ratio): Overall
CHARM programme:
0.91 (0.83-1.00),
P=0.055; Patients with
LVEF < 0.40: 0.88 (0.79-
0.98, P=0.018).
- Clinical benefits
considered for cost-
consequence analyses
were cardiovascular
death and heart failure
admissions
- Clinical health outcome
considered for cost-
effectiveness analyses
was all-cause mortality.
16
Chronic heart failure update (Appendix G)

- Government sources (UK, France, Germany) for all cost inputs (hospitalisation, drug treatment, cardiovascular procedures)
- In the UK: PSSRU; NHS reference cost database; and BNF.
Comments
Source of funding:
The CHARM programme was founded by AstraZeneca

Limitations:
- Within-trial time horizon
- No quality of life estimate incorporated to the analysis
- Use results from a multi-centre and multi-national study. Results applied to specific country contexts.
Overall quality*: Potentially serious limitations Overall applicability**: Directly applicable
Abbreviations: ICER = incremental cost-effectiveness ratio; NHS = National Health Service; BB = Beta-blockers; ACEi = Angiotensin-converting enzyme inhibitors; NYHA = New York Heart
Association Classification; LVEF = Left Ventricular Ejection Fraction; CHARM = Assessment of Reduction in Mortality and morbidity; HF = Heart Failure.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

McMurray 2006 – Cost-consequence results

CHARM Trial Clinical benefits vs placebo UK France Germany
Cardiovascular Heart failure DRG costs Per diem costs DRG costs Per diem costs DRG costs Per diem costs
death admission
Alternative -15% -32% Net increase Savings Savings Savings Net increase Savings
€76±1150/year €391±2192/year €49±1475/year €428±2952/year €117±1164/year €12±1529/year
Added -16% -17% Savings Savings Savings Savings Net increase Savings
€15±314/year €419±825/year €120±841/year €346±1397/year €29±631/year €0.2±805/year
Preserved -1% (ns) -15% Net increase Net increase Net increase Net increase Net increase Net increase
€246±337/year €276±787/year €299±630/year €321±1240/year €327±460/year €372±689/year
Overall -12% -21% Net increase Savings Net increase Savings Net increase Net increase
€116±352/year €122±749/year €73±548/year €88±1045/year €176±421/year €153±563/year

McMurray 2006 – Cost-effectiveness results*

CHARM Trial Life-year gained (LYG) UK France Germany
(95%CI)
Alternative 0.078 (0.003-0.15) €2547 (-18171; 1059150) Dominant** €3881 (-17728; 1105920)
Added 0.061 (-0.002-0.12) Dominant Dominant €1427 (-14479; -984755)

17
Chronic heart failure update (Appendix G)

Reduced LVEF pooled 0.068 (0.02-0.12) €1348 (-16225; 106600) Dominant €2997 (-19183; 121500)
* Unit cost in pound sterling were converted in Euro using 1 Euro = 0.67 GBP
** Dominant means that candesartan added to optimal medical management was more effective and less costly than adding placebo.

Rehabilitation

D. Georgiou, Y. Chen, S. Appadoo, R. Belardinelli, R. Greene, M. K. Parides, and S. Glied. Cost-effectiveness analysis of long-term moderate exercise training in chronic heart
failure. American Journal of Cardiology 87 (8):984-988, 2001.
Study details
Economic analysis:
- Cost-effectiveness analysis
- Reporting cost per life-year
gained
Study design:
- Decision analytic model
- Based on a RCT (Belardinelli
1999)
Perspective:
- US societal perspective
- Included direct medical
costs (Medicare) and
patient-level costs (wage
lost).
Time horizon:
10 years post RCT follow-up
period (1,639 days)
Population & interventions
Population:
- Patients with stable
chronic heart failure
- NYHA class II-III HF patients
between 55 and 64 years
- Population of the
Belardinelli 1999 trial
Intervention 1:
- Physical rehabilitation
program
- 14-month-long moderate
exercise training
(healthcare-based)
- 3 sessions/week for 8
weeks, followed by 12-
month maintenance
program of 2
sessions/week
- 1 hour session
o 20 mins warm-up and
Health outcomes Costs Cost effectiveness
Health outcomes: Cost components Cost-effectiveness results:
Survival incorporated: $1,773 per life-year gained (Incremental
- Survival estimates from Cost of exercise training cost=$3227 per patient; Incremental life-year
Belardinelli 1999 for the - Equipment (bicycle); gained in 15.5 years=1.82 per patient)
within-trial period (1,639 - Rented place (Hospital-
days) based; New York City) Sensitivity analysis:
o Mortality rate of 18% - Trainer salary (1 trainer The following factors were varied (95% CI):
for the treatment supervising 4 patients at a - Survival probabilities of the within-trial
group (end of follow- time) period: ICER from $8,274 to $1,012 per life-
up) - Cardiopulmonary stress test year gained
o Mortality rate of 41% cost including the physician - Survival probabilities post trial varying the
for the control group component of interpretation ACEi survival rate adjustment: ICER from
- 10-year post-trial period and exercise prescription $1,698 to $1,855 per life-year gained
mortality rate - Within-trial rates of hospitalisation: ICER
o Same mortality rate Cost of wage lost from $2,372 to 1,197 per life-year gained
applied for the - 128 hours of wage lost per - Remove cost of wage lost: ICER=$395 per
treatment and patient life-year gained
control groups - Average wage of $19.60 per
o From the National hour (US Census Bureau,
Health and Nutrition 1998)
Examination I –

18
Chronic heart failure update (Appendix G)

Discounting: stretching Epidemiologic follow- Hospitalisation cost

Future costs and benefits were o 40 mins on an up Survey (1982 – Cost of hospitalisation for
discounted at 3% per annum. electronically-braked 1986) patients with chronic heart
cycle ergometer o Mortality rate failure estimated by Delea
(bicycle) adjusted with sex- 1999.
- Total of 128 sessions (128 specific rates using
hours) ratio from the Currency & cost year:
Belardinelli 1999 trial - 1999 US dollars
Intervention 2: o Survival rate
- No physical rehabilitation adjusted upward by
program 23% considering
patients taking ACEi
(introduced after the
national survey)
(Pfeffer 1992; Garg
1995)

Hospitalisation rates
- Same rate of
hospitalisation used
post-trial for both
cohorts (19%;
assumption)
- From difference of rates
in Belardinelli 1999
o 10% and 29%
(training and control
groups)

Data sources
Health outcomes:
See above

19
Chronic heart failure update (Appendix G)

Cost sources:
See above

Comments
Source of funding:
The study was supported in part by a grant from Columbia University (New York), and by Merck & Co. Inc.

Limitations:
- The study only considered patient with NYHA class II-III HF, excluding class IV patients using the assumption that these patients are not appropriate candidates for exercise
training (unable to carry out any physical activities without discomfort).
- The trial was conducted predominantly on a male population; better health outcomes expected for women.
- Trial conducted on a population of patients aged from 55 to 64 years.
- Assumptions for post-trial hospitalisation rate (19%, same for both arms) and survival rate (same for both arms)
- No quality of life measure was incorporated to the analysis: the Minnesota Living with Heart Failure Questionnaire estimates from Belardinelli 1999 cannot be translated in utility
scores
- Cost of exercise training calculated for a New York City Hospital; estimated roughly 20% more expensive than the national average
- Inclusion of wage lost cost (societal perspective); using only direct medical cost is most relevant for decision-making from a public healthcare perspective.

Overall quality*: Potentially serious limitations Overall applicability**: Partially applicable

Abbreviations: ICER = incremental cost-effectiveness ratio; ACEi = Angiotensin-converting enzyme inhibitors; NYHA = New York Heart Association Classification; RCT = Randomized controlled
trial; HF = Heart Failure; CI = Coefficient Interval.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Serial measurement of natriuretic peptide concentration

T. Morimoto, Y. Hayashino, T. Shimbo, T. Izumi, and T. Fukui. Is B-type natriuretic peptide-guided heart failure management cost-effective? International Journal of Cardiology
96 (2):177-181, 2004.
Study details Population & interventions Health outcomes Costs Cost effectiveness
Economic analysis: Population: Health outcomes: Cost components incorporated: Cost-effectiveness result (ICER):
- Cost-effectiveness Symptomatic CHF patients Utility values (symptomatic - BNP measurement once per 3 months - The base-case analysis shows that the BNP
analysis (NYHA class II-IV) aged 35- CHF patients with reduced - Drugs for CHF (digoxin, diuretics, ACEi, group dominates the clinical group (QALYs:

20
Chronic heart failure update (Appendix G)

- Reporting cost per 85, after hospital admission LVEF) from data by and BB) 0.57 vs 0.55; costs: $9577 vs $10,131)
QALY gained because of CHF with Havranek 1999. - Dispensing fee - See table below
reduced LVEF. - Ambulatory care for CHF
Study design: Probability estimates: - Inpatient care for CHF Sensitivity analysis:
- Decision analytic Intervention 1: Variation between cohorts - Non-CHF related care The sensitivity analysis varied all parameters:
model based on a Outpatient BNP-guided were calculated for: - Cost data were all taken from Delea - 95% CI for utility scores
Markov framework heart failure management - Hospitalisation for CHF 1999 (converted in $2002), except the - Ratios of increase in medication and
(proposed by Paul once every 3 months (BNP - Death for CHF cost of BNP measurement which was ambulatory visits in BNP group were varied
1994; for outpatient group) - Frequency of ambulatory derived from a price list of a between 1 and 2 (base-case probabilities of
follow-up after care University hospital in the US 1.5 for ambulatory care and 1.4 for doses of
hospitalisation for (medication usage was - Dose of ACEi (University of Washington) – ref: ACEi and diuretics)
CHF) increased to keep the BNP - Dose of diuretics News Flash (January 1, 2002). - Other parameters were varied ±50%
- Based on Troughton level below 200pmol/l) Department of Laboratory, University - The time horizon was varied from 6 to 18
2000 RCT and on - These probabilities were of Washington. months
Delea 1999 taken from Troughton (http://depts.washington.edu/labweb
economic model Intervention 2: 2000 and Delea 1999 /test/nf/jan02.pdf). Sensitivity analysis results:
(patients with heart No BNP measurement - When varying the time horizon, the BNP
failure) (clinical group) Currency & cost year: group was dominant at 6, 9 and 12 months;
- 2002 US Dollars ICER of $3491 per QALY gained at 15
Perspective: months, and of $7787 per QALY gained at 18
US Medicare months.
- Results were sensitive to the degree of
Time horizon: increase in ambulatory visits for the BNP
- 9 months (median group, the probability of first readmission for
of 9.5 months in the clinical group, the costs of ambulatory
Troughton 2000) visits, and the cost of inpatient care for CHF.
- The time horizon The ICER stayed cost-effective in the
was varied from 6 to majority of simulations.
18 months in the - The BNP group ICER became not cost-
sensitivity analysis effective (threshold of $50,000/QALY) when
were decreased simultaneously the
Discounting: probability of first readmission for the
Future costs and clinical group and the cost of inpatient care

21
Chronic heart failure update (Appendix G)

benefits were for CHF.

discounted at 3% per
annum.

Data sources
Health outcomes:
See above

Cost sources:
See above

Comments
Source of funding:
Supported in part by a grant from the Ministry of Health, Labor and Welfare of Japan.

Limitations:
- The generalisation of these results in a UK context is questionable (this study was conducted from a US perspective, a healthcare system non-comparable to the UK NHS)
- Effectiveness data and probability estimates were taken from a clinical trial which might not illustrate properly the current clinical practice
- Cost data were taken from published studies and not from Government sources which can affect their relevance

Overall quality*: Potentially serious limitations Overall applicability**: Partially applicable

Abbreviations: ICER = incremental cost-effectiveness ratio; NHS = National Health Service; BB = Beta-blockers; ACEi = Angiotensin-converting enzyme inhibitors; NYHA = New York Heart
Association Classification; LVEF = Left Ventricular Ejection Fraction; CHF = Chronic Heart Failure; BNP = Brain Natriuretic Peptide.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

Results - Morimoto 2004

6 months 9 months 12 months 15 months 18 months
(base-case analysis)
BNP Group
QALY 0.38 0.57 0.74 0.91 1.07
Cost $5577 $9577 $13,436 $17,155 $20,737
Clinical Group
QALY 0.38 0.55 0.70 0.83 0.94

22
Chronic heart failure update (Appendix G)

Cost $6230 $10,131 $13,670 $16,861 $19,723

Result
ICER Dominant* Dominant* Dominant* $3491 per QALY $7787 per QALY
*Dominant means that the intervention was more effective and less costly

Tele-monitoring and self-monitoring

S. Scalvini, S. Capomolla, E. Zanelli, M. Benigno, D. Domenighini, L. Paletta, F. Glisenti, and A. Giordano. Effect of home-based telecardiology on chronic heart failure: costs and
outcomes. Journal of Telemedicine & Telecare 11 Suppl 1:16-8, 2005.:16-18, 2005.
Study details Population & interventions
Economic analysis: Population:
Cost-consequence analysis - Patients with stable
chronic hearth failure
Study design: (n=426)
Based on a prospective cohort - Mean age of 59 years
study (SD9)
Perspective: Intervention 1:
Italian public healthcare - Home-based
system telecardiology (n=230)
- Trans-telephonic follow-up
Time horizon: and ECG monitoring
1 year - ECG portable device given
to patients, transferring
Discounting: data by mobile or fixed
Not applicable telephone line
- ECG recording transmitted
to a received station,
where a nurse was
available for reporting and
interactive
Health outcomes
Health outcomes:
See table above ‘Scalvini
2005 – Clinical outcomes’
Costs Cost effectiveness
Cost components Cost-effectiveness results:
incorporated: Cost and clinical outcomes were presented
Home-based telecardiology separately in tables below.
cost:
- Equipment (hardware and Sensitivity analysis:
software, ECG recorders, No sensitivity analysis was undertaken
dedicated telephone line)
- Rental (dedicated lines,
service centre)
- Personnel (physicians,
nurses)
- Other (fax and photocopier,
consumables, software
maintenance, vocational
training)
Hospitalization cost
Currency & cost year:
- Euro (year not reported,

23
Chronic heart failure update (Appendix G)

teleconsultations study published in 2005)

- Tele-assistance for patients
when necessary
- Home visit by paramedical
and medical team if
necessary

Intervention 2:
Usual care (n=196)

Data sources
Health outcomes:
Prospective cohort study

Cost sources:
Not reported

Comments
Source of funding:
Not reported

Limitations:
- No sensitivity analysis
- Short time horizon
- No quality of life estimates
- Source of cost not reported
- The analysis did not considered possible important cost components (drug, intervention, outpatient visit, emergency visit)
- Usual care intervention not described

Overall quality*: Very serious limitations Overall applicability**: Partially applicable

Abbreviations: SD = Standard deviation; ECG = Electrocardiogram.
*Very serious limitations/Potentially serious limitations/Minor limitations; **Directly applicable/Partially applicable/Not applicable

24
Chronic heart failure update (Appendix G)

Scalvini 2005 – One-year clinical outcomes

Usual care Home-based telecardiology Relative risk (95% CI)
(n=179) (n=230)
Hospitalization, n (%) 61 (34) 56 (24) 0.62 (0.43-0.81)
Patients with instability, n (%) 74 (41) 60 (26) 0.50 (0.32-0.68)
Death, n (%) 22 (12) 6 (7) 0.50 (0.20-0.80)

Scalvini 2005 – One-year cost outcomes

Usual care Home-based telecardiology
(n=179) (n=230)
Hospitalisation cost €140,874 €95,688
Telecare service cost NA €11,806
Total cost €140,874 €107,494

25