Estimates of Cost Per Year of Life Saved for Dabigatran

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]

Downlaod the slides here: media:Cost effectiveness of dabigatran final.ppt

Overview

The traditional method of calculating the cost per year of life saved involves:

1. Estimating the cost to save a life
2. Estimating how many years that person will live
3. Dividing the cost to save the life by the number of years the person lives

What are the costs of dabigatran and warfarin?

The annual cost of therapy with Dabigatran are estimated to be $2,884 per patient.

The cost of Warfarin is approximately $0.30 per day, $109.5 per year.

(http://www.pharmacychecker.com/Pricing.asp?DrugName=Warfarin&DrugId=19462&DrugStrengthId=31721 )

There is a cost associated with INR monitoring of Warfarin of $2,134 per year in the first year (http://www.biomedcentral.com/1471-2296/8/6/ ) which drops to $1,170 in the second year once the patient is at a more stable level of anticoagulation. Let’s assume an average cost of $1,652 per year. The cost of Warfarin and its monitoring is therefore $1,761 per year.

Thus, the incremental drug acquisition and monitoring costs of DabigatranDabigatran over Warfarin would be $2,884 - $1,761 per year or $1,123 per year.

How many patients would you need to treat to save a life?

Given the 0.5% mortality reduction (which was of borderline statistical significance for the 150 mg dose), you would need to treat 200 patients for one year to save one life.

How much would it cost to save that life?

That one year of therapy in 200 patients would cost an additional 200 X $1,123 or $224,600.

How long does someone with atrial fibrillation live after you save their life?

The average age of patients in RE-LYwas 71.4 years. The average years of life a female can expect to live beyond 71 years is 14 years and a male is 13 years:

(http://www.census.gov/compendia/statab/cats/births_deaths_marriages_divorces/life_expectancy.html ).

Patients who have atrial fibrillation may live less than the usual 13.5 years for a 71 year old. The survival of an atrial fibrillation patient may depend upon co-morbidities. Although post CABG atrial fibrillation may have a limited impact on survival, some conditions such as systolic heart failure that are associated with atrial fibrillation are associated with a 1.9 fold higher risk of dying.

Given that the risk of dying is doubled in these patients, let’s assume that the survival of a 71-year old atrial fibrillation patient is cut in half to 6.75 years. Using this number will yield a conservative estimate of cost effectiveness.

What is the cost per year of life saved with dabigatran?

Therefore, $224,600 in societal costs divided by those 6.75 years the person whose life was saved goes on to live equates to $33,274 per year of life saved. This is well within the acceptable limits for cost per year of life saved in the United States and similar to the cost per year of life saved for a defibrillator.

What if you factor in the reduction in costly strokes and the reduction in costs due to a reduction in bleeding?

If complications of pharmacotherapy are included in the cost effectiveness analysis, the cost effectiveness of dabigatran may be even more favorable.

The lifetime cost of all cause stroke is estimated by the CDC to be $100,000. The rate of stroke per year was 1.57% for Warfarin and 1.01% for 150 mg of dabigatran. Therefore, there is a 0.56% lower rate of stroke per year. In our population of 200 patients, there would be expected to be 1.12 fewer strokes. If each stroke costs $100,000 over the lifetime of the patient then preventing 1.12 fewer cases with dabigatran would save $112,000.

In addition to the small but statistically significant reduction in mortality associated with Dabigatran therapy, there were also numerically (but not statistically significantly) fewer major bleeds (3.4% vs 3.1% per year). The cost of a major bleed is estimated to be $8000 (personal communication, Dr. David Cohen). The cost of treating 200 patients with dabigatran would be reduced slightly by 0.6 bleeds or $4,800 dollars.

There was an increased risk of MI of 0.21% (p=0.048) for dabigatran vs Warfarin (0.74% - 0.53%/year). The cost of a significant q wave MI is $7000 (Source: Dr. David Cohen). Treatment of the 200 patients with dabigatran would yield 0.4 of an MI at an increased cost of $2800.

Treating 200 patients with dabigatran for a year costs an additional $224,600 over Warfarin. Among 200 treated patients, there is a cost reduction of $112,000 due to 1.12 fewer stroke cases with dabigatran. Among these 200 treated patients, there is a cost reduction of $4,800 due to 0.6 fewer major bleeds. There is also a cost increase of $2,800 among these 200 treated patients due to 0.4 MI cases. The total additional cost of dabigatran treatment inclusive of complications in 200 patients is $110,600. The cost per year of life saved assuming 6.75 years of survival could therefore be as low as $16,385.

If society views $50,000 / year of life saved as cost effective, and' if it costs society $110,600 to save that life when the costs of complications are factored in, then' the survival of atrial fibrillation patients must only exceed 2.2 years for the therapy to be cost effective. (Spending $110,600 to save a life of someone who lives 2.2. years equates to $50,000 per year of life saved.) Most 71 year old atrial fibrillation patients survive more than 2.2 years.

What does it cost to prevent a stroke?

There was a reduction in all cause stroke from 1.57% / year for Warfarin to 1.01% / year for dabigatran, a 0.56% / year reduction. You would need to treat 179 patients with dabigatran to prevent one stroke. This would incur an incremental cost of 179 x $1,123, or $200,535. On the other hand, the lifetime cost of all cause stroke is estimated by the CDC to be $100,000. The net total cost would therefore be $100,535. Patients with stroke have a 2.3 fold higher mortality, so survival was estimated to be 13.5 / 2.3 = 5.8 years. That’s $17,333 per year of stroke-free life saved.

What are the limitations of a cost-effectiveness analysis like this?

These estimates are driven by “point estimates” A 0.5% improvement in mortality is small, and this point estimate may be unstable (p=0.051). Patients with atrial fibrillation may live longer than the 6.75 years assumed here. A longer survival time would reduce the cost per year of life saved.

No quality adjusted life expectancy data is presented. These data mix present cost with future costs. Some cost estimates are old. There may be heterogeneity in costs of stroke, bleeding and MI throughout the world.

How do these costs fit in with what we spend on other things in society?

Last year, in TIME magazine, Sir Michael Rawlins, chairman of the National Institute for Health and Clinical Excellence stated the following (http://www.time.com/time/health/article/0,8599,1888006,00.html#ixzz13Z0tUMuG ):

“Our Department of Transport, for instance, has a cost-per-life-saved threshold for new road schemes of about 1.5 million GBP per life, or around 30,000 GBP per life year gained. The judgment of our health economists is that somewhere in the region of 20,000-30,000 GBP per quality-adjusted life year is the [threshold], but it's not a strict limit. “

This would convert to 1.58 X 20,000 GBP = $31,600 USD to $47,400 USD. Thus, $16,000 to $33,000per year of life saved would fall within or even below this range.

While estimates of what governments are willing to pay for are generally about $50,000 per year of life saved, hemodialysis costs approximately $129,000 per year of life saved. (http://www.time.com/time/health/article/0,8599,1808049,00.html )

Conclusion

The annual difference in costs between dabigatran and Warfarin once the cost of monitoring is accounted for is approximately $1,123 US dollars. The estimated cost per year of life saved ($16,000 to $33,000 per year of life saved) is likely to fall within the range of acceptable cost effectiveness ($50,000 per year of life saved). The cost to prevent all cause stroke appears to be acceptable as well. While significant reductions in stroke and borderline significant reductions in mortality were observed, the point estimates were infrequent (around 0.5%) which may adversely impact the certainty of cost effectiveness estimates.