How much should we be willing to pay for a course of #remdesivir?
A new report by @icer_review suggests $4,460 per hospitalized patient with #COVID19.
As always, the devil is in the details.
THREAD 1/
@EricTopol @Bob_Wachter @KBibbinsDomingo
A new report by @icer_review suggests $4,460 per hospitalized patient with #COVID19.
As always, the devil is in the details.
THREAD 1/
@EricTopol @Bob_Wachter @KBibbinsDomingo
@icer_review's approach is transparent and parsimonious (which is good).
Assumes:
- 66% admitted to med-surg floor ($13K, 4% mort)
- 8% to ICU, no vent ($34K, 10% mort)
- 28% to ICU on vent ($61K, 31% mort).
Seem reasonable. 2/
Assumes:
- 66% admitted to med-surg floor ($13K, 4% mort)
- 8% to ICU, no vent ($34K, 10% mort)
- 28% to ICU on vent ($61K, 31% mort).
Seem reasonable. 2/
It assumes remdesivir shortens recovery time.
It also assumes remdesivir reduces mortality by *31%* based on the NIH announcement re ACCT: "a mortality rate of 8.0% for the group receiving remdesivir versus 11.6% for the placebo group (p=0.059)."
That's a high bar.
3/
It also assumes remdesivir reduces mortality by *31%* based on the NIH announcement re ACCT: "a mortality rate of 8.0% for the group receiving remdesivir versus 11.6% for the placebo group (p=0.059)."
That's a high bar.
3/
If you remember only ONE thing about cost-effectiveness analyses it should be this: models are *exquisitely sensitive* to survival estimates.
That's not a bug - it's a feature. Our conventional willingness-to-pay is based on prolonged survival.
More QALYs = more $$$
4/
That's not a bug - it's a feature. Our conventional willingness-to-pay is based on prolonged survival.
More QALYs = more $$$
4/
Let's assume a $100,000 per QALY threshold.
Never mind, this pandemic is making me feel generous - let's assume a $150,000 per QALY threshold.
Per @icer_review, the "value-based price" for #remdesivir is $52,900 (if 31%⬇️ in mortality) vs $1,200 (if no ⬇️in mortality).
5/
Never mind, this pandemic is making me feel generous - let's assume a $150,000 per QALY threshold.
Per @icer_review, the "value-based price" for #remdesivir is $52,900 (if 31%⬇️ in mortality) vs $1,200 (if no ⬇️in mortality).
5/
When you're talking about treating a LARGE number of patients, the difference between $53K and $1.2K per course is enormous.
For context,
53K = per capita GDP of the Netherlands
1.2K = per capita GDP of Kyrgyztan
6/
For context,
53K = per capita GDP of the Netherlands
1.2K = per capita GDP of Kyrgyztan
6/
Even within the US, $1.2K vs $53K per treatment will make the difference between effectively and equitably trouncing the #COVID19 epidemic, or pricing out a lot of people.
7/
7/
Note, this analysis does NOT consider:
-The societal value of an effective Rx (e.g., may boost consumer confidence; wake-up the economy). If it did, price would be ⬆️⬆️
-The $$$ that the US govt is providing for the high-risk trials right now. if it did, price may be ⬇️⬇️
8/
-The societal value of an effective Rx (e.g., may boost consumer confidence; wake-up the economy). If it did, price would be ⬆️⬆️
-The $$$ that the US govt is providing for the high-risk trials right now. if it did, price may be ⬇️⬇️
8/
This does NOT apply to asymptomatic pts or symptomatic pts managed as outpatients.
If it did, the price would have to be substantially lower. Unless it reduces person-to-person transmission. Lots of ??
For now, let's assume that #remdesivir is restricted to inpatients.
9/
If it did, the price would have to be substantially lower. Unless it reduces person-to-person transmission. Lots of ??
For now, let's assume that #remdesivir is restricted to inpatients.
9/
Key take-away:
Need to understand effect of #remdesivir on survival among hospitalized patients with #COVID19 to better define the value-based price.
Cost-effectiveness, as the saying goes, is first and foremost about effectiveness.
END
Need to understand effect of #remdesivir on survival among hospitalized patients with #COVID19 to better define the value-based price.
Cost-effectiveness, as the saying goes, is first and foremost about effectiveness.
END
