akash
☆

India,
2019-02-23 08:39

Posting: # 19967
Views: 626

## Why BE testing use 90 CI [General Sta­tis­tics]

Why Bioequivalence testing makes use of the 90 CI why not 95 CI which gives more accuracy?

ElMaestro
★★★

Denmark,
2019-02-23 10:18

@ akash
Posting: # 19968
Views: 586

## Why BE testing use 90 CI

Hi Akash,

» Why Bioequivalence testing makes use of the 90 CI why not 95 CI which gives more accuracy?

More accuracy, what does that mean?

We have some more or less empirically justified limits of 80.00%-125.00%.
And we want a 5% risk of making the wrong conclusions in the sense of regulatory (patient's) risk; in practice this means we adopt a policy of a 5% risk of approving a product that is not BE. This is where the 90% CI comes into the equation. There is a (not more than) 5% risk associated with it (1-2*alpha).

A 95% CI would be less risky, ie. up to 2.5% chance of approving a non-BE product. Why would we want that, then alpha=5% seems to work just fine?

``` if (3) 4 x=c("Foo", "Bar") b=data.frame(x) typeof(b[,1]) ##aha, integer? b[,1]+1 ##then let me add 1```

Best regards,
ElMaestro

“(...) targeted cancer therapies will benefit fewer than 2 percent of the cancer patients they’re aimed at. That reality is often lost on consumers, who are being fed a steady diet of winning anecdotes about miracle cures.” New York Times (ed.), June 9, 2018.
Ohlbe
★★★

France,
2019-03-03 23:07

@ ElMaestro
Posting: # 19992
Views: 415

## Why BE testing use 90 CI

Dear Akash,

» This is where the 90% CI comes into the equation. There is a (not more than) 5% risk associated with it (1-2*alpha).

My understanding of the reason why: the true T/R ratio cannot be at the same time lower than 0.8 and higher than 1.25. So even if you test the lower limit with a 5 % risk, and the higher limit with a 5 % risk, the overall risk still remains 5 % for the patients, not 10 %.

Regards
Ohlbe
Helmut
★★★

Vienna, Austria,
2019-03-04 01:20

@ Ohlbe
Posting: # 19996
Views: 416

## one-sided / two-sided

Hi Ohlbe,

» […] So even if you test the lower limit with a 5 % risk, and the higher limit with a 5 % risk, the overall risk still remains 5 % for the patients, not 10 %.

Correct.

@Akash: Maybe you were confused by one-sided superiority testing in phase III (which is performed at an α-level of 5%). In other words, if patients are treated with the originator’s product, there is a 5% risk that it does not perform better (more efficient and/or safer) than placebo.
If we would test for BE at the 2.5% level (95% CI) we would be overly strict and at the same time gain absolutely nothing in terms of the patient’s risk.

Cheers,
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. ☼
Science Quotes
ElMaestro
★★★

Denmark,
2019-03-04 09:04

@ Helmut
Posting: # 19997
Views: 390

## one-sided / two-sided

Hi Hötzi,

» In other words, if patients are treated with the originator’s product, there is a 5% risk that it does not perform better (more efficient and/or safer) than placebo.

There is no more than a 5% risk of approving the test product (=concluding superiority) if the test product is not superior.

» If we would test for BE at the 2.5% level (95% CI) we would be overly strict and at the same time gain absolutely nothing in terms of the patient’s risk.

The patient's risk would be halved (max 2.5% chance of approving a non-BE product; CIs would get wider).

``` if (3) 4 x=c("Foo", "Bar") b=data.frame(x) typeof(b[,1]) ##aha, integer? b[,1]+1 ##then let me add 1```

Best regards,
ElMaestro

“(...) targeted cancer therapies will benefit fewer than 2 percent of the cancer patients they’re aimed at. That reality is often lost on consumers, who are being fed a steady diet of winning anecdotes about miracle cures.” New York Times (ed.), June 9, 2018.
Helmut
★★★

Vienna, Austria,
2019-03-04 11:25

@ ElMaestro
Posting: # 19998
Views: 377

## one-sided / two-sided

Hi ElMaestro,

sure. What I meant is: If a drug-naïve patient is treated with the originator’s product, the risk is 5%. With a 90% in BE the risk is also 5%.

Cheers,
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. ☼
Science Quotes
Bioequivalence and Bioavailability Forum |  Admin contact
19,437 posts in 4,125 threads, 1,325 registered users;
online 7 (1 registered, 6 guests [including 5 identified bots]).
Forum time (Europe/Vienna): 01:49 CEST

On two occasions I have been asked,—“Pray, Mr. Babbage,
if you put into the machine wrong figures,…
will the right answers come out?”

I am not able rightly to apprehend the kind of confusion of ideas
that could provoke such a question.    Charles Babbage

The BIOEQUIVALENCE / BIOAVAILABILITY FORUM is hosted by
Ing. Helmut Schütz