balakotu ★ India, 2012-02-07 10:03 (4433 d 10:48 ago) Posting: # 8063 Views: 5,923 |
|
Dear All, Please help me below query We are looking to conduct Bioequivalence study on an ophthalmic product “Loteprednol Etabonate” Ophthalmic suspension. In this study the OGD recommended 90% CI and the acceptance range for Bioequivalence based upon 90% confidence interval of (Rt ±1.645 St) lies with in (0.8, 1.25). Here, What is St stands for? What is the significance of value 1.645 whether it is constant or not. How to develop Bootstrapping technique by manually or software? |
ElMaestro ★★★ Denmark, 2012-02-07 10:55 (4433 d 09:56 ago) @ balakotu Posting: # 8064 Views: 5,060 |
|
Hello balakotu, ❝ We are looking to conduct Bioequivalence study on an ophthalmic product “Loteprednol Etabonate” Ophthalmic suspension. Great. There's trouble ahead. ❝ In this study the OGD recommended 90% CI and the acceptance range for Bioequivalence based upon 90% confidence interval of (Rt ±1.645 St) lies with in (0.8, 1.25). ❝ What is St stands for? Standard deviation. ❝ What is the significance of value 1.645 whether it is constant or not. Try and google "1.645" and you will quickly get the explanation. Mean plusminus 1.645 st.dev. contains the confidence range from the (any) normal distribution. Consider it a constant as long as you work with something Gaussian. ❝ How to develop Bootstrapping technique by manually or software? By software. You need to sample with replacement "a lot of times", like 500, 5000 or 5 trillion gaziliion times. For that purpose you definitely need proper software with a NLYW. — Pass or fail! ElMaestro |
balakotu ★ India, 2012-02-07 12:12 (4433 d 08:39 ago) @ ElMaestro Posting: # 8066 Views: 4,854 |
|
Hello ElMaestro, how to construct 90% Confidence interval by using bootstrapping technique for bioequivalence assessment. |
ElMaestro ★★★ Denmark, 2012-02-07 13:11 (4433 d 07:40 ago) @ balakotu Posting: # 8068 Views: 4,758 |
|
Hi balakotu, ❝ how to construct 90% Confidence interval by using bootstrapping technique for bioequivalence assessment. Assume the data has the distribution corresponding to what you see in your observed sample. Resample your sample a lot of times. Each times collect the measure you are interested in. Once all resamples have been gathered, derive the 90% CI. If you use R then one place to start would be the 'boot' package. Search the web for examples. The package provides confidence interval options from user-defined models. You can also program it up from scratch, but most people would probably rather have colon surgery done on them with a rusty scissor than do such a thing. Only a guess from my side. — Pass or fail! ElMaestro |
pash413 ★ India, 2015-04-10 09:44 (3275 d 12:07 ago) @ ElMaestro Posting: # 14680 Views: 3,657 |
|
Hello All, I have couple of doubts regarding the discussion on Loteprednol bootstrapping C.I construction method.
Edit: Please don’t start new threads (see the Forum’s Policy). I deleted another post of 2015-04-14. [Helmut] |