Collecting information about mean and median approval times for abbreviated new drug applications (ANDAs) is very difficult, time consuming and frustrating. We looked at the 31 approvals in June 2013 to see how the Office of Generic Drugs (OGD) was doing. What we found is that, due to a few factors, there is really no good way to easily calculate approval times.. The first issue is that FDA does not post many ANDA approval letters in Drugs@FDA. The approval letter is essential to provide an accurate ANDA receipt date. Without access to that letter, those of us on the outside must go through a complicated algorithm to estimate the submission date. This process is further complicated knowing that ANDAs submitted in electronic format are pre-assigned ANDA numbers up to 60 days prior to receipt. In addition, at least 90-100 applications numbers a year are assigned to New Drug Applications (NDAs), ever since FDA began assigning consecutive serial numbers to all NDA and ANDA applications starting with 200000. Remember, ANDAs used to have their own unique numbering system to distinguish them from NDAs. Now, it is not so easy to tell the application type from the serial number. So how did we do it?
We had to rely on the OGD statistical reports to find submission dates associated with tentative approvals and first time approvals, which is sometimes (but not always) reported in the OGD monthly statistical reports. When an ANDA number of one of those applications very close to one of the June approval ANDA numbers was identified, we estimated when the ANDA was submitted by knowing the number of ANDAs submitted each month. When we could not get close enough, we actually searched by ANDA number for an ANDA that was approved that had an approval letter posted on Drugs@FDA and calculated the submission date based on the number of ANDAs submitted that month. It was complicated!
Bottom line is, we may be a month or two off in the calculations, but, for the most part, it is believed that the numbers are fairly reflective of the approval times, at least for the month of June 2013.
In addition, in order not to place OGD at a disadvantage for ANDAs that were first tentatively approved (meaning all of the scientific review was found acceptable, but the ANDA could not be fully approved until a period of patent or exclusivity expired), we used the date of tentative approval as the date to calculate the approval time, and not the date of final approval, which could have been months or years after tentative approval.
Here are the results – The mean (average) approval time of the 31 applications was 36 months from date of receipt. The median time was calculated to be 35 months. The shortest approval time was 17 months and the longest was 89 months. Clearly, this is a brief snapshot in time, but it is something that should be relevant to both industry and FDA representatives. Clearly, if we did the numbers for the entire year, there might be some changes, but if we consider that the backlog of ANDAs is fairly representative and consistent, generally, for all ANDAs from month to month, this is certainly a gauge of reality. It is also consistent with what we have been hearing from industry and FDA at their annual meetings.
We would have liked to provide the July numbers as well; however, we understand that OGD has temporarily stopped releasing its monthly statistics due to an inconsistency between the report OGD releases and another internal FDA report. Hopefully, that will be resolved soon because these numbers are the only concrete view that the industry has into OGD’s performance. This certainly is one piece of information that makes transparency seem alive and well and, now, at least for the short term, that piece of information is not available. As a suggestion to OGD, since the FDA database obviously collects submission dates, as well as approval dates, if OGD published the submission date (or the approval letters) for all full and tentative approvals, industry would have a way to effectively and accurately calculate approval times for all applications.
