FDA drug review time has stayed roughly flat over 20 years, but total time-to-market has grown due to longer clinical trials

Summary

It depends which clock you mean, and the two have moved in opposite directions. The FDA's own review clock -- the months between when a company submits a new drug application and when the FDA decides on it -- fell sharply in the 1990s (median 26.6 months before the 1992 Prescription Drug User Fee Act, PDUFA) and has been roughly flat since the mid-2000s: about 6-9 months for priority-reviewed drugs and 10-13 months for standard-reviewed drugs across 2005-2016, and a 2024 economy-wide median of 356 days (~11.7 months) that is essentially the same order of magnitude as 20 years ago. But the total time to bring a new drug from first-in-human testing to approval -- which is what most people mean by "getting a drug approved" -- has been getting somewhat LONGER over the same period, because clinical trials themselves (patient recruitment, protocol complexity, global multi-site trials) have grown more complex and taken longer, even as FDA's own decision time held steady or shrank slightly. Separately, a much larger share of new drugs now qualify for the faster priority/expedited tracks (59% of 2024 FDA approvals used at least one expedited pathway), so more drugs individually experience shorter review times even though neither track's own timeline has changed much.

What AskAmerica's own data could and could not answer

AskAmerica's health schema carries FDA drug-approval data (health.fda_drug_approvals, health.fda_drug_submissions) sourced from openFDA. I queried health.fda_drug_submissions directly for a known application (Lipitor, NDA020702) and confirmed live: the table records only the date each submission reached a status such as AP (approved) -- e.g. its ORIG submission shows submission_status_date = 19961217 -- with no separate filing/received date column anywhere in the schema or the underlying openFDA feed. That means review-time duration (submission-to-approval) genuinely cannot be computed from this corpus; it is a real data gap, not a query I chose not to write. Every review-time figure below instead comes from FDA's own Prescription Drug User Fee Act (PDUFA) performance reports and from published literature that assembled submission and approval dates directly from Drugs@FDA and FDA approval letters.

The FDA review clock: roughly flat since the mid-2000s

FDA's own official table of CDER approval times for Priority and Standard NDAs/BLAs (calendar years 1993-2016) shows the big drop happened BEFORE the last 20 years, driven by PDUFA (1992): median review time fell from double digits in the early 1990s to roughly 6 months (priority) / 12 months (standard) by 2000, and has stayed in that same band ever since:

A separate 2013-2023 study of small-molecule NME applications found first-cycle (no Complete Response Letter) approvals took a median of roughly 245 days (~8 months) from submission to approval -- consistent with the FDA's own table -- while applications that received a Complete Response Letter and had to be resubmitted took a median of about 709 days (~23 months), roughly 1.5-2 years longer. The post-CRL approval rate improved from 47% (2000-2012) to 64% (2013-2023), meaning fewer of the longest-delayed applications now fail outright, but that is a change in outcome, not in the length of the delay itself.

Cross-agency benchmarking (CIRS R&D Briefing 101, covering 2015-2024) puts FDA's 2024 median approval time (submission to decision, across all new active substances, priority and standard combined) at 356 days, or about 11.7 months -- close to the 2005 standard-review median of 11.8 months, and the shortest of the six major global regulators tracked (ahead of the EU's EMA at 430 days, Switzerland's Swissmedic at 444 days, and behind only Japan's PMDA at 290 days). CIRS's own framing: "Despite improvements over the past 20 years, NAS median approval times still vary across the six authorities" -- signaling that most of the U.S. improvement is a multi-decade story predating the last 20 years, not a within-the-last-20-years one.

What has grown: total time to develop and approve a drug

The FDA review period is only the last stage of a much longer process. Tufts Center for the Study of Drug Development (CSDD), comparing drugs approved in two adjacent cohorts (2008-13 vs. 2014-18), found that the combined clinical-trial-plus-approval phase length grew by 4.8 months on net -- clinical trial time rose by nearly 7 months while FDA approval time actually fell by close to 2 months. In other words, FDA's own piece of the process was getting a little faster over that window, but the growth in how long clinical trials themselves take (larger, more geographically dispersed trials; more complex protocols; more eligibility criteria) more than offset it. Separate Tufts work on trial operations found the duration from protocol approval to first-patient-first-visit rose 27.2% and enrollment duration rose 36.9% between 2010 and 2020, which is the likely mechanical driver.

Net effect for someone asking "how long does it take to get a new drug approved" in the everyday sense (first human dose to market): the honest answer is that it has not obviously gotten shorter over the last 20 years, and by some measures has modestly lengthened, even though the specific number the FDA controls -- its own review clock after a completed application is submitted -- has been essentially stable to slightly improved.

A separate trend: more drugs now qualify for the fast lane

One thing genuinely has shifted over 20 years: the SHARE of new drugs going through the faster expedited pathways (priority review, breakthrough therapy, fast track, accelerated approval) has grown substantially -- CIRS found 75% of FDA new-active-substance approvals in 2020-2024 involved at least one facilitated regulatory pathway, and 59% used at least Priority Review in 2024 alone, versus roughly 20 priority approvals/year in the mid-2010s per FDA's own program briefings. So while the priority-review timeline itself (a ~6-8 month median) hasn't changed much, more drugs now individually experience that faster number rather than the ~10-12 month standard-review number -- a real change in the population-average experience even without the underlying clock moving.

Every query behind this report

1 warehouse call ran in this session, in order. Each is reproducible against the same snapshot.

SELECT application_number, submission_type, submission_number, submission_status, submission_status_date FROM health.fda_drug_submissions WHERE application_number = 'NDA020702' ORDER BY submission_number LIMIT 10