Across 90 European Phase I-containing immunology trials, the median CTIS end-to-end EU submission-to-authorization timeline was 94 days (SD 40.9). Country-specific CTIS Part II took a median 128 days (SD 223.5) across 211 country observations, with a pronounced long-delay tail. The strongest observed delay signals were broader country scope, more sites, country review month, and cell-therapy involvement.
The end-to-end measure runs from the initial CTIS submission to first authorization. The median was 94 days and the SD was 40.9 days. Thirteen of 90 trials (14.4%) were authorized within 30 days, while 74 of 90 (82.2%) were authorized within 120 days.
A 90-day planning assumption covers 44 of 90 trials (48.9%). Expanding the operational allowance to 120 days covers 74 of 90 (82.2%), making 120 days a more defensible portfolio planning point for Phase I immunology CTIS submissions.
Country-specific Part II had a median of 128 days and an SD of 223.5 days. Fifty-one of 211 country reviews (24.2%) finished within 30 days, 90 (42.7%) within 90 days, and 124 (58.8%) within 180 days.
Part II is substantially less predictable than the end-to-end trial metric: 121 of 211 country observations (57.3%) lasted at least 90 days, and 48 (22.7%) exceeded 365 days. Portfolio planning should therefore use country-level contingencies rather than a single EU-wide Part II assumption.
Median country-specific Part II time ranged from 23 days in Norway to 540 days in Ireland. Among countries with at least five observations, Austria had the shortest median at 29 days, followed by Hungary at 54.5 days and Germany at 60 days; Portugal had the longest at 302 days.
| Country | n | Median days | SD | ≤60 days | ≤90 days |
|---|---|---|---|---|---|
| Norway | 3 | 23.0 | 241.5 | 66.7% | 66.7% |
| Austria | 5 | 29.0 | 242.1 | 60.0% | 60.0% |
| Denmark | 4 | 39.5 | 223.9 | 75.0% | 75.0% |
| Hungary | 6 | 54.5 | 213.5 | 50.0% | 66.7% |
| Germany | 32 | 60.0 | 255.0 | 50.0% | 53.1% |
| Czechia | 5 | 75.0 | 170.2 | 20.0% | 60.0% |
| Netherlands | 22 | 93.5 | 209.6 | 45.5% | 50.0% |
| Sweden | 4 | 98.0 | 93.6 | 50.0% | 50.0% |
| Belgium | 15 | 104.0 | 275.7 | 26.7% | 46.7% |
| Romania | 6 | 124.0 | 243.2 | 16.7% | 33.3% |
| Croatia | 3 | 137.0 | 194.0 | 33.3% | 33.3% |
| Poland | 14 | 143.0 | 213.2 | 35.7% | 42.9% |
| France | 26 | 144.5 | 240.4 | 26.9% | 34.6% |
| Italy | 12 | 146.0 | 252.1 | 25.0% | 33.3% |
| Spain | 31 | 161.0 | 228.8 | 29.0% | 35.5% |
| Finland | 1 | 163.0 | — | 0.0% | 0.0% |
| Bulgaria | 9 | 163.0 | 179.0 | 33.3% | 33.3% |
| Latvia | 1 | 171.0 | — | 0.0% | 0.0% |
| Greece | 5 | 188.0 | 142.8 | 0.0% | 0.0% |
| Lithuania | 1 | 235.0 | — | 0.0% | 0.0% |
| Portugal | 5 | 302.0 | 252.8 | 40.0% | 40.0% |
| Ireland | 1 | 540.0 | — | 0.0% | 0.0% |
Country selection materially changes schedule risk. Germany combined a 60-day median with 32 observations, while France, Spain and Italy had medians of 144.5, 161 and 146 days, respectively. The high SDs show that country medians should be paired with contingency buffers.
Geographic and site complexity produced the clearest end-to-end gradient. Single-country trials had a 74-day median, versus 117 days for trials spanning four or more countries. One-site trials had a 70-day median, versus 115 days for trials with six or more sites.
Scope complexity was the most operationally consistent signal: the Spearman correlation with end-to-end time was 0.42 for number of countries and 0.45 for total sites. A compact single-country, single-site EU submission profile was therefore associated with the highest probability of finishing below the 94-day cohort median.
Country site burden, Part II submission month and modality were the strongest country-level signals. One-site country applications had a 60.5-day median, compared with 174 days for country applications covering four or more sites.
For country-specific Part II, local site packaging appears more actionable than total trial size: the median rose by 113.5 days from one site to four or more sites, while participant count showed little monotonic association (ρ=0.03). Cell therapy and mid-year Part II submissions carried the largest observed delay tails.
Submission month showed a cohort pattern without using calendar year. August initial CTIS submissions had the shortest end-to-end median at 41.5 days, while February and March were 115 and 114 days. For country Part II, January was shortest at 22.5 days; June and October were longest at 273 and 195 days.
| Month | End-to-end | Country Part II | ||||
|---|---|---|---|---|---|---|
| n | Median | ≥90 days | n | Median | ≥90 days | |
| January | 6 | 99.0 | 66.7% | 18 | 22.5 | 11.1% |
| February | 5 | 115.0 | 80.0% | 18 | 85.5 | 50.0% |
| March | 6 | 114.0 | 83.3% | 15 | 175.0 | 60.0% |
| April | 6 | 62.5 | 16.7% | 18 | 240.0 | 55.6% |
| May | 7 | 59.0 | 42.9% | 14 | 74.5 | 42.9% |
| June | 4 | 118.5 | 75.0% | 23 | 273.0 | 78.3% |
| July | 10 | 75.5 | 50.0% | 18 | 162.5 | 55.6% |
| August | 8 | 41.5 | 12.5% | 24 | 176.5 | 62.5% |
| September | 8 | 93.5 | 50.0% | 14 | 98.0 | 50.0% |
| October | 9 | 72.0 | 44.4% | 21 | 195.0 | 85.7% |
| November | 6 | 89.5 | 50.0% | 10 | 110.5 | 60.0% |
| December | 15 | 105.0 | 60.0% | 18 | 150.0 | 61.1% |
Month should be treated as a scheduling signal rather than a causal mechanism. Nevertheless, the observed spread was operationally large: 73.5 days between the fastest and slowest end-to-end monthly medians, and 250.5 days between the fastest and slowest Part II monthly medians.
For Phase I immunology CTIS submissions, a 120-day end-to-end baseline covers 82.2% of trials, but country Part II needs a differentiated risk buffer. The most delay-prone profile combined four or more countries, six or more sites, and complex country packages with four or more sites.
The practical lever is scope control: sequence countries, minimize sites in the initial country package where scientifically feasible, and add country-specific buffers for cell therapy and months with historically longer Part II medians. These actions address the strongest observed CTIS delay associations without assuming that any single factor is causal.