Across 228 Phase II haematology trials, median EU/CTIS end-to-end review was 53.5 days (SD 43.7), and 127/228 trials (55.7%) reached first authorization within 60 days. Country-specific Part II was materially longer: median 243.0 days (SD 239.5) across 486 trial-country decisions, with only 150/486 (30.9%) completed within 60 days. August–September submissions, smaller country footprints, and plasma-cell or leukaemia/myeloid trials aligned with faster end-to-end review; November–December submissions, seven or more countries, non-malignant haematology, and complex Part II profiles aligned with delay.
End-to-end timing is an EU/CTIS trial-level measure and is therefore not assigned to individual countries. Part II timing is country-specific and measures the national ethics and local-document pathway from the earliest Part II submission to the latest recorded decision or authorization.
Country comparisons should use Part II only. Reporting end-to-end timing by country would incorrectly attribute the coordinated EU/CTIS authorization pathway to a single member state.
End-to-end review crossed 50% between 30 and 60 days: 79/228 trials (34.6%) were authorized within 30 days and 127/228 (55.7%) within 60. Part II crossed 50% only after 180 days: 215/486 decisions (44.2%) were completed within 180 days and 309/486 (63.6%) within one year.
A 60-day planning assumption covered 55.7% of end-to-end reviews but only 30.9% of Part II decisions. For Part II, even a 90-day assumption covered just 35.8%.
Among countries with at least five observed decisions, Norway had the lowest Part II median at 26.5 days, followed by Finland at 53.0 and the Netherlands at 63.0. Germany had the highest median at 354.5 days, followed by Greece at 332.5 and Belgium at 305.0.
| Country | Decisions | Median days | SD days |
|---|---|---|---|
| Iceland | 1 | 9.0 | — |
| Norway | 16 | 26.5 | 238.9 |
| Lithuania | 2 | 27.5 | 10.6 |
| Romania | 3 | 28.0 | 56.8 |
| Ireland | 1 | 31.0 | — |
| Finland | 11 | 53.0 | 192.6 |
| Netherlands | 29 | 63.0 | 269.3 |
| Bulgaria | 6 | 89.0 | 93.3 |
| Portugal | 1 | 99.0 | — |
| Czechia | 8 | 120.5 | 215.4 |
| Sweden | 21 | 150.0 | 234.8 |
| Austria | 7 | 162.0 | 181.1 |
| Poland | 17 | 168.0 | 255.2 |
| Hungary | 6 | 172.5 | 255.8 |
| France | 88 | 252.0 | 240.6 |
| Italy | 80 | 263.5 | 248.3 |
| Spain | 59 | 268.0 | 223.0 |
| Denmark | 25 | 282.0 | 221.0 |
| Estonia | 2 | 284.5 | 381.1 |
| Belgium | 29 | 305.0 | 242.4 |
| Greece | 16 | 332.5 | 240.0 |
| Germany | 58 | 354.5 | 244.8 |
Country was the dominant Part II differentiator: medians ranged from 9.0 to 354.5 days. For operational benchmarking, the more stable ≥5-observation range was 26.5 days in Norway to 354.5 days in Germany.
Submission month showed the clearest end-to-end timing pattern without using submission year. August and September submissions had medians of 23.5 and 31.0 days; November and December rose to 106.0 and 119.0 days. Across August–September, 39/54 trials (72.2%) were faster than the overall median and only 8/54 (14.8%) lasted at least 90 days.
November–December submissions had a combined median of 119.0 days and 23/34 (67.6%) lasted at least 90 days, versus a 28.5-day median and 8/54 (14.8%) ≥90 days for August–September. In the adjusted model, December, March, November, and February remained 64%–129% longer than September (all p≤0.035).
Country footprint and disease group were the strongest reproducible trial-level factors. Single-country submissions had a 42.0-day median, versus 109.0 days for submissions spanning seven or more countries. Non-malignant haematology had a 111.0-day median, compared with 31.0 days for plasma-cell disorders and 34.0 days for leukaemia/myeloid trials.
| Countries | N | Median | Fast | ≥90 d |
|---|---|---|---|---|
| 1 countries | 141 | 42.0 | 53.2% | 28.4% |
| 2–3 countries | 43 | 56.0 | 48.8% | 37.2% |
| 4–6 countries | 32 | 72.0 | 46.9% | 43.8% |
| 7+ countries | 12 | 109.0 | 25.0% | 66.7% |
| Disease group | N | Median | Fast | ≥90 d |
|---|---|---|---|---|
| Plasma-cell disorders | 36 | 31.0 | 63.9% | 22.2% |
| Leukaemia / myeloid | 77 | 34.0 | 61.0% | 27.3% |
| Lymphoma / WM | 67 | 56.0 | 41.8% | 34.3% |
| Other haematology | 23 | 75.0 | 43.5% | 34.8% |
| Non-malignant haematology | 25 | 111.0 | 24.0% | 72.0% |
Randomization (58.0 vs 51.0 days; p=0.258), open-label design (53.0 vs 54.0; p=0.669), adaptive design (45.0 vs 54.5; p=0.454), biomarker stratification (56.0 vs 53.0; p=0.698), total sites (ρ=0.10; p=0.138), and sample size (ρ=−0.02; p=0.783) showed no consistent end-to-end signal.
Part II delay reflected both national review environment and submission complexity. One-site country packages had a 116.0-day median, rising to 330.5 days for packages with at least ten sites. After accounting for country and the other screened factors, multimodality, pharmaceutical sponsorship, bispecific therapy, CRO involvement, adaptive design, and country site load remained associated with longer Part II timing.
| Sites in country | N | Median | Fast | ≥90 d |
|---|---|---|---|---|
| 1 sites | 108 | 116.0 | 61.1% | 53.7% |
| 2–4 sites | 168 | 253.5 | 49.4% | 66.7% |
| 5–9 sites | 105 | 263.0 | 47.6% | 63.8% |
| 10+ sites | 104 | 330.5 | 42.3% | 72.1% |
The practical Part II risk stack is cumulative: a multi-modality, bispecific, pharma-sponsored, CRO-supported submission with many sites carries several independent markers of delay. These markers should trigger earlier local-document preparation and larger country-specific schedule buffers.
The dataset supports four concrete planning rules for Phase II haematology CTIS submissions.
CTIS is the Clinical Trials Information System used for EU clinical trial applications. End-to-end is the interval from initial CTIS submission to first CTIS authorization at trial level. Part II is the interval from earliest country Part II submission to latest recorded country decision or authorization. “Fast” means shorter than the relevant median; substantial delay is reported at ≥60 and ≥90 days. Medians and sample standard deviations are shown. All analyses pooled the 228-trial cohort and excluded year as an explanatory variable.