21P - Duration of breast cancer trials: Analysis of predicted versus actual completion date

Background

Clinical trial registries record the proposed and actual completion date of studies. Study duration affects budget, personnel, governance and ethics. To our knowledge, no data exists comparing estimated time to study completion with actual completion date. Using breast cancer trials as an exemplar, we analyzed differences between estimated and actual duration of clinical trials.

Methods

A sophisticated artificial intelligence (AI) tool, Risklick AI®^1-3 analysed clinical trial registry fields for proposed trial completion (entered at time of study registration) compared to date entered for study completion. This technology collects and unifies trial data from all 18 public registries, including clinicaltrials.gov and WHO. The comprehensive search performed on 15.3.23 for period 2000 – 2023 used 101 keywords to cover breast cancer subtypes in completed trials. Trials were analyzed according to factors postulated to impact duration: disease extent, histological subtype, phase, number of eligibility criteria, sample size, country and sponsor. Delay was defined as the additional time required to complete the trial compared to the estimated study completion date, with a tolerance of 30 days.

Results

582 trials, involving 104’384 patients, recorded both planned and actual completion dates. Overall, 21.6% were finished early, 7.9% on time, and 70.4% had delay, of mean 31.7 +/- 25.6 months [standard deviation], ranging 1 - 155 months. The mean delay was similar for trials analysed by disease extent, breast cancer subtype, phase, number of eligibility criteria, sample size, country, or sponsor type.

Conclusions

This large real-world dataset revealed an average delay in completion of 2.5-year for 70% of breast cancer trials, independent of multiple factors commonly thought to affect trial duration. This data should be considered when planning, budgeting and setting participant and community expectations for trials. Tools to accurately predict trial duration and efforts to reduce delays are needed.

References: 1. www.risklick.ch 2. Front Digit Health. 2021;3 3. Pharmacology. 2021;106(5-6):244-253.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

University of Bern.

Funding

Has not received any funding.

Disclosure

P. Khlebnikov, F. Meer, Q. Haas: Financial Interests, Personal, Full or part-time Employment: Risklick AG. P. Amini: Financial Interests, Personal, Ownership Interest: Risklick AG. E. Segelov: Financial Interests, Personal, Invited Speaker, Limbic are a medical educational company running various seminars which I have participated in as an invited speaker/Chair: Limbic; Financial Interests, Personal, Invited Speaker, webinar on colorectal cancer 28/02/21: Servier; Financial Interests, Institutional, Local PI, As head of the Oncology Cancer clinical trials at Monash Health (until 15 Aug 2022) where there were >200 trials active, there were the usual commercial arrangements to run the trials with the institution. There was no personal financial benefit from any: > 15 different companies. All other authors have declared no conflicts of interest.