Impact of Clinical Center Variation on Efficiency of Exploratory Umbrella Design

  • Fang LiuEmail author
  • Nicole Li
  • Wen Li
  • Cong Chen


With the rapidly evolving landscape of cancer immunotherapy, traditional oncology trials that investigate one new treatment for one type of cancer within a trial face constraints due to the high cost and slow progress. New strategies have been developed over the last several years to help expedite the drug development process. One of these strategies, umbrella design, tests the effect of multiple investigational treatments in patients with the same type of cancer. When setting up an umbrella trial, a set of clinical centers will be selected for all the investigational treatments. Since trial outcomes may vary across clinical centers, clinical center variation plays a big role in the success of a trial. In this article, we evaluated the impact of clinical center variation on the efficiency of an umbrella trial where clinical centers are shared among all investigational treatments, compared to that of the traditional approach where each experimental drug is evaluated in a separate trial using different clinical centers. We demonstrate mathematically that the umbrella trial setting is more efficient than the tradition trials in terms of identifying more efficacious drugs, as sharing clinical centers among investigational treatments can reduce the impact of clinical center variation. In addition, guidance is provided on center allocation strategies, center selection, and center enrollment caps during the design stage, to further improve the efficiency of the umbrella trial. The conclusions are applicable to the clinical trials with binary endpoints and shed light on the trials with other types of endpoints.


Umbrella trial Multi-center trials Center effects Center variation Clinical center selection Enrollment cap 


  1. 1.
    Akihiro H, Junichi A, Hiroyuki S, Satoshi T (2018) Master protocol trials in oncology: review and new trial designs. Contemp Clin Trials Commun 12:1–8CrossRefGoogle Scholar
  2. 2.
    Chen C, Tipping RW (2002) Confidence interval of a proportion with over-dispersion. Biometr J 44(7):877–886MathSciNetCrossRefGoogle Scholar
  3. 3.
    Clopper CJ, Pearson ES (1934) The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 26:404–413CrossRefGoogle Scholar
  4. 4.
    Goldstein H (1995) Multilevel statistical models. Arnold, LondonzbMATHGoogle Scholar
  5. 5.
    ICH Harmonised Tripartite Guideline (1998) Statistical principles for clinical trials E9.
  6. 6.
    Kahan BC (2014) Accounting for centre-effects in multicentre trials with a binary outcome—when, why, and how? BMC Med Res Methodol 14(1):20CrossRefGoogle Scholar
  7. 7.
    Kahan BC, Harhay MO (2015) Many multicentre trials had few events per centre, requiring analysis via random-effects models or GEEs. J Clin Epidemiol 68(12):1504–1511CrossRefGoogle Scholar
  8. 8.
    Liu F, Li W, Chen C (2018) Impact of clinical center effects on objective response rate. In: JSM proceedings, biopharmaceutical section, pp 644–653Google Scholar
  9. 9.
    McCulloch CE, Searle AFM (2001) Generalized, linear and mixed models. Wiley, New YorkzbMATHGoogle Scholar
  10. 10.
    Ramaswamy G, Sumithra JM, David EG, Geoffrey RO, Suzanne ED, Shakun M, Margaret M, Jeffrey SA, Pasi AJ, David RG, Suresh SR, Everett EV (2015) ALCHEMIST trials: a golden opportunity to transform outcomes in early stage non-small cell lung cancer. Clin Cancer Res 1(24):5439–5444Google Scholar
  11. 11.
    SAS Institute Inc. (2018) SAS/STAT® 15.1 user’s guide. SAS Institute Inc., CaryGoogle Scholar
  12. 12.
    Tang J, Pearce L, O’Donnell-Tormey J, Hubbard-Lucey V (2018) Trends in the global immune-oncology landscape. Nat Rev Drug Discov 17:783–784CrossRefGoogle Scholar
  13. 13.
    Tang J, Shalabi A, Hubbard-Lucey VM (2017) Comprehensive analysis of the clinical immuno-oncology landscape. Ann Oncol 29:84–91CrossRefGoogle Scholar
  14. 14.
    Trusheim MR, Shrier AA, Antonijevic Z, Beckman RA, Campbell RK, Chen C, Flaherty KT, Loewy J, Lacombe D, Madhavan S, Selker HP, Esserman LJ (2016) PIPELINEs: creating comparable clinical knowledge efficiently by linking trial platforms. Clin Pharmacol Ther 100(6):713–729CrossRefGoogle Scholar
  15. 15.
    Woodcock J, LaVange ML (2017) Master protocols to study multiple therapies, multiple disease, or both. N Engl J Med 377:62–70CrossRefGoogle Scholar
  16. 16.
    Yamaguchi T, Ohashi Y (1999) Investigating centre effects in a multi-center clinical trial of superficial bladder cancer. Stat Med 18:1961–1971CrossRefGoogle Scholar
  17. 17.
    Yamaguchi T, Ohashi Y, Matsuyama Y (2002) Proportional hazards models with random effects to examine centre effects in multicentre cancer clinical trials. Stat Methods Med Res 11:221–236CrossRefGoogle Scholar

Copyright information

© International Chinese Statistical Association 2019

Authors and Affiliations

  1. 1.Merck & Co., Inc.KenilworthUSA

Personalised recommendations