Bayesian predictive probabilities can be used for interim monitoring of clinical trials to estimate the probability of observing a statistically significant treatment effect if the trial were to continue to its predefined maximum sample size.
Purpose
We explore settings in which Bayesian predictive probabilities are advantageous for interim monitoring compared to Bayesian posterior probabilities, p-values, conditional power, or group sequential methods.
Results
For interim analyses that address prediction hypotheses, such as futility monitoring and efficacy monitoring with lagged outcomes, only predictive probabilities properly account for the amount of data remaining to be observed in a clinical trial and have the flexibility to incorporate additional information via auxiliary variables.
Limitations
Computational burdens limit the feasibility of predictive probabilities in many clinical trial settings. The specification of prior distributions brings additional challenges for regulatory approval.
Conclusions
The use of Bayesian predictive probabilities enables the choice of logical interim stopping rules that closely align with the clinical decision-making process.
Get full access to this article
View all access options for this article.
References
1.
SimonR. Optimal two-stage designs for phase II clinical trials. Control Clin Trials1989; 10: 1–10.
2.
LeeJJFengL. Randomized phase II designs in cancer clinical trials: Current status and future directions. J Clin Oncol2005; 23: 4450–57.
3.
O’BrienPCFlemingTR. A multiple testing procedure for clinical trials. Biometrics1979; 35: 549–56.
4.
PocockSJ. Group sequential methods in the design and analysis of clinical trials. Biometrika1977; 64: 191–99.
5.
DeMetsDL. Interim analysis: The alpha spending function approach. Stat Med1994; 13: 1341–52.
6.
ThallPFSimonR. Recent developments in the design of phase II clinical trials. Cancer Treat Res1995; 75: 49–71.
7.
ThallPFSimonR. A Bayesian approach to establishing sample size and monitoring criteria for phase II clinical trials. Control Clin Trials1994; 15: 463–81.
SylvesterRJ. A Bayesian approach to the design of phase II clinical trials. Biometrics1998; 44: 823–36.
10.
HeitjanDF. Bayesian interim analysis of phase II cancer clinical trials. Stat Med1997; 16: 1791–802.
11.
TanSBMachinD. Bayesian two-stage designs for phase II clinical trials. Stat Med2002; 21: 1991–2012.
12.
SpiegelhalterDJFreedmanLSParmarMK. Bayesian approaches to randomized trials. J R Stat Soc Ser A Stat Soc1994; 157: 357–416.
13.
BerryDA. Interim analyses in clinical trials: Classical vs. Bayesian approaches. Stat Med1985; 4: 521–26.
14.
BerryDA. A case for Bayesianism in clinical trials. Stat Med1993; 12: 1377–93.
15.
BerryDA. Bayesian statistics and the efficiency and ethics of clinical trials. Stat Sci2004; 19: 175–87.
16.
BerryDA. Bayesian clinical trials. Nat Rev Drug Discov2006; 5: 27–36.
17.
StanglDKBerryDA. Bayesian statistics in medicine: Where are we and where should we be going?Indian J Stat1998; 60: 176–95.
18.
LeeJJChuCT. Bayesian clinical trials in action. Stat Med2012; 31: 2955–72.
19.
LanKKGSimonRHalperinM. Stochastically curtailed tests in long-term clinical terms. Commun Stat Seq Anal1982; 1: 207–19.
20.
BetenskyRA. Early stopping to accept H0 based on conditional power: Approximations and comparisons. Biometrics1997; 53: 794–806.
21.
DmitrienkoAWangMD. Bayesian predictive approach to interim monitoring in clinical trials. Stat Med2006; 25: 2178–95.
22.
EmersonSSKittelsonJMGillenDL. On the Use of Stochastic Curtailment in Group Sequential Clinical Trials. University of Washington, Seattle, WA, 2005.
23.
ChoiSCSmithPJBeckerDP. Early decision in clinical trials when the treatment differences are small. Experience of a controlled trial in head trauma. Control Clin Trials1985; 6: 280–88.
ChoiSCPepplePA. Monitoring clinical trials based on predictive probability of significance. Control Clin Trials1989; 45: 317–23.
26.
BerryDA. Monitoring accumulating data in a clinical trial. Biometrics1989; 45: 1197–211.
27.
LeeJJLiuDD. A predictive probability design for phase II cancer clinical trials. Clin Trials2008; 5: 93–106.
28.
SpiegelhalterDJFreedmanLS. A predictive approach to selecting the size of a clinical trial, based on subjective clinical opinion. Stat Med1986; 5: 1–13.
29.
GrieveAP. Predictive probability in clinical trials. Biometrics1991; 47: 323–30.
30.
JohnsDAndersenJS. Use of predictive probabilities in phase II and phase III clinical trials. J Biopharm Stat1999; 9: 67–79.
31.
HersonJ. Predictive probability early termination plans for phase II clinical trials. Biometrics1979; 35: 775–83.
32.
LecoutreBDerzkoGGrouinJM. Bayesian predictive approach for inference about proportions. Stat Med1995; 14(9): 1057–63.
33.
BroglioKRConnorJTBerrySM. Not too big, not too small: A Goldilocks approach to sample size selection. J Biopharm Stat24(3) : 685–705.
34.
WhiteWBGradyDGiudiceLC. A cardiovascular safety study of LibiGel (testosterone gel) in postmenopausal women with elevated cardiovascular risk and hypoactive sexual desire disorder. Am Heart J2012; 163(1): 27–32.
35.
WilberDJPapponeCNeuzilPPaolaAD. Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with Paroxysmal Atrial Fibrillation: A randomized controlled trial. JAMA2012; 303(4): 333–40.
36.
FriedlinBKornEL. A comment on futility monitoring. Control Clin Trials2002; 23: 355–66.
37.
ConnorJElmJBroglioK. Bayesian adaptive trials offer advantages in comparative effectiveness trials: An example in status epilepticus. J Clin Epidemiol2013; 66(8): S130–37.
38.
JulianTBBlumencranzPDeckK. Novel intraoperative molecular test for sentinel lymph node metastases in patients with early-stage breast cancer. J Clin Oncol2008; 26(20): 3338–45.
39.
DeMetsDL. Futility approaches to interim monitoring by data monitoring committees. Clin Trials2006; 3: 522–29.
40.
EmersonSSFlemingTR. Symmetric group sequential test designs. Biometrics1989; 45: 905–32.
41.
KittelsonJMEmersonSS. A unifying family of group sequential test designs. Biometrics1999; 55: 874–82.
42.
FlemingTRHarringtonDPO’BrienPC. Designs for group sequential tests. Control Clin Trials1984; 5: 348–61.
43.
AndersonJRHighR. Alternatives to the standard Fleming, Harrington, and O’Brien futility boundary. Clin Trials2011; 8: 270–76.
44.
GreenSJDahlbergS. Planned versus attained design in phase II clinical trials. Stat Med1992; 11: 853–62.
45.
EllenbergSSEisenbergerMA. An efficient design for phase III studies of combination chemotherapies. Cancer Treat Rep1985; 69(10): 1147–54.
46.
WeiandSSchroederGO’FallonJR. Stopping when the experimental regimen does not appear to help. Stat Med1994; 13: 1453–58.
47.
SkrivanekZBerrySBerryD. Application of adaptive design methodology in development of a long acting glucagon-like peptide-1 analog (dulaglutide): Statistical design and simulations. J Diabetes Sci Technol2012; 6(6): 1305–18.
48.
BerrySMCarlinBPLeeJJMullerP. Bayesian Adaptive Methods for Clinical Trials. Taylor & Francis Group, Boca Raton, FL, 2011.
49.
YingGSHeitjanDF. Prediction of event times in the REMATCH trial. Clin Trials2013; 10: 197–206.
50.
SpiegelhalterDJAbramsKRMylesJP. Bayesian Approaches to Clinical Trials and Health-Care Evaluation. John Wiley & Sons, Ltd, Chichester, 2004.
51.
KassREWassermanL. The selection of prior distributions by formal rules. J Am Stat Assoc1996; 90: 928–34.
