14.Randomization02.Bayesian adaptive - sporedata/researchdesigneR GitHub Wiki

1. Bayesian adaptive trials

   • Bayesian adaptive trials are used in situations where researchers are interested in: (1) an average reduction in final study sample sizeUsing Bayesian Adaptive Trial Designs for Comparative Effectiveness Research: A Virtual Trial Execution; (2) the flexibility to change the trial (dropping an arm, calling it futile or a successful) [26]; (3) increase the modified intervention dose; (4) have an extended ability to drill down on secondary questions using multilevel Bayesian models; (5) multiple interim analysis without a p-value penalty at the end (for example, through the use of alpha spending functions. Since there is no need for adjustment for multiple comparisons [1] [2].
   • The main disadvantages include a need for extensive simulation before committing to a given design [3], more analytical work with each interim analysis (including an increase in computational capacity for complex models), a lower probability of finding positive associations when analyzing secondary questions, and the requirement for substantial computer power [26].
   • Adaptive seamless trials are a subtype of adaptive trials allowing for a fast transition between phase II and phase III trials, using prior information from phase II for phase III trials.

1. Short-term outcomes or proxies for long-term outcomes.
2. Regulatory agencies willing to accept a Bayesian approach.
3. Ability and funds to conduct a prospective study.
4. Equipoise or willingness to randomize at a ratio different from 1:1, when one of the adaptations involves a change in randomization ratios.

• The HPDI (highest posterior density interval) is the smallest interval with a probability -- for example, 95% -- for a parameter of interest. It is interpreted as 95% sure that the parameter is in this interval.
• An adaptive seamless trial is a subtype of an adaptive trial, combining two or more phases into one adaptive study design [4].
• The adaptation components of these trials are usually applied to the following design choices:
  1. Sample size is reduced [5]-- Using Bayesian Adaptive Trial Designs for Comparative Effectiveness Research: A Virtual Trial Execution. During an interim analysis, one can determine success or futility. With success, the trial is stopped since the intervention is deemed better than the control. With futility, the trial is stopped since simulations determine that the intervention is very unlikely to be better than the control even if the trial is completed. This decision can be made both in relation to completing a phase III trial as well as in relation to the decision to transition from phase II to a phase III trial [26].
  2. Adverse events. Decisions are made in relation to continuing or not the trial as a function of the adverse event rate being above or below a pre-established threshold.
  3. Intervention dose. The trial starts with three or more doses, and a decision is made in relation to dropping one of these arms since it is considered less effective than the remaining. The advantage is that from that point on the remaining sample will be allocated to only two arms, thus accelerating trial completion.
  4. Different interventions. Here we have the same rationale as outlined for intervention doses but applied to trials with three or more arms. In other words, during an interim analysis, one arm is deemed as less effective and therefore dropped from the trial.
  5. Different patient populations In this scenario, trial subpopulations are compared to identify the ones where the intervention might be more or less effective. The decision is then to whether from that point on the trial should make its inclusion/exclusion criteria more stringent. For example, one could eliminate the enrollment of patients with a long time between symptom onset and seeking care, or excluding patients with a higher risk of complications.
  6. Change allocation ratio between intervention and control. An alternative to dropping an arm is to change the allocation ratio. For example, in a two-arm trial where the intervention is demonstrating signs of superior effectiveness, one could, for example, change the allocation ratio from 1:1 to 2:1. This modification would lead to trial completion to be accelerated.
  7. Use information from a trial to its next phase. In seamless trials, there is a pre-planned transition between phase I to II and phase II to III [6].

Of importance, the complete trial adaptations protocol is decided upfront, during the initial trial simulation [7]. In a typical scenario, tens of thousands of different trial design possibilities are simulated to choose the design that is most likely to:

• Improve efficiency and reduce cost
• Maximize the information from the data
• Minimize risk to subjects and sponsors

Reporting guidelines include:

• The JASP Guidelines for Conducting and Reporting a Bayesian Analysis [8].
• The CONSORT guideline and its extensions [9] mostly apply to Bayesian adaptive trials.
• Proposal for CONSORT extension to adaptive trials [10].
• PCORI Standards for Adaptive and Bayesian Trial Designs [11].
• PCORI Standards for Studies of Complex Interventions [12].
• The SPIRIT Initiative (Standard Protocol Items: Recommendations for Interventional Trials) [13].
• Guidelines for the Content of Statistical Analysis Plans in Clinical Trials [14].

• Rstan: R Interface to Stan
• brms: Bayesian Regression Models using 'Stan'
• tidybayes: Tidy Data and 'Geoms' for Bayesian Models

• Platform trials since the advantages of Bayesian adaptive trials are particularly interesting when conducting multiple trials based on a single platform
• Causal analysis of real-world data to feed prior probabilities into a pre-trial simulation.

1. Books

   • Statistical Rethinking: A Bayesian Course with Examples in R and STAN [15].
   • Bayesian Adaptive Methods for Clinical Trials [16].
   • Bayesian Data Analysis [17].
   • Data Analysis Using Regression and Multilevel/Hierarchical Models [18].
   • Doing Bayesian Data Analysis: A Tutorial [19].

2. Articles combining theory and scripts

   • Adaptive designs in clinical trials: why use them, and how to run and report them [20].
   • Bayesian Adaptive Trials for Comparative Effectiveness Research: An Example in Status Epilepticus [21].
   • Using Bayesian adaptive designs to improve phase III trials: a respiratory care example [22] - demonstrates the steps associated with a full trial simulation through the FACTS software by Berry Consulting.
   • Do Bayesian adaptive trials offer advantages for comparative effectiveness research? Protocol for the RE-ADAPT study [23].
   • Treatment effects in multicenter randomized clinical trials [24].
   • Bayesian analysis: Using prior information to interpret the results of clinical trials [25].
   • Bayesian Adaptive Designs.
   • Common references for randomized designs

The frequentist concepts of p-value and confidence intervals are replaced by statements involving credible intervals, which are simpler to interpret.

Since Bayesian multi-level models do not require multiple comparison adjustments, and can also be used for small groups of patients. More extensive drill-down analyses are possible. See the table below. Source.

• Bayesian adaptive

[1] Gelman A, Hill J, Yajima M. Why we (usually) don't have to worry about multiple comparisons. Journal of Research on Educational Effectiveness. 2012 Apr 1;5(2):189-211.

[2] Pallmann P, Bedding AW, Choodari-Oskooei B, Dimairo M, Flight L, Hampson LV, Holmes J, Mander AP, Sydes MR, Villar SS, Wason JM. Adaptive designs in clinical trials: why use them, and how to run and report them. BMC medicine. 2018 Dec;16(1):1-5.

[3] Hansen CH, Warner P, Walker A, Parker RA, Whitaker L, Critchley HO, Weir CJ. A practical guide to pre‐trial simulations for Bayesian adaptive trials using SAS and BUGS. Pharmaceutical statistics. 2018 Nov;17(6):854-65.

[4] Laterre PF, Berry SM, Blemings A, Carlsen JE, François B, Graves T, Jacobsen K, Lewis RJ, Opal SM, Perner A, Pickkers P. Effect of Selepressin vs Placebo on Ventilator- and Vasopressor-Free Days in Patients With Septic Shock: The SEPSIS-ACT Randomized Clinical Trial. Jama. 2019 Oct 15;322(15):1476-85.

[5] Ryan EG, Bruce J, Metcalfe AJ, Stallard N, Lamb SE, Viele K, Young D, Gates S. Using Bayesian adaptive designs to improve phase III trials: a respiratory care example. BMC medical research methodology. 2019 Dec 1;19(1):99.

[6] Laterre PF, Berry SM, Blemings A, Carlsen JE, François B, Graves T, Jacobsen K, Lewis RJ, Opal SM, Perner A, Pickkers P. Effect of Selepressin vs Placebo on Ventilator- and Vasopressor-Free Days in Patients With Septic Shock: The SEPSIS-ACT Randomized Clinical Trial. Jama. 2019 Oct 15;322(15):1476-85.

[7] Hansen CH, Warner P, Walker A, Parker RA, Whitaker L, Critchley HO, Weir CJ. A practical guide to pre‐trial simulations for Bayesian adaptive trials using SAS and BUGS. Pharmaceutical statistics. 2018 Nov;17(6):854-65.

[8] van Doorn J, van den Bergh D, Bohm U, Dablander F, Derks K, Draws T, Evans NJ, Gronau QF, Hinne M, Kucharský Š, Ly A. The JASP Guidelines for Conducting and Reporting a Bayesian Analysis.

[9] Extensions of the CONSORT Statement

[10] Mistry P, Dunn JA, Marshall A. A literature review of applied adaptive design methodology within the field of oncology in randomised controlled trials and a proposed extension to the CONSORT guidelines. BMC medical research methodology. 2017 Dec 1;17(1):108.

[11] PCORI Standards for Adaptive and Bayesian Trial Designs. Posted: November 12, 2015; Updated: February 26, 2019.

[12] PCORI Standards for Studies of Complex Interventions. Posted: November 12, 2015; Updated: February 26, 2019.

[13] SPIRIT Initiative (Standard Protocol Items: Recommendations for Interventional Trials)

[14] Gamble C, Krishan A, Stocken D, Lewis S, Juszczak E, Doré C, Williamson PR, Altman DG, Montgomery A, Lim P, Berlin J. Guidelines for the Content of Statistical Analysis Plans in Clinical Trials. Jama. 2017 Dec 19;318(23):2337-43.

[15] McElreath R. Statistical Rethinking: A Bayesian Course with Examples in R and STAN and the companion https://bookdown.org/connect/#/apps/1850/access. CRC press; 2020 Mar 13.

[16] Berry SM, Carlin BP, Lee JJ, Muller P. Bayesian Adaptive Methods for Clinical Trials. CRC press; 2010 Jul 19.

[17] Gelman A, Carlin JB, Stern HS, Dunson DB, Vehtari A, Rubin DB. Bayesian Data Analysis. CRC press; 2013 Nov 1.

[18] Gelman A, Hill J. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge university press; 2006 Dec 18.

[19] Kruschke J. [Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan](https://www.amazon.com/Doing-Bayesian-Data-Analysis-Tutorial-ebook/dp/B00PYZ2VR6/ref=sr_1_4?dchild=1&keywords=gelman+bayesian&qid=1586718. Academic Press; 2014 Nov 11. 006&sr=8-4)

[20] Pallmann P, Bedding AW, Choodari-Oskooei B, Dimairo M, Flight L, Hampson LV, Holmes J, Mander AP, Sydes MR, Villar SS, Wason JM. Adaptive designs in clinical trials: why use them, and how to run and report them. BMC medicine. 2018 Dec;16(1):1-5.

[21] Connor JT, Elm JJ, Broglio KR, Esett and Adapt-It Investigators. Bayesian Adaptive Trials for Comparative Effectiveness Research: An Example in Status Epilepticus. Journal of clinical epidemiology. 2013 Aug 1;66(8):S130-7.

[22] Ryan EG, Bruce J, Metcalfe AJ, Stallard N, Lamb SE, Viele K, Young D, Gates S. Using Bayesian adaptive designs to improve phase III trials: a respiratory care example. BMC medical research methodology. 2019 Dec 1;19(1):99.

[23] Connor JT, Luce BR, Broglio KR, Ishak KJ, Mullins CD, Vanness DJ, Fleurence R, Saunders E, Davis BR. Do Bayesian adaptive trials offer advantages for comparative effectiveness research? Protocol for the RE-ADAPT study. Clinical Trials. 2013 Oct;10(5):807-27.

[24] Senn SJ, Lewis RJ. Treatment effects in multicenter randomized clinical trials. Jama. 2019 Mar 26;321(12):1211-2.

[25] Quintana M, Viele K, Lewis RJ. Bayesian analysis: Using prior information to interpret the results of clinical trials. Jama. 2017 Oct 24;318(16):1605-6.

[26] Granholm A, Kaas-Hansen BS, Lange T, Schjørring OL, Andersen LW, Perner A, Jensen AK, Møller MH. An overview of methodological considerations regarding adaptive stopping, arm dropping and randomisation in clinical trials. Journal of Clinical Epidemiology. 2022 Nov 17.