Type I error probability spending for post–market drug and vaccine safety surveillance with binomial data

Silva, Ivair Ramos

doi:10.1002/sim.7504

Cited by 12 publications

(20 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, more recently, some results have shown that if the design criterion is expected time to signal, then a concave alpha spending function should be used instead. This is advocated, for example, in an earlier paper by Silva . That paper shows that usage of ρ = 0.5 favors near‐optimal expected time to signal when data follow a binomial distribution.…”

Section: Discussionmentioning

confidence: 99%

“…Studies directed to identify suitable alpha spending shapes for postmarket safety surveillance are just emerging. Seminal works in this direction are the developments of Silva and Kulldorff, where the minimization of expected time to signal was taken as the key design criterion. Their findings suggest that, unlike clinical trials, the alpha spending should have a concave shape for postmarket safety surveillance applications for both Poisson and binomial data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Alpha spending for historical versus surveillance Poisson data with CMaxSPRT

Silva

Lopes

Dias

et al. 2019

Statistics in Medicine

Self Cite

View full text Add to dashboard Cite

Sequential analysis hypothesis testing is now an important tool for postmarket drug and vaccine safety surveillance. When the number of adverse events accruing in time is assumed to follow a Poisson distribution, and if the baseline Poisson rate is assessed only with uncertainty, the conditional maximized sequential probability ratio test, CMaxSPRT, is a formal solution. CMaxSPRT is based on comparing monitored data with historical matched data, and it was primarily developed under a flat signaling threshold. This paper demonstrates that CMaxSPRT can be performed under nonflat thresholds too. We pose the discussion in the light of the alpha spending approach. In addition, we offer a rule of thumb for establishing the best shape of the signaling threshold in the sense of minimizing expected time to signal and expected sample size. An example involving surveillance for adverse events after influenza vaccination is used to illustrate the method.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Alpha spending for historical versus surveillance Poisson data with CMaxSPRT

Silva

Lopes

Dias

et al. 2019

Statistics in Medicine

Self Cite

View full text Add to dashboard Cite

show abstract

“…Although it is possible to begin testing earlier in the influenza season, there can be a substantial amount of noise in sequential testing when the outcome is rare, and there are alternative triggers and pathological mechanisms involved . Further investigation is needed to explore possible solutions including the use of an O'Brien‐Fleming threshold, a minimum number of events to signal, or a different alpha spending plan …”

Section: Discussionmentioning

confidence: 99%

“…23,24 Further investigation is needed to explore possible solutions including the use of an O'Brien-Fleming threshold, a minimum number of events to signal, or a different alpha spending plan. 21,25,26 USPRT's adjustment for both clinical and processing delay may be more suited to insurance claims databases as opposed to Influenza vaccines save lives and prevent many illnesses and their associated effects on morbidity, mortality, and economic losses. 30 Despite the possible risk of GBS due to the influenza vaccines, influenza infection seems to also be a relevant trigger for GBS, and thus, lowering the incidence of influenza by vaccination might protect against GBS.…”

Section: Discussionmentioning

confidence: 99%

A simulation study of the statistical power and signaling characteristics of an early season sequential test for influenza vaccine safety

Forshee

Arya

et al. 2019

Pharmacoepidemiology and Drug

View full text Add to dashboard Cite

Purpose The US Food and Drug Administration monitors the risk of Guillain‐Barré syndrome (GBS) following influenza vaccination using several data sources including Medicare. In the 2017 to 2018 season, we transitioned our near real‐time surveillance in Medicare to more effectively detect large GBS risk increases early in the season while avoiding false positives. Methods We conducted a simulation study examining the ability of the updating sequential probability ratio test (USPRT) to detect substantially elevated GBS risk in the 8‐ to 21‐day postvaccination versus 5× to 30× the historical rate. We varied the first testing week (weeks 5‐8) and the null rate (1×‐3×) and evaluated power. We estimated signal probability and the risk ratio (RR) after signaling when high‐risk seasons were rare. Results Applying fixed alternatives, we found >80% power to detect a risk 30× the historical rate in week 5 for the 1× null and in week 6 for the 1.5× to 3× nulls. Nearly all testing schedules had >80% power for a 5× risk by week 11. To test the robustness of USPRT, we further simulated seasons where 1% were true high‐risk seasons. Using a 1× null led to 10% of seasons signaling by week 11 (median RR approximately 1.4), which decreased to approximately 1% with the ≥2.5× null (median RR approximately 16.0). Conclusions On the basis of the results from this simulation and subsequent consultations with experts and stakeholders, we specified USPRT to test continuously from weeks 7 to 11 using the null hypothesis that the observed GBS rate was 2.5× the historical rate. This helped improve the ability of USPRT to provide early detection of GBS risk following influenza vaccination as part of a multilayered system of surveillance.

show abstract

“…Studies devoted to explore suitable choices of ρ having the expected time to signal as the main design criteria are just emerging. An example is the work of Silva (2018) suggesting that values of ρ around 0.5 are appropriate choices if the design criterion is the minimization of the expected time to signal.…”

Section: Group Sequential Analysis: the Type I Error Probability Funcmentioning

confidence: 99%

Exact conditional maximized sequential probability ratio test adjusted for covariates

Silva

Kulldorff

2019

Sequential Analysis

Self Cite

View full text Add to dashboard Cite

Sequential analysis is now commonly used for post-market drug and vaccine safety surveillance, and a Poisson stochastic process is typically used for rare adverse events. The conditional maximized sequential probability ratio test, CMaxSPRT, is a powerful tool when there is uncertainty in the estimated expected counts under the null hypothesis. This paper derives exact critical values for CMaxSPRT, as well as statistical power and expected time to signal. This is done for both continuous and group sequential analysis, and for different rejection boundaries. It is also shown how to adjust for covariates in the sequential design. A table of critical values is provided for selected parameters and rejection boundaries, while new functions in the R Sequential package can be used for other calculations. In addition, the method is illustrated for monitoring adverse events after pediarix vaccination data.

show abstract

Type I error probability spending for post–market drug and vaccine safety surveillance with binomial data

Cited by 12 publications

References 21 publications

Alpha spending for historical versus surveillance Poisson data with CMaxSPRT

Alpha spending for historical versus surveillance Poisson data with CMaxSPRT

A simulation study of the statistical power and signaling characteristics of an early season sequential test for influenza vaccine safety

Exact conditional maximized sequential probability ratio test adjusted for covariates

Contact Info

Product

Resources

About