Bayesian analysis
Dotaz
Zobrazit nápovědu
BACKGROUND: We provide an overview of Bayesian estimation, hypothesis testing, and model-averaging and illustrate how they benefit parametric survival analysis. We contrast the Bayesian framework to the currently dominant frequentist approach and highlight advantages, such as seamless incorporation of historical data, continuous monitoring of evidence, and incorporating uncertainty about the true data generating process. METHODS: We illustrate the application of the outlined Bayesian approaches on an example data set, retrospective re-analyzing a colon cancer trial. We assess the performance of Bayesian parametric survival analysis and maximum likelihood survival models with AIC/BIC model selection in fixed-n and sequential designs with a simulation study. RESULTS: In the retrospective re-analysis of the example data set, the Bayesian framework provided evidence for the absence of a positive treatment effect of adding Cetuximab to FOLFOX6 regimen on disease-free survival in patients with resected stage III colon cancer. Furthermore, the Bayesian sequential analysis would have terminated the trial 10.3 months earlier than the standard frequentist analysis. In a simulation study with sequential designs, the Bayesian framework on average reached a decision in almost half the time required by the frequentist counterparts, while maintaining the same power, and an appropriate false-positive rate. Under model misspecification, the Bayesian framework resulted in higher false-negative rate compared to the frequentist counterparts, which resulted in a higher proportion of undecided trials. In fixed-n designs, the Bayesian framework showed slightly higher power, slightly elevated error rates, and lower bias and RMSE when estimating treatment effects in small samples. We found no noticeable differences for survival predictions. We have made the analytic approach readily available to other researchers in the RoBSA R package. CONCLUSIONS: The outlined Bayesian framework provides several benefits when applied to parametric survival analyses. It uses data more efficiently, is capable of considerably shortening the length of clinical trials, and provides a richer set of inferences.
Zpracování znalostí zatížených nejistotou je jednou z nejdůležitějších aplikací metod umělé inteligence. Použití technologie bayesovských sítí umožňuje pro tyto ucely využít výsledky po několik století budované teorie pravděpodobnosti a pracovat s mnohorozměrnými pravdepodobnostními distribucemi V tomto případě muže být rozměr distribucí roven stovkám, případně i tisícům. To znamená, že tato technologie může být použita na reálné aplikace, na skutečné problémy, jejichž složitost přesahuje možnosti většiny dalších přístupů pro modelování nejistých znalostí. Vzhledem k tomu, že se jedná o poměrně mladou disciplínu, nelze říci, že všechny teoretické problémy a problémy spojené s návrhem aplikací již byly úspěšně vyřešeny. Nejvíce otevřených problémů je spojeno právě s konstrukcí bayesovských sítu Přesto sejižobjevují aplikace, které naznačují, že bayesovské sítě se stanoujednítn z mocných nástrojů umělé inteligence pro řešení složitých problémů. Proto lze předpokládat, že se s bayesovskými sítěmi budeme v blízké budoucnosti setkávat i v medicíně, která je jednou z oblastí, kde deterministická znalost je spíše výjimkou.
Uncertain knowledge processing is one of the most important applications of artificial intelligence. Bayesian network technology, taking advantage of for several centuries developed results of probability theory, enables processing of multidimensional probability distributions whose dimensionality equals hundreds or even thousands. Therefore, this technology can be applied to real-life problems whose complexity goes beyond cambility of most other approaches for uncertain knowledge processing. It cannot be said that this relatively new discipline has Iready solved all its theoretical and practical problems. Most of still open problems are connected with zonstraction of Bayesian network models for practical applications. Nevertheless, recently published applications suggest that Bayesian network will become one of he most powerful tool of artificial intelligence for uncertain knowledge processing. Therefore, we can assume that in near future we shall meet Bayesian network in medical applications as this field is one of those where deterministic knowledge is exception.
Texts in statistical science
2nd ed. xxv, 668 s.
Texts in statistical science
1st ed. xix, 526 s.
3rd ed. xv, 351 s. : il.
3rd ed. xv, 351 s. : il. ; 24 cm
- MeSH
- statistika jako téma metody MeSH
- Publikační typ
- monografie MeSH
- Konspekt
- Statistika
- NLK Obory
- statistika, zdravotnická statistika
Statistics in practice
1st ed. xi, 266 s.
One of the most common statistical analyses in experimental psychology concerns the comparison of two means using the frequentist t test. However, frequentist t tests do not quantify evidence and require various assumption tests. Recently, popularized Bayesian t tests do quantify evidence, but these were developed for scenarios where the two populations are assumed to have the same variance. As an alternative to both methods, we outline a comprehensive t test framework based on Bayesian model averaging. This new t test framework simultaneously takes into account models that assume equal and unequal variances, and models that use t-likelihoods to improve robustness to outliers. The resulting inference is based on a weighted average across the entire model ensemble, with higher weights assigned to models that predicted the observed data well. This new t test framework provides an integrated approach to assumption checks and inference by applying a series of pertinent models to the data simultaneously rather than sequentially. The integrated Bayesian model-averaged t tests achieve robustness without having to commit to a single model following a series of assumption checks. To facilitate practical applications, we provide user-friendly implementations in JASP and via the RoBTT package in R . A tutorial video is available at https://www.youtube.com/watch?v=EcuzGTIcorQ.
- MeSH
- Bayesova věta MeSH
- experimentální psychologie * metody MeSH
- interpretace statistických dat MeSH
- lidé MeSH
- statistické modely * MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
- přehledy MeSH
BACKGROUND: The outcomes of several randomized trials on extracorporeal cardiopulmonary resuscitation (ECPR) in patients with refractory out-of-hospital cardiac arrest were examined using frequentist methods, resulting in a dichotomous interpretation of results based on p-values rather than in the probability of clinically relevant treatment effects. To determine such a probability of a clinically relevant ECPR-based treatment effect on neurological outcomes, the authors of these trials performed a Bayesian meta-analysis of the totality of randomized ECPR evidence. METHODS: A systematic search was applied to three electronic databases. Randomized trials that compared ECPR-based treatment with conventional CPR for refractory out-of-hospital cardiac arrest were included. The study was preregistered in INPLASY (INPLASY2023120060). The primary Bayesian hierarchical meta-analysis estimated the difference in 6-month neurologically favorable survival in patients with all rhythms, and a secondary analysis assessed this difference in patients with shockable rhythms (Bayesian hierarchical random-effects model). Primary Bayesian analyses were performed under vague priors. Outcomes were formulated as estimated median relative risks, mean absolute risk differences, and numbers needed to treat with corresponding 95% credible intervals (CrIs). The posterior probabilities of various clinically relevant absolute risk difference thresholds were estimated. RESULTS: Three randomized trials were included in the analysis (ECPR, n = 209 patients; conventional CPR, n = 211 patients). The estimated median relative risk of ECPR for 6-month neurologically favorable survival was 1.47 (95%CrI 0.73-3.32) with a mean absolute risk difference of 8.7% (- 5.0; 42.7%) in patients with all rhythms, and the median relative risk was 1.54 (95%CrI 0.79-3.71) with a mean absolute risk difference of 10.8% (95%CrI - 4.2; 73.9%) in patients with shockable rhythms. The posterior probabilities of an absolute risk difference > 0% and > 5% were 91.0% and 71.1% in patients with all rhythms and 92.4% and 75.8% in patients with shockable rhythms, respectively. CONCLUSION: The current Bayesian meta-analysis found a 71.1% and 75.8% posterior probability of a clinically relevant ECPR-based treatment effect on 6-month neurologically favorable survival in patients with all rhythms and shockable rhythms. These results must be interpreted within the context of the reported credible intervals and varying designs of the randomized trials. REGISTRATION: INPLASY (INPLASY2023120060, December 14th, 2023, https://doi.org/10.37766/inplasy2023.12.0060 ).
- MeSH
- Bayesova věta * MeSH
- kardiopulmonální resuscitace * metody normy MeSH
- lidé MeSH
- mimotělní membránová oxygenace metody MeSH
- randomizované kontrolované studie jako téma metody MeSH
- výsledek terapie MeSH
- zástava srdce mimo nemocnici * terapie mortalita MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
- metaanalýza MeSH
