2020
DOI: 10.1002/sim.8563
|View full text |Cite
|
Sign up to set email alerts
|

A Bayesian hierarchical approach for multiple outcomes in routinely collected healthcare data

Abstract: Clinical trials are the standard approach for evaluating new treatments, but may lack the power to assess rare outcomes. Trial results are also necessarily restricted to the population considered in the study. The availability of routinely collected healthcare data provides a source of information on the performance of treatments beyond that offered by clinical trials, but the analysis of this type of data presents a number of challenges. Hierarchical methods, which take advantage of known relationships betwee… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
4
0

Year Published

2020
2020
2021
2021

Publication Types

Select...
2
1

Relationship

2
1

Authors

Journals

citations
Cited by 3 publications
(4 citation statements)
references
References 26 publications
0
4
0
Order By: Relevance
“…A further issue is that it not immediately obvious how to choose a threshold which corresponds to a specific target FDR rate. Theoretical approaches to determining thresholds have been suggested, 20 but, based on previous studies, 16,17,19 we use as the event flagging mechanism for the Bayesian models threshold values of 0.975 and 0.95 for model 1a, and 0.90 and 0.80 for model BB. We chose two values for each model in order to give some idea of the how the estimated power and FDR vary as the threshold values change, and to allow a more accurate assessment when comparing to the error controlling procedures.…”
Section: Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…A further issue is that it not immediately obvious how to choose a threshold which corresponds to a specific target FDR rate. Theoretical approaches to determining thresholds have been suggested, 20 but, based on previous studies, 16,17,19 we use as the event flagging mechanism for the Bayesian models threshold values of 0.975 and 0.95 for model 1a, and 0.90 and 0.80 for model BB. We chose two values for each model in order to give some idea of the how the estimated power and FDR vary as the threshold values change, and to allow a more accurate assessment when comparing to the error controlling procedures.…”
Section: Methodsmentioning
confidence: 99%
“…While these more complicated cases are not specifically investigated, we do investigate how well the models cope with large variations in the AE counts within the SOCs. Bayesian grouped models of this type have also been extended for trial meta‐analysis 16 and for observational data 17 . The error controlling approaches are more general, designed to control error rates at a particular level when performing multiple hypothesis testing.…”
Section: Introductionmentioning
confidence: 99%
“…Statistical analysis tools such as the R programming language are used extensively to draw meaningful conclusions from medical data collected through observation or experimentation (Wang et al, 2007). Applications include diagnosis (Bolboacă, 2019;Ma et al, 2013), comparing different treatments (Huang and Tan, 2016;Stevens et al, 2018), assessing the effectiveness of a treatment (Carragher et al, 2020), and monitoring patient health post-treatment (Ong et al, 2016). Although descriptive analytics can offer insight into historical trends through data summarization and visualization tools, they are not able to pinpoint the underlying patterns and correlations or single out individual features, and require a clinician in the loop to draw conclusions.…”
Section: Current Practicementioning
confidence: 99%
“…A second goal of the package is to provide reference implementations of the methods in Table 1 for use by researchers, both in the area of safety in clinical trials, as well those developing or testing methods for handling error rates when testing multiple hypotheses. Beyond safety in clinical trials, the package will be useful to any project which deals with multiple hypothesis testing, or projects where two groups of comparative data may be modelled by hierarchical Bayesian binomial or Poisson models, with recent extensions of the Bayesian models to observational data being developed (Carragher et al, 2020).…”
Section: Statement Of Needmentioning
confidence: 99%