Diagnosing and dealing with degenerate estimation in a Bayesian meta-analysis

The federal government recently granted emergency approval for the use of antibody rich blood plasma when treating hospitalized COVID-19 patients. This announcement is unfortunate, because we really don’t know if this promising treatment works. The best way to determine this, of course, is to conduct an experiment, though this approval makes this more challenging to do; with the general availability of convalescent plasma (CP), there may be resistance from patients and providers against participating in a randomized trial. [Read More]

Generating data from a truncated distribution

A researcher reached out to me the other day to see if the simstudy package provides a quick and easy way to generate data from a truncated distribution. Other than the noZeroPoisson distribution option (which is a very specific truncated distribution), there is no way to do this directly. You can always generate data from the full distribution and toss out the observations that fall outside of the truncation range, but this is not exactly efficient, and in practice can get a little messy. [Read More]

A hurdle model for COVID-19 infections in nursing homes

Late last year, I added a mixture distribution to the simstudy package, largely motivated to accommodate zero-inflated Poisson or negative binomial distributions. (I really thought I had added this two years ago - but time is moving so slowly these days.) These distributions are useful when modeling count data, but we anticipate observing more than the expected frequency of zeros that would arise from a non-inflated (i.e. “regular”) Poisson or negative binomial distribution. [Read More]

A Bayesian model for a simulated meta-analysis

This is essentially an addendum to the previous post where I simulated data from multiple RCTs to explore an analytic method to pool data across different studies. In that post, I used the nlme package to conduct a meta-analysis based on individual level data of 12 studies. Here, I am presenting an alternative hierarchical modeling approach that uses the Bayesian package rstan. Create the data set We’ll use the exact same data generating process as described in some detail in the previous post. [Read More]

Simulating multiple RCTs to simulate a meta-analysis

I am currently involved with an RCT that is struggling to recruit eligible patients (by no means an unusual problem), increasing the risk that findings might be inconclusive. A possible solution to this conundrum is to find similar, ongoing trials with the aim of pooling data in a single analysis, to conduct a meta-analysis of sorts. In an ideal world, this theoretical collection of sites would have joined forces to develop a single study protocol, but often there is no structure or funding mechanism to make that happen. [Read More]

Consider a permutation test for a small pilot study

Recently I wrote about the challenges of trying to learn too much from a small pilot study, even if it is a randomized controlled trial. There are limitations on how much you can learn about a treatment effect given the small sample size and relatively high variability of the estimate. However, the temptation for researchers is usually just too great; it is only natural to want to see if there is any kind of signal of an intervention effect, even though the pilot study is focused on questions of feasibility and acceptability. [Read More]

When proportional odds is a poor assumption, collapsing categories is probably not going to save you

Continuing the discussion on cumulative odds models I started last time, I want to investigate a solution I always assumed would help mitigate a failure to meet the proportional odds assumption. I’ve believed if there is a large number of categories and the relative cumulative odds between two groups don’t appear proportional across all categorical levels, then a reasonable approach is to reduce the number of categories. In other words, fewer categories translates to proportional odds. [Read More]

Considering the number of categories in an ordinal outcome

In two Covid-19-related trials I’m involved with, the primary or key secondary outcome is the status of a patient at 14 days based on a World Health Organization ordered rating scale. In this particular ordinal scale, there are 11 categories ranging from 0 (uninfected) to 10 (death). In between, a patient can be infected but well enough to remain at home, hospitalized with milder symptoms, or hospitalized with severe disease. [Read More]

To stratify or not? It might not actually matter...

Continuing with the theme of exploring small issues that come up in trial design, I recently used simulation to assess the impact of stratifying (or not) in the context of a multi-site Covid-19 trial with a binary outcome. The investigators are concerned that baseline health status will affect the probability of an outcome event, and are interested in randomizing by health status. The goal is to ensure balance across the two treatment arms with respect to this important variable. [Read More]

Simulation for power in designing cluster randomized trials

As a biostatistician, I like to be involved in the design of a study as early as possible. I always like to say that I hope one of the first conversations an investigator has is with me, so that I can help clarify the research questions before getting into the design questions related to measurement, unit of randomization, and sample size. In the worst case scenario - and this actually doesn’t happen to me any more - a researcher would approach me after everything is done except the analysis. [Read More]