Thread by @DrBrocktagon

My latest for @natureindex is about Living Systematic Reviews.

Come for the Zika virus research. Stay for the robots.
natureindex.com/news-blog/livi…

@NatureIndex A systematic review is where you define a protocol that says how you're going to survey the literature, decide what is relevant, and how you're going to synthesise the data.

It's slow and laborious. Reviews are often out of date before they're published.

@NatureIndex An extreme example I stumbled on is @nicolamlow's work on the Zika virus and microcephaly. Her team did a systematic review but they were getting 200-300 new papers a month.

So they decided to turn it into a *living* systematic review.

@NatureIndex @nicolamlow The basic idea with a living systematic review is that you use your original protocol to continually update the review. And you commit to doing that at regular intervals. Say every 3 months.

@NatureIndex @nicolamlow I had a great chat to @jhjelliott who first had the original idea for living systematic reviews.

Julian is now leading @cochranecollab's attempts to make systematic reviews easier and more efficient, which should make LSRs increasingly feasible

@NatureIndex @nicolamlow @jhjelliott @cochranecollab I also spoke to @mavergames about the technical innovations at @cochranecollab

They're using AI and crowd-sourced tasks to create an "evidence pipeline".

@NatureIndex @nicolamlow @jhjelliott @cochranecollab @mavergames Chris also introduced me to RobotReviewer, an AI tool for automated evaluation of risk of bias in randomised control trials. Still a work in progress but you can check out the demo. robotreviewer.net/about

@NatureIndex @nicolamlow @jhjelliott @cochranecollab @mavergames Thanks to @nicolamlow @jhjelliott and @mavergames for their time.

Thanks also to @mjpages and @adamgdunn for background chats and idea bouncing.

Here's the link again natureindex.com/news-blog/livi…