Ability to specify iteration order of references#449

It would be really helpful if I could sort unscreened references.

For example, sorting by publication date (newest to oldest) would help me find the latest version of a study’s publication and then I could mark any older iterations (study design posting, baseline characteristics, conference abstracts, etc) as duplicates. Currently, whenever I come across an interrim study publication, I push it through as a precaution and often find myself with 4+ copies of the study in its different stages of completion. If I could sort unscreened references by year, I would have a better sense of what the latest publication is.

This can be accomplished with an Inspector filter + sort, see below. But perhaps you mean elsewhere, such as Duplicate Review or (sequential) Screening?

Thanks for submitting! Quick question– are you averse to sorting in Inspector? You can filter in Inspector and then make Publication Date a column (see screenshot):

I’ll also note that we are coming out with a feature where you can Link publications of a single study (https://nested-knowledge.nolt.io/142) soon; its fully designed, just need to develop it. Feel free to comment there so you get updates when it gets launched :)

Oh, forgot to say– the way you sort columns is by clicking the up/down arrows next to the column title; the ‘sort’ will be added to the Inspector bar in gold. Thx!

Is it currently possible to sort by date in the Screening module? For example, I would like to first screen the trials published in 2022 and then move my way backwards.

In the Dual Screening Study Inspector module, I don’t see an option to click and sort by Final Screening Status.

Thanks for sharing that Kevin, this feature request is kind of a roundabout way of getting to the same thing.

I played around with the Study Inspector and found a way to show only unscreened and then sort by year. However, when I screen one it doesn’t move me to the next most recent references; it requires me to manually click to the next reference. It would be nice to have something like this baked in the original Screening module.

@Kevin.kallmes Sorting by Final Screening Status is not an option. It rquires adding the Final Screening Status filter

So, are you looking for Inspector to recognize when an action is ‘complete’? We added the navigation arrows on the far left and right (https://nested-knowledge.nolt.io/24) to make this as simple as possible; I’d worry about an auto-complete because, in some Inspector actions, you may NOT want to move on right after, e.g. selecting an Exclusion Reason. I’m open to suggestions, but want to avoid an auto-jump that may surprise a lot of users!

Let me know, especially, if the little arrows on the far left and right are hard to use/making it difficult to go between articles!

Based on the above (let me know if the navigation arrows aren’t sufficient); should we repurpose this request for Add Prelim vs Final Screening filter options in Inspector?

Ideally, I’d like an option to sort by publication year within the Dual Screening module so that I can screen without the friction of having to go between each study through Study Inspector.

On a related note, one another factor that would be nice to be able to sort unscreened references by is P (Inclusion). So far, I’m not sure if it provides any utility beyond just being able to glance to see if the AI will think it’s worthy of inclusion.

So if I could sort unscreened refs by likelihood of model inclusion, that would be really useful.

The software automatically sorts by P (inclusion)! as soon as a model is trained, we re-order the references that way :)

Thanks Dmitriy, hear you on the request to specify iteration order for the Screen/Tag/Extract modules. I’d like to get a sense for how common of a pain point this is before committing to developing it.

I’ll note that we used to auto-advance records in Inspector when screening decision was changed, but enough users complained about the lack of stickiness that we had to change!

Ah! I thought this seemed familiar. Linking to #251