Auto-generated methods section#400

Researchers that plan on reporting their findings often spend a considerable amount of time writing up detailed descriptions of the NK methods. While certain sections of the NK process are very easily written up (for instance, the search strings used), other sections may be more difficult to write up (for instance, quantitative synthesis). To both make it easier on the researcher and so that we can accurately represent NK in published reports, it would be very helpful to generate a structured methods page for the methods that were used.

Maybe we could work on some generic language that best represents the NK process and that easily transfers across different study designs/nests? I was thinking of structured headings that are identical to the modules (Literature Search, Screening, etc.), and automatically gets reported under a “Methods” section within “Manuscript Editor”.

Thanks John– I think we could use the Templating function to provide template text for all the parts that are replicable across ALL nests. This could start out as an importable template that we make available to all nests (regardless of Org association), and Orgs could edit it and turn it into their own customized template. Users would still need to import it into their blank page, but thats a lot less work than writing it!

Yeah, I think it actually might be better to have this as a separate attachment that you could import. Easy enough to create a template where someone could copy in values, but for clarification I was specifically thinking of a generic methods section that is automatically updated based on nest-specific things like which search strings/databases were specifically used. For instance, under the “Literature Search” section, we could provide something like the following:

METHODS

Literature Search

A systematic literature search was conducted using the AutoLit platform [insert citation]. Search strings used for each database are provided in Table 1 [insert table with search strings/databases/dates searched]

Studies identified in the literature search were screened automatically using the AutoLit platform for the following preconfigured automated exclusion criteria: [insert automated reasons]. Two independent reviewers then used the AutoLit dual screening module to evaluate the remaining studies [change this depending on whether dual screening was used] using the following selection criteria:

Manual inclusion criteria included: [insert table or bullet point list of manual inclusion criteria]

Manual exclusion criteria included: [insert table or bullet point list of manual exclusion criteria table]

A third reviewer resolved any disagreements between the two independent reviewers [update depending on whether dual screening was used].

Agreed, language would need to be prepopulated on the basis of the nest’s configuration. Where this gets tricky is that surely researchers will want to modify our language, but we also would like to be able to update the methods automatically (as we do with PRISMA diagrams, tables, etc. in manuscript), e.g. as exclusion reasons are added.

For the record, I really like this idea, as it takes us a step closer to providing concrete outputs automatically. Methods feels like the easiest section to start, followed by Results (which we have much of the needed data for). Intro/Discussion are more open-ended narration we’d probably need to stay away from to start.

Yeah for sure agree that automatic Methods is the necessary first step.

It will be difficult to generate results sections that automatically summarize the important quantitative results (especially for very large nests). However, it would be relatively straightforward to create generic templates for certain results like PRISMA diagrams and network plots.

Intro/Discussion will be very difficult. I think it is ok to stay away from auto-generated Intro/Discussion for the foreseeable future. However, some elements of each might be automatically updated. For instance, a 1-paragraph summary of findings (typically first paragraph of a discussion) and 1-paragraph limitations section (typically last paragraph of a discussion).

Everything would of course need to be editable by the researcher.

Would it be helpful if I started on templates for Methods?

It absolutely would be! I’ll send you a link for where :)

Great, thanks!