Thread by @adambress

THREAD (1) In preparation for the upcoming @AHAMeetings #QCOR2019 early career programming, @mad_sters, @MikeTPhD, and I, wanted to start a conversation around best-practices for collaboration on manuscripts to maximize success and to avoid common pitfalls and missteps

(2) Although this thread is focused on collaboration on manuscripts, we feel that some of the principles can be extended/adapted to grants and other kinds of collaborations

(3) We'd love input and your stories on best practices and also what are your pet peeves when being engaged as a co-author on a manuscript?
@rwyeh @boback @JeremySussman @hmkyale @MuntnerPaul @DaichiShimbo @berthahidalgo @angiefagerlin @rhessmd @JDodsonMD @jordy_bc @kejoynt

(4) What are some ways in which you've seen collaborations on manuscripts go wrong? @bnallamo @UmeshKhotMD @JeremySussman @SpatzErica @TYWangMD @slcrubes @brentnreed @DaveDixonPharmD @wbaker0621

Here are some things we came up with:

(5)
- Leaving off an important co-author by mistake

- Sending a draft manuscript to co-authors that is not polished enough (eg, "this is a mess")

-Sending a draft manuscript that is “too polished” or too far along (eg, "I would have like to have been engaged earlier")

(6)
- Pinging/engaging coauthors enough so that they feel engaged and included but not too much where they are overburdened/annoyed with the amount of communication

-Making sure you have everyone’s approval before submitting a manuscript to a journal (this is important)

(7)
- How to handle conflicting co-author comments

- How to handle non-responsive co-authors

(8) It is important to note that there is an etiquette and culture that can vary across collaborators, institutions, people, disciplines, types of researchers etc. Even the most well-intentioned collaborations can go wrong.

(9) We’ve oversimplified and distilled the complicated collaboration process down to some steps which we will go through now. This is geared towards outcomes/epidemiology/population type research of existing data and manuscript preparation.

(10) Following these steps to a "T" is A LOT of WORK. Some people don't some of these simply because it’s a ton of work and too slow. We feel that good collaboration practices takes time and requires patience.

(11) Step 1, of course, is to flesh out the research question and corresponding methodology, and most appropriate data source. For early-stage investigators, this is an iterative process usually done with a committed and engaged senior mentor, which is an important prerequisite.

(12) Step 2 is to draft a proposal/analysis plan.

This includes background/rationale, aims/hypothesis, data sources and rationale for why the data are well-suited for the particular question, and brief statistical analysis plan and table shells.

(13) At step 2, we find it crucial to have an engaged and committed mentor who will help develop this document before it goes to co-authors and provide feedback quickly. This is an iterative process. There is a trade-off between rapid response and quality of comments sometimes.

(14) After the analysis plan is well thought out and vetted by your mentor, step 3 is to invite people to participate as co-authors by sending out the clean manuscript proposal/analysis plan and ask them to review and provide comments. Example email below

(15) This is the kind of email we usually send to invite co-authors.

What kinds of things do you say when inviting co-authors to participate in a manuscript at this stage?

You can even ask co-authors the extent to which they would like to be involved.

(16) We feel that some of the key purposes of this step are to:

-Establish author order

-Engage co-authors early so that they had a chance to provide their input on the hypothesis and overall direction of the work before any data analysis have begun

(17) Step 4 is to review the co-author comments with your mentor and decide what to do next. A few of the more common outcomes at this step are:

- Major feedback/changes are requested

- Minor feedback/changes are requested

(18) If major changes requested, we usually would re-work the aims based on the feedback and consider setting up a conference call to discuss.

Then, after major comment are addressed, we would send the revised analysis plan back out to co-authors for a 2nd round of comments.

(19) If only minor feedback/changes are requested, we usually would simply incorporate the feedback and send out a new clean draft to co-authors thanking them from their time and input and say something like the screenshot below:

(20) After all-co authors have signed off on the aims/hypothesis and analysis plan, the next step (#5) usually is to just begin the analyses. The lead author/analyst will then perform the analysis per the agreed upon analysis plan.

(21) Make sure to get all other necessary approvals have been obtained (P&P approval for example if needed)

Our experience is that step 5 is an iterative process working with a senior mentor or perhaps 1-2 other close colleagues to work out initial bug/snags in the analysis

(22) At this step, we feel that the outcome you are driving towards is creating a "to the point", easy to read and interpret display of the results so that your co-investigators can understand everything that is going on even if they don't remember the initial proposal

(23) It is helpful to assume co-authors don't remember details exactly (aims, study population), having lots of reminders is helpful. Also, good now is usually better than perfect later. Advisable to not hold on to initial results too long in an effort to display them “perfect”

(24) The next step, (#6) usually is to send out the first draft of results tables and figures to co-authors and ask for their review and comments.

(25) What we usually do at step #6 is send out a document with a manuscript title page, and abstract, and the populated table shells and figures and nothing else.

(26) We feel that this gives coauthors the opportunity to engage and provide input very early in the process after data analysis has begun, in the case a change of direction is recommended before you go down one path too far.

(27) We usually ask to set up a conference call at this point to present the results to co-authors and get their feedback. It’s helpful to have a slide titled “desired feedback” to point co-authors to places where you need their input the most.

(28) This step can be impt because it protects against co-authors feeling that they weren’t engaged early enough in the process after results were generated

It also helps re-solidify author order and roles (thanks @JeremySussman) that should have been done in the proposal phase

@JeremySussman (29) We often say things like “We are happy to re-analyze anything you think we need to or do any new analyses that you think would be helpful.”

But you have to be careful here; you don't want to go down too many rabbit holes regarding too many secondary anaylses

@JeremySussman (30) After you're happy with the analysis and completed and new analyses requested by co-authors, the next step (#7) for us is to send out a **draft outline** of the manuscript

(31) We emphasize draft outline so that co-authors are engaged early in manuscript drafting so they can provide input on the direction and messaging before a full first draft is developed. Example email below:

(32) After you've received input on the outline of the manuscript, the next step (#8) is to incorporate feedback received on the outline and produce and then send out a *very clean* first draft of the manuscript.

(33) We emphasize very clean b/c this is your time to shine and show co-authors your best work. Also, co-authors do not want to read poorly written manuscripts. This is one of the hardest parts of the process. Again, crucial to have engaged/committed mentor to help with this step

(34) Again, this usually takes several iterations of going back and forth on revisions with a committed, engaged, and invested senior author or another dedicated author 3-5 times to get a nice polished first draft. Writing=hard for most of us

(35) We usually give co-author 2 weeks to review first complete drafts for non-topical manuscripts and one week for time-sensitive manuscripts.

How much time do you usually give co-authors for comments on 1st drafts of manuscripts? @KBibbinsDomingo @DaichiShimbo @berthahidalgo

@KBibbinsDomingo @DaichiShimbo @berthahidalgo (36) Step 9 is to manage and integrate the co-author comments that you receive. This can be a tricky step based on our experience.

Depending on the number and style of co-authors, a few things could have happened:

@KBibbinsDomingo @DaichiShimbo @berthahidalgo (37) You could have received very few comments and have little work to do to revise the manuscript. If this is the case, it may be reasonable to go straight to the next step (i.e., the penultimate draft step)

@KBibbinsDomingo @DaichiShimbo @berthahidalgo (38) However, if major edits were provided or major changes or new analyses requested, we usually repeat step #8 and send the manuscript out for another round of comments.

(39) Based on our experience, it is not atypical for this phase to go through 2-3 cycles, and sometimes more to get the manuscript in good shape.

(40) Step 10 is the penultimate draft phase meaning the manuscript draft that is in good enough shape and is about to be submitted to the journal. How do you know if your manuscript is at the penultimate draft stage? This can be challenging.

What do people think about this?

(41) Usually, we base it off the # and substantiveness of comments received on the last manuscript draft

If only a few minor comments-->we usually move on to penultimate draft phase

If several, more substantive comments--> we usually send out for another rd of co-author review

(42) The penultimate draft step involves emailing co-authors to ask them to review and provide comments on a penultimate draft of the manuscript, if they'd like, and ask for their approval to submit the manuscript as is. Example email here:

(43) This step serves as the final approval before submitting to the Journal-- ***this is a very important step***

Journal selection conversations can begin here or, ideally, earlier, perhaps when the first draft is sent out.

(44) Next step is submitting to the journal and wait for an initial decision. For collaboration steps, the next two steps are to notify co-authors of submission and then the initial decision when ready (e.g., reject, revise with de novo resubmission, revise and resubmit etc)

(45) For a rejection, here is an example of how we have communicated about a journal rejection and navigating next steps with co-authors:

(46) Example email to co-authors for a revise and resubmit decision:

(47) A lot of the guidance provided above probably can be adapted for engaging your co-authors when drafting the response to reviewers document and revising the manuscripts when you have a "revise and resubmit"

(48). After the paper has been accepted; we usually send an email list this to let co-authors join in on the acceptance celebration and congratulations and give them an easy to copy and paste citation so they can easily add the new paper to their CV.

(49) The last thing I have is around notification and dissemination of proofs

Impt point here is to give co-authors the opportunity to review and confirm their name and affiliations are correct.

I like to provide a screenshot at the btm of the email for easy viewing

(50) That's all I have. Thanks so much for @MikeTPhD and @mad_sters for helping me develop this.

Thanks to all that participated!

And thanks to my role-model mentors who taught me these things @rhessmd @DaichiShimbo @MuntnerPaul @slcrubes (others who are not on twitter)

@MikeTPhD @mad_sters @rhessmd @DaichiShimbo @MuntnerPaul @slcrubes (51) And huge thanks to the VPCAT program at the University of Utah for teaching a lot of this stuff as well @UofUHealth @slcrubes @carrie_byington

medicine.utah.edu/faculty-dev/pr…

@MikeTPhD @mad_sters @rhessmd @DaichiShimbo @MuntnerPaul @slcrubes @UofUHealth @carrie_byington I wanted to tag one of my favorite resources for early career advice and guidance, @EdgeforScholars available here: edgeforscholars.org and also #epitwitter #cardiotwitter #AcademicTwitter