Research Workflow | From Idea to Publication: The Clinical Research Li
Most research workflow advice describes an idealized version of the process — linear, orderly, each stage neatly completing before the next begins. The actual research workflow looks nothing like that. It loops back on itself, stalls, gets restarted, and occasionally produces a paper.
What's useful isn't a perfect map. It's a working map — one that tells you which stage you're in, what you should be doing, and when it's legitimate to move forward.
The Six Stages of a Clinical Research Lifecycle
After running several studies from initial concept through publication, the process consistently breaks into six distinct stages. Each has a different purpose, different failure modes, and a different decision that ends it.
Stage 1 — Idea Capture The idea arrives as an observation, a gap in rounds, or a question that keeps recurring. At this stage, the work is to articulate it precisely enough to evaluate — not to research it. A useful output here is a single-sentence problem statement: what phenomenon, in what population, compared to what, with what outcome. If it can't be written in one sentence, it isn't specific enough yet.
Stage 2 — Feasibility Check Before designing anything, three questions need answers: Is it answerable with available data? Has it already been answered? Does the answer actually matter? A quick scoping search (not a full literature review) and an honest look at data access usually takes a few hours. Many ideas that felt compelling on Wednesday fail this check by Friday, which is exactly the point.
Stage 3 — Protocol Development If the idea survives feasibility, the protocol translates the research question into operational decisions: study design, inclusion/exclusion criteria, primary outcome, statistical analysis plan. This stage produces a written protocol before any data is touched. The protocol isn't bureaucracy — it's the commitment that separates planned analysis from post-hoc rationalization.
Stage 4 — Data Collection and Management Prospective data is collected per protocol. Retrospective data is extracted against pre-defined variables. Either way, this stage should be boring: following the extraction form, documenting decisions, flagging ambiguous cases. Interesting deviations from protocol need to be noted and later disclosed, not silently corrected.
Stage 5 — Analysis and Writing Analysis follows the pre-specified plan. Writing begins from the methods — because what was done is clear and shouldn't change. Results follow. Introduction and discussion are written last, when it's clear what the paper actually found rather than what was hoped for. This order prevents the common problem of building an argument before knowing if the data supports it.
Stage 6 — Submission and Revision Journal selection, submission, managing reviews, revision, resubmission. This stage is often underestimated in time and cognitive load. It typically spans months. Having a clear response strategy and understanding the review system helps — academic publishing operates on different incentives than most researchers expect.
Why the Workflow Breaks Down
Most workflow problems don't look like workflow problems. They look like writing problems, or analysis problems, or motivation problems.
A paper where the introduction keeps changing is often a stage 2 problem — the research question was never specific enough. A discussion that can't land a clear conclusion usually signals a mismatch between what the protocol promised and what the data produced. A study that keeps expanding its scope mid-collection is a stage 3 problem arriving late.
Naming which stage a problem belongs to often clarifies what actually needs to be done — and what to stop doing.
The Workflow as a Decision Framework
Each stage ends with a specific decision, and that decision is binary.
Stage
Ends with
Idea Capture
Is this worth evaluating?
Feasibility
Is this worth designing?
Protocol
Is this worth executing?
Data Collection
Is the dataset clean enough to analyze?
Analysis & Writing
Is this finding worth submitting?
Submission
Is this revision worth doing?
The decision at each stage is a gate. The purpose of a gate is to stop things that shouldn't proceed — which is most ideas, at various stages. That's not failure. That's the system working correctly.
A Note on Shortcuts
The stages that most commonly get skipped are Stage 2 (feasibility) and Stage 3 (protocol). Both feel like overhead when there's momentum behind an idea. Both generate the most downstream problems when skipped.
A study that began without a written protocol is almost impossible to write cleanly. The methods section that should take a day to draft instead takes weeks, because analytical decisions that should have been made in advance are being made retroactively. That process also produces a paper that reviewers can detect — because the internal logic of the methods section reveals when decisions were made.
Writing is a system, not an isolated skill — and the quality of the writing reflects the quality of the process that generated it.
End-to-End Research Workflow Checklist
Use this at the start of any new study to confirm you're beginning the right stage with the right outputs.
Stage 1 — Idea Capture - [ ] One-sentence problem statement written - [ ] Target population defined - [ ] Comparator or context specified - [ ] Primary outcome named
Stage 2 — Feasibility - [ ] Scoping search completed (≥3 recent papers reviewed) - [ ] Data source identified and access confirmed - [ ] Sample size rough estimate done - [ ] Decision made: proceed or stop
Stage 3 — Protocol - [ ] Study design selected and justified - [ ] Inclusion/exclusion criteria written - [ ] Primary outcome defined (with measurement method) - [ ] Statistical analysis plan written - [ ] Protocol registered (if applicable) or dated and filed
Stage 4 — Data Collection - [ ] Extraction form finalized before data entry - [ ] Deviations from protocol documented - [ ] Data dictionary maintained - [ ] Dataset locked before analysis
Stage 5 — Analysis & Writing - [ ] Primary analysis follows SAP - [ ] Methods written first - [ ] Results written without interpretation - [ ] Introduction and discussion written last
Stage 6 — Submission - [ ] Journal selected before manuscript finalized - [ ] Submission checklist completed - [ ] Response strategy prepared before reviews arrive
Want the complete checklist from idea to submission in one document? — Stage-by-stage checklist for every step from research idea to journal submission ($5)
Checklist: Idea to Submission
Stage-by-stage checklist from research idea to journal submission.
$5
If you found this helpful for your manuscript, you might want to check out my Checklist: Idea to Submission.