Research
Managing your studies and cohorts
Ongoing data collection, site initiations, analysis plans, submission deadlines: one table per study and milestones planned backwards.
The problem
A closed cohort waiting to be published, another still recruiting, a multicentre study to set up, a risk calculator whose data collection must finish before a conference deadline: each project has its own timeline, partners and regulatory documents.
Step-by-step method
-
One sheet per study: status (in preparation, site initiation, data collection, closed, under analysis), partners, target and achieved sample size, next visit, regulatory documents.
-
Milestones planned backwards: late-breaking abstract deadline, end of data collection, database freeze, analysis plan.
-
A versioned statistical analysis plan (v1.0, v1.1…) and a decision log.
-
Spin-off projects: for a cohort, list the secondary questions (biomarkers, remodelling, machine learning) and who leads them.
-
Site initiation: extract the list of documents to provide and the steps, and turn it into a work item.
-
Prepare meetings (data warehouse, steering committee): Claude drafts the status report from the study sheets.
-
Link to the database: candidates for a study are identified by querying aggregates, never by exporting identifying data.
-
Multicentre registries: procedure for sending exams to core laboratories, tracking of quality and of reported missing data.
-
Assign each study to the right hat: academic, or a service provided to an industry sponsor, depending on who funds it.
Deliverable
A table of studies, milestone alerts, and a status report ready before each meeting.
Safeguards
Regulatory framework for research involving human participants; sensitive data isolated and processed locally; approval by the clinical research department.
Starter prompt
Here are my ongoing studies (short name, status, next deadline). Build a management table with backward-planned milestones for each, and flag those whose data collection will not meet the next submission deadline.