Course levels

TipWhere you are: Level 1

This website is Level 1: Foundations, the whole 8-session course, from a clinical question to a draft manuscript. Level 2 and Level 3 are separate courses that build on it.

Level 1, Level 2 and Level 3

Download the course levels overview (PDF)

Organisation: Neudata · #ClearDataClearImpact

Neudata’s clinical-research training runs in three levels. Each level builds on the one before, and each department can stop at the level that fits its needs.

Level 1: Foundations Level 2: Doing the analysis Level 3: Advanced topics
In one line Understand research from question to publication Fit, check and interpret the common statistical models yourself, in R Reproducible research, journal-quality output and specialised methods for complex data
Who A whole department (about 30 health professionals: doctors, nurses, etc.) An interested subgroup (8 to 12) who want to run their own analyses Researchers with a specific project or dataset
Prerequisites None Level 1 and the free Neudata course Clinical Data Analysis in R (or equivalent basics) Level 2 (or equivalent experience), plus a real question or dataset
Format 8 online sessions of 3 hours, every Saturday for 8 weeks 8 online sessions of 3 hours over about 4 to 6 weeks, with homework Online workshops of 1 or 2 sessions, or consultancy, on request
Software None for participants (instructor demos) R and RStudio on own laptops R (and specialised packages)
Hands-on? Concepts, reading and interpreting; analyses demonstrated live on FLUID-ICU Every session is hands-on coding; the class works together, exercises are individual Hands-on on the participant’s own data
Outcome An individual research proposal and an individual draft manuscript R scripts for an analysis question and a draft Results section A reproducible analysis, or an analysis plan, for a real project, with Neudata support
Documents This website (the full Level 1 course) Level 2 Level 3

How the levels connect

  • From Level 1 to Level 2. In Level 1, participants read the FLUID-ICU analyses (Table 1, crude vs adjusted ORs, Kaplan-Meier, multiple imputation) that the instructor runs. Level 2 is a statistical-modelling course: participants clean and describe the data, then fit, check and interpret the models that clinical papers use most (linear, logistic, Poisson, negative binomial, zero-inflated and Cox), each session starting with the exploratory data analysis (EDA) suited to that model. It ends with an individual analysis question and a written Results section. The FLUID-ICU dataset and R scripts in Course/Data/ are shared by both levels.
  • From Level 2 to Level 3. Level 3 adds the research-craft topics moved out of Level 2 (reproducible projects, version control, Quarto reports, figures that meet journal requirements, and missing-data and sensitivity analyses in depth) and the methods that complex data demand (meta-analysis, prediction models, diagnostic accuracy, advanced survival analysis, longitudinal and multilevel models, large or clustered datasets). Every Level 3 lecture includes an EDA section and journal-ready figure code for its analysis problem.

Progression at a glance

Skill Level 1 Level 2 Level 3
Research question, design, ethics Taught in depth Assumed Assumed
Data management and cleaning Concepts; live demo Hands-on in R scripts: import, clean, validate, cleaning log Reproducible projects and version control (Module G); large, linked or clustered data (Module F)
Describing data, Table 1, figures Read and judge Table 1; EDA plots suited to each model, from course code Figures that meet journal requirements (vector vs raster, dimensions, resolution, fonts) and figure code in every module
Regression Read and interpret outputs Fit, check, interpret and report linear, logistic, count (Poisson, negative binomial, zero-inflated) and Cox models Mixed models, competing risks, time-varying covariates, prediction models
Missing data Mechanisms and ideas Describe missingness; complete-case vs multiple imputation, touched on Multiple imputation in practice and sensitivity analyses in depth (Module H)
Survival analysis Read Kaplan-Meier and HRs Kaplan-Meier and Cox, including the proportional-hazards check Advanced: time-varying covariates, competing risks, restricted mean survival time (Module D)
Writing and publishing Full manuscript workflow (individual draft) Results section written from own R output Quarto reports in which every number comes from code; Methods and Results for complex analyses

Status: Level 1 materials are in Course/. Level 2 and Level 3 are outlines. Their materials will be developed after Level 1 has run.