← All case studies ◆ Clinical research · Cardiovascular

Cardiovascular registry analysis for the PROFID programme

PROFID is an EU Horizon 2020 programme coordinated by Charité Berlin, improving the prevention of sudden cardiac death after a heart attack. Umbizo was lead analytical partner across nine pre-specified sub-studies drawing on more than 140,000 patients from 20 international registries.

ECG heart rhythm trace with one beat highlighted

The challenge

PROFID is an EU Horizon 2020 programme (Grant 847999) coordinated by Charité-Universitätsmedizin Berlin. Its aim is to improve the prevention of sudden cardiac death after a heart attack, and in particular to work out which patients benefit from an implantable cardioverter-defibrillator (ICD). The programme brings together data on more than 140,000 post-myocardial infarction patients from 20 registries in Europe, the USA and Israel.

The programme called for nine pre-specified sub-analyses, each answering a different clinical question. Every study needed its own statistical plan, methods suited to the question, and a manuscript fit for a leading cardiology journal. The registries had been collected in different countries with different coding and follow-up, so careful harmonisation and clear endpoint definitions were essential. Umbizo was appointed as lead analytical partner to Charité for these analyses.

What we set out to do

  • Deliver nine pre-specified sub-analyses of sudden cardiac death after myocardial infarction
  • Work from a common data model harmonised across 20 international registries
  • Write a statistical analysis plan for each study
  • Produce publication-ready manuscripts formatted for the European Heart Journal
  • Support the programme's data governance obligations

Our approach

Data harmonisation and governance

We built the analysis on the PROFID common data model and pre-processing rules, covering both ICD and non-ICD cohorts, with consistent endpoint definitions for sudden cardiac death, SCD proxies and death from other causes. We worked endpoint coding questions through with Charité researchers, including how follow-up is handled after ICD therapy, and supported data use agreements with contributing registries and an audit of grant expenditure.

Statistical analysis

Each sub-study had its own statistical analysis plan. Methods were chosen to suit the question: time-dependent Cox regression and competing risks models, seasonal ARIMA models for patterns across the year, restricted cubic splines for the U-shaped relationship between BMI and risk, multilevel models of variation between countries and healthcare systems, and segmented regression and interrupted time series to track trends between 2002 and 2020.

Manuscript development

Studies were scheduled in order of complexity, with the standard analyses completed first. Manuscripts were drafted in European Heart Journal format and taken through review cycles with the PROFID coordinators and study leads at Charité.

Results and impact

All nine pre-specified sub-analyses were delivered to publication-ready standard, covering seasonal patterns in sudden cardiac death, age-stratified incidence, heart failure therapy, BMI, geographic disparities, atrial fibrillation and temporal trends, across more than 140,000 patients from 20 international cohorts. The work gives new evidence on which post-MI patients should be offered an ICD, and having one analytical partner across all nine studies kept the methods consistent so findings can be compared between them. Data quality and endpoint questions were raised early and resolved jointly with the Charité team, and governance support kept data access and grant reporting in order.

Published

The temporal trends analysis has been published in the European Heart Journal: Mortality reduction with implanted defibrillator for primary prevention of sudden death after myocardial infarction: temporal trends in the PROFID study (2026). See all of the team's papers on our publications page.

In their words

“Fabulous work and considerable time pressure, which resulted in publication-ready outputs.”

Study Coordinator, PROFID programme

Timeline

Six months to complete nine sub-studies.

Looking forward

The manuscripts are moving through coordinator review towards submission. The same approach, a shared data model, a written analysis plan for each question and a consistent set of methods, can be applied to other multi-centre registries and cardiovascular outcomes programmes where several related studies need to be delivered together.