Biometry and database analysis

An experienced team of biostatisticians/data managers following rigorous statistical procedures established for over 10 years and regularly updated.

Pathway modeling, target population estimation, survival analysis, effectiveness comparison, and unit or annual cost estimation—whatever your objectives may be, our expertise in real-world studies enables us to assist you in addressing them.



CEMKA it’s:

  • Over 15 years of experience in analyzing medical-administrative data from the French National Health Data System (SNDS), including PMSI, SNIIRAM, and EGB, ensuring strong expertise in handling these complex databases.
    • Construction and validation of algorithms
    • Identification of target populations
    • Sensitivity analyses
    • Direct or indirect matching
    • Estimation of unit costs
  • Training and staying informed: keeping abreast of access developments, coding subtleties, the emergence of new algorithms, and maintaining continuous communication with Health Insurance or other stakeholders in the Health Data Hub. CEMKA actively participates in knowledge-sharing events and regularly offers training on these topics to those interested.
  • Expertise in current statistical methodologies based on academic and professional backgrounds, collaboration with a methodologist scientific advisor, and a commitment to exploring the most innovative methods: machine learning, survival analyses with latent variables, mixed models, logit, tobit, propensity score, meta-analysis, indirect comparisons. When, why, and how to use them? We can advise you, conduct these analyses, interpret the results, and also train your teams on these subjects.
  • Precision in writing protocols, statistical analysis plans, regulatory documents, study reports, abstracts, and posters.

Key points

Proven statistical expertise in:

  • Modelling techniques:
    • Linear, logistic regressions
    • Mixed models (repeated measurements, splines, fractional polynomials)
    • Semi-parametric survival model (Cox model) and parametric model (Weibull, Exponential, log-logistic, log-Normal, Gompertz).
  • Use of propensity scores
  • Carrying out meta-analyses and indirect comparisons
  • Analysis of medico-administrative databases.