MATRAM Project

Explanation of the acronym MATRAM 

The Matram project is aiming at setting up a mathematical based model (Bayesian Network ) through a six steps methodology for the purpose of appraising, measuring and comparing performances of public health programmes.

There are six  compulsory steps  of the MATRAM methodology, which are gathered into three intertwined expertises: 

Modeling   –   Reinterpretation   –   Analytics




Expertise 1 : MODELING

STEP 1: M for Modelling Based on field experiences and literature available with regard to current health programme & system, draft models within its operational environment are devised

STEP 2: A for Acquisition of data There are data gold mines buried into national information and programme systems, those database could easily and purposefully feed any model. A vast array of information from local surveys or studies could also be considered. Multidimensional approach is fundamental to gaining a full mastering with analytics (step 5)

NOTA: These two steps are much intertwined. In case theoretical  knowledge from experts be needed, the BEKEE (Bayesia Expert Knowledge Elicitation Environment) process from Bayesia company can be activated.



STEP 3: T for Transposition Transfer of models and their data is processed into BayesiaLab software, which provides an easy way to set up model and transpose data aiming at accelerating research workflows

STEP 4: R for Reinterpretation So, whatever the (causal in the best case) connections (probabilistic relationships) among parameters (variables of interest) of the model, duly validated by statistical tests, it will need a “rationale” reinterpretation against initial model (step 1) , which means the reasoning exercise could lead to a recurrent interpretation

NOTA: These two steps are managed by Raphael Girod independently from other project expertise as only BayesiaLab functionalities are utilized.


Expertise 3 : ANALYTICS

STEP 5: A for Analytics The analytical capacity of the software is optimized in terms of identification of  (i) classification of contributing factors to targeted indicators (ii) quantified effects among parameters  (iii) target optimization leading to specific and tailored recommandations for the sake of improving public health programme performances. That step is carried out in close collaboration with experts of the domain (peer review approach), who were involved from the beginning of the Matram project (step 1 &2).

STEP 6: M for Management for decision making Future performances in relation to an array of parameters & assumptions are forecasted. Deep understanding means knowing, not merely how programme behaved yesterday, but also how things will behave under new hypothetical circumstances tomorrow.

Practically speaking, the model thus could become a “playing instrument” for formal reasoning that is entirely transparent to the experts and researchers, and lastly, decision makers.

NOTA: Steps 5 & 6 are much iterative. Scenario and simulation are laid down in taking due consideration of the strategic, financial, economic and/or epidemiological constraints or targets from  all  stakeholders and lastly, endorsed by decision makers