Introduction:

A new methodology for the use of contextualized statistical implicative analysis in causality medical research provides rules and meta-rules that must correspond to medical practice, which is unknown.

Objective:

To evaluate the quality of the rules derived from the application of the statistical implicative analysis in causality medical research.

Methods:

An observational study was carried out applying qualitative techniques based on expert judgement. Using a five-point Likert scale, the 33 experts with the highest score expressed their judgment about the frequency of the fulfilment of the 23 rules derived from the application of the statistical implicative analysis in three studies in order to identify prognostic factors for breast cancer. Opinions were processed using the two-round Delphi method, and the rational validity of these rules was triangulated.

Results:

The experts considered that about 74% of the rules are always or frequently followed in medical practice; only one rule was absurd. In the second round, the three rules in which there was no perfect agreement were discussed; one rule was ratified as absurd and there was an increase in the level of agreement. The triangulation allowed us to find contradictions between literature and experts, which facilitated the identification of problems in the drafting of the interpretation of the rules.

Conclusions:

The rules evaluated generally reflect what happens in medical practice and those that are absurd warn the researcher about the possibility of a confusion bias.

        · resumen en Español     · texto en Español     · Español ( pdf )