Skip to main content

Archived Comments for: ask MEDLINE: a free-text, natural language query tool for MEDLINE/PubMed

Back to article

  1. Lack of standard methods for evaluation of askMedline

    Julio Bonis, Research Group on Biomedical Informatics IMIM (

    16 March 2005

    In the paper of Fontelo et als, a information retrieval system is described.

    The objective that the authors aim to reach is clever and important for

    allowing the practice of Evidence Based Medicine.

    However, the criteria for evaluation and validation of the tool is in some

    way esoteric, and does not fit with the standar criteria for information

    retrieval systems evaluation.

    Specifically, the authors do not refer to the precision of the system as is

    described eslewhere[1]. It could be desiderable to know, at least the mean

    of the precision of the system for the testing set (POEMs and CAT), in order

    to be able to compare with other information retrieval systems.

    In conclusion, I miss some of the standard methods for information retrieval

    evaluation in the, in the other hand very well oriented, work of Fontelo P

    et als.

    Conflicts of interest: None.

    Julio Bonis MD

    Research Group on Biomedical Informatics (

    Biomedical Research Park of Barcelona (

    Barcelona - Spain

    1 - Kagolovsky Y, Moehr JR. Current status of the evaluation of information

    retrieval. J Med Syst. 2003 Oct;27(5):409-24.

    Competing interests