In the same way that when a human encounters a word that he does not know, IOMED tries to infer its meaning based on the context of the word. The probability that the word(s) or abbreviation(s) refer to a known entity or concept is calculated based on the context, allowing one or more entity proposals to be obtained that are later reviewed. For this, word embedding is used in the search for synonyms, a semi-supervised learning technique that allows using the context of a concept to find synonyms or abbreviations that may be unique in a hospital.

Yes, it is legal, since IOMED complies with the General Data Protection Regulation (GDPR) of the EU and the Spanish data protection regulations for the processing of these data. Moreover, all the data that IOMED processes are always anonymized.

No. The processing of data in text format or unstructured to a structured format is framed within what the data protection regulations understand as the legitimate interest of the hospital and, therefore, does not require the informed consent of each patient.

The output is a structured database that is housed in the hospital.

IOMED has a medical team that, on a constant and recurring basis, reviews a sample of the medical records that have been processed.

IOMED has two ways to ensure quality. On the one hand, and more generally, there is a medical team that, constantly and recurrently, reviews a sample of the Clinical Histories that have been processed, and corrects the outputs of the tool. On the other hand, for each study a review of a specific sample of the study data is made, where the variables that are of interest to the study in question are manually reviewed.

Yes, in order to remove any personal information, the texts are processed at the hospital with the IOMED anonymization tool. This tool is the first layer of anonymization and is capable of finding personal data included in the clinical text, such as names or addresses, being replaced by other false data with similar characteristics (for example, a mention of the name “Maria” is changed to the of “Andrea”). Additionally, the second layer of anonymization finds any personal data that is already structured by the hospital and masks it (with “*”) to avoid possible re-identification.

IOMED can access both but not link them patient to patient because the data it works with is anonymized.

Yes, the tool processes both structured and unstructured information.

The tool can work with all specialties.

No, but IOMED can process image reports.

Currently, IOMED works with Spanish and Catalan and it is expected to cover English and German by 2021.

Yes, IOMED converts both unstructured data and data already structures in the hospital into OMOP format, regardless of whether the latter is in the OMOP format or not.

The Ethics Committee has a fundamental role. All uses of data produced by IOMED must be approved by it.

No, usually a study is approved with a certain number of collections, which are approved by the entire Ethics Committee.

The inclusion or exclusion criteria (or any variable too be extracted in the framework of a study) can be changed but must be reviewed and approved again by the Ethics Committee.

The data is extracted for a specific purpose that is presented and approved by the Ethics Committee. Analysis of the data should not go beyond this approval.

The database is regularly updated with the incremental data generated by the hospital.

No, IOMED can work with any Electronic Health Record System.

No, it only works with those Medical Records that are electronic.

No, the only limitation is that the data must be from Electronic Health Records (EHR).

No, it sticks only to data extraction

IOMED can access all hospital databases that contain clinical data.

OMOP’s Common Data Model (CDM) enables systematic analysis of disparate observational databases. The concept behind this approach is to transform the data contained within those databases into a common format (data model), as well as common representations (terminologies, vocabularies, coding schemes). Once all the data is in a common format, it is easier and faster to perform systematic analyzes.

It has no cost to the hospital. The installation, maintenance, and use of the data will be free of charge for the hospital.

A hospital information system (HIS) that stores hospital records, as well as a data processing center, either owned or outsourced, where virtual co-hosting or server assignment can be performed.

All data is available to the hospital, which can access it through the technical team. In addition, interfaces are being developed that facilitate the use of data by the hospital.

Yes, the data is available to the hospital at no cost.

IOMED makes available to its partners and users a team dedicated to solving doubts of a technical nature or the use of our technology.