Real-life data in research

Real-life data collection is an integral part of clinical research

So-called “real-life” data are defined in opposition to data collected in interventional clinical trials: unlike the latter, they concern all patients (the selection criteria are generally much broader) at the cost of much more data. missing. However, they allow to provide an additional level of evidence, useful in decision-making. Moreover, in its definition of the different levels of scientific evidence, the Haute Autorité de Santé includes them according to their relevance.

There are 2 types of so-called “real life” data:

  • Those who are from medical records (and / or administrative) patients
  • Those who are directly from patients

Different levels of proof

This method of systematic scientific analysis aggregates the results of several independent studies which have answered the same problem, and which have used a similar methodology. The meta-analysis allows for a broader analysis of the data by increasing the number of cases studied and drawing an overall conclusion.

In this type of clinical study, one group of participants is exposed to a substance (for example, a medicine) while the other group (called a "control" or "control") is not exposed. The results are then compared between these two groups to determine the specific health effects of the intervention studied. These are studies whose data are collected prospectively and which usually require a large number of participants.

A cohort study brings together subjects who together share a number of common characteristics. They are then monitored over time, at the individual level by collecting information about them at different times. Most often, cohort studies collect data prospectively. The number of participants depends mainly on the characteristics of the population to be included.

Cohorts are particularly well suited to assess a risk associated with exposure to substances hazardous to health, for example. After a certain observation period, the incidence rates of the effects of the substance are compared between exposed and non-exposed: these are then "exposed / non-exposed studies".

Case-control studies are used to highlight factors that may contribute to the development of a disease by comparing subjects who have the disease (the cases) with subjects who do not have the disease but who are similar moreover (the witnesses). Most often these are studies with retrospective data collection; they are well suited to the analysis of rare cases.

This method uses the description of one (or more) particular case (s) to clarify certain points not previously known. It allows hypotheses to be developed before carrying out other studies.

Expert opinions are a type of publication that makes it possible to highlight concepts known to certain recognized experts in a specific scientific field but which have not yet been covered by clinical studies or whose results have not yet been unanimously recognized. by the scientific community. These expert opinions make it possible to alert and / or guide towards an improvement in practices but need to be compared with the results of subsequent clinical studies.

* Use real life data

What data SKEZI can collect?

The tools developed by SKEZI make it possible to collect all types of real life data directly from patients. Secondly, these data can be analyzed with regard to data from the medical records of the same patients (this is referred to as “data chaining”, after agreement of the participants).


PROMs are questionnaires, accompanied by documentation specifying information such as methods of administration, scoring, analysis and interpretation. They provide important information that is not collected with traditional clinical measurements. Are questionnaires that explore important points for patients: they can be general and concern all patients, regardless of their health problem. or more specific to a given pathology. They are particularly interested in questions concerning the impact on their quality of life (quality of life, QoL or Health related QoL - HRQoL) or on more specific dimensions such as physical capacities. They can also measure parameters relating to a particular pathology.


PREMs are interested in how the patient experiences the care experience, his subjective (for example, attention to his pain) and objective (such as waiting times), or the relationships he has had with caregivers.


Studying satisfaction is an important step in adapting a service offer like care. Satisfaction surveys ask individuals about various points that are important for improve the quality of the service provided. It is possible to adapt the questions according to the situations.

Other data obtained directly from the patient

For routine examinations, it is possible to ask patients to photograph medical reports (prescription, scanner, ultrasound, etc.) or enter this information themselves, which makes it possible to share real-life data usually present in medical files but, here, acquired directly from the patient.

New e-health technologies are generating new types of data that will soon be possible to collect also thanks to SKEZI: for example, connected watches or smartphones also generate health data that it can be useful to collect, in particular on the physical activity of the participants.

Real-life data

Real-life data is data generated in the everyday life of patients: they areoppose to clinical trial data that define in advance the data they will collect in a selected population. The latter have the advantage of being in general exhaustive but their generalization is limited: the results of these trials make it possible to better explain certain clinical situations but are not transposable to all patients. For example, children, pregnant women, the very old or those with a lot of co-morbidities are often excluded from clinical trials. It is therefore difficult later to apply the results of these same tests to them.

Real-life data complements these tests by providing another type of information: they better represent the reality on the ground, even if they are less exhaustive.

For example, they can describe the use of a drug in an environment that does not necessarily correspond to the conditions of the studies that led to its marketing: these data can describe a need not covered and not identified during the development of the drug. drug. These real-life data may come from clinical data from the medical file (medical history, prescribed treatments, medical and paramedical procedures, results of biology or imaging analysis, etc.),administrative information (socio-demographic data, medical coverage), of data from connected objects (saturometer, blood pressure monitor, glucometer, bathroom scale…) or finally data collected directly from the patient. 

All of this collected data must be transmitted in secure databases in real time on approved servers hosting health data.

SKEZI® has created the tools allowing all stakeholders to generate results that can safely support data collection directly from patients.

Personal data

We talk about personal data when it comes to information that can be reported to a natural person, whether he or she is identified at the outset or can be subsequently identified. We then distinguish in particular health data, the definition of which is broad: can be considered "health" data which by nature has a link with health (such as medical history, treatments, etc.), data that will be subsequently cross-referenced with health data by nature (for example, the number of steps and blood pressure), or even any data from the moment when their use is medically planned.

The collection of personal data makes it possible to have more precise details of the patient's overall health. By collecting this information without going through a caregiver (medical record often incomplete or focused on a particular medical situation), we can thus collect more precise and reliable data. For example, with regard to data on the care pathway (relations with caregivers, access to care, support, patient involvement in medical decisions, etc.), it is essential to know the personal point of view of the patients.

How does an e-cohort accelerate clinical research?

The e-cohort, an infrastructure for simpler and faster research.

A cohort is a group of individuals who will be followed over time. It can be selected on a given pathology (diabetes, hypertension), or be general population, and include everyone. These cohorts will define data, which will be collected over time.

An e-cohort is characterized by collected data on an internet platform. Its participants consent to the collection of their data in advance.

There are two types of e-cohorts:

  • Some collect data from medical records
  • Others collect data directly from patients

These two types of e-cohorts are collaborative : in the first case, several doctors (investigators) collect their data to advance in research; and in the other case, it is the patients themselves who become actors in the research by entering their data. These types of e-cohorts can be combined: patients and doctors can collect the data.

Accelerating research thanks to e-cohorts

Cohort studies consist of observe the occurrence of health events over time within a defined population that shares common characteristics. They allow, in particular, toevaluate the links between factors so-called exposure (demographic, biological, behavioral, environmental, genetic, etc.) and the occurrence of health events (disease, biological marker, etc.). An e-cohort makes it possible to constitute a community of patients wishing to participate in research on a theme by contributing to the generation of a large database enriched secondarily by other data sources.

Involving patients in research allows the results to be more relevant and ultimately to be used by society.

The central concept of an e-cohort is to lead to a collaborative infrastructure for research. This unique and shared research infrastructure, involving a shared participant and a data pool, facilitates and accelerates research for all. In this community, patients give responses to online questionnaires. Depending on the wishes of the investigator, they can become active members of the research community and be involved in all stages of the research, such as the choice of projects to be carried out within an e-cohort, the design of all the materials intended for participants, for data analysis or for the dissemination of the project.

When new questions emerge about a specific disease, they can be asked directly of the population carrying that disease, if a specific cohort already exists. These responses provided by the many individuals concerned thus accelerate the answer to a new research question.

If investigators need to test a new molecule in a given disease, it is possible to contact the participants of these e-cohorts, and to propose via the e-cohort the participation in a clinical trial. The e-cohort thus makes it possible to recruit potential candidates more quickly. in large scale. In addition, some e-cohorts offer participants to interact with other individuals and so create a community : they can then suggest new directions for current research.

A few examples of e-cohorts