Developing learning organizations around health data
Developing learning organizations around health data
By Romain Finas, Vice President Real-World Evidence, Alira Health
Much has been written about the sensitivity of health data, which sits at the crossroads of ethical issues and protection of privacy and sovereignty. However, beyond these concerns lies the subject of the quality of the health data quality itself, which is often inconsistent or incomplete. At its core, health data can provide a world of possibilities for research and Artificial Intelligence (AI) mechanisms, making the quality challenge a critical issue to address and overcome. Alira Health, a healthcare and life sciences consultancy with expertise in health data, approaches the development of health data in an integrated way, bringing together clinicians, software publishers, and AI to improve reliability, discretion, and patient care.
Produce quality data for research and artificial intelligence
For scientists, manufacturers, and regulators, accessing the data produced during medical treatments opens new fields of research and innovative applications. For Alira Health, this is a competitive issue, as the richer and broader the database, the more scientific value the data has. It is also an opportunity to analyse “real-life” to understand and compare the health care lifecycle and better identify the difficulties that patients face throughout their journey. The access to quality data enables new methods of remuneration for health outcomes and provides for faster access to innovations by measuring the effectiveness of treatments as soon as they are on the market, instead of after the comparative clinical trial phase. Eventually, and above all, this data is the raw material for artificial intelligence algorithms that will make medicine even more personalized, preventive, and predictive.
However, “high quality and in-depth data are necessary for current analyses but difficult to obtain,” says Giacomo Basadonna, Chief Medical Officer at Alira Health. “The information that comes from healthcare information systems is still too heterogeneous: the
Fortunately, the technological framework is evolving in France towards greater connectivity of information systems under the leadership of the Agence du Numérique en Santé. We are only at the beginning of a deep transformation of the healthcare system, which must also tackle the way the data is produced.
Few alternatives to get real-life data
For research, the standard remains the controlled database. Data is entered manually, and the data coming from information systems is verified by an operator. This solution often has organizations financing their own real-life research in partnership with one or more hospitals. The company then bears the infrastructure and data control costs alone, in addition to the development of personalized medicine algorithms and the performance of clinical trials.
The alternative to the controlled database is to use registers. These are “ready-to-use” databases, but the clinical data set is often too rigid and does not always correspond to the variables of interest to the clinical trial. While this is a faster solution, it is not very agile.
For medico-economic or epidemiological analyses, manufacturers can have access to medico-administrative databases such as the National Health Data System (Système National des Données de Santé, SNDS) in France. This database is unique in the world, and now brings together information from health insurance, medico-administrative hospitalization data, and the death register. It gives a true picture of the healthcare consumption of all French patients and covers more than ten years of patient data. Additionally, the data is coded in a homogeneous manner. It is one of the few databases that can be used as a source of proof of effectiveness and is enforceable against the regulator. This database is a virtual gold mine for anyone who wants to understand the care pathways and measure total costs of care. However, without clinical data, it remains an approximation of the patient’s actual state of health.
Moving towards healthcare data learning organizations
What lessons can we learn from the SNDS? All stakeholders should consider providing high-performance coding to guarantee that the data is useful, usable, and meaningful. Health professionals and institutions are encouraged to produce this quality information because it impacts their remuneration, while the insurer seeks to control costs.
However, it seems difficult to ask medical and nursing teams to “properly” code medical information in addition to providing care. The New Yorker recently concluded1 that computing has already induced such an overload that it is considered a cause of burnout in the medical professions.
“We have to innovate through a short cycle approach between the producer, the user of the data, and the developers,” says Bernard Castells, PhD, Director of Innovation & Transformation at the Hospital Valenciennes (France) and Coordinator of GHT medical project. “It is about creating a learning ecosystem, focused on solving the problems faced by health professionals during data entry.” This means starting from the use of the data (during and for care), working with publishers on the integration of new technologies (such as the reports coding, including during the consultation), and then monitoring the quality of the data set in relation to its end users. Eventually, the difficulties encountered in obtaining the missing data must be studied again and questioned.
From our perspective at Alira Health, a virtuous circle is taking place. These new organizations must be built around a pathology, like the registers. It is about securing the mobilization of clinicians around research projects and limiting IT developments in each field. This sets the stage for providing enough agility to researchers, manufacturers, publishers, and developers to constantly increase the data’s reliability and nurture the data sets they need.
By acquiring Care Factory, the international healthcare and life sciences consulting firm, Alira Health has extended its unique portfolio of expertise and offers to support, partner, and promote these new real-life research environments. Alira Health is now a recognized partner in data management and analysis of health data (SNDS, etc.) in France, as well as in Europe and in the United States, which complements the firm’s services in clinical development, regulatory affairs, healthcare innovation market access, transaction advisory, and equity investments in innovative companies.
- Medical data is key to understand and improve patient care. It will be at the center of the algorithms for more predictive, preventive, and personalized medicine.
- Medical data is currently difficult to use due to its heterogeneity and its random entries.
- Alira Health proposes to create ecosystems of clinicians and researchers, as well as publishers, developers, and manufacturers with the common objective of producing meaningful data to understand a pathology.
About the author:
Romain Finas, Vice-President Real-World Evidence, Alira Health and Director, Care Factory. Romain Finas is an expert in strategy consulting with a specialization in healthcare organizations. The exploitation of real-life data has been at the center of recent projects he has led, including the development of tools for managing the population of diabetic patients in collaboration with a pharmaceutical manufacturer and medical software publishers.