Hospitals in general and IT departments in particular are not known for having overflowing budgets. Yet despite this, they possess abundant riches, real treasures even. Sadly, in many places these remain unremarked, because they are not dazzling and eye-catching, but exist as contemptible bits and bytes. We are, of course, referring to medical data without which modern medical care hardly seems possible. When will healthcare facilities finally mine this wealth of treasure?
In the private sector, people are just dying to get their hands on data gold, because it enables enormous efficiency and quality improvements to be made in medical care and forms the basis of progress, for example, in Artificial Intelligence or in building scientific databases. And the only ones who possess this gold in huge quantities and in a suitable quality are hospitals.
Digitization is the foundation, not the goal.
Managers in healthcare facilities must concede that digital data makes them the 'nouveau riche', because digitization in medicine is only now reaching a level where it can create potential added value. Yet now is exactly the right time to recognize that this cannot be achieved purely by digitizing individual processes alone. It requires strategies and holistic solutions that go beyond simply changing from paper-based processes to digital ones.
Users see it like that too. In December 2017, the German Marburg Bund campaigning association published the results of a survey of 1,800 employed physicians on the subject of "Digitization in Hospitals". One survey comment was as follows: "If digitization was thought through, understood by management to be a core competence and obtained the appropriate support and adequate funding, it would have huge potential to simplify workflows in hospitals and remove the burden of bureaucracy from medical staff". This assessment matches the essentially positive general tone of the survey fairly well. 46 percent of those surveyed detected an improvement in the medical quality of their own work as a result of digitization, with around 40 percent claiming that their work has speeded up. Positive emphasis was placed, for example, on faster access to patient data, simultaneous accessing data by several people and avoidance of transfer errors.
It also emerged from the survey that the potential of digitization has not yet been fully exploited, because for example, too often digital records have to be duplicated or are incomplete or the data search is too complicated. One survey respondent hit the nail on the head: "Sometimes we have to call up two separate programs to work through a file. We're producing huge data cemeteries and thus wasting our time". One important task of clinics is not to let the available data go to waste, but broaden the information content and improve the information distribution. What needs to happen then? The available data must be consolidated, linked to other data, collected, and made available.
Turn your own processes into gold
From the Marburg Bund survey, it is clear that there is a great willingness to use IT systems from the user's side and the advantages are clearly indisputable in theory. They just need to be implemented in practice. For physicians, comprehensive access to medical information—and even case or patient data— is the be-all and the end-all. To meet these demands, it is necessary to harmonize a clinic's data in the different subsystems with the various formats in such a way that it can be called up from a single system. This combining of data has an additional advantage in that the data corpses in the individual subsystems are given a new lease of life, because they are available clinic-wide and can be assigned to new contexts. Also, the clinic is finally master of its own data again, data which was previously captured in isolated systems now and then.
Once it has been merged, this data is not only available internally for ongoing therapy, it ultimately also makes invoicing and collaboration with the medical service easier—simply because it can be called up on a case-by-case basis from a single system via a single viewer. That saves time and potentially hard cash. And it facilitates collaboration with external medical facilities—referring physicians or partner clinics—to whom data can be made available subject to country-specific data protection regulations. Duplicate examinations or communication misunderstandings across sectoral boundaries are avoided, improving the quality of medical care significantly. With a view to building up a patient's case or creating other ways for the patient to access their own data, clinics cannot avoid processing medical information in the medium term in such a way as to enable it to be bundled separately.
Shaping medical progress
Merging data obtained from a clinic's individual systems and areas is, however, not only valuable in terms of increasing the efficiency of its own processes. In a wider context, it is also essential to mine this medical gold. Developments in Artificial Intelligence show this (see page 20) as well as the large amounts of funding that are invested by politics in projects to improve data usage. For example, the funding in the MIRACUM project. The MIRACUM (Medical Informatics for Research and Care in University Medicine) consortium will be funded from 2018 for four years by 32.1 million euros as part of the Medical Informatics Initiative Germany (Medizininformatik-Initiative, MI-I) of the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF). MIRACUM is currently supported by eight universities with university clinics, two universities and one industry partner. In the foreseeable future, up to three more university clinics will join the consortium. The aim is to merge the currently very different data islands obtained from patient care and research into data integration centers so that data can be utilized centrally for research projects and specific therapy decisions using innovative IT solutions. As part of the MIRACUM project, it is intended that a 'data integration center' will exist at the university hospital Universitatsmedizin Mainz in Germany.
Gunther Honing, Co-PI (PI - Principal Investigator) of this project at Universitatsmedizin Mainz, explains the background: "Roughly speaking, the BMBF's aim is to supply the data gathered in clinical routines to research. And the knowledge gained from research should then flow back into medical care. The following requirement is emerging: there must be an exchange of data across all contexts and a common use of data and knowledge. In this case, across all contexts means across all medical facilities".
How this task is solved in technical terms remains a matter for the consortium itself. At MIRACUM, the implementation of open sources has been agreed, because "non-university healthcare facilities are also intended to be connected after the end of the project. To make it realistic, the barriers must be as low as possible," says Gunther Honing. In principal, there should be data warehouses concealed behind the data integration centers and those warehouses will on the one hand bundle together all of a healthcare facility's medical data in databases. On the other, they will coordinate the consortium partners' data searches. In this way, healthcare facilities can request data to answer specific questions posed by partners, for example, about specific gene variants. Searches are initially checked by a commission and—so long as they receive a positive decision—the data integration center will tie up the corresponding data package in this way and archive the original data record.
"Unstructured data, that is, the contents of evaluation letters, represents a challenge to implementation. For data that requires semantic and organizational interoperability, we rely on standards like LOINC or IHE profiles which we recommend that individual hospitals use although we cannot insist. Ultimately, it is left to the partners themselves how to ensure their data provision", explains Mr Honing.
If we think a couple of years ahead then non-university healthcare facilities will therefore soon be faced with the task of making their data available for clinical purposes too. And for this scenario it is also crucial to first of all relocate the data out of the individual silos and to consolidate and standardize them so they can make their contribution to medical progress. The first published results of the MIRACUM project which, for example, have yielded new knowledge relating to bowel cancer based on shared data usage show that these are no empty promises.
Anyone who would therefore also like to practice high-quality, financially efficient, and modern medicine in the future must already be searching for tools to mine their wealth of data.