Big Data, Big Trouble? How hospitals master the flood of data

Digitization brings many advantages, but it also generates massive amounts of data. Especially in healthcare. The editorial team asked Swiss hospitals how they deal with it. There are many possible solutions – but there are also hurdles.

The flood of data in the Swiss healthcare system is greater than ever. Precision medicine is an example of this. It is based on personal medical data and quickly generates 1 terabyte per treatment case. “But Switzerland has some catching up to do, a large amount of data is hardly used at all,” says Martin Pfund, CIO of the Cantonal Hospital of Graubünden. The effort to make the often unstructured data usable is very high. Some hospitals have already failed because of this.

The Lindenhof Group has also recognized the problem. In Switzerland there is no stringent national initiative for the collection and use of data, it is said on request. The data of the service providers are often too incomplete to be used meaningfully. This currently works best with the oncological data registers.

Alexander Urs Thomann, Dr. sc. ETH and Head of Digitization and Architecture at the University Hospital Zurich. He says that internal medical and operational processes have been digitally supported for decades. In terms of data-driven cooperation across company boundaries, Switzerland actually has some catching up to do. The reasons for this are the difficult margin situation and the high need for investment funds for structural adjustments. There is therefore often a lack of funds for investments in IT and in data and digital topics.

What would be possible if the collection and use of data in the Swiss healthcare system were better? Thomann says this would lead to more efficiency and effectiveness along the supply chain. The quality of treatments would increase and the patient experience would improve. With the recognition of patterns one could diagnose emerging diseases early on, adds Pfund.

The paper has to go

One reason for the inadequate collection of data is that many medical professionals still work paper-based and do not use digital patient files. “Nothing is as fast as a paper note,” says Stefan Hunziker, head of IT at the Lucerne Cantonal Hospital (LUKS). He digitized this process in LUKS in September 2019 with the LUKiS clinic information system. “Some had problems with the fact that they now have to enter their notes into the system themselves,” says Hunziker. But they would be supported with speech recognition, text modules and ongoing optimizations.

The Cantonal Hospital of Graubünden hardly ever relies on paper files. Only when it comes to historical files that have not yet been digitized, says Pfund. For general practitioners, it is often a question of age and habit. Since the effort for an electronic file is very high, older doctors in particular often come to the conclusion that a change is not worthwhile. The Lindenhof Group also notes that many doctors are not even ready to invest in digital systems due to their advanced age.

“In our experience, at most a third of all resident doctors work without a practice information system, and we don’t know any hospital or aftercare facility without a clinical information system,” says Thomann. The fragmentation of the market makes cooperation more difficult. The actors in the health care system would improve their internal processes with data, but there were hardly any incentives for optimizing the overall processes. Neither monetary as in the Anglo-Saxon countries nor regulatory as in the Nordic countries or in Eastern Europe.

“The paper processes have established themselves in clinical processes over the years,” says Christian Schraner, CIO of the Baselland Cantonal Hospital. New procedures and digital processes would temporarily always lead to additional work in everyday clinical practice and only add value after a while. It is important to find the right balance between quick success and sustainable solutions.

Adrian Grob, the media spokesman for the Insel Gruppe, points out that a digital patient file is not enough to successfully digitize a hospital. It is always necessary to adapt the clinical processes. “Data alone has no value.” Valuable information can only be obtained from the data if all data silos are brought together and the data obtained from them are combined sensibly.

Brake block federalism

“I don’t have the impression that Switzerland is lagging behind other countries when it comes to integrating the systems within health care facilities,” says Schraner. Pound sees it that way too. The interfaces are mostly international standards such as HL7, DICOM, IHE and FHIR. Rather, the problem lies in the complex and highly heterogeneous system and equipment landscape in Swiss hospitals. The federal structure of Switzerland also means that there are no central requirements regarding systems and technology. In the Scandinavian countries, however, all hospitals would use the same clinical information system.

Thomann also states that federalism is paralyzing progress. “The concepts and technologies for the electronic patient record will no longer be state of the art when they are introduced at the end of 2020.” Regulatory requirements for standards, monetary incentives or tax rebates on digitization investments could improve the situation. There is potential wherever digitization is top-down, writes the Lindenhof Group. The Swiss approach to the EPD is only conditionally expedient, the double voluntariness an obstacle.

To tackle the problem, the State Secretariat for Education, Research and Innovation and the Federal Office of Public Health have launched the “Swiss Personalized Health Network” (SPHN) initiative. It should create standards for formats, semantics, and governance and thus increase interoperability. The archipelago participates in SPHN. She hopes that this will improve the prevention, diagnosis, therapy, and recovery of diseases. Transdisciplinary research should also benefit. The University Hospital Zurich also finds the SPHN useful.

Another example of more interoperability is the “eMediplan” of the eHealth Central Switzerland association. It contains the current medication of patients in various formats, such as paper, QR code, and machine-readable code. This simplifies the exchange of medicines between general practitioners, pharmacies, and hospitals. The Lucerne Cantonal Hospital wants to join the “eMediplan” in 2021.

The cloud is finding its way into hospitals

One way to deal with the flood of data in the healthcare sector is through the cloud and big data technologies. The Insel Gruppe uses these for the automated recognition of patterns and for the analysis and structuring of large amounts of data. The use of cognitive computing is still not widespread. According to Grub, such applications are only used in specific clinical expert systems.

The Cantonal Hospital of Graubünden also relies on the cloud. However, only if it is possible to keep the data exclusively in Switzerland, says Pfund. Big data technologies and cognitive computing, however, are not yet in use.

“Where data security requirements are guaranteed, cloud solutions are already being used selectively today,” says Schraner. On the other hand, cognitive computing is only just beginning. At the Lucerne Cantonal Hospital, patient data is stored in its own data center. Today, for reasons of cost and the complexity of large data repositories, it is no longer possible to keep all data locally, according to Hunziker.

“The cloud is demanding from a compliance point of view,” warns Thomann from Zurich University Hospital. For the preparation of data, the process optimization, and the support of clinical-diagnostic decisions, however, artificial intelligence is used. However, there are few big data applications with billions of data points. The Lindenhof Group announces that it is not yet using any cloud or big data technologies. The hurdles are high due to the legal situation.

What do wearables bring?

Another trend is the private recording of activity and health data. Can hospitals benefit from the quantified self-movement? The Insel Group hopes to gain knowledge about the clinical pictures of the patients. In order to use the data, however, investments in data development and analytics are required, says Thomann. He hopes that the patients themselves will now take responsibility for avoiding typical lifestyle diseases. That would redefine the role of the service provider – from therapist to coach for health and quality of life.

Wearables could play a role in the treatment of chronic diseases such as diabetes and cardiovascular diseases in the future, says Pfund. The devices should, however, be certified as medical devices. The Lindenhof Group also identifies the greatest potential of wearables in dealing with chronic diseases.

When patients block the data

Handling personal health data is particularly tricky. “There are resistance and fears on the part of the patient,” says Grob. In the case of rare diseases in particular, however, it is important that the data can be used for research purposes in order to increase the sample size. “It makes a difference whether you are dealing with digital natives or digital nomads,” adds Pfund. The natives would usually be more relaxed about passing on personal data. “Skepticism can only be countered if the digitized solutions meet all data protection requirements,” says Schraner.

Thomann demands that the hospitals create trust through transparency and education. In the research context, the “General Consent” for the further use of data and samples and the “Informed Consent” for specific projects are important. One adheres to the requirements of the Human Research Act and Ordinance. In this way, the hospital can guarantee that no data is sent to researchers before the ethics committee has approved the project. On average, 85 percent of patients would agree to the general consent. Incidentally, the coronavirus had little effect on reducing resistance, says Thomann.

Original Article Source from