Escape from the data dilemma

  • Good data bad data

The more data the better the therapy decision and greater success with treatment. Sounds good in theory. But rarely happens in practice. At least not without more factors. Because – and this is something health institutions prefer not to discuss – much of the medical data circulating in hospitals have no additional benefit. In the worst case the data are detrimental. How can this happen? Very simple: data as such have no value. Value must first be added. 

To make one thing clear: the digitalization and availability of medical information is an essential step on the road to high-quality and also affordable medicine. We need digital medical data to offer good medical care into the future. However, health institutions are currently in a rather uncomfortable transition phase in which data are available in increasing volume but not necessarily in a secure or useful form. 

The case of Bernd Bochum and his digital data

The problem can be considered in more detail by describing the case of the patient Bernd Bochum and his constricted coronary arteries as it plays out in the not too distant future. Bernd Bochum is among the digital pioneers and is keen to have as much data as possible for his treatment. For this reason, he has a digital patient record, he consults doctors who are involved in telemedicine networks, and he himself collects digital health data, such as his blood pressure. 

Therefore, when he is admitted to the cardiology ward in the hospital of his choice a lot of his health information is available in digital form. His cardiologist has uploaded radiological images to the hospital PACS and has sent laboratory results by email. Bernd Bochum still has the left cardiac catheter images from a previous catheter procedure on a data medium and he brings that with him. His general practitioner conducted another ECG before his admission to the hospital and the results are also available in digital form. He still has some information on paper and of course there is also the medical data in the electronic patient record (EPR) of his health insurer. Bernd Bochum would of course like to make this information available – but not completely. He would prefer to keep any information about his mental health treatment to himself. 

Frustration more than enjoyment: responsibility for the completeness of data

The volume of data is not always welcomed by the treating doctors in the hospital but tends to be more a source of frustration. For example, how are they supposed to know that radiological images are in the PACS? Are they allowed access to the data media with the catheter image? When were the ECG data recorded? And how useful are the laboratory values if they can't be sure that they include all the (medication) information from the EPR that is relevant to the treatment? 

Instead of simplifying their work and saving time, the treating doctors find themselves responsible for searching out all the information that may be relevant to the treatment, from inside the hospital and from external sources. What happens if they overlook something is not currently precisely defined. And also the fact that Bernd Bochum does not release all the data of his EPR is another problem for doctors, because they have no idea that some information is not available and, if they do know, they don't know what it is. The result is that examinations may be repeated to make sure that they are not missing anything. Therefore, that digitalization and data availability will avoid duplicate examinations may be true only to a limited extent. In the case of the missing information on Bernd's mental health, the result may even be medication-related complications in the catheter laboratory. Responsibility status: not clear.

What is important? The problem of data validity

Apart from data completeness, defective data validity may also cause problems with treatment. The validity describes the nature, the quality and thus the usefulness of the data. The following example demonstrates why this may be a problem: at first glance the doctors find it reasonable that ECGs taken by Bernd Bochum's general practitioner are available. Only on a second look is it clear that the last ECG was recorded ten days ago and that therefore its value for the current assessment is limited. This brings up the question of whether it makes sense for doctors to search this information themselves. This will certainly take additional time and effort when considering the number of patients, treatments and data. 

Validity may be affected by methodological and technical errors, such as with the creation of radiological data and particularly with ultrasound examinations. With the increasing attention to external data, whether from patients themselves or non-medical sources, there is also the question of the value of the technology and its application. Examples include blood pressure or pulse data recorded by patients themselves. It is difficult to judge whether the blood pressure values recorded by Bernd's carer at home are relevant to his treatment. After all, nothing is really known of the instrument, the skill of the carer and the circumstances of the measurement. Who is to decide whether these records should even be considered?

A question of liability: data security and data protection

A particularly sensitive topic for doctors and patients can be liability in the case of accidental infection by malware, such as ransomware, whether from an external data medium or even the EPR. What happens if Bernd Bochum exposes the hospital of his choice to a hacker attack from outside by the transmission of his data from the EPR? Initial discussions on this topic are already under way and at least hospitals are considering the question of data security with emphasis on EPR data received by the hospital. 
How realistic any liability risk for patients is remains to be seen. In any case, this debate draws attention to a previously unresolved problem to which there is still no answer.
And ultimately the medical professionals treating Bernd Bochum and others are confronted with the, very significant, interest in data protection. Access to the various data sources is mostly subject to various protective measures such as passwords and authorizations. In the everyday work of the hospital these protective measures may hinder or even prevent access to all available data. The task here is to establish a reasonable compromise between data protection and a positive user experience. 

Who is responsible for good data?

Not all the current problems with the variety of data can be resolved at a stroke, but some can. However, in practice there is always the question of who is responsible for good data. Is it the medical practice? Is it the insurers? Or is it the hospitals as the place where large volumes of information come together? 

In actual fact it makes the most sense for hospitals to take on this work and to establish order in the data. For ultimately this effort will benefit the efficiency, the quality of the treatment and the satisfaction of all involved. There are some points where hospitals can make good data from bad data. This starts with the consolidation of all data within one system and the view with a common viewer, continues from sorting the data to assigning metadata and structuring the data through to communication of data over secure paths, such as external servers.

The deciding factor is finally open communication on the problems of data variety and also open discussion of effective solutions.