New browser can identify disease trajectory patterns among 7.2 million Danes

Diet and lifestyle 13. nov 2020 3 min Research programmer Troels Siggaard Written by Kristian Sjøgren

A newly developed browser enables researchers, doctors and lay persons to explore patterns in disease trajectories among millions of people in Denmark.

Interested in Diet and lifestyle? We can keep you updated for free.

Since 1976, the Danish National Patient Registry has collected data on all residents of Denmark who have ever been in contact with a hospital. Researchers have now turned these countless millions of data entries into a browser that people can explore.

The newly developed Danish Disease Trajectory Browser, which anyone with a computer can access through the website dtb.cpr.ku.dk, enables people to enter a disease such as Alzheimer’s and search for statistics on what typically precedes a person being diagnosed with the disease and what may follow.

The search example above identifies that people diagnosed with Alzheimer’s are often diagnosed previously with unspecified dementia.

“The Browser is a tool for identifying two or more sequential diseases involving one person with a statistically significant direction. This is useful in medicine because it provides statistical insight into the other diseases people have when diagnosed with a given disease or what they might develop in the future,” explains Troels Siggaard, a research programmer at the Novo Nordisk Foundation Center for Protein Research at the University of Copenhagen who developed the software behind the new browser.

The article about the browser has been published in Nature Communications.

Data from 7.2 million people

The Browser allows anyone to search for statistical associations between diseases in hospital data from 1994 to 2018, the period in which Denmark’s healthcare system used version 10 of WHO’s International Classification of Diseases to code diagnoses (ICD-10).

The Danish National Patient Registry registers people who are hospitalized and are diagnosed using ICD-10. For example, the ICD-10 code for Down syndrome is Q90, and the code for high blood pressure – “essential (primary) hypertension” – is I10.

The tool is based on data from 7.2 million patients and 122 million hospitalizations over 25 years. Anyone can search for associations between diseases but not in data at the individual level. The idea is to enable people to analyse disease trajectories without having direct access to person-sensitive registry data.

When a user enters a specific ICD-10 code, the Browser spits out a result and identifies sequential trends in disease trajectory patterns linked to other ICD-10 codes for up to six diseases.

For example, entering the ICD-10 code for type 2 diabetes (E11) into the Browser will produce many strong and weak associations with other diseases that typically precede or follow a diagnosis of type 2 diabetes.

The associated diagnoses may include high blood pressure (I10), overweight and obesity (E66) or various types of cancer (C00–C97).

“This provides insight into what comorbidities or multimorbidities people with certain diagnoses have and how they involve the same people over time,” says Troels Siggaard.

Advancing knowledge on associations between diseases

By searching the Browser, users can also obtain insight into the probability of a given type of disease trajectory.

A search of Down syndrome shows a frequent association with epilepsy (G40) and that 367 people diagnosed with Down syndrome have also been diagnosed with epilepsy. Further, 112 have been diagnosed with blood poisoning, and 59 have been dying within 5 years of diagnosis.

The Browser also shows how many people have been diagnosed with pneumonia, common colds, ischaemic heart disease, hearing loss and any other disease.

The Browser solely provides disease trajectory steps if a minimum of 20 people have the disease, and otherwise shows only the number of people with the disease.

“We have collected all the data and entered it into a supercomputer to enable the creation of these trajectory networks that you can search,” explains Troels Siggaard.

The Browser was created by a project group with many qualifications: bioinformatics, medical science, human biology and computer science.

Troels Siggaard explains that he is not a health professional but a programmer and that his primary task in developing the Browser was to create the software that enable people to search the huge data set meaningfully and clearly – without direct access to person-sensitive registry data.

May benefit research into new drugs

Troels Siggaard says that the Browser is primarily intended for researchers who study diseases and disease progression.

A possible example would be researchers trying to develop drugs for a specific disease. They can use the Browser to investigate whether people with this disease often have other diseases that may confound the effectiveness of a drug in a clinical trial.

Today, people are also often interested in developing drugs that are effective in treating several diseases, and the Browser can also inspire such work.

Anyone can use the browser to get an indication of how a given disease typically develops among average Danes.

“If you are diagnosed with a disease, you can see what typical disease trajectories patients on average might have in the future, and they can often be quite different. For example, if you have type 2 diabetes, you may develop quite different types of complications later on. This is the background to the whole concept of personalized medicine, in which people with the same disease should not necessarily get the same treatment. Everyone can use the Browser, including hospital personnel, who may have an interest in identifying which multimorbidity subgroups exist in the Danish population,” says Troels Siggaard.

The Brunak Group aims for understanding multi-morbidity disease progression patterns and their relation to treatment events. The group integrates hete...

English
© All rights reserved, Sciencenews 2020