Home > All articles > Unstructured healthcare data contributes to solving new issues
Unstructured healthcare data contributes to solving new issues
Finland has unique data for pharmaceutical purposes and even unstructured data can be effectively utilised.
Approximately half of the real-world evidence (RWE) studies conducted by Medaffcon utilise unstructured healthcare data. Unstructured data refers to, among other issues, text-form entries in, for example, medical reports and patient records. Other forms of unstructured data include images, videos and audio files, the use of which is much less common in RWE studies.
The use of unstructured data is necessary, because structured data from, for example, laboratory tests and diagnosis codes is not sufficient to respond to all research questions of pharmaceutical companies.
A lot of information has been entered into patient records in free-form text, or data has to be gathered from multiple sources. Such data include, for example, the metastasis of cancer. This data is needed, for example, when different patient groups are created for an RWE study on the basis of whether the patient has metastatic or local cancer.
In addition, data on the patient’s smoking status, height, weight, body mass index or similar issues is usually only available in unstructured form.
Utilising unstructured healthcare data
Healthcare data systems are widely discussed in terms of how well the data from different systems can be utilised and what could be done to make it more usable.
Medaffcon’s Data Analysis Lead Iiro Toppila and Data Scientist Juhani Aakko consider the Finnish social and healthcare data already excellent and unique even at an international scale. There is always room to improve, but already now the data can be very well used to support research, as demonstrated, for example, by a study on smoking and surgical complications conducted by Medaffcon.
– Finnish registers and data lakes are unique when thinking about what data is accumulated and how they can be utilised. In addition, legislation provides a unique opportunity to utilise data.
According to Toppila and Aakko, a good example is a data lake study that Medaffcon has conducted with a Finnish hospital.
– One might think that a data lake study focusing on one hospital is very small, but in reality the number of patients in the study is larger than in a review based on patient records of the same topic covering the entire Europe.
Careful consideration is key
According to Toppila, international pharmaceutical companies operating in Finland should actively bring up the uniqueness of Finnish data lakes and data. He also encourages them not to hesitate to contact Medaffcon’s experts already when they are planning a study. The use of unstructured data increases the duration and costs of a study. That is why its use should be carefully considered, says Juhani Aakko.
– Unstructured data should be used in a study when it actually helps to include critical data. It should not be lightly introduced into any study.
Iiro joined Medaffcon in March 2017 as a Biostatistician. For the preceding four years, he has worked as a research assistant in an academic study group, analyzing clinical and genetic patient data. Iiro holds a Master of Science degree in Technology in Bioinformation Technology.
Iiro’s strengths include his strong expertise in statistics and data-analysis, hands-on experience in working with sensitive patient data, and strong interdisciplinary communication skills with experts from various fields. In the field, he is particularly interested in the large data amounts made available with the revolution of technology and how the information received such data can potentially be utilized to draw concrete conclusions, both in order to understand the nature of diseases and to advance the goals of the pharmaceutical industry and patient treatment.
“Machine learning and AI-based solutions will have a major impact on the healthcare sector now and in the future. However, effectively utilizing the already collected and available health-data will have a higher importance in order to improve health-care”.
Juhani joined Medaffcon in October 2020 as a data scientist. Prior to joining Medaffcon, Juhani has worked as a data scientist in a global IT company as well as a scientist at the University of Turku in the Medical Bioinformatics Centre (MBC) and Functional Foods Forum (FFF). Juhani holds a Doctor of Science in Technology degree (2017) and the topic of his thesis was the development of human gut microbiota in early infancy.
Juhani has experience from applying statistical and machine learning methods in medicine and due to his multidisciplinary background, he can easily communicate with people with varied expertise ranging from clinicians to IT-professionals. “Knowledge management and business intelligence have become hot topics also in the social and healthcare sectors. It is very interesting to be involved in harnessing the vast amounts of data available in the systems to actual usable information to support decision making. Both traditional statistics as well as advanced analytics and artificial intelligence will be in a key role in this job.”