Building a dashboard with unstructured data for prostate cancer treatment

How do you build a dashboard when half of the data has to come from free text fields? Lars van der A used CTcue to create a dashboard for the treatment of prostate cancer for his graduation research as a Clinical Informatics Specialist. He is working at the I&O (information and organisation) department at ZGT, where he directs the information provision of the hospital on a strategical level. Lars shares how he used CTcue in his project and what it brought him.

The analysis of free text fields

To build the dashboard, Lars needed to be able to search in free text fields: “Analysis showed that our EHR contains a lot of relevant information for health care professionals, for the most part this is enclosed in the free text fields. Doctors and health care professionals indicated that the most interesting indicators to them, can be found in oncology letters, letters from general practitioners and in medical results.” Lars became familiar with CTcue via a colleague: “Within ZGT we already used CTcue and we decided to use it in the project as a text-mining tool. We first ran a pilot and it turned out to be successful.” The dashboard he created showed indicators and KPIs in the context of prostate cancer, supplied with data from HiX, the EHR. Of the eight collected indicators, four had to come from free text fields. “That is where CTcue really adds value for me,” Lars comments.

Easy to use interface

It was rather easy for Lars to get started with the application: “The added value of CTcue for me is the very accessible interface, it is easy to use.” That you do not need knowledge of programming to use CTcue is a plus for Lars. When looking at his colleagues, he also notices this advantage: “What I saw, for example, was that research nurses were building their own queries for a trial research. If you have just a bit of affinity with data analysis, you can quite quickly and easily use the application.” Being able to successfully use the tool also depends on personal knowledge and competences according to Lars. He emphasizes: “To build a proper query, you need to have knowledge of the data model within your hospital domain. Also, substantial knowledge of the medical process is of the greatest importance.” That’s where he sees a challenge in the future for hospitals: “The more uniform the registration at the source is, the easier it will become to use the tool and the higher the quality of the data will be.”

Meeting the information need of the health care professional

The dashboard Lars made was received positively. Colleagues from the tumour-research-unit use the dashboard to support their quarterly consultations. “It meets the information need of the health care professionals. Quality of health care can only be improved when there is insight, the dashboard provides that.” He also noticed that the dashboard led to enthusiasm among colleagues: “You see that when health care professionals start talking about it, the demand also increases. With the help of CTcue you can analyse unstructured data in a self-service manner, enabling you to answer ad hoc questions much faster. It’s a driving force.” Within ZGT, he sees great opportunities to make the dashboard part of daily practice: “The challenge for health care staff is to derive the right actions from that insight to improve health care. Do we see trends emerging in, for example, the number of complications after the OR? Then we will take action to reduce the percentage of complications.”

Shorter lead time for research questions

The fact that CTcue provides Lars with independence is of absolute significance for him. “Before it took up a great deal of time to answer the relatively smaller questions. When you offer CTcue as a self-service tool to users, they can rather quickly find answers to those questions themselves. That without being dependent upon others within the hospital.”
He sees that a dashboard, with the use of CTcue, can have an impact on the outcomes of health care. Different kinds of questions can now be answered more easily: “For example, you can make clinically reported outcomes, such as complications, more directly visible. It also helps to determine whether the standards have been achieved. Have certain surgeries been performed often enough this year? The dashboard can also provide insight and direction to determine, together with the patient which intervention fits best. The ambition Lars has with the dashboard is clear: ‘Eventually, with this insight you want to improve the quality of health care and thus the patient’s quality of life. CTcue helps to make that possible.”


CTcue B.V.
Science Park 406
1098XH Amsterdam
The Netherlands
+31 (0)85 600 1037