Cloud Based Decision Support System for Personalised and Participatory Cancer Therapy Selection

DECIDE: Better decisions with smarter data.

Clinicians are currently not able to integrate valid prediction models into their clinical workflow. This is unfortunate, as these models would provide clinicians with a comprehensive overview of prognosis and treatment options, leading to more informed decisions in personalised cancer therapy. Most of the required data and models for this purpose are included in the patient‘s electronic health records (EHR) and medical imaging data. But due to the many different data formats, it has not been possible yet to access them in a selective and specific manner and integrate them into the clinical workflow.

But now, for the first time ever, DECIDE introduces a standardised platform that provides the base for semantic data processing, secure data exchange and the interoperability of this important data. Using this semantic architecture, clinicians can now access the prediction models through a cloud based interface (cUI) and use them to make more profound and faster treatment choices and supporting their clinical decisions (CDS).

DECIDE also provides an additional interface for shared decision making (SDM) that has been specially developed for the use by patients. This interface enables the integration of patients‘ individual preferences and life style information into the decision-making process, if required. In addition, the patient interface (pUI) informs the patient in detail about his or her illness, the various treatment options and their side effects.

In summary, DECIDE enhances clinical decision support and shared decision making with prediction models and optimal treatment options and provides clinicians and patients with crucial information in the decision making process of therapy planning.

Helping clinicians help patients:
Clinical Decision Support (CDS) with DECIDE cUI

Treatment decisions are usually based on the individual experience, medical knowledge and skill of a physician. But with the advent of personalised patient-centered treatment and its many different aspects this poses a major challenge for clinicians, especially in the field of oncology.

For instance, to choose the appropriate cancer therapy, an oncologist must consider patient specific biology such as genetics, epigenetics, and micro-environmental factors and balance it with various treatment options, median survival, treatment toxicity, comorbidities, patient preferences, and cost-effectiveness. In addition, the increasing focus on shared decision making (SDM) places additional demand on the physicians time and effort to engage and inform patients. And as the field of medicine is getting more and more complex, there are constantly new guidelines and rules in how to conduct specific treatments that are released on a continually basis. This enormous number of different tasks is far too intricate for a clinician to handle single-handed, whether it‘s too time consuming or due to the fact that the limited human cognitive capacity can only process five factors per decision.

For this reason, various CDS tools have been developed to help reduce diagnostic errors, improve quality of care and alleviate intense time demand placed on clinicians. Most of these tools focus on information management and situational awareness, and include features such as automated information retrieval, alerts, summaries, and reminders. However, the demand for these CDS tools is limited, mainly due to the fact that they do not show an improvement in outcome and disrupt existing clinical workflows.

To be truly useful for treatment selection, a CDS should provide enhanced insights and contain epidemiological and research data in combination with electronic health records, imaging data, and patient genetic information. This would give the clinicians a comprehensive picture of the patient‘s disease and its treatment options, right up to distinct treatment outcome prediction models. However, a CDS with this range would require the interoperability of patient data with predictive models and the uniformity of medical ontology. And there is the catch: The many different data standards have made this impossible so far.

But with DECIDE‘s cross-data technology these most significant data sources can finally be accessed and used for clinical decision support.

Architecture of DECIDE

DECIDE is utilizing the SeDI 2.0 technology developed by SOHARD SOFTWARE to retrieve and search existing data from various sources for specific attributes. After querying the patient data from PACS and EHR via SEDI 2.0 and creating a personalised patient locker for storage of the queried patient information, the data is further processed via feature extraction and the prediction can be derived.

DECIDE provides a clinician interface (cUI) that enables healthcare professionals to view these results in clinically validated models for predicting treatment outcome. After reviewing the outcome, the clinician can then select one of the various treatment options to best suit the patient.

With DECIDE, clinicians are supported with the right information and knowledge at the right time to choose the best treatment for any given patient condition.

Architecture of DECIDE

No decision about you without you:
Shared Decision Making (SDM) with DECIDE pUI

A cancer diagnosis changes the lives of those affected significantly. Existential fears and uncertainty regarding the disease and its treatment determine everyday life. The hope for a cure is linked to the fear of fundamental changes in lifestyle, stressful therapies with often uncertain outcomes, and patients‘ concern that they will have no influence on the course of treatment. Many patients perceive this situation as uncontrollable, generating even more fears and helplessness. In an approach to get back a feeling of control and have a voice in the whole therapeutic process, affected patients seek to find out more information about their illness and the treatment options at hand. It shows that if patients are involved in clinical decision making, they no longer feel that they are at the mercy of treatment and the course of the disease.

For physicians, this is a crucial point given that the promotion of patients‘ autonomy in decision-making causes more acceptance of treatment strategies. We argue that the most balanced treatment decisions result from the combination of the clinicians medical knowledge combined with the wishes and views of the patient. However, the process of shared decision making can also be challenging, especially when the clinicians and patients do not share the same view regarding the risk of the disease.

Another challenge lies in different evaluations of therapy-related restrictions impacting the patient‘s quality of life. For example, a patient may have strong concerns about a drug due to potential serious side effects, no matter how rarely they occur. In contrast, the clinician may not see any cause for concern because of the rare manifestation of these side effects. The perception of side effects is highly individual, as some patients deem side effects as more disturbing than others.

The role of transparent, mutual communication between physicians and patients becomes evident. It is important that patients are welcome to express their personal preferences, desires and fears as regards the therapy. Only in this way can the information about his lifestyle be incorporated objectively into treatment planning.

For this reason, DECIDE offers the patient an informative and user-friendly interface. Developed by the experts at MAASTRO Clinic, the DECIDE pUI gives the patient a valid source of information and makes it possible to include his or her preferences and lifestyle choices into treatment planning.

The many explanatory texts, illustrative animations and interviews with clinical experts in DECIDE‘s pUI provide the patient with all the necessary information on the disease and its various treatment options. In this way, the patient receives a comprehensive overview of the advantages and disadvantages of the various treatment options and their side effects, both short and long term. With all this information provided, the patient can now confidently enter his or her personal treatment preferences and lifestyle choices.

Based on this information and further prediction models, the patient is shown the expected results and the effects of the proposed treatment(s) on his preferences and lifestyle in DECIDE‘s pUI. By assessing the risks of the various options and incorporating their personal values, the patient can now make better informed decisions about his therapy together with the attending physician.

Prof. Andre Dekker
Prof. Andre DekkerMAASTRO Clinic, NL

„Thanks to DECIDE, our oncologists can now access a condensed overview of all clinical information on the patients‘ disease and the resultant treatment options, including valid prediction models.

This helps them to make faster and more profound treatment decisions. DECIDE‘s patient interface also allows the integration of the patients‘ individual preferences into treatment planning, whilst the patients receive thorough information about their disease and the various treatment options.

This also saves a lot of time and supports our clinicians in the complex clinical decisions they have to make every day.“

Architecture of DECIDE

Benefits at a glance

  • Reduce misdiagnoses and the risk of medication errors

  • Provide the entire care team with consistent, reliable information

  • Access all information in one place

  • Improve efficiency and patient throughput

  • Patient and Physician Satisfaction

  • Improved Communication and Outcome

  • Raised Patient Empowerment

Technical information

DECIDE contains the following components:

2. Cloud based services consisting of

cUI – A clinician specific user interface for clinical decision support (CDS) with imagebased cancer phenotyping and outcome predictions performed through an application service : cUI allows a selection of appropriate clinical factors, proposed treatment(s), diagnostic imaging examinations and displays the predicted treatment outcomes, supporting the medical staff in their clinical decisions.

pUI – Q patient specific user interface for preference input and shared decision making (SDM): pUI enables the collection of patient preferences and lifestyle information and displays the predicted outcomes and effects of the proposed treatment on the preferences and lifestyle habits based on.

1. A semantic architecture consisting of

SeDI 2.0 – A semantic platform for consolidating data, queries and interoperability across different resources including electronic health records, genetic information of the patient and prediction models.

pLock – A personalised patient locker for storage of queried patient information.

Interested in DECIDE? Contact us.