Ductal Carcinoma In Situ (DCIS) Treatment Decision Tool
A world leader in cancer research and treatment, our client Dana Farber Cancer Institute relies on data science to translate cutting-edge research into frameworks and decision tools that stakeholders can understand and use. They needed one such tool to leverage output from a microsimulation model that predicts outcomes for patients choosing from different treatment options for DCIS. The web-based decision tool has two audiences: (1) clinicians treating patients diagnosed with DCIS, and (2) patients (or family of patients) diagnosed with DCIS.
Dana Farber Cancer Institute is world-renowned for the care they provide. It should come as no surprise then that one of their paramount goals was to ensure that patient users—who, after all, were facing invasive treatment options for a life-threatening disease—could easily access the information they needed from the tool without becoming overwhelmed with statistical jargon or visualizations that, while perfectly appropriate for clinicians, could confuse patients or increase their anxiety. At the same time, the decision tool’s clinical audience needed relatively information-dense visualizations. In total, our client aimed to provide both audiences with appropriate visualizations of the model outputs based on information about the patient such as age, recurrence risk, ER status, and patient preferences.
To address the needs of both audiences and in collaboration with the primary researchers, we helped design and build an interface for each—one for healthcare professionals to provide an overview of all the outcomes on a single page and another to provide patients and their families with the same information but spread over several pages and with more appropriate visualizations and text explanations. Both interfaces presented estimated outcomes a patient might face in order to help her make a treatment choice that is best for her. Specifically, the decision tool predicts the likelihood of recurrence (DCIS and invasive), breast preservation, and survival as a result of selected treatment options (lumpectomy, radiation, tamoxifen, and mastectomy) and the age, recurrence risk, and Estrogen Receptor status.
In keeping with the high value our client places on relieving the burden of disease, a significant portion of the total effort was spent making the patient-facing interface easy to use and understandable for patients and their families. We helped researchers test the patient-facing interface with several patient-advocates who provided valuable feedback that was then used to improve the interface in an iterative approach. Color schemes and photo artifacts were chosen, evaluated, and tested to ensure that the site was as welcoming and inclusive as possible. Great care was taken to mesh different methods of communicating the concept of probability to patients, as studies have shown that they can easily be misunderstood or misinterpreted.
Examples of the decision tool interfaces:
The deliverables for this project are as follows:
- Decision tool source code and deployment instructions
- Website build and deployment source code
- Website deployment documentation
- User interface for clinicians
- User interface for patients and their families
In support of our client’s mission to advance the understanding, diagnosis, treatment, cure, and prevention of cancer and related diseases, CSNW leveraged expertise in stochastic modeling, microsimulation, data-visualization, high-performance computing, web-based standards for accessibility, and software engineering to deliver the DCIS Treatment Decision Tool.