How Analytics Can Help CPMC?
Who is CPMC & What is Avatar?
Our MIP is California Pacific Medical Center, commonly referred to as CPMC. It is a leading Research Institute in the medical field of cancer research, particularly in precision medicine. Precision Medicine is a huge market. In 2018 the estimated global precision medicine market was 80 billion dollars and is expected to grow over 200 billion dollars by 2028. Cancer is, of course, a major focus for most organizations working towards precision medicine. In 2018, 2 million new cancer cases occurred in the US.
One of the key goals of CPMC is to revolutionize cancer care by providing individualized treatment for patients. In this regard, ‘Cancer Avatar Project’ came into existence. It was launched in 2016 to create ‘Mouse Avatars’ by transplanting cancer tissue from patient to mouse to test drug efficacy. Researchers studies patient’s genetic makeup and tumor response to drugs in different concentrations, then collect data for analysis and visualization. What’s more, with the information, physicians can find more effective treatment for a certain patient.
Although precision medicine has a bright future, the current research process can be time-consuming, prone to error and not scalable. That’s why it’s critical to apply analytics on the project.
Associated application of analytics
1. Project proposal
Based on the analysis of the background, CPMC’s current workflow, data type and sources, and objectives of the Avatar project, we came up with a project proposal including our understanding of the opportunities based on the challenges in current workflow, proposed workflow and project timeline.
2. Visualization analysis
One of our opportunities is constructing easily understood visualizations that provide enough information for scientists to have more intuitive but in-depth understanding of test results, as well as for physicians to make an accurate assessment for patients according to visualization reports provided. To achieve these goals, in addition to designing the prototype of the dashboard, we also need to build an analytic platform to carry out the calculation of metrics needed.
1) Analytics platform
The analytics platform is one of the most important deliverables in this quarter. In order to obtain the desired calculations results based on experiment data, certain algorithms used on our automatic pipeline in Python are required. And we must ensure that the calculation is consistent with the results obtained from the external paid software used currently.
2) Visualization report
First of all, we focused on the analysis of demand characteristics of different users. The two types of reports require disparate visualizations because they serve different purposes for the two audiences. The scientists’ report needs more granular data exhibition. However, the physicians’ report needs to be clear and concise. After that, we start to create two reports, a less technical physicians’ report and a more comprehensive scientists’ report, to demonstrate the efficacy of various drugs separately, using visualization techniques such as bar plots, probability density curves, heat maps and so on.
Impact & risks
1. Impacts:
The proposed workflow provides us with a clear instruction during the project. What’s more, the utilization of scientific analysis allows us to help CPMC to reduce workflow time, human errors and costs. Physicians and scientists can also benefit from our team’s analysis because they will receive an optimal version of the metrics measuring the effectiveness of drugs toward the specific tumor. They can easily interpret and view these visualizations more quickly.
2. Risks
Scientist and physicians have relatively very little knowledge about the analysis tools and methods. When our team is no longer directly working with them, these non-technical users may run into bottleneck without clear instructions about how to leverage the analytics platform and the visualization analysis tool: Tableau. Thus, to matigate such risk, explicit documentation really matters.
