What are your current research interests and/or what is a project you are currently working on?
We just started a NCI-funded R01 (#1R01CA246418) project recently on understanding the benefits and harms of lung cancer screening in Florida using large collections of real-world data from the OneFlorida Clinical Research Consortium. Lung cancer screening using low-dose computed tomography is a promising technique to reduce the burden of lung cancer, the leading cause of cancer-related death in the United States. Nevertheless, concerns over the high false positives, invasive diagnostic procedures, post procedural complications and downstream health care costs impede developing and promoting lung cancer screening program. Our research is to leverage the linked electronic health record and administrative claims data to understand contemporary use of lung cancer screening and the associated health care outcomes and costs in a real-world setting.
Why did you decide to focus on cancer?
Personal experience. A close family relative was diagnosed with breast cancer when I was little (in the ’80s in China). She lived with us when seeking treatment and care and survived the initial treatment. But then the cancer reoccurred after 15 years, and she passed away. Also, three years ago I experienced an episode myself. Even though it was just a benign pituitary tumor, the experience wasn’t pleasant. I’m still on treatment.
I truly think screening and cancer prevention is better than having to be put on treatment. A lot of my work has focused on mining real-world data from OneFlorida to understand utilization patterns of resources like lung cancer screening and palliative care, but I have also done quite a bit of work on using social media data to understand patients’ knowledge and attitudes toward cancer-related health behaviors.
What excites you about your work? What is exciting to you about your field right now?
OneFlorida is a great data infrastructure that contains rich, linked electronic health records and claims. The data have great breadth, but we need more depths. The advancements in areas like imaging and genetics are giving us more data to “play with.” Being able to apply modern data science approaches to these datasets will be exciting. Nevertheless, there are huge barriers in linking and integrating these heterogeneous datasets from multiple aspects across different levels (e.g., from patient privacy concerns to technological barriers). I think some of our work on record linkage and data integration will bring more value to these datasets.