Key Takeaways:
– Severe flu outbreaks increase the chance of death for lung cancer patients by 25%.
– The study involved researchers from Columbia Medical School and data scientists from Virtualitics.
– The analysis required correlating two data variables – incidence of flu and lung cancer.
– The data sets used were from the SEER database and CDC’s ILINet.
– The study showed that higher vaccination rates can help reduce death rates amongst lung cancer patients.
A recent joint research project conducted by experts at the Columbia Medical School and Virtualitics, a data science firm, has revealed new insights into the connection between flu outbreaks and the mortality rates of lung cancer patients. The study, backed by statistical data, demonstrates that during severe flu outbreaks, lung cancer patients have a 25% higher chance of death compared to periods of low flu spread.
Unveiling the Connection
For years, oncologists have suggested that influenza outbreaks could pose a significant risk to lung cancer patients, given the virus’s propensity to attack the upper respiratory tract. However, there was a gap in concrete, empirical evidence to back this theory. Through an extensive data analysis project, this gap has been filled, confirming previous inklings and presenting new challenges for healthcare practices.
The Columbia and Virtualitics team published an academic paper in the journal Nature in 2023, wherein they discussed their research techniques and findings. Dr. Simon Cheng, the Columbia Medical School oncologist who led the study, voiced surprise and satisfaction at the clear and strong results yielded by the project.
Involvement of Virtualitics
Having originated from a California Institute of Technology astronomy data lab, Virtualitics made its mark by applying its blend of analytics, machine learning, and visualization tools to military and government sectors. The latest study reinforces the company’s potent contribution to the medical field.
Virtualitics brought their software and AI expertise to the table, collaborating to execute a rich, robust analysis. As their CEO and co-founder, Michael Amori, emphasised, the successful application of their platform to healthcare data affirms its power and versatility.
The Unique Challenges of the Study
Despite the comprehensive nature of the study, various challenges marked its execution. The study aimed to identify correlations between two data variables – the incidence of flu and lung cancer – across populations over time and space. This was particularly demanding as it was difficult to consider all variables concurrently.
Data Resources and Analysis
The lung cancer data for the study was sourced from the government’s Surveillance, Epidemiology, and End Results (SEER) database, known for its high data quality. The team also used flu data collected weekly by the CDC’s U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet).
The actual process of merging these two extensive datasets required a considerable amount of skill and time. A Master’s degree computer science student helped in wrangling with the data and integrating it onto a cloud-based cluster. Once the data was standardized, the Virtualitics data scientists could apply the company’s proprietary software to analyse and extract insights effectively.
Valuable Visualizations and Solutions
The Virtualitics software, originally designed to analyse astronomical data, employs several built-in machine learning algorithms. A unique feature of the software platform is its ability to explain its workings clearly, making AI-guided visualizations more understandable and quicker to query.
Their contribution helped provide powerful visualizations that showed the increase in mortality amongst lung cancer patients during high flu periods and developing a time-lapsed map that showed data over time. This movie can be viewed in the supplemental materials section of the Nature study.
Implications and Future Potential
As a result of the study, the teams suggest that the death rate of lung cancer patients could be decreased with regular flu vaccinations, an area that is currently on par with general adult population figures. The integration of AI in medical research still has considerable potential and could contribute majorly to future advancements in healthcare. This project and its findings mark a significant stride forward in this domain.
