Article Image

CNS: Exploring Data and How to Better Access It

The 2019 Congress of Neurological Surgeons meeting was held in San Francisco. The theme was “The Age of Reason for Neurosurgery.” The distinguished speaker list included Lucy Kalanithi, Shankar Vedantam, Bret Stephens, Doris Kearns Goodwin, Rebecca Skloot, and Carl Zimmer. The theme focusing on the utilization of reason in this “post-truth” era is especially relevant. 

Dr. Lucy Kalanithi, the widow of neurosurgeon and "When Breath Becomes Air" author Paul Kalanithi, partook in a moving conversation. Joined by Anil Nanda and Julie Pilitsis, Kalanithi’s remarks set the tone for the four-day conference. She spoke of her late husband, mortality, and physician burnout. Being a physician herself, she is aware of what burnout is doing to our profession. Not only is it an expensive crisis, but it is leading to growing numbers of physician suicides. 

Giving her perspective on the burnout crisis, she gave some insight into its causes. She touched on the administrative burden facing physicians. This is reaching critical mass as physicians now spend two hours of administrative time for every hour of patient care. The amount of time and money spent reporting data for quality metrics has also reached absurd levels

The presidential address by Dr. Ganesh Rao highlighted how misinformation threatens modern medicine. Pushing back against purveyors of false information, such as the anti-vaccination movement, Dr. Rao’s address implored us, as neurosurgeons, to adhere to a scientific rigor in our practices and our lives. 

Many has hoped that a proliferation of quality data will allow our field to enter “The Age of Reason.” Despite much progress, the rise of “big data” alone will not save us from a proliferation of misleading studies. It is worth exploring how “bad data” can, in fact, exacerbate the spread of misinformation. Given the massive amount of data being collected with EMRs and administrative databases, combining this with the computing power to run advanced statistical models, it’s possible to come up with a multitude of associations. Because of differing statistical methods, different variables have completely opposite effects. This explains the stream of contradictory observational studies, many of which make it to the lay press. 

Additionally, poor data can lead to misleading conclusions. With the recent metric fixation in health care, it is vital that accurate, high-quality data be used. Metrics are not only used for public reporting of outcomes, but with the rise of value-based payments ensuring metrics have a fiscal impact on medical care. Bad data can have profound downstream effects. For example, the use of imprecise data has led a widely used commercial health care prediction algorithm to demonstrate a racial bias against African-American patients. 

Datasets have varying quality. There are many publications utilizing administrative datasets, despite studies showing their inferiority to registry datasets. However, even in highly-regarded registries, simple inputs can have questionable accuracy. For example, our research shows that a relatively straightforward data point, the Glasgow Coma Scale at admission, can have variable accuracy in the trauma registry of a level one center. The downstream effects of risk-adjustment, performance metrics, and process measures can be profound if the baseline data are inaccurate.

Dr. Rao also cautioned against disreputable journals promoting misleading conclusions. It’s easy for us, as physicians, to dispute the “evidence” promoting the use of dubious therapies such as homeopathy, reiki, and naturopathy. Yet these methods are abundantly used by the public. Poor science is abound in the lay media, with widely-read publications spreading misinformation using these dubious statistical methods in observational datasets. Of the most shared health articles in 2018, three quarter of them contained false or misleading information

This misinformation on dubious treatments can cause serious harm. Dr. Rao told the story of the questionable reasoning in Steve Jobs’ resistance in using established methods in treating his pancreatic cancer. Bringing the conversation around to the most personal of notes, Dr. Rao showed a similar tumor in his own wife. 

Despite the focus on reason, value, data, and metrics, things that matter often cannot be quantified. As neurosurgeons, we are notorious in our neglect of these immeasurable aspects of life. Speaking of deeply personal family tragedies, Dr. Kalanithi and Dr. Rao remind us of the importance of these things. I’m grateful that I was able to listen to them speak and participate in the 2019 CNS annual meeting. 

More from Op-Med