How can data analytics to improve the care and outcomes of cancer patients?

I want to share some thoughts about artificial intelligence, or as I prefer to call it “data analytics.” Fundamentally: How can we capture the capability of analytics to improve the care and outcomes of cancer patients? And more importantly: How can we harness this technology to help bring back the human touch in cancer care?

Admittedly that’s a large focus covering lots of opportunities. Speak to one expert, and you will get one idea of how analytics could improve care. Speak to another, and you will get another entirely different view of what that means and how we should be using our rapidly advancing capabilities to harness machines and their capacity to learn and engage health care, specifically cancer care.

My real concern is that there are so many sophisticated opportunities to choose from that we may be missing some of the forest for the trees: Our care system is in such a mess that we be overlooking an opportunity to apply even the simplest, most rational use of data to improve everyday care.

Doing that, which we are currently not — perhaps because we are focusing on other, lofty ideas — might well have the most impact getting us down the road to using technology as an enabler of efficiency in the health care system. Finding a simpler way to apply what we know may improve not only clinical outcomes, but help patients and families better understand their care and treatment options while offering the clinicians caring for them more satisfaction in their professional lives as well.

A recent presentation at the plenary session of the ASCO meeting — where thousands of cancer clinicians and researchers gather in one room to hear the four abstracts chosen from thousands submitted for presentation — somewhat makes my point.

The researchers asked a simple question: What was the impact of Medicaid expansion on reducing the disparity between black and white cancer patients to receive timely treatment after an advanced cancer diagnosis in states that expanded Medicaid as part of the Affordable Care Act?

The answer was straightforward: Patients diagnosed with cancer in Medicaid expansion states after implementation of the Affordable Care Act had an overall improvement in the time from diagnosis to treatment. Before expansion, black patients had a longer delay to treatment compared to white patients. After Medicaid expansion, that difference was almost completely resolved.

The takeaway: Medicaid expansion improved care for advanced cancer, at least with respect to this group of people (patients on Medicaid insurance) and one variable, namely the time to treatment after diagnosis. In turn, the result teaches a very important lesson: Health policy can have a genuine impact on improving the quality of cancer care.

That may sound simple, but the implications are huge: Medicaid expansion worked.

Now about the data part: This study accessed information gleaned from a medical records system deployed in oncology offices across the country, looking at very specific pieces of information contained in those records to find the data that could answer the question. In other words, the wisdom of the masses offered real-world evidence to answer a question of national importance, concluding as noted that Medicaid expansion made an important difference in the care of cancer.

This may not sound as fancy as having a neural network read a radiology image to potentially augment or perhaps in some cases replace a skilled radiologist. It is not as fancy as financial algorithms that now appear to be ruling (and roiling) our financial markets, looking for tiny correlations unseen by human calculation to give signals whether to buy or sell a stock.

Nope, it ain’t fancy. It’s just practical and it’s important.

Later in the day, I participated in a panel discussion at an IBM Watson Health event on the role of artificial intelligence in cancer care.

During my comments on the panel I focused on a theme: Maybe in this world of gee-whiz computational capabilities we need to focus on some more basic questions, such as how do we help our patients and caregivers get better access and understand their care, and how can we reduce the immense burden currently placed on clinicians to deliver quality cancer care?

I wish I could put into words how badly physicians and other health professionals are getting burned out and overwhelmed by the demands health information technology has placed on them. To paraphrase what I said to the audience last evening: “There isn’t enough cheese you can give to the hamster to keep them running on this wheel faster and faster. We need technologies that are going to help us, not make the lives of clinicians and patients even more complicated.”

Let me not be misunderstood: There is an incredible amount of value in using artificial intelligence, machine learning, and data analytics to solve very complicated problems. No doubt whatsoever.

However, while we have our eyes on the sky, the moon, the stars, and the sun, let’s not forget to see the trees in the forest. Let’s take some of that capacity to help technology become an enabler and not a barrier. Let’s take some of that data and answer some straightforward questions, such as how to measure the quality of care, make certain patients don’t get lost in the system, help address access to the appropriate medication for the appropriate patient, reduce some of the excessive administrative burdens that clinicians and hospitals face every day and that consume millions if not billions of dollars.

Let’s unleash technology to do what it is supposed to do and return us in part to a day where oncology health professionals can look their patients and their loved ones in the eye, have a heart to heart discussion, perhaps hold a hand. Let’s recreate a human dynamic and interaction that most of us crave when facing the most serious threats to our health and our lives, especially when that threat is cancer.

Wouldn’t it be great if technology could help us to become human again? That, my friends, would be intelligent — and not artificial.

J. Leonard Lichtenfeld is deputy chief medical officer, American Cancer Society. He blogs at Dr. Len’s Cancer Blog.

Image credit: 

Leave a Comment

Most Popular

✓ Join 150,000+ subscribers
✓ Get KevinMD's most popular stories