In my side-hustle podcast RED HOT HEALTHCARE, I have the opportunity to interview senior-level and cutting-edge healthcare leaders from all segments of the industry. As A.I. in healthcare has become a top-of-mind effort and discussion, I wanted to learn more about the collaboration between IBM Watson Health and the Mayo Clinic*.
This is a summary of my interview and takeaway.
Dr. Tufia Haddad is the breast cancer chair at the renowned Mayo Clinic. She’s also a world-class researcher whose work focuses on drug development, biomarker discovery, and transforming cancer care. Perhaps her most promising work is in being the physician leader for Mayo’s collaboration with IBM Watson Health in its Clinical Trial Matching or CTM program.
“I was born and raised here in Rochester, Minnesota, and I’ve always joked that Mayo clinic is literally in my blood, having been born and raised here.”
We had originally met a HIMSS in early 2017, after she gave a powerful set of personal comments to a room of top health executives in a Mayo-Watson-PwC collaborative. Several months later, I sat down for a remote interview with Dr. Haddad.
I was impressed at her commitment to career, support for her staff, as well as leading efforts with Watson.
“My commitment is really undiminished. I remain a busy oncologist and I haven’t decreased any effort or time to my patient care responsibilities or to my clinical translational research program here at Mayo.
In addition to that, I still am very engaged with the education of our medical trainees. But I must say, you know, my initial exposure to the AI technology, and specifically Watson, my initial exposure was just over 2 years ago.”
She discussed her earliest experience with Watson, likening it to a newborn. She recognized and took on a responsibility to training its algorithms – an ongoing and extremely important part of a man-AND-machine process, with significant long-term expected value.
Lately, there has been some media and industry pushback on Watson Health. While I won’t get into the support, detraction and varying opinions, it remains vitally important to recognize the distinction between ‘augmented intelligence’, ‘general purpose A.I.’, ‘consumer A.I.’ and ‘business A.I.’ – and the different positioning that technology companies have taken toward these respective roles and offerings in healthcare.
For its part, IBM’s CEO Ginny Rometty has made a bold prediction that by years’ end Watson Health will be trained on 80% of all cancers. While that won’t bring an end to the War on Cancer, it seeks to help more healthcare providers bring better, more personalized levels of data and best care options into timely doctor-patient-family communications.
The amount of new studies and data surrounding disease detection and treatment is staggering. Dr. Haddad spoke further on the challenges surrounding care providers today:
“By 2020, medical knowledge will double at a rate of every two and a half months. How can we augment and enhance the intelligence of our clinicians and our providers?
It is currently impossible for us to cognitively keep up with the pace of knowledge growth now.”
Tabling outcomes and quality for just a moment, much of the patient care experience has soured in the last 10 years. The necessity and implementation of Electronic Health Records (EHRs) have led weary providers into two hours of computer work for every three hours of patient time.
I bring up to Dr. Haddad the point that Dr. John Glaser of mega-EHR company Cerner had previously stated to me. Specifically, that clinicians are tired of ‘pajama time’– that is, working on their home computers after dinner to finalize daily notes.
Moreover, providers often have to spend significant amounts of time, driving through large amounts of EHR data, to accumulate and pull together the right data to come to the best decisions for each individual patient and their circumstance.
She chimes in:
“The goal with cognitive computing systems in terms of developing robust clinical decision support is really to reduce that time that it takes a physician to go mine the electronic health record and find those critical pieces of information or data points for each individual patient that they then need to cast against all of what we know about medicine then, and then to formulate a treatment recommendation.”
IBM Watson and Mayo seek to bring the following together: the EHR system data, the National Library of Medicine, and also the patient’s DNA and their individual profile and care needs. Plus in the case of cancer patients, the ability to match up the patient with the potential benefit of engaging in most-appropriate clinical trials.
Surprisingly, only 3-5% of all patient in the U.S. with cancer participate in clinical trials – and according to Dr. Haddad, that number hasn’t changed over decades. I reel back in my chair, wondering why someone hasn’t come up with real-time matching before – it seems so incredibly necessary.
Much of the A.I. world today is promised results. So I take the conversation to actual results.
“So what are Mayo’s results with Watson Health – specific to Clinical Trial Matching (CTM)?”
She jumped in:
“We have seen a significant decrease in the amount of time that it takes to screen a patient very systematically at a high level of detail across all the different inclusion and exclusion criteria for each study. That time has gone down from approximately 30 minutes WITHOUT Watson to about 8 minutes [a 73% drop]with a clinical research coordinator doing the screen.
So really, we’ve been able to markedly improve the efficiency of screening.
And just to give some perspective, at our Cancer Center we have over 1,000 clinical trials that are active at any given time. And if I look at myself as a breast cancer specialist, we have 105 systemic therapies of medical therapy trials that are active and enrolling patients today. We have 9,000 patient visits each year, men and women with breast cancer at Mayo Clinic, and we do over 1,000 new-patient consultations.”
I then divert our conversation from clinical into administrative.
After all, administrative services and back-end processes account for 25% of all hospital costs. It’s an operational segment still carrying large amount of human employment (salaries + benefits), manual processes, and a big need to increase efficiency and decrease time waste.
Though her training and expertise is largely clinical, I ask Dr. Haddad if she’s considered Watson as a potential ‘replacement tool’ for the often inefficient, manual efforts in admin services. In other words, I broached the taboo subject of replacing human jobs with smart A.I. technology and related systems.
Having a clinical background, I was surprised at the amount of thought she had given toward this area:
“You know, initial efforts for leveraging A.I. and cognitive systems in healthcare have indeed been focusing on how can we better augment medical knowledge for healthcare providers. There is increasing attention and focus now on how such systems can also be utilized to improve the overhead, to run the hospitals, the clinics, and these complex healthcare organizations.
You know, one of our earliest observations in working with Watson Health is that some of the data being derived from the EHR can absolutely be re-purposed into new solutions that address some of these administrative tasks.
You know, supply chain would be another example. Cognitive systems, you know, can be used to assess drug and medical supplies, ensuring that, that our stock is sufficient, but wouldn’t it be great if we could have a cognitive system that could develop predictive algorithms that are going to anticipate the needs before the stock is running low?”
Now I move our conversation into the healthcare world of tomorrow. A future that paints a challenging picture.
She tells me that over the next 15 years there’s going to be a 45% increase in the incidence of cancer in the United States – by virtue of our baby boomers coming of cancer age. As breast cancer survival rates improve by 1 to 1 ½ percent each year, that increase in cancer survivors coupled with an expected shortage of oncologists, causes a necessary thought-shift to care for these patients under new models.
Dr. Haddad takes a bold step in moving away from the common thoughts of her oncology peers. Frankly, I appreciate her candor.
She believes the role of the primary care physician will be key for cancer survivors, especially in post-treatment and surveillance management. As we move into value-based care, plus an increasing need for greater efficiency and cost-effectiveness, it is more vital than ever to reduce the fragmentation and duplication of care. Two factors that weigh heavily on piling up cost and waste.
IBM Watson and many technology companies have taken on a mission to redefine the long-term expectations and results of our health system. One where the highest levels of investment, costs, employment and profitable growth have not been equally matched with care quality and outcomes.
As demand and chronic disease rates rise, Watson Health positions itself as a collaborative partner in a journey with its provider clients.
Along the way, I suspect they will uncover and embrace more doctors such as Tufia Haddad. Professionals whose vocational training has not limited their innate interest in being part of solutions that will help shape next generation health and wellness.
Matching cancer patients to the right clinical trials at the right time is powerful and life changing. Along the way, Watson Health has also discovered the strength in matching their mission and technology with powerful efforts from dedicated healthcare providers.
As IBM’s Rometty states, it’s not ‘Man Vs. Machine’…but ‘Man AND Machine’.
* Remote interview with Dr. Haddad on Red Hot Healthcare Show (June 25, 2017). You can listen to the entire conversation at this link.