People Analytics in Medicine
Find the right care with People Analytics
It is 11:23 PM on a Tuesday night at the emergency department of the Brigham and Women’s Hospital in Boston. It has been an unusually calm night with only three patient boxes occupied, and most of the staff is watching a new Netflix show on the largest monitor of the coffee corner. Suddenly, the ambulance beeper goes off, alarming the night shift that an unstable patient is on his way with heavy abdominal pain. Upon arrival at the hospital, the internal medicine resident examines the patient and discovers clinical signs of an acute abdomen with fever, a high pulse, and low blood pressure. Inflammatory markers are high and infectious signs are unmistakable. The resident suspects a bowel perforation and calls the radiologist for additional medical imaging. The CT scan shows a perforated diverticulitis with abscesses in the lower abdomen, indicating a left colonic resection with a stoma (known as a Hartman procedure). The resident contacts the surgeon on call, a general surgeon specialized in hepato-pancreatic surgery. The last time the surgeon did a left colonic resection was four months ago… Luckily, the patient doesn’t know, and to be fair, this is a procedure every general surgeon should know. But is this (still) true?
Luckily, the patient doesn’t know, and to be fair, this is a procedure every general surgeon should know. But is this (still) true?
The case described above is a fictive but realistic scenario. I have seen it with my own eyes. I have assisted surgeries with surgeons who were not optimally qualified for the procedures they were performing. Leading with people analytics made me reflect on the enormous potential the novel analytical discipline could have within healthcare organizations and delivery in general.
In healthcare, there is a significant information asymmetry (i.e. the physician has more or better information than the patient). When a patient arrives at the hospital, you actually don’t know who will treat you. What is their expertise, what are they good at, what are their colleagues good at? How many procedures did they perform before you? Imagine that an algorithm could help a patient find the right physician, and conversely, an algorithm could find the right patient for the physician.
Would our patient still want the surgery at night if the algorithm would recommend someone else? Or maybe the patient would like to wait till the next morning and choose the more experienced surgeon. Of course, every case is different, and sometimes, there is just no time to wait. But data-driven decisions in medicine are, in my opinion, still scarce. The question is not if people analytics will be used in healthcare. The question is when and to what extent?
In the article “How People Analytics Are Helping Healthcare Firms Increase Profitability“, the CEO of Arena Analytics explains how the company applies predictive analytics and machine learning to talent acquisition challenges in the healthcare sector. This is people analytics in its most modest form, which is a data-driven hiring process in hospitals. However, the full potential of people analytics in healthcare is yet to be unlocked! I envision various potential exciting applications ranging from selecting the right teams to perform complex procedures, to selecting the best night workers and the best day workers for an emergency department. Hiring the right professionals for your healthcare facility is one thing, data-driven clinical decisions based on people analytics (or physician analytics) is another. Data-driven decisions in medicine are the future, and it is time to jump on the train of innovation before it’s too late.
Reference
Whitler, K. A. (2016, January 14). How people analytics are helping healthcare firms increase profitability. Forbes. Retrieved April 18, 2022, from https://www.forbes.com/sites/kimberlywhitler/2016/01/14/how-people-analytics-are-helping-healthcare-firms-increase-profitability/?sh=e953e92a5936
Dr. Gauthier Willemse graduated Magna Cum Laude from KU Leuven Medical School. He devoted the past four years to surgical training during which he also published scientific articles and presented research at conferences on international surgical training. Gauthier was born in Brussels, Belgium to an internationally-minded family. A polyglot, Gauthier cultivated a global mindset and gained extensive medical expertise in low- and middle-income countries. These experiences sparked a deep passion for social advocacy, which led to the co-founding of Residents Abroad, an organization facilitating international opportunities for care workers. Until July 2021, he acted as the medical coordinator of the largest COVID-19 vaccination center in Belgium. He has medical experience in Belgium, Burkina Faso, Brazil and speaks Dutch, French, English, Spanish, Portuguese, and German fluently. Gauthier is a Fulbright scholar and a current MPH student at the Harvard T.H. Chan School of Public Health where he focuses on innovations and entrepreneurship in healthcare.
Interesting and (to be honest) slightly intimidating proposal. Challenging to quantify a worker’s quality (physician, nurse, phlebotomist, janitor, everyone)! Hard to get data and the necessary surrounding data. A classic example is patient outcomes without accounting for their comorbidities. Patient survey scores, without patient demographics, SES, education level, health literacy? Phlebotomy success rates without accounting for patient access, age, comorbidities (central line? dialysis patient?). Restroom cleanliness scores without time of day, nearby patient volume, hospital/clinic location, accessibility?). Which, if any, of these scores would be significantly associated with the patient experience and patient outcomes?
Anecdotally, some surgeons track their volume, outcomes, and complication rates (on their own or as part of their department’s quality metrics) and tell patients who are “doctor shopping.” But is that the only metric that matters? How accurate is it?
I worry about “garbage in, garbage out” and the potential to limit access to care or damage the patient-healthcare system relationship and trust. If patients in a certain region don’t have access to a 5 out of 5-star provider/hospital, but don’t have the resources to go to a different provider/hospital, then will they lack trust and start to question their quality of care? Confirmation bias for people with higher scores who may not actually be good or become complacent? Will people with lower scores be fired or required to remediate certain skills? What if there is a lack of healthcare workers (or at least, a geographic distribution issue with supply) — is it better to hire someone with a low score or to be understaffed?
Clearly I have a gut response to this, and in part because you seem so enthusiastic about it without acknowledging some of the challenges. Would be very interested in learning more about [potential] applications in the US and internationally.
Addendum:
Could the scores be “weaponized” and be used for other applications, like health insurance agreements or CMS reimbursements?
Totally agree, there are many caveats in using people analytics the way I described it. How to use the data and how to interpret it will be a big part of the challenge and the debate! Nevertheless, I do believe there could be a consensus on how to use people analytics in clinical medicine. But the debate has, to my knowledge, not been introduced yet…
Hi Gauthier,
What a well written piece – and unfortunately, so true. As I read it, I was reminded of my time in India with Johnson and Johnson when we helped build surgical and nursing capacity in villages through dedicated training and skill-upgrade programs. Given the high variability in medical education within the country, the disconnect between the CV (academic credentials) and practice was exceptionally high. Your insights make me wonder if we could parameterize practice components for a physician/surgeon/nurse which can help us identify high performing human resources in healthcare. Given the high information asymmetry and supplier induced demand, I worry what these measures could be? and would the medical association ever agree to these?
This was such an interesting use case of people analytics to consider, and I agree that there is wide-ranging potential for the application of people analytics to healthcare and treatment. One hesitation in my mind (as I considered the hypothetical scenario with the physician who hadn’t performed this procedure for 4 months) was how newer and younger professionals would ever gain the necessary experience without the exposure to ‘new’ procedures or ones that they have less experience with. Of course, as a patient, I would prefer to have the most experienced healthcare professional possible – but at the same time I recognize the need for physicians to “learn on the job” if we are to continue developing the pipeline of future surgeons. [I recently read the book by Henry Marsh – ‘Do No Harm: Stories of Life, Death and Brain Surgery’ which chronicles his career in a series of cases, and discusses this very issue of allowing junior doctors to learn on the job, while balancing optimal outcomes for patients]. Another concern I have is whether this sort of information would lead to treatment disparity across socio economic groups if a situation developed where patients paid extra to get the ‘most experienced’ surgeon, according to the data. Sadly, in a country with a healthcare system that is heavily geared towards wealthier patients, I fear this could quickly become a bigger issue.
Thanks for sharing this blog post – I really enjoyed thinking more about this topic!
Hi Gauthier,
It was an interesting read because it related to your personal experience as a surgeon, who could see with his own eyes how informational asymmetry was skewed towards the patient and resulted in suboptimal treatments. In the spirit of sharing my personal experience, I did have at least one surgery performed on me, which I wished had been done by a different doctor or not done at all. There is tremendous potential for use of people’s analytics in various medical fields, but I agree that there are major challenges there also — many of which we touched upon during class i.e., how far should data analytics go w/o having adverse outcomes (many of these were touched upon Ellen in the comments above). In the context of your article, playing the devil’s advocate, as a young surgeon without much experience in “hepato-pancreatic surgery,” I would be worried about my professional outcomes in a highly data driven environment like the one you envisioned. My medical malpractice insurance premium will go up, patients will always opt to have a surgery performed by a more experienced surgeon, then how will I ever acquire appropriate training to be deemed ready to perform hepato-pancreatic surgeries or any other types of surgeries? I think there are many things that can go wrong here, hence the use of data in medicine should be and — as a matter of fact — is regulated. Maybe some level of informational asymmetry here is needed to ensure we have enough aspiring future doctors enter the field w/o getting paralyzed with fear of how big data and ML can prevent them from succeeding long term.
Hi Gauthier,
This was a fascinating — albeit scary — post! The implications of an unqualified surgeon in certain scenarios is certainly a real one, and brings the implications of people analytics to the forefront in a really important way.
I’d be curious to know your thoughts as to whether you believe surgeons in general would be supportive or unsupportive of increased data analytics and monitoring in their profession. Do you think there’ll be a generational divide? Further, will patients need to consent to provide this information to hospitals?
Thanks for this great post. A data-driven culture needs to be internalized in the healthcare sector, which I see as a challenge, especially in developing countries. There is an opportunity to partner with academia, so programs and graduates can know beforehand what the industry needs and is expecting. I’m full of hope that both scenarios, the healthcare industry and higher education, can realize this, as there can be cultural factors potentially undermining this process.