Have you ever noticed how a patient’s body mass index (BMI) in the EHR seems to be placed alongside a patient’s vital signs, as if it is a vital sign?
I don’t think BMI is a vital sign. For me, there are only four vital signs (blood pressure, heart rate, respiratory rate, and body temperature) which, in the right clinical context—such as a life-threatening illness—are essential to know, and which in acute illnesses can be important to know.
And BMI sure as heck shouldn’t be considered a vital sign because it’s particularly useful. Even BMI’s numerator (body weight, the “W” of W/H2) usually isn’t clinically useful (or, rephrased, vital to know) for most adult office visits. For example, when an adult comes in with a fever, I have to know their temperature, their heart rate, their blood pressure, and their respiratory rate or I am working at a disadvantage, but for that visit I don’t necessarily need to know if weight is 220 lbs, 240 lbs, or 260 lbs, or if their BMI is 25, 30, or 35. (Don’t get me wrong. Measuring a patient’s weight can be essential in some acute settings, such as congestive heart failure management and dialysis, and it is often necessary to know for dosing of some medications in some scenarios, such as in chemotherapy or anesthesia, but as a general rule, it’s not medically necessary to measure weight at every office visit. In fact, I believe that measuring weight when it is not necessary for that patient’s explicit benefit, but instead measuring it to just populate a healthcare organization’s data pool, may be deontologically unethical [see my post Healthcare AI, Kant’s Deontology, and My Cat ]).
And finally, I certainly don’t think BMI should be considered a vital sign just because when I do a PubMed search on “BMI” I get 68,959 results. I suspect the large volume of papers using BMI is more of a reflection on the ease of the calculation and measurement, rather than often more useful measures, such as waist circumferences, or measures of visceral obesity.
But in most EHRs, BMI visually appears alongside the vital signs, giving it a “vital sign aura”. Why? Because it’s a hard data point, which, when calculated (from height usually measured once per year, and weight measured every office visit), makes a patient’s data-set more valuable.
But valuable for whom?
Here’s a hint. I saw a presentation recently in which BMI was listed under “vital signs” by a healthcare corporation (one which generates over 100 billion dollars a year in revenue) in the context of an analysis of 60 million EHR records.
(Pause here for just a moment, and think what it means for a healthcare organization to have deep digital access to 60 million EHRs.)
You see, if you believe as I do—that healthcare, because of EHR design and data architecture, has shifted from a people-centric focus to a data-centric focus—and if you understand that hard data points (such as weight and height) make unstructured data (such as nursing and physician notes and the like, which compromise 80% of patient’s data) much more useful, it’s pretty clear why large healthcare companies, which have invested billions of dollars in artificial intelligence and EHR infrastructure, consider BMI as “vital.” It’s because they think it will help them better make predictions, or gain more “insights”. Rephrased: it will help them increase revenue. (See my recent post Is Healthcare Really a Service Industry?)
Look, no physician thinks BMI is a “vital sign.”
But every large healthcare organization and insurance company using artificial intelligence has decided that weight and height—and weight and height’s W/H2 algorithmic child, BMI—is “vital.”
To not recognize the multibillion-dollar healthcare industry’s financial investment in EHRs, artificial intelligence, and data-mining, and to not understand their decision to determine what data they chose to mine—and their power to name and elevate in importance the data as they see fit, including categorizing what is and isn’t a vital sign—suggest a lack of understanding of how much leverage and power healthcare IT has in “defining” terms.
So keep in mind that individuals have only limited control over their own EHR, and physicians have only limited control over their patient panel EHRs ( perhaps 2000-4000 EHRs?). But the large healthcare organizations have deep control over tens of millions of patients EHRs—how it is structured, how it is accessed, what is mined, and if I had to guess, this disparity is only going to get worse over the next twenty to thirty years.