We’re not talking baseball, exactly.
We talking about something that started 776 years before Christ, when a record keeper whose name is lost to history wrote that a cook named Coroebus won the 192-meter footrace to become the first Olympic champion.
The most important stat that day was 1. As in #1. First place.
Two thousand six hundred years after Coroebus’s race, Henry Chadwick invented the box score complete with batting averages, runs scored and runs allowed. Chadwick’s game was cricket.
After Chadwick came Ernie Lanigan, John Heydler, George Moreland, Al Munro, Walter Elias, Allan Roth, Harlan Mills and Bill James—who collectively intertwined sports and statistics to the point where today’s fan can’t imagine one without the other.
Arguably, the most valuable member of any professional sports team is the statistician.
“Moneyball”, the book, was published in 2003. The Journal of Quantitative Analysis in Sports published its first issue two years later, in 2005. The movie, “Moneyball”, starring Brad Pitt didn’t make it out to theaters until 2011.
By then, the first conference for sports analytics had been held on MIT’s campus (that was in 2006). In 2015, that same conference had become so popular it was being held in the Boston Convention Center. Thousands of sports data professionals and researchers attended. Oh…and the number of companies exhibiting at the now named, The Sloan Sports Analytics Conference, was in the hundreds.
The sports statistical revolution—for it is nothing less than that—brings new statistical measures for the musculoskeletal system, new technologies for capturing and quantifying biologic phenomena and, finally, new ways of converting that data into treatment paradigms and predictive statistics to the physician.
Medtronic plc is in with both feet.
Soon. Maybe sooner than anyone expects, it will be in your clinic or office.
From Box Scores to Predictive Musculoskeletal Analytics
The first studies in the first peer review journal of sports statistics tells us a lot about the gestalt of this new discipline.
- A Procedure for Prediction of Sports Records
- Determinants of Success in the Olympic Decathlon: Some Statistical Evidence
- Hybrid Paired Comparison Analysis, with Applications to the Ranking of College Football Teams
- Scientific Football 2005 (by KC Joyner)
Later, studies like this:
- Creating space to shoot: quantifying spatial relative field goal efficiency in basketball
- Gasping for air: soccer players’ passing behavior at high-altitude
- openWAR: an open source system for evaluating overall player performance in major league baseball
- Modeling team compatibility factors using a semi-Markov decision process: a data-driven approach to player selection in soccer
- A finite mixture latent trajectory model for modeling ultra-runners’ behavior in a 24-hour race
- Effect of position, usage rate, and per game minutes played on NBA player production curves
Notice the obsession with predictive statistics. And the powerful advanced analytics and modelling tools they use to deconstruct athletic performance.
Sports medicine is becoming more data than jock.
At the recently concluded American College of Sports Medicine meeting in Boston, approximately 40% of the exhibitors were data collection/analytic companies.
Here’s a sampling of the kinds of companies who now inhabit the sports medicine community:
 |
Wearable sensor solutions that address some of healthcare’s biggest challenges. |
 |
Wrist-worn actigraphy. GENEActiv—broad range of raw data, tri-axial accelerometers, designed and built to measure human movement and behaviors in free living studies and clinical trials. |
 |
Software tools for research biomechanics which enhance 3D motion capture systems. Improves patient outcomes and sports performance, reduce injuries, and generally improve lives. Provides mathematical tools that researchers and other professionals need to improve their decision making. |
 |
Spiroergometry systems and mobile respiratory gas analysis systems. These systems record and measure physical fitness and performance—as precisely and conveniently and in an efficient and quantifiable manner. |
 |
Advanced sEMG, dEMG technologies and software solutions that capture, analyze and deliver biomechanical, cardiovascular and performance data. |
 |
Motion capture technologies to capture musculoskeletal movement data, create a visual rendering, analyse it and then use that data for treatment protocols and advanced orthotic and prosthetic designs. |
 |
Data capture using physical medicine, nuclear medicine and molecular imaging. Also advanced tools for assessment and analysis of musculoskeletal biomechanics. |
 |
Gait analysis systems for clinical research, education, hospitals, and clinics. Software solutions for scientifically valid and clinically relevant metrics for a variety of gait and balance assessments that are utilized throughout the healthcare industry |
Source: Logos courtesy of company web sites and table creation RRY Publications LLC
One exhibitor, MuscleSound, which was founded 2011, uses ultrasound to measure glycogen levels in the muscles of athletes. The inventors, Drs. San Millán and John Hill teamed with businessman Stephen Kurtz to commercialize the technology.
Kurtz, a distance runner and fitness enthusiast, saw significant sports medicine potential in a non-invasive muscle glycogen measurement tool when combined with software to read the ultrasound images. Their first customer ready product was introduced in 2013. In May and July 2014 independent validation studies of MuscleSound (directly comparing to muscle biopsies to measure glycogen levels) were completed at Appalachian State University and the University of Colorado Medical Center. Their conclusion, MuscleSound’s non-invasive tools provided comparable measures of muscle performance to biopsy.
This is huge.
The initial product measured the ability of the body to store carbohydrates for energy in the muscles in the form of glycogen. By mapping out the muscular glycogen levels the physician can measure muscle fatigue, healing rates or injury severity.
The company’s next product scanned seven muscle sites in the body to provide an overall picture of the tissue layers and body composition beneath the skin. It tracks muscle efficiency over time.
Clearly MuscleSound is a sports medicine product. But in the world of large joint, trauma, extremity post-op recovery this has the ability to quantify, non-invasively, what has heretofore been patient self-reported.
Finally, another exhibitor named Medtronic (!) told us that they are the largest data collection, analytics and integrator of data systems in this nascent industry.
Medtronic, also known in the medical data business as Zephyr Technology Corporation, now a subsidiary of Covidien, now a subsidiary of Medtronic…err, buried in there somewhere, provides a full range of data collection products for patients at home or athletes in sports, the military or in industry. Zephyr’s products are real time physiological and biomechanical data capture tools which can stream that data remotely to doctors, hospitals or researchers.
Data Is Not Information
The famous computer scientist Clifford Stoll was once quoted as saying, “Data is not information. Information is not knowledge. Knowledge is not understanding. Understanding is not wisdom.”
Sports medicine appears to be in the data, information and understanding parts of Stoll’s framework. The reason sports medicine is so far ahead of any other sector in medicine is because it is so results oriented. Sports is unerringly focused on results.
In sports what is the relevant data which, in turn, can become knowledge, understanding and treatment wisdom? Seems like there are three main categories.
- Training routines
- Nutritional regimens
- Biomechanic and physiologic data capture technologies
Benjamin Alamar and Vijay Mehrotra in their excellent book “Beyond Moneyball,” defined this new discipline of sports data analytics as follows: “The management of structured historical data, the application of predictive analytic models that utilize that data, and the use of information systems to inform decision makers and enable them to help their organizations in gaining a competitive advantage on the field of play.”
What is revolutionary about this definition is that it incorporates a broader information value chain into statistical models. Given the inherently variability of each athlete and their biology, this is the only framework that will work.
The Revolution on Four Tracks
The sports statistics revolution appears to be moving along four parallel tracks that interact with each other.
Track One: Data Acquisition. Wearable sensor technologies and innovative uses of existing non-invasive technologies combined with remote transmission via smart phones are driving Stage One.
Track Two: Data Clean Up. Software tools which find missing, incomplete and/or inaccessible data, correct it, verify it and prepare it for integration and analytics.
Track Three: Data Integration. Integrate with other information systems. Several firms are offering tools that compare individual patient data with the other databases of clinical, registry or insurance data. No longer will information exist in isolated silos.
Track Four: Create Patient Specific Treatment Plans. Apply artificial intelligence and advanced analytics to create treatment paradigms and predictive disease or healing models. IBM’s Watson—among several other new AI initiatives—is setting up intelligent information systems to create truly predictive and patient specific analytics.
So, is Moneyball coming to your practice?
Yes. It is. And the companies that are driving it are, by and large, in tiny little booths at the sports medicine meetings.
Finally, if you attend these meetings, you can dress much more casually than you do at AAOS.