Language Processing Algorithms Can Boost Orthopaedic Research
Manufacturing, farming, and shopping…These are just 3 diverse examples of how technology is advancing daily and automating tedious tasks, decreasing costs, and improving efficiencies. Orthopaedics and orthopaedic research are not being left behind in this progression. In the November 6, 2019 edition of JBJS, Wyles et al. evaluate the accuracy of natural language processing (NLP) tools in automating the extraction of orthopaedic data from electronic health records (EHRs) and registries. The findings suggest that NLP-generated algorithms can indeed reliably extract data without the labor-intensive and costly process of manual chart reviews.
First, using an open-source NLP “engine,” the researchers developed NLP algorithms focused on 3 elements of >1,500 total hip arthroplasty (THA) procedures captured in the Mayo Total Joint Registry: (1) operative approach, (2) fixation technique, and (3) bearing surface. They then applied the algorithm to operative notes from THAs performed at Mayo and to THA-specific EHR data from outside facilities to determine external validity.
Relative to the current “gold-standard” of manual chart reviews, the algorithm had an accuracy of 99.2% in identifying the operative approach, 90.7% in identifying the fixation technique, and 95.8% in identifying the bearing surface. The researchers found similar accuracy rates when they applied the algorithm to external operative notes.
The findings from this study strongly suggest that properly “trained” NLP algorithms may someday eliminate the need for manual data extraction. That, in turn, could substantially streamline future research, policy, and surveillance tasks within orthopaedics. As Gwo-Chin Lee, MD predicts in his Commentary on this study, “When perfected, NLP will become the gold standard in the initial data mining of patient records for research, billing, and quality-improvement initiatives.” Dr. Lee is quick to add, however, that “no machine learning can occur…without the integral and indispensable input of the human element.”
Orthopaedic surgeons are already using robots to assist them in performing total joint arthroplasties. Wyles et al. show how we can use technology to reliably expedite research on that same subject. I believe the future holds much promise for the use of ever-advancing technologies in orthopaedic surgery and research.
Matthew R. Schmitz, MD
JBJS Deputy Editor for Social Media
JBJS 100: THA Registries, Bone-Repair Growth Factors
Under one name or another, The Journal of Bone & Joint Surgery has published quality orthopaedic content spanning three centuries. In 1919, our publication was called the Journal of Orthopaedic Surgery, and the first volume of that journal was Volume 1 of what we know today as JBJS.
Thus, the 24 issues we turn out in 2018 will constitute our 100th volume. To help celebrate this milestone, throughout the year we will be spotlighting 100 of the most influential JBJS articles on OrthoBuzz, making the original content openly accessible for a limited time.
Unlike the scientific rigor of Journal content, the selection of this list was not entirely scientific. About half we picked from “JBJS Classics,” which were chosen previously by current and past JBJS Editors-in-Chief and Deputy Editors. We also selected JBJS articles that have been cited more than 1,000 times in other publications, according to Google Scholar search results. Finally, we considered “activity” on the Web of Science and The Journal’s websites.
We hope you enjoy and benefit from reading these groundbreaking articles from JBJS, as we mark our 100th volume. Here are two more, both from 2002:
The Swedish Total Hip Replacement Register
H Malchau, P Herberts, T Eisler, G Garellick, P Soderman: JBJS, 2002 November; 84 (Suppl 2): S2
In this 19-page analysis of data from the Swedish Total Hip Replacement Register, which was initiated in 1979, Malchau et al. pinpoint the striking clinical and socioeconomic effects of the Register’s first 20 years. The information captured by joint registries, especially in regions that provide universal health care coverage and thus maintain robust databases, has helped orthopaedic surgeons refine indications, surgical techniques, and implant choices.
The Role of Growth Factors in the Repair of Bone: Biology and Clinical Applications
J R Lieberman, A Daluiski, T A Einhorn: JBJS, 2002 June; 84 (6): 1032
Countless studies related to tissue engineering and the musculoskeletal system have been published in the 16 years since this Current Concepts Review appeared in JBJS. Yet this article remains an essential primer for understanding how growth factors affect cells and tissues—and the possible applications for using growth factors to accelerate fracture healing, treat nonunions, and enhance spinal fusion.
Linking Registry and Claims Data—Not as Boring as it Sounds
It is not often that The Journal of Bone & Joint Surgery publishes an article about data-linkage efforts. To even raise the topic with most readers of The Journal would elicit a yawn and quick dismissal of the abstract without a second thought. With this fact duly noted, the possibility of linking health-system joint-replacement registries with Medicare claims data is a first step in a potentially game-changing approach to achieving the long-term clinical research our specialty needs.
In the June 20, 2018 issue of JBJS, Raman et al. detail their successful linkage of a total ankle arthroplasty (TAA) registry with Medicare data without the use of unique patient identifiers. Among 280 TAA patients over the age of 65, 250 had their registry data linked with their Medicare record with exact matches for date of procedure, date of birth, and sex. Of the linked records, 214 (76.4%) had ≥3 years of postoperative claims data.
Why are these findings so important? The answer is follow-up. Every clinician and/or researcher who has attempted to follow patients beyond the first year after a procedure understands how difficult long-term follow-up is. We live in a mobile society in which informative posttreatment data is easily lost. The younger the patient group, the more difficult it usually is to locate patients as time passes. If patients are doing well, many stop coming to our offices, no matter how strongly we recommend annual follow-ups. Everyone is busy—including retirees—and most have better things to do than drive to their surgeon’s office or even complete a web-based questionnaire. Additionally, some patients care only about their own outcomes; they are not as focused as we are on contributing to the advancement of the profession and improving outcomes at the population level.
By linking patient data from a local health-system registry to nationwide claims data, we can gain a better understanding of long-term patient progress. We can use the patient- and implant-specific data housed in the registry and essentially substitute the information from follow-up visits that did not take place within the registry system with the data contained within the Medicare system, which follows beneficiaries wherever they live.
The marriage of registry and claims data is not perfect, though, because patients who are still working probably have private insurance coverage that is not captured by the Medicare system. (Of course, if universal coverage were to come to pass, that issue would be eliminated.) Furthermore, any time claims data are used, uncertainty about the accuracy of coding must be considered. These real-world limitations notwithstanding, the linkage of registry data with claims data does have great potential for enhancing our ability to analyze—and improve—long-term orthopaedic outcomes.
Marc Swiontkowski, MD
JBJS Editor-in-Chief
Maximizing Specialized Health Registries Under MIPS
OrthoBuzz occasionally receives posts from guest bloggers. This guest post comes from Jason Weisstein, MD, MPH, FACS.
I want to expand on my previous posts (Tips to Excel Under MIPS and Why EHR Data & Analytics Matter) and focus on another differentiating factor when it comes to electronic health record (EHR) systems and your success with Medicare’s Merit-based Incentive Payment System (MIPS).
The ability to interact with specialized health registries is another functionality your EHR system should have. Active engagement with a clinical data registry falls under the Advancing Care Information (ACI) component of MIPS. In general, having EHR-enabled access to such specialized health registries can make MIPS compliance easier and help you earn bonus points, which translates into increased practice income.
Some examples of orthopaedic-specific registries could include the following:
- Medial Meniscus Tear, Acute Registry
- Plantar Fasciitis Registry
- Low Back Pain
- Herniated Disc, Cervical Registry
In addition to having the ability to interact with orthopaedic-specific registries in order to participate in ACI and improve your MIPS score, registry engagement through your EHR system will help to improve population health by collecting and reporting on data about musculoskeletal treatment effectiveness and disease trends. Public health reporting can be very complicated and time-consuming, but having an EHR system that automatically and seamlessly collects and transmits the data to the registry, without manual intervention, is a robust advantage.
Finally, access to registry data will help your practice with the Improvement Activities component of MIPS, which, during the so-called transition year of 2017, is weighted at 15% of the total MIPS score.
Jason Weisstein, MD, MPH, FACS is the Medical Director of Orthopedics at Modernizing Medicine.
Recent Comments