2| Is Telehealth Here to Stay in Orthopedics?

Is Telehealth Here to Stay in Orthopedics?

Orthopedics does not lend itself well to telehealth. Unlike other specialties, where diagnoses are made from a description of symptoms and treatment plans are as easy as following the instructions on a pill bottle, orthopedics is unique. Diagnoses require comprehensive imaging and nuanced physical exam maneuvers requiring finesse. Treatment plans are comprehensive and often require shared decision-making between provider and patient, as well as the input of physical therapists, dieticians, and social work. Like every other specialty, orthopedics was forced to adapt at the onset of the COVID pandemic. How orthopedics was able to adapt during the pandemic sheds some light on the future of telehealth in orthopedics.

There is a large financial incentive to make telehealth work. Medicare now reimburses the same for telehealth visits as normal visits, and telehealth theoretically would require less staff, time, and improved patient satisfaction [1]. As the field moves to bundle payments, saving money on routine office visits could provide much higher margins to different practices. The underlying question remains – are telehealth visits as good as in-person ones? Can we both collect and disseminate information as well over camera as we do in-person?

I personally saw telehealth work well when on service. For a large part of last year, we got rid of 2-week post-operative visits where we would traditionally look at a patient’s wounds and take-out sutures; instead, we had patients send pictures of their wounds to us and take out their sutures on their own. We started calling patients in their longer-term follow-ups and were able to answer most complaints and questions on the phone preventing patients from coming to their appointment altogether. Finally, we were able to conduct physical exams from the window of patient rooms with COVID rather than have to perform physical exam tasks in person. It was encouraging to see the amount of information we could collect by having patients do exam maneuvers themselves while following instructions.

In the last several months, research on the reliability of telehealth has emerged in ortho trauma, spine, and sports. The treatment of many non-operative orthopedic trauma injuries completely via telehealth was found to be successful [2]. In spine surgery, 94% of surgical decisions made during an initial telehealth visit were confirmed when seeing the patients in person [3]. In addition, spine surgeons believe in general that telehealth can even be better than traditional visits for treatment planning and discussions [4]. In sports care, post-operative rehabilitation and progress with metrics such as range of motion and gait pattern were easily measured from a remote setting [5]. Provocative diagnostic physical exam tests, however, were unable to be done remotely [6].

Progress has been made to show the current limitations and advantages of telehealth. Even though a lot of orthopedic is currently made easier by in-person visits, my experience and the current research on telehealth paint a different picture: telehealth without a doubt has a role to play in orthopedics and could make many orthopedic visits cheaper, more efficient, and less time-consuming. There is a clear need for innovation to improve the way we do remote imaging and physical exam maneuvers, as well as teach patients about their disease and their recovery from their own homes. This will be the next wave of telehealth companies.

Using Machine Learning for Orthopedic Data Modeling

It seems like recently we have broken through the glass ceiling of using machine learning and neural networks for any sort of model that we would initially use traditional regressions. Essentially, we have found the same thing repeatedly: machine learning is better at making diagnoses and predicting complications when compared to regressions with the same data points.

Cue a massive onslaught of papers using machine learning to improve a data model in orthopedics. Just from the last month, here is a bunch spanning subspecialties:

1. Predicting complications after total hip arthroplasty

2. Predicting hip fracture mortality

3. Predicting functional improvement after hip arthroscopy

4. Predicting survival after Ewing’s sarcoma

5. Predicting orthopedic readmission rates

6. Predicting the diagnosis of septic arthritis

7. Predicting prolonged opioid use after orthopedic surgery

8. Predicting injuries for MLB players

All this research has made it clear – there is a real market opportunity here in using machine learning modeling to improve data analytics in orthopedics. I hope startups start commercializing on these tools in this sector. 2 specific needs that I hope to see soon include:

1. Improving Risk-Adjustment. Medicare uses traditional models to calculate how many complications a hospital “should” have based on their patient mix, and then either penalizes or rewards a hospital if it has more/less complications. These models are heavily criticized for their lack of effective risk-adjustment. I would think machine learning models would do a better job.

2. Risk-Calculators. Several companies have created tools to predict a patient’s risk of a series of complications after several different procedures. However, these models are old, based on traditional regressions, and rarely accurate. Machine learning can help improve these models.
   

What else caught my attention this month

1. AR and VR are revolutionizing surgical education. A study showed that using VR in surgical education can reduce errors by 50% in residents and that 45 minutes of VR training is equivalent to 1 hour of actual OR training. Read more here.

2. Robots have been able to make total joint replacements more precise and mirror normal joint kinematics more accurately, but are they worth the increased cost? Biosensors may allow for similar improvements at a much lower cost. Read more about the tech here, and how Stryker took a large jump into that market recently here.

3. 85% of orthopedic cases that were scheduled since the pandemic have been delayed, postponed, or canceled. Of the new patients, 1 in 5 have planned to delay their care. Once the pandemic is over, I would seriously consider buying shares in orthopedic implant companies.

An Interesting Data Trend

Public perceptions of technology heavily depend on where you live