Objective performance indicators bring data-driven feedback to robotic surgery

Objective performance indicators captured during robotic surgery are opening a window into surgical quality that has never before existed—and the data emerging from early work at Hackensack Meridian Health suggest the technology may ultimately transform not only how surgery is taught, but how it is performed in real time, according to Michael Stifelman, MD, professor and chair of urology at Hackensack Meridian Health.

Stifelman described the fundamental shift that objective performance indicators represent.

"When we think about what good surgery looks like, there was never an objective way to quantify it," he said. Intuitive Surgical's platform processes surgical video using artificial intelligence to segment procedures into discrete steps, then captures more than 150 quantitative metrics across those steps—including tissue force, instrument smoothness, camera movement frequency, energy delivery, and fourth-arm utilization.

"For the first time in the history of surgery, we actually have this new technology that can objectively quantify what we're doing," Stifelman said.

Early signal data from a machine learning analysis of more than 100 prostatectomy cases with linked pre- and postoperative outcomes have identified performance indicators associated with sexual function and continence recovery. The left and right neurovascular bundles and the posterior dissection emerged as the steps most predictive of sexual function outcomes—confirming clinical intuition with quantitative evidence for the first time. Among the OPI signals emerging from those steps, lower tissue force, smoother instrument movements, and reduced energy application during neurovascular bundle dissection appear to correlate with better functional recovery.

"We thought we knew what parts of the operation were important—but now we can prove it," Stifelman said.

The commercially available Case Insights platform—distinct from the research-phase OPI work—is already in use at Hackensack for surgical education. Every week, Stifelman reviews a specific operative segment with a resident, comparing video side by side with approximately 10 available performance metrics. The process has surfaced consistent patterns: Residents move the camera less than attendings, underutilize the fourth arm, and apply substantially more energy.

"It's not now subjective—we're able to pick a very specific part of the operation, review it with them, and layer on the data," he said. Residents engaging with the platform appear to be progressing faster and producing better early outcomes, with targeted simulation exercises addressing specific identified deficiencies.

Stifelman also confronted the governance questions the technology raises head-on. When the platform launched at Hackensack, surgeons immediately asked about data ownership and discoverability. The institution convened compliance, legal, IT, strategy, clinical, and perioperative stakeholders to establish a framework, ultimately placing OPI data under a patient safety organization structure—limiting its use to performance improvement and shielding it from discovery.

"We are not using it for credentialing, for PHP, for OP, or during complication rounds," he said. "We're using it only for the purpose of letting surgeons develop their skills more quickly."

Looking ahead, Stifelman said he envisions the technology evolving from a tool for post-procedure reflection to one that informs surgeons intraoperatively in real time.

"Rather than doing it post-procedure, we're going to do it intraoperatively—use the data to help us in terms of augmented dexterity and augmented knowledge in real time," he said. He described a future where real-time alerts—analogous to a lane-departure warning—could notify a surgeon that tissue force is exceeding safe thresholds during neurovascular bundle dissection, or that energy has been applied beyond an established limit.

"That," he said, "is going to be the next step."